**Nanomedicine Based Drug Delivery Systems: Recent Developments and Future Prospects**

Editor

**Faiyaz Shakeel**

MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade • Manchester • Tokyo • Cluj • Tianjin

*Editor* Faiyaz Shakeel Pharmaceutics King Saud University Riyadh Saudi Arabia

*Editorial Office* MDPI St. Alban-Anlage 66 4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access journal *Molecules* (ISSN 1420-3049) (available at: www.mdpi.com/journal/molecules/special issues/Nanomedicine Drug).

For citation purposes, cite each article independently as indicated on the article page online and as indicated below:

LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. *Journal Name* **Year**, *Volume Number*, Page Range.

**ISBN 978-3-0365-7761-6 (Hbk) ISBN 978-3-0365-7760-9 (PDF)**

© 2023 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications.

The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND.

## **Contents**


Reprinted from: *Molecules* **2022**, *27*, 4366, doi:10.3390/molecules27144366 . . . . . . . . . . . . . . **121**


## **About the Editor**

## **Faiyaz Shakeel**

Prof. Faiyaz Shakeel received his M. Pharm. and Ph.D. in Pharmaceutics from Jamia Hamdard (Hamdard University, New Delhi, India). At Jamia Hamdard, he worked on nanoemulsion-based drug delivery systems for some poorly soluble drugs. Then, he became a lecturer at the University of Benghazi (Libya), where he worked on nanoemulsion and self-nanoemulsifying drug delivery systems of some biologically active molecules. In 2011, he was awarded the Young Scientist Award from the Association of Pharmacy Professionals (APP). One of his group's research articles was awarded with the most cited paper award from the *European Journal of Pharmaceutics and Biopharmaceutics* in March 2012. Currently, he is working as a professor at the Department of Pharmaceutics, College of Pharmacy, King Saud University. At King Saud University, he developed several nanocarrier-based formulations of various drugs. He also developed a double nanoemulsion for a self-nanoemulsifying drug delivery system of 5-fluorouracil. He has very good expertise in the solubilization of drug molecules using cosolvency models. He developed various analytical methods for the determination of various drugs in a variety of sample matrices. His research interests lie in the general area of pharmaceutics and novel drug delivery systems. He is the author of more than 400 journal articles and several book chapters. He also has a US patent. He is Editor/Editorial Board Member of several journals such as *Pharmaceutics*, *Molecules*, *Separations*, *Current Drug Delivery*, and *Pharmaceutical Sciences*, among others. He has 10601 total citations with an H-index of 48 and an i10 index of 230. In 2020, 2021, and 2022, he was named to the Stanford/Elsevier list of the top 2% scientists in the world for both his career (coveted) and a single year.

## *Editorial* **Editorial: Nanomedicine-Based Drug Delivery Systems: Recent Developments and Future Prospects**

**Faiyaz Shakeel**

Department of Pharmaceutics, College of Pharmacy, King Saud University, Riyadh 11451, Saudi Arabia; fsahmad@ksu.edu.sa

Since the discovery of nanomedicine-based drug delivery carriers such as nanoparticles, liposomes, and self-nanoemulsifying drug delivery systems (SNEDDS), enormous progress has been achieved in the field of innovative active biomolecule drug delivery systems. The use of nanomedicines as drug delivery carriers has received lot of interest in recent years for the therapeutic targeting of specific cells. Biocompatibility, biodegradability, low toxicity, drug delivery efficiency, drug targeting efficiency, and improved solubility, bioavailability, and bioactivities are all advantages of these nanosized drug delivery carriers. Furthermore, these carriers can encapsulate a diverse range of active therapeutic biomolecules. These nanomedicine-based drug delivery carriers can also improve the pharmacokinetic and pharmacodynamic efficiency of active therapeutic biomolecules, allowing for a more sustained, targeted, and controlled drug delivery system. Various studies have recently shown progress in nanomedicine-based drug delivery systems for future therapeutic targeting. The aim of this Special Issue was to collect papers on recent advances, developments, and future prospects in the design, development, characterization, and biological evaluation of nanomedicine-based drug delivery systems of active therapeutic biomolecules.

This Special Issue starts with the paper by Shakeel et al. [1], who developed SNEDDS formulations of a bioactive compound, luteolin, in order to enhance its dissolution rate and hepatoprotective effects. Different SNEDDS formulations of luteolin were developed using an aqueous phase titration method, characterized physicochemically, and evaluated for in vitro drug release and hepatoprotective effects. The findings of this study indicate the potential of SNEDDS for the enhancement of the dissolution rate and hepatoprotective effects of luteolin.

Al-Joufi et al. [2] next enhanced the ocular bioavailability and antibacterial effects of ciprofloxacin using colloidal lipid-based carriers (liposomal drops) for the management of post-surgical infection. The liposomal drops of ciprofloxacin were characterized for various physicochemical parameters and evaluated for in vitro drug release, antibacterial effects, and pharmacokinetic studies. The results showed significant enhancement in the ocular bioavailability of ciprofloxacin using the liposomal drops compared to its commercial formulation.

Novel turmeric rhizome extract nanoparticles were developed and evaluated by Auychaipornlert et al. [3] in the next article. The prepared nanoparticles were characterized well with an optimized experimental design technique. The anticancer potential of nanoparticles was evaluated against the human hepatoma cells, HepG2. The proposed nanoparticles showed significant anticancer effects compared to pure curcumin. The results suggested the potential of nanoparticles of turmeric rhizome extract in the treatment of hepatocellular carcinoma.

Thermosensitive liposomes were developed using the QbD approach by Dobo et al. [4] in the next article. Thermosenistive liposomal formulations were produced using different phospholipids and PEGylated lipids and optimized using the QbD approach. The findings showed that the application of different types and ratios of lipids influences the thermal properties of liposomes.

**Citation:** Shakeel, F. Editorial: Nanomedicine-Based Drug Delivery Systems: Recent Developments and Future Prospects. *Molecules* **2023**, *28*, 4138. https://doi.org/10.3390/ molecules28104138

Received: 10 May 2023 Accepted: 15 May 2023 Published: 17 May 2023

**Copyright:** © 2023 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Kalam et al. [5] developed and evaluated noninvasive chitosan nanoparticles of tedizolid phosphate for the treatment of MRSA-related ocular and orbital infections. The release profile of the studied drug was sustained release from the chitosan nanoparticles compared to its aqueous suspension. The transcorneal flux and antibacterial effects of tedizolid phosphate-loaded chitosan nanoparticles were significant compared to its aqueous suspension. The findings of this work indicate the potential of chitosan nanoparticles in the treatment of MRSA ocular infections and related inflammatory conditions.

Rajput et al. [6] developed a liposome-loaded microneedle array patch of levonorgestrel for contraception. A levonorgestrel-loaded liposomal formulation was obtained using a solvent injection method, characterized for various physicochemical parameters and studied well. The findings of this study showed the better contraceptive effects of the levonorgestrel liposome-loaded microneedle array patch compared to the drug-loaded microneedle array patch.

The SNEDDS formulations of apremilast were developed and evaluated for the treatment of psoriatic arthritis in the next article [7]. Thermodynamically stable SNEDDS formulations were characterized physicochemically and then subjected to in vitro drug release and pharmacokinetics studies. The optimum formulation showed excellent physiochemical parameters and an excellent drug release profile. The significant enhancement in the drug release and bioavailability of apremilast SNEDDS was recorded compared to its suspension. The findings of this study suggest that apremilast SNEDDS is a possible alternative delivery system for apremilast. However, further studies exploring the major factors that influence the encapsulation efficiency and stability of apremilast SNEDDS were suggested.

In another article, the antibacterial and cytotoxic properties of a novel fullerene derivative composed of C<sup>60</sup> fullerenol and standard aminoglycoside antibiotic–gentamicin (C<sup>60</sup> fullerenol–gentamicin conjugate) were evaluated [8]. In vitro assays suggested that the developed C<sup>60</sup> fullerenol–gentamicin conjugate possessed the same antibacterial activity as standard gentamicin against various bacterial strains. The in vitro cytotoxicity assessment indicated that the fullerenol–gentamicin conjugate did not decrease the viability of normal human fibroblasts compared to control fibroblasts. The findings of this study suggested that the developed C<sup>60</sup> fullerenol–gentamicin conjugate could have biomedical potential.

Alshememry et al. [9] evaluated the successful utilization of the positively charged nanocrystals of tedizolid phosphate for topical ocular applications. The developed nanocrystals of tedizolid phosphate showed significant antibacterial activity against *B. subtilis*, *S. pneumonia*, *S. aureus*, and MRSA strains as compared to pure drug. Various pharmacokinetics parameters of nanocrystals were also increased significantly in rabbits compared to the pure drug in the ocular pharmacokinetic study. The nanocrystals of tedizolid phosphate were identified as a promising substitute for the ocular delivery of tedizolid phosphate, with better performance as compared to pure drug.

Two different PAMAM dendrimer generations, G4 and G5 dendrimers, were developed and evaluated in the next article [10]. Developed G4 and G5 dendrimers were characterized well and evaluated for in vitro drug release and cytotoxic effects in human lung adenocarcinoma cells. The findings of this study highlighted the potential anticancer effects of cationic G4 dendrimers as a targeting-sustained-release carrier for erlotinib.

Kenchegowda et al. [11] explore the potential of smart nanocarriers as an emerging platform for cancer therapy. In this exhaustive review, they focus on current advances made through the use of smart nanocarriers such as dendrimers, liposomes, mesoporous silica nanoparticles, quantum dots, micelles, superparamagnetic iron-oxide nanoparticles, gold nanoparticles, and carbon nanotubes. Various topics such as drug targeting, surfacedecorated smart-nanocarriers, and stimuli-responsive cancer nanotherapeutics responding to temperature, enzyme, pH and redox stimuli have been covered in this review.

In the next review article, Kumar et al. [12] explore the potential of natural productbased nanomedicine for maintaining oral health. In their exhaustive review, the potential

of natural products obtained from different sources for the prevention and treatment of dental diseases is discussed and summarized in the form of nanomedicines.

In the next review, the application of the QbD approach is utilized in the robust product development of liposomal formulations [13]. This review discusses and summarizes the current practices that employ QbD in the robust development of liposomal-based nanopharmaceuticals.

In recent years, there has been tremendous research on nanomedicine-based drug delivery systems. This Special Issue has brought together prominent scientists who have explored a diverse applications range of nanomedicine-based drug delivery systems. I believe that further clinical and toxicological studies on both animal and human models are still required to explore the complete potential and commercial exploitation of nanomedicine-based drug delivery systems. The diverse and critical perspectives within this Special Issue provide sufficient information on the development, characterization and evaluation of nanomedicine-based drug delivery systems.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Article* **Hepatoprotective Effects of Bioflavonoid Luteolin Using Self-Nanoemulsifying Drug Delivery System**

**Faiyaz Shakeel 1,\* , Moad M. Alamer <sup>1</sup> , Prawez Alam <sup>2</sup> , Abdullah Alshetaili <sup>3</sup> , Nazrul Haq <sup>1</sup> , Fars K. Alanazi <sup>1</sup> , Sultan Alshehri <sup>4</sup> , Mohammed M. Ghoneim <sup>5</sup> and Ibrahim A. Alsarra <sup>4</sup>**


**Abstract:** Luteolin (LUT) is a natural pharmaceutical compound that is weakly water soluble and has low bioavailability when taken orally. As a result, the goal of this research was to create selfnanoemulsifying drug delivery systems (SNEDDS) for LUT in an attempt to improve its in vitro dissolution and hepatoprotective effects, resulting in increased oral bioavailability. Using the aqueous phase titration approach and the creation of pseudo-ternary phase diagrams with Capryol-PGMC (oil phase), Tween-80 (surfactant), and Transcutol-HP (co-emulsifier), various SNEDDS of LUT were generated. SNEDDS were assessed for droplet size, polydispersity index (PDI), zeta potential (ZP), refractive index (RI), and percent of transmittance (percent T) after undergoing several thermodynamic stability and self-nanoemulsification experiments. When compared to LUT suspension, the developed SNEDDS revealed considerable LUT release from all SNEDDS. Droplet size was 40 nm, PDI was <0.3, ZP was −30.58 mV, RI was 1.40, percent T was >98 percent, and drug release profile was >96 percent in optimized SNEDDS of LUT. For in vivo hepatoprotective testing in rats, optimized SNEDDS was chosen. When compared to LUT suspension, hepatoprotective tests showed that optimized LUT SNEDDS had a substantial hepatoprotective impact. The findings of this investigation suggested that SNEDDS could improve bioflavonoid LUT dissolution rate and therapeutic efficacy.

**Keywords:** bioflavonoid; droplet size; hepatoprotective effects; luteolin; SNEDDS

## **1. Introduction**

The chemical name of luteolin (LUT) is 2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-4*H*chromen-4-one [1,2]. Celery, perilla leaf, chamomile tea, and green pepper all contain this poorly soluble bioactive flavonoid [3,4]. It has antioxidant [5], anti-inflammatory [6–8], antiallergic [6], anti-amnestic [9], hepatoprotective [10], cardioprotective [3], neuroprotective [8,9], and anticancer [11–14] properties. Although it has been shown to be a good bioactive chemical for treating liver problems, due to its limited solubility and bioavailability after oral administration, substantial doses are necessary [10].

Various formulation approaches, including complexation with cyclodextrin [15,16], complexation with phospholipid [10,17], complexation with cyclosophoraoses [18], cocrystal technology [19], palmitoylethanolamide/LUT composite [20], and microparticles [21,22], were investigated to modify its physicochemical properties, which would finally results in enhancement in solubility, dissolution rate, therapeutic activity, and bioavailability. The

**Citation:** Shakeel, F.; Alamer, M.M.; Alam, P.; Alshetaili, A.; Haq, N.; Alanazi, F.K.; Alshehri, S.; Ghoneim, M.M.; Alsarra, I.A. Hepatoprotective Effects of Bioflavonoid Luteolin Using Self-Nanoemulsifying Drug Delivery System. *Molecules* **2021**, *26*, 7497. https://doi.org/10.3390/ molecules26247497

Academic Editor: Jessica Rosenholm

Received: 18 November 2021 Accepted: 8 December 2021 Published: 11 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

solubility of LUT in water was reported to be 1.0 mg/mL at ambient temperature [4,23]. Because LUT has a low aqueous solubility, it has a low in vitro dissolution rate, which means it has a low oral bioavailability [23].

The development of nanocarrier-based drug delivery systems for bioactive compounds/nutraceuticals in order to improve bioavailability and therapeutic efficacy while minimizing side effects has sparked a great deal of attention recently [24–27]. SNEDDS can encapsulate hydrophobic bioactive compounds/nutraceuticals into their internal oil phase, boosting medication solubility, therapeutic efficacy, and bioavailability, and minimizing side effects [26,27]. SNEDDS can generate very tiny nanoemulsions (less than 100 nm in size) when diluted with an aqueous media such as gastrointestinal (GI) fluids or water [28–30]. SNEDDS have been utilized for a long time to increase the solubility, GI permeability, bioavailability, and therapeutic effects of a number of poorly soluble bioactive natural compounds [29–37]. LUT has recently been explored using nanotechnology-based drug carriers such as copolymer micelles [38], solid-lipid nanoparticles [39], zein-based nanoparticles [40], and liposomes [41] to improve its bioavailability and bioactivity in animal models. The antioxidant and anti-inflammatory potential of LUT SNEDDS has also been studied [29]. Despite this, the hepatoprotective effects of LUT when it is encapsulated in SNEDDS have not been studied. As a result, these studies were conducted in order to develop multiple SNEDDS formulations of LUT using pseudo-ternary phase diagrams and low energy emulsification techniques in order to increase its hepatoprotective properties. Capryol-PGMC (oil phase), Tween-80 (surfactant), Transcutol-HP (co-emulsifier), and water (aqueous phase) were used to create different SNEDDS formulations of LUT.

### **2. Results and Discussion**

#### *2.1. Equilibrium Solubility Data of LUT in Different Components*

The major criterion for component screens were the equilibrium solubility data of LUT in different components [42,43]. Table 1 shows the equilibrium solubility values of LUT in various components at 25 ◦C. Among the several oil phases tested, Capryol-PGMC had the highest solubility of LUT (25.72 ± 1.74 mg/g), followed by Capryol-90, Lauroglycol-90, Lauroglycol-FCC, Triacetin, and sesame oil. Among the several surfactants studied, Tween-80 (18.52 ± 0.81 mg/g) had the highest solubility of LUT, followed by Cremophor-EL and Labrasol. Transcutol-HP, on the other hand, had the highest solubility of LUT (68.32 ± 2.83 mg/g), followed by isopropanol (IPA), ethanol, propylene glycol (PG), and ethylene glycol (EG), among the several co-emulsifiers studied. Equilibrium solubility of LUT in water was found to be 0.03 ± 0.00 mg/g. Equilibrium solubility of LUT in water at 37 ◦C was estimated as 50.60 µg/mL by Qing et al. (2017) [38]. The solubility of LUT as mole fraction in water at 25 ◦C was recorded as 1.75 <sup>×</sup> <sup>10</sup>−<sup>6</sup> (converted as 27.80 µg/g) elsewhere [23]. Equilibrium solubility of LUT in water at 25 ◦C was recorded as 30 µg/g in the present work, which was very close to the literature values. The solubility of LUT as a mole fraction in ethanol and IPA at 25 ◦C has been reported as 1.85 <sup>×</sup> <sup>10</sup>−<sup>3</sup> (converted as 11.50 <sup>µ</sup>g/g) and 1.94 <sup>×</sup> <sup>10</sup>−<sup>3</sup> (converted as 9.25 mg/g), respectively, by Peng et al. (2006) [3]. The solubility of LUT as mole fraction in ethanol, IPA, EG, PG, and Transcutol-HP at 25 ◦C has been reported as 1.88 <sup>×</sup> <sup>10</sup>−<sup>3</sup> (converted as 11.70 mg/g), 2.51 <sup>×</sup> <sup>10</sup>−<sup>3</sup> (converted as 12.00 mg/g), 1.30 <sup>×</sup> <sup>10</sup>−<sup>3</sup> (converted as 6.00 mg/g), 2.12 <sup>×</sup> <sup>10</sup>−<sup>3</sup> (converted as 8.00 mg/g), and 3.09 <sup>×</sup> <sup>10</sup>−<sup>2</sup> (converted as 68.00 mg/g), respectively, by Shakeel et al. (2018) [23]. Equilibrium solubility of LUT in ethanol, IPA, EG, PG, and Transcutol-HP was obtained as 11.84 mg/g, 12.13 mg/g, 6.07 mg/g, 8.24 mg/g, and 68.32 mg/g, respectively, in the preset research work, which were also very close to the literature values [3,23]. Based on the equilibrium solubility data of LUT, Capryol-PGMC (oil phase), Tween-80 (surfactant), and Transcutol-HP (co-emulsifier) were selected as the optimum components for the preparation LUT SNEDDS. Water was selected as the aqueous phase due to its "availability, cost effectiveness, and frequent use" in the literature [42–44].


SNEDDS. Water was selected as the aqueous phase due to its "availability, cost effective-

**Components Equilibrium Solubility (mg/g) \*** 


**Table 1.** Equilibrium solubility data of luteolin (LUT) in different excipients at 25 °C.

*Molecules* **2021**, *26*, x FOR PEER REVIEW 3 of 16

ness, and frequent use" in the literature [42–44].

\* Values are presented as mean ± SD (*n* = 3). \* Values are presented as mean ± SD (*n* = 3).

#### *2.2. Construction of Pseudo-Ternary Phase Diagrams for the Preparation of LUT SNEDDS 2.2. Construction of Pseudo-Ternary Phase Diagrams for the Preparation of LUT SNEDDS*

The "low energy emulsification technique" was used to create different SNEDDS formulations of LUT by creating pseudo-ternary phase diagrams with Capryol-PGMC (oil), Tween-80 (surfactant), Transcutol-HP (co-emulsifiers), and water (aqueous phase) [43,44]. Figure 1 depicts phase diagrams for various surfactant to co-emulsifier (Smix) ratios. The "low energy emulsification technique" was used to create different SNEDDS formulations of LUT by creating pseudo-ternary phase diagrams with Capryol-PGMC (oil), Tween-80 (surfactant), Transcutol-HP (co-emulsifiers), and water (aqueous phase) [43,44]. Figure 1 depicts phase diagrams for various surfactant to co-emulsifier (Smix) ratios.

**Figure 1.** Pseudo-ternary phase diagrams for the preparation of the SNEDDS zones of LUT for oil phase (Capryol-PGMC), surfactant (Tween-80), co-emulsifier (Transcutol-HP), and aqueous phase (water) at Smix ratios of (**A**) 1:0, (**B**) 1:2, (**C**) 1:1, (**D**) 2:1, (**E**) 3:1, and (**F**) 4:1. **Figure 1.** Pseudo-ternary phase diagrams for the preparation of the SNEDDS zones of LUT for oil phase (Capryol-PGMC), surfactant (Tween-80), co-emulsifier (Transcutol-HP), and aqueous phase (water) at Smix ratios of (**A**) 1:0, (**B**) 1:2, (**C**) 1:1, (**D**) 2:1, (**E**) 3:1, and (**F**) 4:1.

When compared to the other Smix ratios tested, the 1:0 Smix ratio (Figure 1A) showed the lowest SNEDDS zones. However, when compared to Smix ratio of 1:0, the Smix ratio of 1:2 (Figure 1B) revealed slightly greater SNEDDS zones. In contrast to the other Smix ratios studied, the 1:1 Smix ratio (Figure 1C) produced the highest SNEDDS zones. When the Smix ratio of 2:1 (Figure 1D) was investigated, the SNEDDS zones began to shrink once more. The SNEDDS zones of 2:1 Smix ratio were slightly lower than 1:1 ratio but higher

than the other Smix ratios studied. The SNEDDS zones of Smix ratio of 3:1 (Figure 1E) and 4:1 (Figure 1F) were further decreased compared with Smix ratios of 1:1 and 2:1. Smix ratios of 1:1 were used to indicate the maximal SNEDDS zones (Figure 1C). As a result, different SNEDDS formulations for LUT were chosen from Figure 1C. The whole SNEDDS zones in Figure 1C were taken into consideration for formulation selection, keeping in mind the solubilization of the oil phase (Capryol-PGMC) with respect to Smix. From Figure 1C, varied concentrations of Capryol-PGMC (12, 16, 20, 24, and 28 percent *w/w*) were combined with constant amounts of Tween-80 (20 percent *w/w*) and Transcutol-HP (20 percent *w/w*) to make SNEDDS. Each SNEDDS included 20 mg of LUT, and the resulting formulations were labeled LSN1-LSN5. Table 2 shows the LSN1-LSN5 chemical compositions.


**Table 2.** Composition of self-nanoemulsifying drug delivery system (SNEDDS) prepared using Capryol-PGMC, Tween-80, Transcutol-HP, and deionized water.

#### *2.3. Thermodynamic Stability Tests*

For the elimination of unstable or metastable formulations during the low energy emulsification method, various thermodynamic tests were performed. Centrifugation, heating and cooling cycles, and freeze-pump-thaw cycles were all used in these studies [28,29,35]. Table 3 shows the qualitative findings of these tests. After centrifugation, heating and cooling cycles, and freeze-pump-thaw cycles, all of the SNEDDS formulations were confirmed to be stable. As a result, these formulations were chosen for self-nanoemulsification testing.

**Table 3.** Qualitative results of thermodynamic stability and self-nanoemulsification efficiency of LUT-SNEDDS in the presence of different diluents.


\* All the formulations passed this test with Grade-A in the presence of deionized water, 0.1 N HCl, and phosphate buffer (pH 6.8); X(passed the test); C/F (centrifugation); H/C (heating and cooling); F/T (freeze-pump-thaw).

#### *2.4. Self-Nanoemulsification Tests*

The goal of this experiment was to see if there was any phase separation or precipitation with three different diluents: water, 0.1 N HCl, and phosphate buffer (pH 6.8) [35]. Table 3 displays the qualitative outcomes of this examination. In the presence of all three diluents, the results revealed that all LUT SNEDDS (LSN1-LSN5) passed this test with grade A. Furthermore, there was no evidence of LUT precipitation during a selfnanoemulsification test in the presence of water, 0.1 N HCl, or phosphate buffer (pH 6.8), implying that LUT in the form of SNEDDS was stable under aqueous, stomach, and intestinal pH conditions.

#### *2.5. Physicochemical Characterization of SNEDDS*

In order to ensure the proper formation of LUT SNEDDS in nanosized range, prepared LUT SNEDDS were characterized physicochemically. Table 4 displays the results for these parameters. The droplet size (Z-average) of various LUT SNEDDS (LSN1-LSN5) was recorded as 48.58–124.58 nm using a Malvern Zetasizer. The Z-average value of SNEDDS was found to be enhanced significantly with an increase in the amount of Capryol-PGMC/oil phase (*p* < 0.05). The Z-average value was inversely proportional with the amount of Capryol-PGMC in the formulations. The maximum Z-average value was recorded in formulation LSN5 (124.58 ± 9.41 nm). This result was most likely attributable to the existence of the highest concentration of Capryol-PGMC (28.0 percent *w/w*) in LSN5. The lowest Z-average value (48.58 ± 2.47 nm) was attained in formulation LSN1. LSN1 had the lowest Z-average value, which was most likely owing to the existence of the lowest concentration of Capryol-PGMC (12.0 percent *w/w*). The impact of Smix concentrations was not studied in this work. It has been frequently reported in the literature that the Z-average value of SNEDDS/nanoemulsions is increased with increases in the concentration of the oil phase in the formulation [42,43]. Our findings were consistent with those previously published in the literature.



The polydispersity indices (PDIs) of various LUT SNEDDS (LSN1-LSN5) were obtained in the range of 0.168–0.293. The lower PDI values for all formulations showed droplet homogeneity. The lowest PDI was obtained for formulation LSN1, indicating the most uniform size distribution. However, the formulation LSN5 yielded the highest PDI (0.293).

The zeta potential (ZP) values for various LUT SNEDDS (LSN1-LSN5) range from −23.74 to −30.58 mV. These values were not substantially different amongst SNEDDS (*p* > 0.05). The stability of prepared SNEDDS was shown by ZP values in the magnitude of ±30.0 mV [28,35].

For various LUT SNEDDS (LSN1-LSN5), the refractive indices (RIs) were recorded as 1.344–1.349. The RIs of various SNEDDS were not substantially different (*p* > 0.05). The recorded RIs for all SNEDDS were very near to the RI of water (RI = 1.33), showing that various LUT SNEDDS have a "transparent nature and oil-water (*o/w*) type behavior" [28].

For various LUT SNEDDS (LSN1-LSN5), the turbidity/percent of transmittance (percent T) values were recorded as 94.27–98.94 percent (Table 4). Formulation LSN1 yielded the highest percent T (98.94 ± 0.53 percent). Formulation LSN5, on the other hand, had the lowest percent T (94.27 ± 1.09 percent). The highest and lowest percent T of formulations LSNI and LSN5 were possible due to the lowest and highest droplet size of formulations LSN1 and LSN5, respectively. The "optical clarity and translucent nature" of the prepared SNEDDS was demonstrated by the greater percent T values in all SNEDDS.

The surface texture/morphology and size distribution of an optimized SNEDDS LSN1 were studied using transmission electron microscopy (TEM). Figure 2 shows a TEM picture of the SNEDDS LSN1. The droplets of SNEDDS LSN1 were spherical and scattered within a nanometer range. The presence of Tween-80 and Transcutol-HP was most likely responsible for the droplets' spherical shape.

responsible for the droplets' spherical shape.

**Figure 2.** Transmission electron microscopy (TEM) image of optimized LUT-SNEDDS (LSN1) showing spherical-shaped droplets within nanometer range. **Figure 2.** Transmission electron microscopy (TEM) image of optimized LUT-SNEDDS (LSN1) showing spherical-shaped droplets within nanometer range.

LSNI and LSN5 were possible due to the lowest and highest droplet size of formulations LSN1 and LSN5, respectively. The "optical clarity and translucent nature" of the prepared

LSN1 were studied using transmission electron microscopy (TEM). Figure 2 shows a TEM picture of the SNEDDS LSN1. The droplets of SNEDDS LSN1 were spherical and scattered within a nanometer range. The presence of Tween-80 and Transcutol-HP was most likely

The surface texture/morphology and size distribution of an optimized SNEDDS

SNEDDS was demonstrated by the greater percent T values in all SNEDDS.

## *2.6. In Vitro Drug Release Studies*

*2.6. In Vitro Drug Release Studies*  In vitro drug release tests were performed to assess the release profile of LUT from LUT SNEDDS (LSN1-LSN5) and LUT suspension via "Dialysis Bag" in order to find the best formulation*.* Figure 3 shows the results of LUT release from various SNEDDS (LSN1- LSN5) and LUT suspension. The initial release of LUT from all SNEDDS and LUT suspension was observed as rapid (i.e., immediate drug release profile). When compared to LUT suspension, the rate of LUT release from all SNEDDS (LSN1-LSN5) was greater (*p* < 0.05). For up to 6 h, the immediate release profile of LUT from all SNEDDS and LUT suspension were recorded (Figure 3). After 6 h, all SNEDDS and LUT suspensions showed slow LUT release (i.e., sustained release drug profile). The highest release of LUT was seen in the SNEDDS formulation LSN1 (Figure 3). After 24 h of investigation, the cumulative percent release of LUT from LSN1 was 96.6 percent, compared to 36.8 percent from LUT suspension. Within 6 h of the trial, more than 81 percent of LUT was released from LSN1. In LUT suspension, the minimal drug release profile was observed. The cumulative percent re-In vitro drug release tests were performed to assess the release profile of LUT from LUT SNEDDS (LSN1-LSN5) and LUT suspension via "Dialysis Bag" in order to find the best formulation. Figure 3 shows the results of LUT release from various SNEDDS (LSN1-LSN5) and LUT suspension. The initial release of LUT from all SNEDDS and LUT suspension was observed as rapid (i.e., immediate drug release profile). When compared to LUT suspension, the rate of LUT release from all SNEDDS (LSN1-LSN5) was greater (*p* < 0.05). For up to 6 h, the immediate release profile of LUT from all SNEDDS and LUT suspension were recorded (Figure 3). After 6 h, all SNEDDS and LUT suspensions showed slow LUT release (i.e., sustained release drug profile). The highest release of LUT was seen in the SNEDDS formulation LSN1 (Figure 3). After 24 h of investigation, the cumulative percent release of LUT from LSN1 was 96.6 percent, compared to 36.8 percent from LUT suspension. Within 6 h of the trial, more than 81 percent of LUT was released from LSN1. In LUT suspension, the minimal drug release profile was observed. The cumulative percent release of LUT increased considerably with the reduction in droplet size of the formulation among distinct SNEDDS (LSN1-LSN5) (*p* < 0.05). The highest release profile of LUT from formulation LSN1 was the most likely because of its small droplet size and the presence of the minimum amount of Capryol-PGMC. The presence of the minimum amount of oil phase, i.e., Capryol-PGMC, would result in reduction in droplet size. This results in increased surface area for the release of LUT in the dissolution media [28]. The release profile of LUT from different SNEDDS (LSN1-LSN5) and LUT suspension in two steps (i.e., immediate release profile in first step and sustained release profile in second step) suggested the "diffusion controlled dissolution rate" of LUT [28,43].

gested the "diffusion controlled dissolution rate" of LUT [28,43].

lease of LUT increased considerably with the reduction in droplet size of the formulation among distinct SNEDDS (LSN1-LSN5) (*p* < 0.05). The highest release profile of LUT from formulation LSN1 was the most likely because of its small droplet size and the presence

phase, i.e., Capryol-PGMC, would result in reduction in droplet size. This results in increased surface area for the release of LUT in the dissolution media [28]. The release profile of LUT from different SNEDDS (LSN1-LSN5) and LUT suspension in two steps (i.e., immediate release profile in first step and sustained release profile in second step) sug-

**Figure 3.** In vitro drug release profile of LUT via dialysis bag (mean ± SD, *n* = 3) from various SNEDDS (LSN1-LSN5) and aqueous suspension of LUT. **Figure 3.** In vitro drug release profile of LUT via dialysis bag (mean ± SD, *n* = 3) from various SNEDDS (LSN1-LSN5) and aqueous suspension of LUT.

#### *2.7. Drug Release Kinetics 2.7. Drug Release Kinetics*

Various parameters of release kinetics for LUT SNEDDS (LSN1-LSN5) and LUT aqueous suspension were obtained using various computational models such as zero order, first order, Higuchi, Hixon-Crowell, and the Korsemeyer–Peppas model [45,46]. Table 5 displays the results of this analysis. Based on the values of correlation coefficients (R2) obtained for different models, the release pattern of LUT from formulations LSN1- LSN5 and LUT suspension followed the Korsemeyer–Peppas model because the R2 values were recorded as maximum for this model. The mechanism of drug release was evaluated based the recorded values of diffusion coefficients (n). Based on n values, formulations LSN1-LSN4 followed the Korsemeyer–Peppas model with non-Fickian diffusion mechanism because the value of n was less than 1.0 for these formulations. However, the formulation LSN5 and LUT suspension followed the Korsemeyer–Peppas model with supercase II transport mechanism because the n value was greater than 1.0 for LSN5 and LUT suspension (Table 5). Various parameters of release kinetics for LUT SNEDDS (LSN1-LSN5) and LUT aqueous suspension were obtained using various computational models such as zero order, first order, Higuchi, Hixon-Crowell, and the Korsemeyer–Peppas model [45,46]. Table 5 displays the results of this analysis. Based on the values of correlation coefficients (R<sup>2</sup> ) obtained for different models, the release pattern of LUT from formulations LSN1-LSN5 and LUT suspension followed the Korsemeyer–Peppas model because the R<sup>2</sup> values were recorded as maximum for this model. The mechanism of drug release was evaluated based the recorded values of diffusion coefficients (n). Based on n values, formulations LSN1-LSN4 followed the Korsemeyer–Peppas model with non-Fickian diffusion mechanism because the value of n was less than 1.0 for these formulations. However, the formulation LSN5 and LUT suspension followed the Korsemeyer–Peppas model with supercase II transport mechanism because the n value was greater than 1.0 for LSN5 and LUT suspension (Table 5).

0.5 but n < 1.0, it suggests the non-Ficknian diffusion mechanism. On the other hand, if n > 1.0, it suggests the supercase II transport mechanism [45,46]. The n value in formulations **Table 5.** Correlation coefficients and kinetics of drug release from SNEDDS (LSN1-LSN) and LUT suspension.

If the value of n = 0.5, this suggests the Fickian diffusion mechanism. However, if n >


Correlation coefficient (R<sup>2</sup> ), Zero order rate constant (K0), first order rate constant (k1), diffusion coefficient (n).

If the value of n = 0.5, this suggests the Fickian diffusion mechanism. However, if n > 0.5 but n < 1.0, it suggests the non-Ficknian diffusion mechanism. On the other hand, if n > 1.0, it suggests the supercase II transport mechanism [45,46]. The n value in formulations LSN1–LSN4 was obtained as 0.978–0.972, suggesting a non-Ficknian diffusion mechanism. However, the n value for formulation LSN5 and LUT suspension was obtained as 1.201 and 1.368, respectively, suggesting the supercase II transport mechanism. In the literature, many hydrophobic compounds followed the Korsemeyer–Peppas model with non-Fickian diffusion mechanism from SNEDDS or oral nanoemulsions [43,47,48]. Therefore, the results of drug release kinetics of LUT from most of the studied SNEDDS were in good agreement with those reported in the literature.

#### *2.8. Hepatoprotective Effects*

Formulation LSN1 was chosen as an optimized SNEDDS of LUT for further hepatoprotective study based on maximum drug release (96.6 percent), minimum Z-average value (48.58 nm), and the presence of a minimum concentration of Capryol-PGMC. Table 6 displays the results of hepatoprotective evaluation. The animals in the control group (Group I) had normal levels of serum alanine aminotransferase (ALT), serum aspartate aminotransferase (AST), γ-glutamyl transpeptidase (γ-GGT), serum bilirubin, and serum alkaline phosphatase (ALP), among other group biochemical markers. Oral CCl<sup>4</sup> dosing (Group II animals) resulted in a substantial increase in AST, ALT, ALP, γ-GGT, and bilirubin (*p* < 0.01). Standard (i.e., silymarin; Group III), LUT suspension (Group IV), and optimized LUT-SNEDDS LSN1 (Group V) oral delivery resulted in substantial reductions in AST, ALT, ALP, γ-GGT, and bilirubin levels compared to the toxic control (Group II animals) (*p* < 0.01).

**Table 6.** Influence of LUT-SNEDDS (LSN1) and LUT suspension administrations on different biomarkers of rat serum.


Table 7 shows the findings of the hepatoprotective evaluation of liver tissue. It was discovered that the animals in the control group had normal levels of catalase (CAT), glutathione peroxidase (GSH), superoxide dismutase (SOD), and melondialdehyde (MDA) in their tissues. GSH levels in the control group were assessed to be 1.19 ± 0.03 nmol/mg. In Group II rats, however, oral treatment with CCl<sup>4</sup> lowered the GSH value to 0.45 ± 0.01 nmol/mg. Furthermore, when compared to the control group, silymarin suspension, LUT suspension, and SNEDDS LSN1 treatment decreased GSH levels (Table 7). The concentrations of CAT, MDA, and SOD in liver tissues were calculated as 47.71 ± 1.29 U/mg, 3.25 ± 0.07 nmol/mg, and 24.86 ± 1.34 U/mg, respectively, in the control. In Group II rats, oral treatment with CCl<sup>4</sup> resulted in substantial reductions in CAT, MDA, and SOD levels in liver tissues (*p* < 0.05). When compared to the control, oral treatment of silymarin suspension, LUT suspension, and SNEDDS LSN1 lowered CAT, MDA, and SOD levels (Table 7). In comparison to control, the hepatoprotective efficacies of silymarin, LUT suspension, and SNEDDS LSN1 were likewise significant (*p* < 0.05). Table 6 summarizes the typical serum values of AST, ALT, ALP, γ-GGT, and bilirubin in rats [49]. Silymarin suspension, LUT suspension, and SNEDDS LSN1 have all been found to be effective in lowering AST, ALT, and ALP levels. However, silymarin was found to be the most effective in reducing the levels of AST, ALT, and ALP. In comparison to the control, oral delivery with CCl<sup>4</sup> resulted in a significant increase in γ-GGT and bilirubin levels. However, the

oral treatment with silymarin, LUT, and SNEDDS LSN1 resulted in a marked reduction in γ-GGT and bilirubin levels compared with Group II rats.


**Table 7.** Influence of LUT-SNEDDS (LSN1) and LUT administrations on different biomarkers ofrat liver.

The liver contains many forms of transaminases, such as serum AST, ALT, and ALP, and their levels in the blood have been seen to increase in individuals with liver diseases [35]. In the examination of hepatocellular injury, serum AST, ALT, and ALP have been identified as specific indicators. The most common method for assessing hepatocellular damage is to measure serum bilirubin and ALP levels [50,51]. The levels of bilirubin, AST, ALT, and ALP in the proposed study were considerably higher than in the control group. Oral dosing of silymarin, LUT, and SNEDSS LSN1 resulted in significant hepatoprotective effects. Table 7 lists the typical amounts of CAT, GSH, MDA, and SOD found in rat liver tissues [49]. GSH was found to be vital in the liver's cellular activity [35]. It also detoxifies organic chemicals to control gene expression, apoptosis, and cellular transport. Free radicals and other reactive oxygen species are efficiently scavenged by enzymes. GSH is critical for the regular functioning of cells and tissues. Hepatic damage was produced by its significant depletion [35,52–54]. In the suggested investigation, GSH depletion was seen after CCl<sup>4</sup> delivery compared to a control group. Oral treatment of silymarin, LUT, and SNEDDS LSN1 restored hepatotoxicity by raising liver GSH levels. It is well known that CCl4's free radical causes peroxidative breakdown, resulting in MDA formation and membrane damage. MDA levels in the liver are connected to lipid peroxidation, which causes tissue damage and the failure of antioxidant defense mechanisms [55–57]. In the planned investigation, MDA levels were drastically lowered after oral treatment with silymarin, LUT, and SNEDDS LSN1. In comparison to hazardous CCl4, silymarin, LUT, and SNEDDS LSN1 dramatically boosted CAT and SOD levels following oral treatment. Hepatoprotective effects are aided by elevated SOD and CAT levels. Overall, the findings of this investigation indicated that the developed SNEDDS LSN1 had stronger hepatoprotective effects compared to the toxic control (Group II animals) against CCl4-induced liver injury.

#### **3. Materials and Methods**

#### *3.1. Materials*

LUT was obtained from Beijing Mesochem Technology Pvt. Ltd. (Beijing, China). Lauroglycol-90, Lauroglycol-FCC, Capryol-90, Capryol-PGMC, Labrasol, and Transcutol-HP were obtained from Gattefosse (Lyon, France). Triacetin was obtained from Alpha Chemica (Mumbai, India). Chromatography grade acetonitrile, Tween-80, and sesame oil were obtained from Sigma-Aldrich (St. Louis, MO, USA). Cremophor-EL was obtained from BASF (Cheshire, UK). EG, PG, ethanol, and IPA were obtained from E-Merck (Darmstadt, Germany). Chromatography grade water was collected from Milli-Q water purification. All other chemicals and reagents used were of analytical/pharmaceutical grades.

#### *3.2. HPLC Method for LUT Estimation*

The estimation of LUT in all studied samples was carried out using a validated HPLC method [23]. The chromatographic identification of LUT was achieved at room temperature (25 ± 1 ◦C) using a Waters HPLC system (Waters, USA) attached to a 1515 isocratic HPLC pump, 717 plus Autosampler, quaternary LC-10A VP pumps, and a programmable

2487 dual λ absorbance UV-visible detector. The separation of LUT was carried out using a Nucleodur (150 mm × 4.6 mm) RP C<sup>8</sup> column filled with 5 µm filler as a static phase. The binary mixture of 0.1 % formic acid and acetonitrile (7:3 % *v/v*) was used as the mobile phase. The mobile phase was flowed with a flow rate of 1.0 mL/min and detection was performed at 348 nm. The injection volume for analytes was set at 20 µL. Millennium (version 32) software was used for the data analysis.

#### *3.3. Solubility Study of LUT in Different Components*

Based on LUT's equilibrium solubility data, various components were chosen. Solubility studies were performed in order to select suitable components for the preparation of LUT SNEDDS instead of optimizing LUT dose. Using a saturation shake flask method, the equilibrium solubility of LUT in various oils (Lauroglycol-90, Lauroglycol-FCC, Capryol-90, Capryol-PGMC, Triacetin, and sesame oil), surfactants (Labrasol, Tween-80, and Cremophor-EL), co-emulsifiers (Transcutol-HP, EG, PG, ethanol, and IPA), and water was estimated [58]. These experiments were carried out in triplicate with an excess amount of LUT mixed in with known proportions of various components. After being vortexed for approximately 5 min, the samples were then transferred to a WiseBath® WSB Shaking Water Bath (Model WSB-18/30/-45, Daihan Scientific Co. Ltd., Seoul, Korea) for continuous shaking. These tests were carried out at a temperature of 25 ± 0.5 ◦C; 100 rpm and 72 h were chosen as the speed and equilibrium time, respectively. The samples were removed from the shaker and centrifuged at 5000× *g* rpm once equilibrium was reached. The supernatants were carefully removed, diluted (as needed) with the mobile phase, and utilized to analyze LUT using the HPLC technique at 348 nm [23].

#### *3.4. Construction of Pseudo-Ternary Phase Diagrams for Preparation of SNEDDS*

Capryol-PGMC (oil), Tween-80 (surfactant), and Transcutol-HP (co-emulsifier) were chosen for the manufacture of LUT SNEDDS based on the equilibrium solubility data of LUT in various oils, surfactants, and co-emulsifiers. Because water is frequently utilized in the literature [42–44], it was chosen as the aqueous phase. Low-energy emulsification was used to create phase diagrams [35,43]. As a result, the surfactant (Tween-80) and co-emulsifier (Transcutol-HP) were properly blended in different mass ratios, such as 1:0, 1:2, 1:1, 2:1, 3:1, and 4:1 mass ratios. The oil phase, i.e., Capryol-PGMC, was combined with a specified Smix ranging from 1:9 to 9:1. By slowly adding water, the oil phase and certain Smix mixes were titrated, and visual observations were made based on their transparency/clarity. Upon each addition of water, the physical appearance was recorded. The formulations showing clear/transparent and easily flowable behavior were selected, and other formulations such as turbid emulsions, turbid gels, and translucent gels were discarded based on visual observations. At this stage, visual observations were made. However, selected formulations were fully characterized for thermodynamic stability, selfnanoemulsification efficiency, and various physicochemical parameters, which are detailed in the next sections. For each Smix ratio, the SNEDDS zones were built individually on pseudo-ternary phase diagrams [28,44], where one axis representing oil phase, second representing aqueous phase, and third representing a specific Smix ratio (Figure 1).

The highest SNEDDS zones were represented by a 1:1 mass ratio of Tween-80 and Transcutol-HP, according to phase diagrams. As a result, utilizing a 1:1 Smix ratio, multiple SNEDDS formulations for LUT were chosen. Different SNEDDS with assigned codes of LSN1–LSN5 were accurately picked from a 1:1 Smix ratio. In the phase diagram, the complete region of SNEDDS zones was considered. Different concentrations of Capryol-PGMC (12, 16, 20, 24, and 28 percent *w/w*) with consistent amounts of Tween-80 (20 percent *w/w*) and Transcutol-HP (20 percent *w/w*) were selected from the phase diagram for the manufacture of LUT SNEDDS. In each SNEDDS, 20 mg of LUT was included.

#### *3.5. Thermodynamic Stability Tests*

LUT SNEDDS (LSN1-LSN5) underwent various thermodynamic stability tests in order to eliminate metastable/unstable SNEDDS. Centrifugation, heating and cooling cycles, and freeze-pump thaw cycles were used in these tests [28,35]. For approximately 30 min, LUT SNEDDS were centrifuged at 5000× *g* rpm and observed for any physical changes. SNEDDS that were centrifugally stable were exposed to further heating and cooling cycles. Four heating and cooling cycles were carried out at temperatures ranging from 4 to 45 ◦C for a total of 48 h at each temperature. Freeze-pump-thaw cycles were used on SNEDDS that were stable during the heating and cooling cycles. Four freeze-pump-thaw cycles were carried out at temperatures ranging from −21◦ to 25 ◦C for a total of 24 h at each temperature. Finally, those SNEDDS that passed all three steps of thermodynamic stability tests were chosen for future study.

#### *3.6. Self-Nanoemulsification Test*

Using an A–E grading systems, a self-nanoemulsification test was performed to investigate any drug precipitation or phase separation after dilution with various diluents [28,35]. For this test, three different diluents were used: deionized water, 0.1N HCl, and phosphate buffer (pH 6.8). Each SNEDDS (LSN1-LSN5) had its self-nanoemulsification efficiency assessed visually using the A–E grading systems, as reported previously [28,35]. Only those SNEDDS who received an A or B on this test were chosen for further evaluation.

#### *3.7. Physicochemical Characterization of SNEDDS*

Droplet size distribution, PDI, ZP, surface morphology, RI, and percent T were all measured in LUT SNEDDS (LSN1-LSN5). A Malvern Zetasizer (Nano ZS90, Malvern Instruments Ltd., Holtsville, NY, USA) was used to determine the droplet size and PDI of the LUT SNEDDS. The experiments were conducted at a temperature of 25 ◦C and a scattering angle of 90◦ . Approximately 1 mL of each LUT SNEDDS was diluted with water (1:200) and 3 mL of each diluted SNEDDS was transferred to an acrylic plastic cuvette to determine droplet size and PDI. A Malvern Zetasizer (Nano ZS, Malvern Instruments Ltd., Holtsville, NY, USA) was used to determine the ZP of each LUT SNEDDS. The process for determining ZP was the same as that for determining droplet size and PDI, with the exception that samples were taken into glass electrodes for ZP measurement.

An Abbes type Refractometer (Precision Standard Testing Equipment Corporation, Darmstadt, Germany) was used to determine the RI of each LUT SNEDDS. Castor oil was utilized as the standard, and RI measurements were performed on undiluted samples. The turbidity/percent T of each LUT SNEDDS was determined using a UV-Visible Spectrophotometer (SP1900, Axiom, Germany) at 550 nm according to the literature [28].

Transmission Electron Microscopy (TEM) was used to investigate the surface morphology and form of an optimized SNEDDS (LSN1). JEOL TEM (JEOL JEM 2100 F, USA) was used to conduct TEM on LSN1. Central Laboratory, Research Center, College of Science, King Saud University, Riyadh, Saudi Arabia conducted the TEM evaluation. Optimized SNEDDS LSN1 was diluted as stated for droplet size and PDI assessment. On a carboncoated grid, a drop of diluted SNEDDS LSN1 was placed and left to dry. The experiment was carried out using a TEM at 80 KV.

#### *3.8. In Vitro Drug Release Evaluation*

Using a dialysis bag (MWCO: 12-14 KDa; Spectrum Medical Industries, Mumbai, India), in vitro drug release studies of LUT from five distinct SNEDDS (LSN1-LSN5) were carried out and compared to a control (LUT suspension) [28]. These tests were carried out using 200 mL of phosphate buffer (pH 6.8) as a dissolution medium. The dissolution media was filled in suitable glass beakers. Approximately 1.0 mL of each formulation of LUT and LUT suspension (each formulation containing 20 mg of LUT) were transferred to a dialysis bag, which was clamped using plastic clips. The dialysis bags clamped with plastic clips were immersed into glass beakers containing 200 mL of dissolution

media. The whole assembly was placed into a WiseBath® WSB Shaking Water Bath (Model WSB-18/30/-45, Daihan Scientific Co. Ltd., Seoul, Korea) at 37 ± 1.0 ◦C for shaking at 100 rpm. At different time periods, 1.0 mL of samples from each formulation were carefully withdrawn and replaced with the same volume of freshly made LUT free dissolution media. The concentration of LUT in each SNEDDS and LUT suspension was measured using the HPLC technique at 348 nm at each time interval [23].

#### *3.9. Hepatoprotective Effects*

Thirty male Wistar Albino rats weighing 200–250 mg/kg were donated by the Experimental Animal Care Center (EACC) at Prince Sattam bin Abdulaziz University, Al-Kharj, Saudi Arabia. Before the experiment began, all the animals were acclimatized and kept in plastic cages in typical laboratory settings for animal care and storage, and they were fed a regular pellet diet with water ad libitum. The EACC, Prince Sattam bin Abdulaziz University, Al-Kharj, Saudi Arabia provided the recommendations for all experimental protocols and procedures. The Animal Ethics Committee of the EACC Board (Prince Sattam bin Abdulaziz University, Al-Kharj, Saudi Arabia with approval number: BERC-017-10-21) gave its approval to these investigations. Animals were used, and experimental techniques followed the European Union (EU) directive 2010/63/EU.

Because CCl<sup>4</sup> has been identified as a suitable toxicant for hepatotoxicity [59], it was used to induce hepatotoxicity. The rats were put into five groups at random, with six rats in each group. For 5 days, Group I animals were given a daily dose of 1 mL aqueous solution of 0.5 percent w/w carboxymethyl cellulose (CMC) (p.o.) as a control. As a toxic control, Group II animals were given a daily dosage of aqueous solution containing 0.5 percent *w/w* CMC (p.o.) and a single dose of CCl<sup>4</sup> (1 mg/kg, i.p.) on day 1. The standard group consisted of animals that were given an oral suspension of standard silymarin (10 mg/kg) on all 5 days and CCl<sup>4</sup> (1 mg/kg, i.p.) on days 2 and 3 following 1 h of silymarin administration. The test LUT group consisted of animals that were given LUT suspension (20 mg/kg) for all 5 days and CCl<sup>4</sup> (1 mg/kg, i.p.) on days 2 and 3 following 1 h of LUT suspension administration. The test LUT SNEDDS LSN1 group consisted of rats that were given optimized LUT SNEDDS LSN1 (containing 20 mg/kg of LUT) on all 5 days and CCl<sup>4</sup> (1 mg/kg, i.p.) on days 2 and 3 after 1 h of receiving LUT SNEDDS LSN1. All animals were starved overnight on day 6, and 1.5–2.0 mL of blood was collected in sterile Eppendorf tubes from the tail vein and maintained at 37 ◦C for around 45 min. After centrifugation at 3000 rpm for 15 min, the serum was separated using a sterile micropipette. In order to assess liver function, serum samples were subjected to biochemical analysis. Various biochemical markers, such as serum AST, serum ALT, serum ALP, serum γ-GGT, and serum bilirubin were calculated following protocols described in the literature [60–62].

#### *3.10. Estimation of Biomarkers of Liver Tissue*

Fresh livers from rats were obtained and weighed precisely. In 1.5 M KCl, liver homogenates were produced at 10 percent *w/v*. In liver homogenates, biochemical markers such as CAT, GSH, SOD, and MDA were measured, as reported in the literature [63,64].

#### *3.11. Statistical Evaluation*

In vitro experiment results are presented as the mean ± SD of three independent experiments. The results of hepatoprotective studies, on the other hand, are given as the mean ± SD of six independent experiments. All the results were statistically assessed by one way analysis of variance (ANOVA) followed by Dennett's test using GraphpadInstat software (San Diego, CA, USA). The significant value was defined as the *p* value at the 5% level of significance (*p* < 0.05).

#### **4. Conclusions**

Hepatoprotective effects of a weakly soluble bioactive flavonoid LUT were evaluated using SNEDDS formulations. Different SNEDDS formulations of LUT were created using

a low energy emulsification approach, characterized physicochemically, and tested for in vitro drug release utilizing a dialysis bag. The hepatoprotective effects of SNEDDS formulation LSN1 in rats were explored further based on the minimal Z-average value, maximum drug release profile, and the existence of a minimum concentration of Capryol-PGMC. When compared to LUT suspension and other SNEDDS examined, optimized SNEDDS LSN1 revealed a considerable in vitro drug release profile of LUT. Furthermore, as compared to LUT suspension, the hepatoprotective effects of optimized SNEDDS LSN1 were found to be considerable. Overall, the findings of this study revealed that SNEDDS has the potential to improve LUT's in vitro dissolution rate and hepatoprotective effects.

**Author Contributions:** Conceptualization, F.S. and I.A.A.; methodology, M.M.A., P.A., N.H. and F.S.; software, M.M.G.; validation, P.A., I.A.A. and S.A.; formal analysis, A.A.; investigation, A.A., S.A. and F.K.A.; resources, F.K.A.; data curation, M.M.G.; writing—original draft preparation, F.S.; writing—review and editing, I.A.A., P.A., F.K.A. and S.A.; visualization, F.K.A.; supervision, I.A.A. and F.S.; project administration, I.A.A.; funding acquisition, F.K.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by King Saud University (KSU), Vice Deanship of Research Chairs, Kayyali Chair for Pharmaceutical Industries (Grant Number FS-2021). The APC was funded by KSU.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Animal Ethics Committee of the EACC Board at Prince Sattam bin Abdulaziz University, Al-Kharj, Saudi Arabia (approval number: BERC-017-10-21).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** This study did not report any data.

**Acknowledgments:** Authors are thankful to King Saud University, Vice Deanship of Research Chairs, Kayyali Chair for Pharmaceutical Industries for supporting this project through Grant Number FS-2021.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples of the LUT compounds are available from the authors.

#### **References**


## *Article* **Enhancing Ocular Bioavailability of Ciprofloxacin Using Colloidal Lipid-Based Carrier for the Management of Post-Surgical Infection**

**Fakhria A. Al-Joufi <sup>1</sup> , Mounir M. Salem-Bekhit 2,3,\*, Ehab I. Taha <sup>2</sup> , Mohamed A. Ibrahim <sup>2</sup> , Magdy M. Muharram 4,5, Sultan Alshehri <sup>2</sup> , Mohammed M. Ghoneim <sup>6</sup> and Faiyaz Shakeel <sup>2</sup>**


**Abstract:** Conjunctivitis and endogenous bacterial endophthalmitis mostly occurred after ophthalmic surgery. Therefore, the present study aimed to maximize the ocular delivery of ciprofloxacin (CPX) using colloidal lipid-based carrier to control the post-surgical infection. In this study, CPX was formulated as ophthalmic liposomal drops. Two different phospholipids in different ratios were utilized, including phosphatidylcholine (PC) and dimyrestoyl phosphatidylcholine (DMPC). The physiochemical properties of the prepared ophthalmic liposomes were evaluated in terms of particle size, entrapment efficiency, polydispersity index, zeta potential, and cumulative CPX in-vitro release. In addition, the effect of sonication time on particle size and entrapment efficiency of CPX ophthalmic drops was also evaluated. The results revealed that most of the prepared formulations showed particle size in nanometer size range (460–1047 nm) and entrapment efficiency ranging from 36.4–44.7%. The antibacterial activity and minimum inhibitory concentration (MIC) were investigated. Ex vivo antimicrobial effect of promising formulations was carried out against the most common causes of endophthalmitis microorganisms. The pharmacokinetics of the prepared ophthalmic drops were tested in rabbit aqueous humor and compared with commercial CPX ophthalmic drops (Ciloxan®). Observed bacterial suppression was detected in rabbit's eyes conjunctivitis with an optimized formulation A3 compared with the commercial ophthalmic drops. CPX concentration in the aqueous humor was above MIC against tested bacterial strains. The in vivo data revealed that the tested CPX drops showed superiority over the commercial ones with respect to peak aqueous humor concentration, time to reach peak aqueous humor concentration, elimination rate constant, half-life, and relative bioavailability. Based on these results, it was concluded that the prepared ophthalmic formulations significantly enhanced CPX bioavailability compared with the commercial one.

**Keywords:** bioavailability; ciprofloxacin; lipid-based colloidal carriers; ocular delivery

## **1. Introduction**

Ciprofloxacin (CPX) is one of the most effective antibiotics, active against the vast range of infection causing ophthalmic pathogens. It blocks the bacterial deoxyribonucleic acid (DNA) synthesis via inhibition of DNA gyrase [1,2]. CPX lipophilicity is high enough to permeate via ocular humors and it is the preferred therapy for intraocular infections [3].

**Citation:** Al-Joufi, F.A.; Salem-Bekhit, M.M.; Taha, E.I.; Ibrahim, M.A.; Muharram, M.M.; Alshehri, S.; Ghoneim, M.M.; Shakeel, F. Enhancing Ocular Bioavailability of Ciprofloxacin Using Colloidal Lipid-Based Carrier for the Management of Post-Surgical Infection. *Molecules* **2022**, *27*, 733. https://doi.org/10.3390/ molecules27030733

Academic Editor: Jessica Rosenholm

Received: 24 December 2021 Accepted: 22 January 2022 Published: 23 January 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

When CPX is dispensed locally into the conjunctival sac at specific concentration, it infiltrates into the aqueous humor and its concentration is reliant on the dose number [4]. Therefore, it has been verified to be an effective local antimicrobial for the application as a sole drug for conjunctivitis and keratitis treatment [5]. CPX has also been proposed for prophylactic use in cases of endogenous bacterial endophthalmitis. Different studies reported the most common causes of endophthalmitis include *Staphylococcus aureus*, *Bacillus* spp., *Escherichia coli*, *Pseudomonas* spp., and *Klebsiella* spp. [6]. CPX interacts with infectious bacteria at the site of infection, not in the blood stream. Therefore, in order to prevent bacterial growth, CPX must be delivered at a high concentration to the infected area [7].

For marketed CPX ophthalmic drops to be effective, frequent dosing to eye sac must be applied because all marketed CPX ophthalmic drops utilize aqueous solutions [8,9]. The solubility of CPX is optimum in acidic pH (around 4.5); however, the pH of the tears is in the neutral range (almost 7) [10]. This way of administering the commercial CPX ophthalmic drops is always accompanied by burning sensation and itching [11]. In addition, because of the negative charge in the corneal surface (due to presence of a thin layer of negatively charged mucin), a positively charged carrier would result in increasing the drug resident time in the eye [12]. Stearylamine (STA) is a well-known positive charge-inducing agent. Thus, it was utilized in the formulations to increase the contact time, prolong the drug release, and improve CPX bioavailability.

Liposomes are vesicular systems (usually within the size range of 10 nm to 1 µm or greater) composed of an aqueous core enclosed by phospholipid bilayers of natural or synthetic origin. Liposomes are advantageous in encapsulating both lipophilic and hydrophilic molecules. Hydrophilic drugs are entrapped in the aqueous layer, while hydrophobic drugs are stuck in the lipid bilayers [13]. Liposomes have been considered for ocular administration because they pose ophthalmic drug delivery advantages. They are biocompatible and biodegradable nanocarriers and can enhance the permeation of poorly absorbed drug molecules by binding to the corneal surface and prolonging residence time [14].

This study aims to design CPX loaded colloidal lipid-based carriers (ophthalmic liposomal drops) to enhance its ocular bioavailability in order to reduce and control surgical site infections after ophthalmic operation. The pharmacokinetics parameters such as peak aqueous humor concentration (Cmax), time to reach peak aqueous humor concentration (Tmax), half-life (t1/2), and area under curve from time zero to t (AUC0–t) were used to evaluate the prepared CPX drops in comparison with the marketed one in the rabbit model. In order to achieve this goal, several vesicular formulations utilizing phospholipids such as phosphatidylcholine (PC) and dimyrestoyl phosphatidylcholine (DMPC) were used. The prepared CPX ophthalmic drops were evaluated in terms of particle size, entrapment efficiency, polydispersity index, zeta potential, and in vitro CPX release rate.

#### **2. Results and Discussion**

#### *2.1. Polydispersity Index (PDI) and Zeta Potential of CPX Colloidal Lipid-Based Formulations (CPX-CLBFs)*

The PDI and zeta potential values of different CPX-CLBFs formulations were determined as described in the experimental section. The PDIs of different formulations were found to be 0.128–0.767. The minimum PDI was recorded for formulation B3 (0.128), indicating the maximum uniformity in particle size distribution compared to other liposomal formulations evaluated. The maximum PDI was recorded for formulation A0 (0.767), indicating broad size distribution in formulation A0. The zeta potential of different formulations was obtained as 3.6–25.7 mV. The minimum and maximum zeta potential values were recorded for formulations A0 (3.6 mV) and A3 (25.7 mV), respectively. The zeta potential value in the range of ±30 mV indicated the maximum stability of formulations [15]. The zeta potential of liposomal formulation A3 was found to be much closed with ±30 mV, indicating the maximum stability of liposomal formulation A3 compared to the other formulations studied.

#### *2.2. Influence of Phospholipids Type on Ophthalmic CPX-CLBFs 2.2. Influence of Phospholipids Type on Ophthalmic CPX-CLBFs* Figures 1 and 2 show the effect of phospholipids type on particle size and the percent

formulations studied.

*Molecules* **2022**, *27*, x FOR PEER REVIEW 3 of 13

Figures 1 and 2 show the effect of phospholipids type on particle size and the percent of entrapment efficiency (EE%) of ophthalmic CPX-CLBFs. It was obvious that the presence of DMPC phospholipid (formulations B0 and B1–B3) in large amounts produced larger particle sizes compared with the incorporation of PC in the formulations (A0 and A1– A3). Among the different PC formulations studied (A0 and A1–A3), the particle size of formulation A3 was smallest compared with formulations A0, A1, and A2. These observations indicated that the amount of PC had a definite impact on particle size of PC formulations. The particle size of PC formulations was found to be decreased with increase in the concentration of PC in the formulations (Figure 1). Among different DMPC formulations studied (B0 and B1–B3), the particle size of formulation B0 was smallest compared with formulations B1–B3. These observations also indicated that the amount of DMPC had a definite impact on the particle size of DMPC formulations. The particle size of DMPC formulations was also found to be decreased with increase in the concentration of DMPC in the formulations (Figure 1). On the other hand, phospholipids type had little effect on EE% of the prepared ophthalmic CPX-CLBFs, which could be due to the 36 carbon atoms for the hydrophobic part of PC molecules compared with 28 carbon atoms for hydrophobic part of DMPC. Among different PC formulations studied (A0 and A1–A3), the EE% of formulation A0 was highest compared with formulations A1–A3. However, the EE% of different PC formulations was not significantly different. These observations indicated that the amount of PC had little impact on EE% of PC formulations (Figure 2). Among different DMPC formulations studied (B0 and B1–B3), the EE% of formulation B3 was highest compared with other DMPC formulations studied. However, the EE% of different DMPC formulations was not significantly different. These observations also indicated that the amount of DMPC had little impact on EE% of DMPC formulations (Figure 2). Gulati et al. observed that drug entrapment was increased by increasing the carbon chain length of the phospholipids used [16]. Our results were in good agreement with those reported by Gulati et al. [16]. of entrapment efficiency (EE%) of ophthalmic CPX-CLBFs. It was obvious that the presence of DMPC phospholipid (formulations B0 and B1–B3) in large amounts produced larger particle sizes compared with the incorporation of PC in the formulations (A0 and A1–A3). Among the different PC formulations studied (A0 and A1–A3), the particle size of formulation A3 was smallest compared with formulations A0, A1, and A2. These observations indicated that the amount of PC had a definite impact on particle size of PC formulations. The particle size of PC formulations was found to be decreased with increase in the concentration of PC in the formulations (Figure 1). Among different DMPC formulations studied (B0 and B1–B3), the particle size of formulation B0 was smallest compared with formulations B1–B3. These observations also indicated that the amount of DMPC had a definite impact on the particle size of DMPC formulations. The particle size of DMPC formulations was also found to be decreased with increase in the concentration of DMPC in the formulations (Figure 1). On the other hand, phospholipids type had little effect on EE% of the prepared ophthalmic CPX-CLBFs, which could be due to the 36 carbon atoms for the hydrophobic part of PC molecules compared with 28 carbon atoms for hydrophobic part of DMPC. Among different PC formulations studied (A0 and A1–A3), the EE% of formulation A0 was highest compared with formulations A1–A3. However, the EE% of different PC formulations was not significantly different. These observations indicated that the amount of PC had little impact on EE% of PC formulations (Figure 2). Among different DMPC formulations studied (B0 and B1–B3), the EE% of formulation B3 was highest compared with other DMPC formulations studied. However, the EE% of different DMPC formulations was not significantly different. These observations also indicated that the amount of DMPC had little impact on EE% of DMPC formulations (Figure 2). Gulati et al. observed that drug entrapment was increased by increasing the carbon chain length of the phospholipids used [16]. Our results were in good agreement with those reported by Gulati et al. [16].

were recorded for formulations A0 (3.6 mV) and A3 (25.7 mV), respectively. The zeta potential value in the range of ±30 mV indicated the maximum stability of formulations [15]. The zeta potential of liposomal formulation A3 was found to be much closed with ±30 mV, indicating the maximum stability of liposomal formulation A3 compared to the other

**Figure 1.** Effect of sonication time on particle size of ophthalmic ciprofloxacin (CPX) colloidal lipidbased formulations (CPX-CLBFs) (mean ± SD, *n* = 3.0). **Figure 1.** Effect of sonication time on particle size of ophthalmic ciprofloxacin (CPX) colloidal lipid-based formulations (CPX-CLBFs) (mean ± SD, *n* = 3.0).

**Figure 2.** Effect of sonication time on percent entrapment efficiency (EE%) of ophthalmic CPX-CLBFs (mean ± SD, *n* = 3.0). **Figure 2.** Effect of sonication time on percent entrapment efficiency (EE%) of ophthalmic CPX-CLBFs (mean ± SD, *n* = 3.0). CLBFs (mean ± SD, *n* = 3.0). *2.3. Impact of Sonication on Ophthalmic CPX-CLBFs*

#### *2.3. Impact of Sonication on Ophthalmic CPX-CLBFs* Figures 1 and 2 show the impact of sonication time on ophthalmic CPX-CLBFs parti-

*Molecules* **2022**, *27*, x FOR PEER REVIEW 4 of 13

*2.3. Impact of Sonication on Ophthalmic CPX-CLBFs* Figures 1 and 2 show the impact of sonication time on ophthalmic CPX-CLBFs particle size and EE%. It was clear that 2 min of sonication time decreases the particle size of all formulations. In addition, it was observed that CPX-CLBFs containing cholesterol (CL) are not able to increase EE% due to high membrane rigidity of CL. As a result, the values of EE% for formulations containing CL (A1–A3 and B1–B3) was decreased after sonication for 2 min (Figure 2) due to the presence of CL which minimized CPX leakage from CLBFs Figures 1 and 2 show the impact of sonication time on ophthalmic CPX-CLBFs particle size and EE%. It was clear that 2 min of sonication time decreases the particle size of all formulations. In addition, it was observed that CPX-CLBFs containing cholesterol (CL) are not able to increase EE% due to high membrane rigidity of CL. As a result, the values of EE% for formulations containing CL (A1–A3 and B1–B3) was decreased after sonication for 2 min (Figure 2) due to the presence of CL which minimized CPX leakage from CLBFs upon sonication. cle size and EE%. It was clear that 2 min of sonication time decreases the particle size of all formulations. In addition, it was observed that CPX-CLBFs containing cholesterol (CL) are not able to increase EE% due to high membrane rigidity of CL. As a result, the values of EE% for formulations containing CL (A1–A3 and B1–B3) was decreased after sonication for 2 min (Figure 2) due to the presence of CL which minimized CPX leakage from CLBFs upon sonication.

#### upon sonication. *2.4. In Vitro Release Study from Ophthalmic CPX-CLBFs* Figure 3 shows the in vitro release profile of CPX form ophthalmic CPX-CLBFs. All

*2.4. In Vitro Release Study from Ophthalmic CPX-CLBFs*

*2.4. In Vitro Release Study from Ophthalmic CPX-CLBFs* Figure 3 shows the in vitro release profile of CPX form ophthalmic CPX-CLBFs. All formulations showed small initial gradual CPX release. Both B0 and A3 showed a faster release compared with other formulations with cumulative CPX percent release of 73% and 87%, respectively. CPX release was found to depend mainly on formulations particle Figure 3 shows the in vitro release profile of CPX form ophthalmic CPX-CLBFs. All formulations showed small initial gradual CPX release. Both B0 and A3 showed a faster release compared with other formulations with cumulative CPX percent release of 73% and 87%, respectively. CPX release was found to depend mainly on formulations particle size. Formulations with small particle size showed high cumulative drug percent release. formulations showed small initial gradual CPX release. Both B0 and A3 showed a faster release compared with other formulations with cumulative CPX percent release of 73% and 87%, respectively. CPX release was found to depend mainly on formulations particle size. Formulations with small particle size showed high cumulative drug percent release.

**Figure 3.** In vitro release profile of CPX from different ophthalmic CPX-CLBFs (mean ± SD, *n* = **Figure 3.** In vitro release profile of CPX from different ophthalmic CPX-CLBFs (mean ± SD, *n* = 3.0).

**Figure 3.** In vitro release profile of CPX from different ophthalmic CPX-CLBFs (mean ± SD, *n* =

3.0).

3.0).

#### *2.5. Antibacterial Activity of Different CPX-CLBFs Formulations*

The antibacterial activity and the minimum inhibitory concentration (MIC) were evaluated using the cup plate method and microdilution method, respectively, against bacterial strains frequently responsible for endophthalmitis. The MIC values of liposomal formulation samples containing CPX were investigated and the most effective are presented in Table 1. All formulations showed weak antibacterial activity (MIC ranged from 64–128 µg/mL) against all the tested strains except for formulations A3 and B0, which displayed a potent antibacterial activity (MIC ranged from 16–32 µg/mL) against all the tested strains. These formulations showed obvious activities in comparison with the other formulations against Gram-positive and -negative strains. For the *S. aureus*, *B. subtilis*, and *P. aeruginosa* bacterial strains, A3 revealed the best strength with MIC values of 16 µg/mL and exhibited minimum bactericidal concentration (MBC) values lower than that of the reference CPX. The higher in vitro antibacterial activity of formulations A3 and B0 were possibly due to the faster release of CPX from formulations A3 and B0 compared with other formulations studied.

**Table 1.** Antibacterial activity of selected ophthalmic ciprofloxacin (CPX) colloidal lipid-based formulations (CPX-CLBFs) and the commercial one.


The microbiological susceptibility testing of released CPX from the lipid-based entrapped CPX formulations was tested against different strains of Gram-negative (*E. coli*, *P. aeruginosa*, and *K. pneumonia*) and Gram-positive (*S. aureus*, *E. faecalis*, and *B. subtilis*) bacteria. These organisms are implicated in endophthalmitis [17]. Donnenfeld et al. alleged that the most common bacteria accountable for endophthalmitis were *S. aureus*, *S. epidermidis*, and *Streptococcus* sp. (50%, 30%, and 10%, respectively) [18]. The mean concentrations of antibiotic achieved in the current study were less than 1 µg/mL. The reported MIC of CPX required to inhibit the growth of *S. aureus* and *P. aeruginosa* ranged from 0.125–1 µg/mL [19]. This result was in agreement with Cutarelli et al. who reported a MIC90 of 1 µg/mL for CPX against tested strains [20]. Lesk et al. also estimated as low as 0.4 µg/mL for *S. epidermidis* and up to 1 µg/mL for *B. cereus* [21]. Therefore, the most effective two lipid-based entrapped CPX formulations were selected for the in vivo evaluation (B0 and A3). Both CPX-CLBF and Ciloxan® eye drop improved MIC and MBC of *P. aeruginosa* and *S. aureus*. Jain and Shastri reported a comparable result of liposomal formulation on MIC and MBC of CPX [22].

The results of the antimicrobial effectiveness and the level of CPX in the conjunctival sac of tested rabbits following the topical application of B0 and A3 formulations and Ciloxan® eye drops are shown in Figure 4. The reduction in bacterial count was detected and better improvement was observed in rabbit's eyes conjunctivitis. The ex vivo results were consistent with the results of the in vitro release, where the lipid-based entrapped CPX formulations A3 showed the uppermost CPX levels in the conjunctival sac in comparison to retailed Ciloxan® drops. B0, when compared to A3 and commercial eye drops, significantly failed to reduce the bacterial growth, which could be due to lack of positively charging agent. Effectually, cationic molecules such as STA are deemed to be notable bio-adhesives owing to an ability to produce forces of molecular attraction through the electrostatic interactions with the negative charges of mucin coating corneal surface as well as the secretions at the site of infection. to retailed Ciloxan® drops. B0, when compared to A3 and commercial eye drops, significantly failed to reduce the bacterial growth, which could be due to lack of positively charging agent. Effectually, cationic molecules such as STA are deemed to be notable bio-adhesives owing to an ability to produce forces of molecular attraction through the electrostatic interactions with the negative charges of mucin coating corneal surface as well as the secretions at the site of infection.

consistent with the results of the in vitro release, where the lipid-based entrapped CPX formulations A3 showed the uppermost CPX levels in the conjunctival sac in comparison

*Molecules* **2022**, *27*, x FOR PEER REVIEW 6 of 13

**Figure 4.** Colony count of the infected aqueous humor after the topical application of B0 and A3 formulations and Ciloxan® eye drops against *S. aureus* (**A**), *K. pneumonia* (**B**), and *P. aeruginosa* (**C**), (mean ± SD, *n* = 3.0). **Figure 4.** Colony count of the infected aqueous humor after the topical application of B0 and A3 formulations and Ciloxan® eye drops against *S. aureus* (**A**), *K. pneumonia* (**B**), and *P. aeruginosa* (**C**), (mean ± SD, *n* = 3.0).

#### *2.6. Pharmacokinetic Study 2.6. Pharmacokinetic Study*

Figure 5 shows CPX concentration in aqueous humor following application of CPX-CLBF A3 compared with Ciloxan® . Table 2 shows the pharmacokinetic parameters of A3 and Ciloxan® . One–way ANOVA was utilized to analyze the calculated pharmacokinetic parameters after application of CPX-CLBF A3 and Ciloxan® . The Cmax of 4.2 μg/mL was obtained from A3 compared with 2.7 μg/mL for Ciloxan® . There were no significant differences in values of Tmax between A3 and Ciloxan® . The values of AUC**0-t** for A3 (34.9 μg min/mL) were significantly greater than those obtained from Ciloxan® (12 μg min/mL). These results confirmed a greater ocular bioavailability of CPX-CLBF A3. High values of AUC0–<sup>t</sup> indicated high extent of drug absorption. Regarding t1/2, A3 shows increased t1/2 value compared with that of the commercial formulation which indicated the presence of CPX for longer time in aqueous humor. The relative bioavailability valued significantly emphasize that A3 increased drug bioavailability by 2.9-folds compared with the commercial formulation. Based on the bioavailability results, the CLBF delivered extra CPX into the aqueous humor and enhanced therapeutic effects in comparison with Ciloxan® drop. Figure 5 shows CPX concentration in aqueous humor following application of CPX-CLBF A3 compared with Ciloxan®. Table 2 shows the pharmacokinetic parameters of A3 and Ciloxan®. One–way ANOVA was utilized to analyze the calculated pharmacokinetic parameters after application of CPX-CLBF A3 and Ciloxan®. The Cmax of 4.2 µg/mL was obtained from A3 compared with 2.7 µg/mL for Ciloxan®. There were no significant differences in values of Tmax between A3 and Ciloxan®. The values of AUC0-t for A3 (34.9 µg min/mL) were significantly greater than those obtained from Ciloxan® (12 µg min/mL). These results confirmed a greater ocular bioavailability of CPX-CLBF A3. High values of AUC0–t indicated high extent of drug absorption. Regarding t1/2, A3 shows increased t1/2 value compared with that of the commercial formulation which indicated the presence of CPX for longer time in aqueous humor. The relative bioavailability valued significantly emphasize that A3 increased drug bioavailability by 2.9-folds compared with the commercial formulation. Based on the bioavailability results, the CLBF delivered extra CPX into the aqueous humor and enhanced therapeutic effects in comparison with Ciloxan® drop.

**Table 2.** Pharmacokinetics parameters of ophthalmic CPX- CLBF (A3) and commercial one (mean ± SD, *n* = 3.0).


Cmax: maximum plasma concentration; Tmax: time to reach Cmax; K: elimination rate constant; t1/2: half-life; AUC0–t: area under curve from time 0 to t.

**Figure 5.** CPX concentrations in rabbit aqueous humor for ophthalmic CPX-CLBF (A3) compared with commercial one (mean ± SD, *n* = 6.0). **Figure 5.** CPX concentrations in rabbit aqueous humor for ophthalmic CPX-CLBF (A3) compared with commercial one (mean ± SD, *n* = 6.0).

#### **Table 2.** Pharmacokinetics parameters of ophthalmic CPX- CLBF (A3) and commercial one (mean ± **3. Materials and Methods**

#### SD, *n* = 3.0). *3.1. Materials*

**Parameters A3 Commercial CPX** Dose (µg) 115 150 Cmax (µg/mL) 4.2 ± 0.1 2.7 ± 0.1 CPX was obtained from Fluka Biochemika (Busch, Switzerland). PC, CL, DMPC, and STA were obtained from Avanti Polar Lipids Inc. (Alabaster, AL, USA). Chloroform was obtained from BDH Laboratory Ltd. (Poole, England, UK). The solvents used in this study were of chromatography grade.

#### Tmax (h) 2 2 K (h−<sup>1</sup> ) 0.2 ± 0.03 0.4 ± 0.01 *3.2. Bacterial Culturing*

t1/2 (h) 3.5 ± 0.2 1.7 ± 0.2 AUC 0–<sup>t</sup> (µg.h/mL) 34.9 ± 2.1 12 ± 0.5 Relative bioavailability 2.9 - Cmax: maximum plasma concentration; Tmax: time to reach Cmax; K: elimination rate constant; t1/2: half-life; AUC0–t: area under curve from time 0 to t. **3. Materials and Methods** *3.1. Materials* CPX was obtained from Fluka Biochemika (Busch, Switzerland). PC, CL, DMPC, and STA were obtained from Avanti Polar Lipids Inc. (Alabaster, AL, USA). Chloroform was The growing bacterial cultures were determined by controlling the turbidity changes with the measurements of absorbance at 600 nm. Applying a titration curve, the changes in the absorbance were then turned correspondingly into CFU/mL (colony forming units). The titration curve was created by considering the bacterial samples with optical density of 0.05–2.0 at 600 nm. Serial dilution for the cells was performed in trypticase soy broth (Oxoid, Lenexa, KS, USA) and then culturing the diluted cells on Mueller–Hinton agar (MHA) (Oxoid, Lenexa, KS, USA). The plates were incubated for overnight at 35 ± 1 ◦C. Bacterial colonies were calculated, and result obtained from triplicate assays were applied to generate a titration curve that compared optical density measurements with CFU/mL for the bacterial strains growth.

#### obtained from BDH Laboratory Ltd. (Poole, England, UK). The solvents used in this study *3.3. Preparation of CPX-CLBFs*

were of chromatography grade. *3.2. Bacterial Culturing* The growing bacterial cultures were determined by controlling the turbidity changes with the measurements of absorbance at 600 nm. Applying a titration curve, the changes in the absorbance were then turned correspondingly into CFU/mL (colony forming units). The titration curve was created by considering the bacterial samples with optical density CPX-CLBFs were prepared utilizing reverse evaporation technique [23]. The molar ratios of CPX-CLBFs are shown in Table 3. The exact amount of lipids was dissolved in 13 mL of chloroform. Five mL of CPX was dissolved in phosphate buffer (pH 4.5) and added to the mixture. Chloroform was evaporated from the mixture by rotary evaporation at 40 ◦C. The mixture was centrifuged for 40 min at 30,000 rpm. Free CPX content in supernatant was detected and the lipid capsules were dispersed in phosphate buffer (pH 7.4) [24]. Each formulation was subjected to sonication for 2 min right after preparation.

#### of 0.05–2.0 at 600 nm. Serial dilution for the cells was performed in trypticase soy broth (Oxoid, Lenexa, KS, USA) and then culturing the diluted cells on Mueller–Hinton agar *3.4. Characterization of CPX-CLBFs*

*3.3. Preparation of CPX-CLBFs*

#### (MHA) (Oxoid, Lenexa, KS, USA). The plates were incubated for overnight at 35 ± 1 °C. 3.4.1. Particle Size, PDI, and Zeta Potential Measurements

Bacterial colonies were calculated, and result obtained from triplicate assays were applied to generate a titration curve that compared optical density measurements with CFU/mL for the bacterial strains growth. The mean particle size and PDI for each CPX-CLBFs were determined at ambient temperature using laser diffraction analysis (NiComp Particle Size system ZW380 Santa Barbara, CA, USA). Each CPX-CLBF formulation was diluted with phosphate buffer before analysis.


**Table 3.** Composition of ophthalmic CPX-CLBFs in molar ratios.

PC: phosphatidylcholine; DMPC: dimyrestoyl phosphatidylcholine; CL: cholesterol; STA: stearylamine.

#### 3.4.2. Evaluation of CPX EE%

Each CPX-CLBF formulation was centrifuged for 40 min at 30,000 rpm to get rid of free CPX content and EE% was calculated using the following equation [25,26]:

$$\text{EE\%} = \frac{(\text{W}\_{\text{t}} - \text{W}\_{\text{f}})}{\text{W}\_{\text{t}}} \times 100\tag{1}$$

where, W<sup>t</sup> and W<sup>f</sup> are the total CPX and the free CPX content, respectively.

#### 3.4.3. Effect of Sonication on CPX-CLBFs Particle Size and EE%

To study the impact of sonication on particle size and EE% of CLBFs, all formulations were sonicated for 2 min. The data were recorded before and after sonication.

#### *3.5. CPX Assay Method*

Reversed-phase high performance liquid chromatographic (HPLC) method was used for the quantification of CPX. A mobile phase was a mixture of acetonitrile: methanol: 0.05 acetate buffer (10:20:70, *v*/*v*/*v*) and 1% (*v*/*v*) of acetic acid. The acetate buffer was at pH 3.6. The mobile phase was degassed using sonication and filtered through 0.45 µm Millipore filter immediately after preparation. Samples of 20 µL were injected in Shimadzu HPLC system consists of Fluorescence detector, pump, Kromasil 100, C18, 5 µm (250 × 4.6 mm) column, and Rheodyne sample injector were used in the assay method. CPX and anthranilic acid (internal standard) was detected at λexc = 300 nm and λemi = 458 nm, respectively, using mobile flow rate of 0.8 mL/min at room temperature [27].

#### *3.6. In Vitro CPX Release Study from CLBFs*

Franz diffusion cells were used to carry out this experiment. The in vitro release of CPX was performed in release medium of artificial tears (pH 7.4) placed in receptor cells [28]. A semipermeable membrane with a molecular weight cut-off range of 12,000–14,000 Da was used [29]. The membrane was soaked in the dissolution medium for about 12 h before conducting the experiment. The utilized membrane was able to permeate CPX molecules and hold CLBF vesicles. The temperature was kept at 34 ± 0.5 ◦C and the mixture was stirred at 300 rpm. A specific amount of each CPX ophthalmic formulation was placed in the dialysis donor cell. Samples of 1 mL each were periodically collected at time intervals of 0, 0.25, 0.5, 1, 2, 3, 4, and 6 h and replaced by same volume of artificial tears to maintain constant volume and sink condition. Then, the amount of CPX was determined in each sample using HPLC method described above [27]. All tests were repeated in triplicate. Based on the characterizations of ophthalmic CPX-CLBFs formulations, A3 (after sonication) and B0 (without sonication) were selected for further investigation.

#### *3.7. Evaluation of the In Vitro Antibacterial Activity of CPX-CLBFs*

Antibiogram was carried out for the prepared 8 colloidal lipid-based entrapped CPX formulations, as well as the standard CPX-hydrochloride, by the agar well diffusion

method [30,31]. CLBF vesicles were separately investigated against a set of Gram-positive and -negative bacterial strains as proposed by the Clinical Laboratory Standard Institute (CLSI) guidelines [32]. For this purpose, different strains of Gram-negative bacteria (*E. coli* ATCC 9637, *P. aeruginosa* ATCC 27953, and *K. pneumonia* ATCC 10031) as well as three different strains of Gram-positive bacteria (*S. aureus* ATCC 29213, *E. faecalis* ATCC 29212, and *B. subtilis* ATCC 10400) were selected. Bacterial suspension (100 µL) comprising <sup>1</sup> <sup>×</sup> <sup>10</sup><sup>6</sup> CFU/mL of bacterial strain were mixed with MHA medium. After waiting for the media to solidified, wells of 6.0 mm diameter were made through the solidified MHA using a cork borer and burdened equal quantities of each tested liposomal formulation samples. The implanted plates were then overnight incubated at 35 ± 1 ◦C. Negative control was prepared using free MHA. The zone of inhibition against the tested organisms was measured after incubation time for evaluating the antimicrobial activity. Each test was performed in triplicate and inhibition zone average was calculated. The average diameters of the inhibition zones were recorded and expressed according to the following score guide in mm as follows: + = ≥ 6 mm; ++ = ≥ 10 mm (moderate); +++ = ≥ 15 mm (strong); and ++++ = ≥20 mm (very strong).

#### *3.8. Determination of In Vitro MIC and MBC*

The MIC of the most effective CPX-CLBF formulations (3 efficient formulae) and commercial ophthalmic formulation containing 0.3% CPX-hydrochloride (Ciloxan® drop, Alcon Laboratories Inc., Fort Worth, TX, USA) were assayed against mentioned bacteria that are frequently responsible for endophthalmitis [33]. The test was performed by broth microdilution method [34] in conformity with the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [35]. Overnight bacterial suspensions were adjusted to <sup>1</sup> <sup>×</sup> <sup>10</sup><sup>6</sup> CFU/mL as described above. The CLBF samples and CPX commercial formula were adjusted to give appropriate concentration by sequential dilutions in 2X MH broth (Merck, Darmstadt, Germany). Equal bacterial inoculums (5 µL) were dispensed into wells and the microtiter plate was incubated overnight at 35 ± 1 ◦C. Subsequently, the MIC results were listed and explicated consistent with the EUCAST guidelines. Positive control (inoculum and MH media, devoid of liposomal entrapped CPX) was used. Free culture medium was used as a negative control. To ensure reproducibility, each assessment was completed in triplicates on three different days. Wells displaying no bacterial growth were streaked on MH agar and incubated overnight for MBC. Tested MBC was considered as the lowest concentration of either free or liposomal CPX that able to cause reduction (more than 99.9%) of the initial inoculum.

## *3.9. Ex Vivo Antibacterial Effect of CPX-CLBFs and Ciloxan®*

Male New Zealand white Albino rabbits (weighed 2–2.3 kg) were obtained from the Experimental Animal Care Center (EACC) at Prince Sattam bin Abdulaziz University, Al-Kharj, Saudi Arabia and utilized to conduct the in vivo/ex vivo antibacterial study. All animals were provided free access to water but were starved for 24 h before the experiment. The in vivo antibacterial studies were approved by the Animal Ethics Committee of the EACC Board (Prince Sattam bin Abdulaziz University, Al-Kharj, Saudi Arabia with approval number: BERC-005-04-21). The animal use and experimental procedures followed EU directive 2010/63/EU.

The bacterial suspensions were prepared by overnight growing at 35 ± 1 ◦C in trypticase soy broth and adjusted to achieve a concentration of 1 <sup>×</sup> <sup>10</sup><sup>3</sup> CFU/30 <sup>µ</sup>L, proper for ocular application. Rabbits were randomly selected and divided into three equal groups for each bacterial strain: group A for *S. aureus*, group B for *K. pneumonia*, and group C for *P. aeruginosa*. Each group contained 9 rabbits with the same race, gender and average weight of 2.0–2.3 kg. To induce intraocular infections in the rabbit eyes, 5 µL of each bacterial suspension was inoculated into the conjunctival sac through sterile dropper for each group. After infection verification, the right eyes were treated by CLBF and Ciloxan® drops for the assessment of the in vivo antibacterial efficacy of different CPX-CLBF formulations. On the other hand, the left eyes were kept un-treated. From anterior chamber paracentesis, the aqueous humor specimens were withdrawn via the translimbal pathway by a 27-gauge needle mounted on a tuberculin syringe at a dose of 0.1 mL, at zero (before pouring the eye drops), 0.5, 1, 2, 3, 4, 5, and 6 h after application of the formulated drug. The CPX-CLBF in aqueous humor specimens were measured by culturing on TSA media for the microbial count and estimation of MIC and MBC. Antibacterial activities were expressed as inhibition diameter zones in mm. MIC was expressed in µg/mL and evaluated by microbroth dilution method according to EUCAST.

## *3.10. Ex Vivo Antibacterial Effect of Selected CPX-CLBFs and Ciloxan® Drop*

Both MIC and MBC of the prepared CLBF formulations in ex vivo media (media enriched by infected aqueous humor specimens) against *S. aureus*, *K. pneumonia*, and *P. aeruginosa* were examined (data not presented). CPX entrapped in A3 formulation decreased MIC and MBC against *K. pneumonia* and *S. aurous* more than B0 and Ciloxan®. A3 significantly enhanced the susceptibility of *K. pneumonia* and *S. aurous* to CPX compared with B0 and Ciloxan®. On the other hand, only A3 decreased MIC and MBS for *P. aeruginosa* and no significant difference was noticed between B0 and Ciloxan®.

#### *3.11. Study of CPX-CLBFs Treatment Effect*

For the evaluation of the treatment outcome, colony count was determined in treated and un-treated rabbit's eyes. Selected rabbits infected by *S. aureus*, *K. pneumonia*, and *P. aeruginosa* were treated by B0, A3, and Ciloxan®. After 6 h of treatment, the samples withdrawn from the infected aqueous humor and streaked for colony count. According to these results, A3 displayed significantly reduced colony counts for all tested organisms after 6 h. Although Ciloxan® showed reducing effect in colony count, but it is insignificant compared to A3 effect. Blank CLBF control did not show any effect on demonstrated bacterial growth. On the other hand, bacterial count was significantly higher than that observed with A3 and commercial ophthalmic drops, which means B0 failed to reduce the bacterial infection.

#### *3.12. Pharmacokinetic Studies*

Based on the data obtained from the physiochemical and antimicrobial evaluations of the prepared CPX ophthalmic formulations, formulation A3 was selected for in vivo pharmacokinetic experiment along with the commercial CPX eye drops, Ciloxan®. Fifty µL dose of CPX ophthalmic formulation was applied in the lower conjunctival sac of the rabbit's right eye; however, Ciloxan® (Riyadh Pharma, Riyadh, Saudi Arabia) was applied in the left one. The animals were euthanized after withdrawing the aqueous humor at specific time intervals. The samples were centrifuged for 20 min at 4 ◦C and 20,000 rpm and supernatant were frozen at –20 ◦C until assay. The amount of CPX in each sample was analyzed using the HPLC method described above [27].

#### *3.13. Pharmacokinetic Parameters*

The Cmax and Tmax values were obtained directly from rabbit aqueous humor concentrationtime plot. The AUC0–t (µg min/mL) was calculated using the trapezoidal method. The elimination rate constant (k) was determined from the final segment of the curve by fitting the data in first order elimination graph. Relative bioavailability of CPX-CLBF was obtained in comparison to that of the commercial product.

#### *3.14. Statistical Analysis*

All experiments in this study were subjected to one way analysis of variance with Tukey's multiple comparisons test at a 5% level of significance (*p* ≤ 0.05).

#### **4. Conclusions**

The MIC for CPX lipid-based formulations was noticeably lower than that of the retailed and free drug, indicating that CLBF systems may participate in the increased antimicrobial activity of CPX. The pharmacokinetic parameters of the selected ophthalmic formulation showed significant improvement in ocular bioavailability of CPX in comparison with Ciloxan® ophthalmic drops. This could be due to several factors that characterize the tested CPX formulation, including that it produces low lacrimation due to its almost neutral pH value of 7.4. In addition, the presence of a positive charge inducing agent which increases the drug contact time with corneal surface. Finally, the liposomal formulation improves drug permeability to the cornea which maximizes drug concentration in the affected area. Based on these results, the prepared CPX formulation could provide a better alternative to the marketed one with respect to enhancing ocular drug activity, especially in the management of post-surgical infections.

**Author Contributions:** Conceptualization, M.M.S.-B. and F.S.; methodology, F.A.A.-J., E.I.T., M.A.I., M.M.M., F.A.A.-J. and S.A.; software, M.M.G.; validation, F.S., M.M.G. and S.A.; formal analysis, M.M.G.; investigation, M.M.M., E.I.T., F.A.A.-J. and F.S.; resources, M.A.I.; data curation, M.M.G.; writing—original draft preparation, E.I.T.; writing—review and editing, M.M.S.-B., F.S. and M.A.I.; visualization, M.A.I.; supervision, M.M.S.-B. and F.S.; project administration, M.M.S.-B.; funding acquisition, M.A.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Researcher Supporting Project (Number RSP-2021/171) at King Saud University, Riyadh, Saudi Arabia. The APC was funded by RSP.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Animal Ethics Committee of the EACC Board at Prince Sattam bin Abdulaziz University, Al-Kharj, Saudi Arabia (approval number: BERC-005-04-21).

**Data Availability Statement:** This study did not report any data.

**Acknowledgments:** Authors are thankful to the Researcher Supporting Project (Number RSP-2021/171) at King Saud University, Riyadh, Saudi Arabia for supporting this work.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples of the CPX compounds are available from the authors.

#### **References**


## *Article* **Design of Turmeric Rhizome Extract Nano-Formula for Delivery to Cancer Cells**

**Sakchai Auychaipornlert <sup>1</sup> , Pojawon Prayurnprohm Lawanprasert 1,\*, Suchada Piriyaprasarth <sup>2</sup> , Pongtip Sithisarn <sup>3</sup> and Supachoke Mangmool <sup>4</sup>**


**Abstract:** Novel turmeric rhizome extract nanoparticles (TE-NPs) were developed from fractions of dried turmeric (*Curcuma longa* Linn.) rhizome. Phytochemical studies, by using HPLC and TLC, of the fractions obtained from ethanol extraction and solvent–solvent extraction showed that turmeric rhizome ethanol extract (EV) and chloroform fraction (CF) were composed mainly of three curcuminoids and turmeric oil. Hexane fraction (HE) was composed mainly of turmeric oil while ethyl acetate fraction (EA) was composed mainly of three curcuminoids. The optimal TE-NPs formulation with particle size of 159.6 ± 1.7 nm and curcumin content of 357.48 ± 8.39 µM was successfully developed from 47-run D-optimal mixture–process variables experimental design. Three regression models of z-average, d50, and d<sup>90</sup> could be developed with a reasonable accuracy of prediction (predicted r<sup>2</sup> values were in the range of 0.9120–0.9992). An in vitro cytotoxicity study using MTT assay demonstrated that the optimal TE-NPs remarkably exhibited the higher cytotoxic effect on human hepatoma cells, HepG2, when compared with free curcumin. This study is the first to report nanoparticles prepared from turmeric rhizome extract and their cytotoxic activity to hepatic cancer cells compared with pure curcumin. These nanoparticles might serve as a potential delivery system for cancer therapy.

**Keywords:** turmeric rhizome extract; turmeric oil; curcumin; HepG2; nanoparticles; anticancer

#### **1. Introduction**

Turmeric is a dried rhizome of Curcuma longa Linn. of the family Zingiberaceae. It is mostly cultivated in Southern and Southeast Asia [1]. A number of pharmacological activities, especially anticancer activities, of compounds contained in turmeric were reported [2]. Most of them showed the pharmacological activities of curcumin, the major active compound found in the turmeric rhizome [2–5]. Anticancer activities of its analog compounds and turmeric oil were also reported [6–9]. Turmeric and curcumin can be considered as safe [10,11]. However, the low aqueous solubility and poor stability of curcumin led to limitations of its use as a therapeutic agent. Many advanced technologies were proposed to overcome this limitation [12–14].

Nanotechnology had been the one potential strategy for treatment of cancer diseases [15–19]. The nanoscale materials are currently being investigated to improve their specificity towards cancer cells and towards subcellular compartments in order to reduce systemic toxicity [15–20]. Curcumin has also been developed in nanoscale [21–37].

**Citation:** Auychaipornlert, S.; Lawanprasert, P.P.; Piriyaprasarth, S.; Sithisarn, P.; Mangmool, S. Design of Turmeric Rhizome Extract Nano-Formula for Delivery to Cancer Cells. *Molecules* **2022**, *27*, 896. https://doi.org/10.3390/ molecules27030896

Academic Editor: Faiyaz Shakeel

Received: 20 December 2021 Accepted: 24 January 2022 Published: 28 January 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

For example, Shaikh and coworkers (2009) [29] prepared curcumin-loaded poly(lactideco-glycolide) (curcumin loaded PLGA) nanoparticles by using an emulsion–diffusion– evaporation method. Curcumin-loaded PLGA nanoparticles demonstrated at least 9-fold increase in oral bioavailability when compared with curcumin administered with piperine as an absorption enhancer. Anand and colleagues (2010) [21] prepared curcumin-loaded PLGA-PEG nanoparticles by a nanoprecipitation technique. In vitro study showed that curcumin nanoparticles exhibited rapid cellular uptake and induced apoptosis in human chronic myeloid leukemia (KBM-5). Additionally, curcumin nanoparticles could inhibit cell proliferation of various tumor cells, i.e., human leukemia (KBM-5 and Jurkat), prostate (DU145), breast (MDA-MB-231), colon (HCT116), and esophageal (SEG-1) cancer cells. Mohanty C. and Coworkers (2010) [27] prepared curcumin nanoparticles by an emulsifying method with a group of surfactants, i.e., glycerol monooleate (GMO), polyvinyl alcohol (PVA), and Pluronic®. These curcumin nanoparticles were more effective than curcumin against different cancer cells. In addition, Zhao L and coworkers (2012) [37] prepared curcumin-loaded mixed micelles (Cur-PF) that were composed of Pluronic P123 and Pluronic F68. They found that Cur-PF presented a sustained release property. O/W nanoemulsion containing curcumin was prepared by using high-speed and high-pressure homogenization [34]. Medium chain triacylglycerols (MCT) and Tween 20 were used as oil phase and emulsifier, respectively. This 1% curcumin o/w nanoemulsion exhibited an inhibition effect of 12-O-tetradecanoyl- horbol-13-acetate (TPA)-induced edema of mouse ear.

The enhancements of anticancer activities were found when nanoparticles of an anticancer drug were coated with hyaluronic acid [38–40]. This may be because the high binding affinity of hyaluronic acid to the CD44 receptor, which overexpresses in tumor cell [39,41–43].

The spontaneous nanoemulsion formed by the solvent displacement method called the Ouzo effect was originally found in anise-flavored alcoholic beverages [44–55]. Vitale and Katz explained that the effect occurs when solutions are rapidly brought into the metastable region by the addition of water. When the solubility of some of solutes decreases more rapidly, supersaturation is then large, and homogeneous nuclei form spontaneously [47].

A number of studies in the literature have focused on the nanoparticle formation of curcumin. The study of nanoparticle formation from turmeric rhizome extract has not yet been reported. In this study, we aimed to develop nanoparticles from various turmeric rhizome fractions by using the solvent displacement method and investigated the cytotoxic activity of the obtained turmeric rhizome extract nanoparticles toward HepG2 cells.

#### **2. Results and Discussion**

#### *2.1. Curcuminoids Content of Turmeric Rhizome Fractions*

The chromatograms of turmeric rhizome fractions analyzed by thin layer chromatography (TLC) detected by ultraviolet (UV) and spray reagent are shown in Figure 1. It was found that turmeric rhizome fractions except aqueous fraction (AQ) (track 9) developed chromatographic bands with hRf values corresponding to the standard three curcuminoids (Figure 1). The hRf values of standards, curcumin (CM, track 1), desmethoxycurcumin (DCM, track 2), and bisdesmethoxycurcumin (BDCM, track 3) determined by spraying with 10% phosphomolybdic spray reagent were 18.13, 16.88, and 10.00, respectively. In addition, the turmeric rhizome ethanol extract (EV), hexane fraction (HE), and chloroform fraction (CF) developed blue bands above three bands of standard curcuminoids at hRf values of 78.75, 79.38, and 80.00, respectively. This blue band was clearly seen for HE detected under UV 254 nm and 10% phosphomolybdic spray reagent (track 5). The remaining two bands of EV (track 4) and CF (track 6) showed pale blue bands under UV 254. The band at an hRf value of approximately 80 might be dl-turmerone, one of turmeric oil's components, as specified in the Thai Herbal Pharmacopoeia 2020 vol 1 [56].

**Figure 1.** TLC chromatograms of standard curcuminoids and all turmeric rhizome fractions. Track 1 = standard curcumin (CM); 2 = standard desmethoxycurcumin (DCM); 3 = standard bisdesmethoxycurcumin; 4 = EV; 5 = HE; 6 = CF; 7 = EA; 8 = BU; 9 = AQ. Solvent system, benzene/chloroform/ethanol (49:49:2 by volume). Detection: (**A**) = UV 254 nm; (**B**) = UV 366 nm; (**C**) = 10% phosphomolybdic spray reagent; all heated at 105 °C for 5 min, detected under white light. **Figure 1.** TLC chromatograms of standard curcuminoids and all turmeric rhizome fractions. Track 1 = standard curcumin (CM); 2 = standard desmethoxycurcumin (DCM); 3 = standard bisdesmethoxycurcumin; 4 = EV; 5 = HE; 6 = CF; 7 = EA; 8 = BU; 9 = AQ. Solvent system, benzene/chloroform/ethanol (49:49:2 by volume). Detection: (**A**) = UV 254 nm; (**B**) = UV 366 nm; (**C**) = 10% phosphomolybdic spray reagent; all heated at 105 ◦C for 5 min, detected under white light.

Turmeric rhizome fractions were also analyzed by the validated modified high performance liquid chromatographic method (HPLC) [57]. The HPLC fingerprints of the standards and turmeric rhizome fractions are shown in Figure 2. For ethanol extract (EV), three peaks that have retention time corresponding to standard curcumin, desmethoxycurcumin, and bisdesmethoxycurcumin at 20.63, 19.36, and 17.60 min, respectively, are shown. Similar results were found for the other turmeric rhizome fractions except AQ, i.e., hexane fraction (HE), chloroform fraction (CF), ethyl acetate fraction (EA), and n-butanol fraction (BU). The quantitative analysis results of turmeric rhizome fractions are shown in Table 1. The results show that all turmeric rhizome fractions except AQ contained various amounts of three main curcuminoids. The highest total curcuminoid content was found in EA fraction (421.41 mg/g of dried extract), which contained bisdesmethoxycurcumin (BDCM) as the major component. Low curcuminoid content was found in HE and BU fractions (4.88 and 12.14 mg/g of dried extract, respectively). TLC and HPLC analysis suggested that the phytochemical profile of turmeric rhizome fractions prepared in this study were mainly composed of curcuminoids and turmeric oil. This result is consistent with the results reported previously [56,58–61]. Turmeric rhizome fractions were also analyzed by the validated modified high performance liquid chromatographic method (HPLC) [57]. The HPLC fingerprints of the standards and turmeric rhizome fractions are shown in Figure 2. For ethanol extract (EV), three peaks that have retention time corresponding to standard curcumin, desmethoxycurcumin, and bisdesmethoxycurcumin at 20.63, 19.36, and 17.60 min, respectively, are shown. Similar results were found for the other turmeric rhizome fractions except AQ, i.e., hexane fraction (HE), chloroform fraction (CF), ethyl acetate fraction (EA), and n-butanol fraction (BU). The quantitative analysis results of turmeric rhizome fractions are shown in Table 1. The results show that all turmeric rhizome fractions except AQ contained various amounts of three main curcuminoids. The highest total curcuminoid content was found in EA fraction (421.41 mg/g of dried extract), which contained bisdesmethoxycurcumin (BDCM) as the major component. Low curcuminoid content was found in HE and BU fractions (4.88 and 12.14 mg/g of dried extract, respectively). TLC and HPLC analysis suggested that the phytochemical profile of turmeric rhizome fractions prepared in this study were mainly composed of curcuminoids and turmeric oil. This result is consistent with the results reported previously [56,58–61].


**Table 1.** Curcuminoids content of turmeric rhizome fractions analyzed by HPLC (Mean ± SD, *n* = 3). **Table 1.** Curcuminoids content of turmeric rhizome fractions analyzed by HPLC (Mean ± SD, *n* = 3).

AQ ND ND ND ND CM = curcumin; DCM = desmethoxycurcumin; BDCM = bisdesmethoxycurcumin; EV = ethanol extract; HE = hexane fraction; CF = chloroform fraction; EA = ethyl acetate fraction; BU = n-butanol CM = curcumin; DCM = desmethoxycurcumin; BDCM = bisdesmethoxycurcumin; EV = ethanol extract; HE = hexane fraction; CF = chloroform fraction; EA = ethyl acetate fraction; BU = n-butanol fraction; AQ = aqueous fraction; ND = cannot be detected.

fraction; AQ = aqueous fraction; ND = cannot be detected.

**Figure 2.** HPLC fingerprint of standard curcumin (**A**); standard desmethoxycurcumin (**B**); standard bisdesmethoxycurcumin (**C**); and turmeric rhizome ethanol extract (EV, (**D**)). **Figure 2.** HPLC fingerprint of standard curcumin (**A**); standard desmethoxycurcumin (**B**); standard bisdesmethoxycurcumin (**C**); and turmeric rhizome ethanol extract (EV, (**D**)).

#### *2.2. Determination of the Optimal Turmeric Rhizome Extract Nanoparticles Formulation*

The curcuminoids content and particle characteristic results of 47 designed TE-NP formulations are shown in Table S1 (Supplementary Materials).

The significant regression model for dependent variables (*p*-value < 0.05) with a high degree of model fitness (r<sup>2</sup> = 0.8961–0.9635) were obtained for curcumin content and %LA, z-average, d50, and d90, defined as Ym1, Ym2, Ym5, Ym6, and Ym7. This indicates that these regression models have power to explain the effect of independent variables on the dependent variables of TE-NPs. The regression models for curcumin analysis created in this work were the combined reduced quadratic × linear models. For particle analysis, the combined reduced quadratic × 2FI, quadratic × linear, and linear × 2FI models were obtained for z-average, d50, and d90, respectively.

The regression model equations in terms of actual components and actual factors of all dependent variables were obtained as follows:

$$\sqrt{\mathbf{Y}\_{\text{m1}}} = 0.072\mathbf{A} + 5.218\mathbf{B} + 4.520\mathbf{C} + 0.825\mathbf{A}\mathbf{B} + 1.930\mathbf{A}\mathbf{C} + 85.035\mathbf{A}\mathbf{D} \tag{1}$$

$$\begin{array}{l} \sqrt{\text{Ym}\_{\text{m}}} = 0.155 \text{A} & +3.576 \text{B} + 2.914 \text{C} + 1.699 \text{AB} + 2.883 \text{AC} + 0.174 \text{AD} \\ & + 0.301 \text{B} - 197.093 \text{BD} - 0.164 \text{BE} - 29.685 \text{CD} \\ & + 0.039 \text{CE} - 0.388 \text{AB} - 0.311 \text{AC} + 219.773 \text{BD} \\\\ \text{Ym}\_{\text{m}5} = 70.592 \text{A} & + 61.265 \text{B} + 62.690 \text{C} - 2.196 \text{AB} - 10.592 \text{AC} + 9.530 \text{AE} \\ & - 7.599 \text{BC} + 242.050 \text{BD} + 2.246 \text{BE} + 3.625 \text{CE} \\ & + 3.484 \text{AB} \text{BE} + 742.110 \text{BD} \text{E} \\\\ \sqrt{\text{Ym}\_{\text{6}}} = 5.443 \text{A} & + 4.603 \text{B} + 4.962 \text{C} - 0.582 \text{AC} + 0.828 \text{AE} + 102.780 \text{BD} \\ & + 0.520 \text{BE} + 0.380 \text{CE} \\\\ \sqrt{\text{Ym}\_{7}} = 7.128 \text{A} & + 6.324 \text{B} + 6.273 \text{C} + 1.402 \text{AE} - 12.783 \text{BD} + 0.443 \text{BE} \\ & + 1.008 \text{CE} + 154.056 \text{BD} \text{E} \end{array}$$

where Ym1 = curcumin content (µM), Ym2 = % label amount of curcumin (%LA), Ym5 = zaverage (nm), Ym6 = d<sup>50</sup> (nm), Ym7 = d<sup>90</sup> (nm), A = HE (%*w*/*w*), B = CF (%*w*/*w*), C = EA (%*w*/*w*), D = external curcumin (%*w*/*w*), and E = sodium hyaluronate (NaHA) (%*w*/*w*). To assess the predictability of the regression model, all models obtained were validated. The predictive root mean square error (predictive RMSE) and predictive r<sup>2</sup> of all regression models calculated using Equations (1)–(5) are shown in Table 2.

**Table 2.** The predictive root mean square error (predictive RMSE) and predictive r<sup>2</sup> of all regression model equations.


Predicted r<sup>2</sup> close to one and low predictive RMSE should be obtained for the model with good predictability. In this study, it was found that regression models for z-average, d50, and d<sup>90</sup> had predicted r<sup>2</sup> higher than 0.9 (0.9120–0.9992), and the predictive RMSEs of these models were 12.00, 26.30, and 88.24, respectively. This indicates the good predictability of these models. However, the predictabilities of models for curcumin content showed low power (r<sup>2</sup> of 0.8673 and 0.7140 for CM content and %LA of CM, respectively).

The optimal TE-NPs formulation was selected from the optimal region (Figure 3). To determine the optimal region, the acceptance limits of desired dependent variables were specified first. The highest curcumin content obtained in MPV design was 348.67 µM. Thus,

the acceptance lower limit of curcumin content of 300 µM was used. The acceptance limit of 80–120% LA for curcumin was chosen [62]. Particle size plays a crucial role in the delivery of nanoparticles to tumor cells. The nanoparticles of appropriate size can be selectively delivered to tumor cells and can escape from the defensive system of body. Angiogenesis in cancer cells results in abnormalities—namely, hypervascularization, aberrant vascular architecture, extensive production of vascular permeability factors stimulating extravasation within tumor tissues, and lack of lymphatic drainage. It allows the passive accumulation of the nanoparticles in tumor tissue, which is known as an enhanced permeability and retention effect (EPR effect) [63–66]. To achieve the extravasation into a tumor by the EPR effect, nanoparticles' size should be below 200 nm [67]. Moreover, nanoparticles must have an appropriate circulation half-life, avoiding the action of the mononuclear phagocyte system (MPS) and the reticuloendothelial system (RES). To overcome these effects, the nanoparticles' size must not exceed 400 nm to escape from the MPS effect [67]. Thus, the acceptance upper limit of 200 nm for z-average and d50, and 400 nm for d<sup>90</sup> were specified in this study. The optimal region (black area) that was obtained by overlaying between dependent variable plots is shown in Figure 3. Each individual point in this optimal region represents an appropriate TE-NPs formulation. In this study, the optimal TE-NPs formulation consisting of 1.3697 %*w*/*w* CF, 1.2970 %*w*/*w* EA, and 0.0067 %*w*/*w* external curcumin was selected. The physicochemical properties of the optimal TE-NPs formulations are shown in Table 3. onuclear phagocyte system (MPS) and the reticuloendothelial system (RES). To overcome these effects, the nanoparticles' size must not exceed 400 nm to escape from the MPS effect [67]. Thus, the acceptance upper limit of 200 nm for z-average and d50, and 400 nm for d90 were specified in this study. The optimal region (black area) that was obtained by overlaying between dependent variable plots is shown in Figure 3. Each individual point in this optimal region represents an appropriate TE-NPs formulation. In this study, the optimal TE-NPs formulation consisting of 1.3697 %*w*/*w* CF, 1.2970 %*w*/*w* EA, and 0.0067 %*w*/*w* external curcumin was selected. The physicochemical properties of the optimal TE-NPs formulations are shown in Table 3. **Table 3.** Physicochemical properties of the optimal turmeric rhizome extract nanoparticles stored at 5 °C for 3 months (Mean ± SD, *n* = 3). **Dependent Variables Acceptance Limit Initial 3 Months**  CM content (µM) ≥300 µM 357.48 ± 8.39 358.84 ± 4.65 %LA of CM (%LA) 80–120 %LA 92.74 ± 2.18 93.09 ± 1.21 Z-average (nm) ≤200 nm 159.6 ± 1.7 166.6 ± 0.6 d50 (nm) ≤200 nm 169.7 ± 2.1 177.0 ± 1.0 d90 (nm) ≤400 nm 272.3 ± 9.1 276.7 ± 2.5 CM = curcumin. onuclear phagocyte system (MPS) and the reticuloendothelial system (RES). To overcome these effects, the nanoparticles' size must not exceed 400 nm to escape from the MPS effect [67]. Thus, the acceptance upper limit of 200 nm for z-average and d50, and 400 nm for d90 were specified in this study. The optimal region (black area) that was obtained by overlaying between dependent variable plots is shown in Figure 3. Each individual point in this optimal region represents an appropriate TE-NPs formulation. In this study, the optimal TE-NPs formulation consisting of 1.3697 %*w*/*w* CF, 1.2970 %*w*/*w* EA, and 0.0067 %*w*/*w* external curcumin was selected. The physicochemical properties of the optimal TE-NPs formulations are shown in Table 3. **Table 3.** Physicochemical properties of the optimal turmeric rhizome extract nanoparticles stored at 5 °C for 3 months (Mean ± SD, *n* = 3). **Dependent Variables Acceptance Limit Initial 3 Months**  CM content (µM) ≥300 µM 357.48 ± 8.39 358.84 ± 4.65 %LA of CM (%LA) 80–120 %LA 92.74 ± 2.18 93.09 ± 1.21 Z-average (nm) ≤200 nm 159.6 ± 1.7 166.6 ± 0.6 d50 (nm) ≤200 nm 169.7 ± 2.1 177.0 ± 1.0 d90 (nm) ≤400 nm 272.3 ± 9.1 276.7 ± 2.5 CM = curcumin.

specified first. The highest curcumin content obtained in MPV design was 348.67 µM. Thus, the acceptance lower limit of curcumin content of 300 µM was used. The acceptance limit of 80–120% LA for curcumin was chosen [62]. Particle size plays a crucial role in the delivery of nanoparticles to tumor cells. The nanoparticles of appropriate size can be selectively delivered to tumor cells and can escape from the defensive system of body. Angiogenesis in cancer cells results in abnormalities—namely, hypervascularization, aberrant vascular architecture, extensive production of vascular permeability factors stimulating extravasation within tumor tissues, and lack of lymphatic drainage. It allows the passive accumulation of the nanoparticles in tumor tissue, which is known as an enhanced permeability and retention effect (EPR effect) [63–66]. To achieve the extravasation into a tumor by the EPR effect, nanoparticles' size should be below 200 nm [67]. Moreover, nanoparticles must have an appropriate circulation half-life, avoiding the action of the mon-

specified first. The highest curcumin content obtained in MPV design was 348.67 µM. Thus, the acceptance lower limit of curcumin content of 300 µM was used. The acceptance limit of 80–120% LA for curcumin was chosen [62]. Particle size plays a crucial role in the delivery of nanoparticles to tumor cells. The nanoparticles of appropriate size can be selectively delivered to tumor cells and can escape from the defensive system of body. Angiogenesis in cancer cells results in abnormalities—namely, hypervascularization, aberrant vascular architecture, extensive production of vascular permeability factors stimulating extravasation within tumor tissues, and lack of lymphatic drainage. It allows the passive accumulation of the nanoparticles in tumor tissue, which is known as an enhanced permeability and retention effect (EPR effect) [63–66]. To achieve the extravasation into a tumor by the EPR effect, nanoparticles' size should be below 200 nm [67]. Moreover, nanoparticles must have an appropriate circulation half-life, avoiding the action of the mon-

*Molecules* **2022**, *27*, x FOR PEER REVIEW 6 of 16

*Molecules* **2022**, *27*, x FOR PEER REVIEW 6 of 16

**Figure 3.** Overlay plot. Black area = the optimal region; A = HE; B = CF; C = EA; D = external curcumin; and E = sodium hyaluronate (NaHA). Lines represent the TE-NPs formulations with minimum and maximum dependent variable values according to the acceptance limit. The spots (●) represent the design points. **Figure 3.** Overlay plot. Black area = the optimal region; A = HE; B = CF; C = EA; D = external curcumin; and E = sodium hyaluronate (NaHA). Lines represent the TE-NPs formulations with minimum and maximum dependent variable values according to the acceptance limit. The spots ( **Figure 3.** Overlay plot. Black area = the optimal region; A = HE; B = CF; C = EA; D = external curcumin; and E = sodium hyaluronate (NaHA). Lines represent the TE-NPs formulations with minimum and maximum dependent variable values according to the acceptance limit. The spots (●) represent the design points. ) represent the design points.

**Table 3.** Physicochemical properties of the optimal turmeric rhizome extract nanoparticles stored at 5 ◦C for 3 months (Mean ± SD, *n* = 3).


It was found that the optimal TE-NPs had physicochemical properties within the acceptance limits. The optimal TE-NPs were stable for up to 3 months when stored at 5 ◦C. Furthermore, the results show that the optimal TE-NPs had curcumin content higher than TE-NPs prepared from the ethanol extract (EV) obtained directly from turmeric rhizome powder extraction. The TEM study confirmed that the optimal TE-NPs had a spherical shape with a size below 200 nm (Figure 4). Additionally, it was noticed that the TE-NPs nanoparticles had a special structure that looked like a polyp inside the particle (Figure 4B, white arrow). This special structure may be the agglomeration of the solid particles containing the turmeric extract. The result of this study shows that stable TE-NPs formulation containing increased curcumin content was successfully developed. Furthermore, the results show that the optimal TE-NPs had curcumin content higher than TE-NPs prepared from the ethanol extract (EV) obtained directly from turmeric rhizome powder extraction. The TEM study confirmed that the optimal TE-NPs had a spherical shape with a size below 200 nm (Figure 4). Additionally, it was noticed that the TE-NPs nanoparticles had a special structure that looked like a polyp inside the particle (Figure 4B, white arrow). This special structure may be the agglomeration of the solid particles containing the turmeric extract. The result of this study shows that stable TE-NPs formulation containing increased curcumin content was successfully developed.

It was found that the optimal TE-NPs had physicochemical properties within the acceptance limits. The optimal TE-NPs were stable for up to 3 months when stored at 5 °C.

*Molecules* **2022**, *27*, x FOR PEER REVIEW 7 of 16

**Figure 4.** The particle morphology of turmeric rhizome extract nanoparticles of CTOP formulation observed by TEM. (**A**,**B**) = sample coated with 2% uranyl acetate: 8000× and 40,000×, respectively; (**C**,**D**) = sample coated with 2% osmium tetroxide: 8000× and 40,000×, respectively. **Figure 4.** The particle morphology of turmeric rhizome extract nanoparticles of CTOP formulation observed by TEM. (**A**,**B**) = sample coated with 2% uranyl acetate: 8000× and 40,000×, respectively; (**C**,**D**) = sample coated with 2% osmium tetroxide: 8000× and 40,000×, respectively.

#### *2.3. Cytotoxicity of Turmeric Rhizome Extract Nanoparticles in HepG2 Cells 2.3. Cytotoxicity of Turmeric Rhizome Extract Nanoparticles in HepG2 Cells*

The cytotoxicity of TE-NPs compared with free curcumin is shown in Table 4 and Figure 5. It was found that free curcumin and four TE-NPs exhibited a cytotoxicity effect on the human hepatoma HepG2 cells. The IC50 values were 43, 40, 37, 41, and 42 µM for free curcumin, CTEV, CTEVHA, CTOP, and CTOPHA, respectively. Although, the IC50 values for TE-NPs were shown to be slightly lower than those of free curcumin, all TE-NPs formulations showed a significantly stronger inhibition effect than free curcumin at the equivalent curcumin concentrations of 50–100 µM (*p*-value < 0.05). The higher inhibition The cytotoxicity of TE-NPs compared with free curcumin is shown in Table 4 and Figure 5. It was found that free curcumin and four TE-NPs exhibited a cytotoxicity effect on the human hepatoma HepG2 cells. The IC50 values were 43, 40, 37, 41, and 42 µM for free curcumin, CTEV, CTEVHA, CTOP, and CTOPHA, respectively. Although, the IC50 values for TE-NPs were shown to be slightly lower than those of free curcumin, all TE-NPs formulations showed a significantly stronger inhibition effect than free curcumin at the equivalent curcumin concentrations of 50–100 µM (*p*-value < 0.05). The higher inhibition effect of TE-NPs might be due to other compositions contained in the turmeric rhizome in addition to curcumin. These compositions were DCM, BDCM, and turmerone compounds [6,7,9].

**Curcumin Concentration** 

Ethanol was also reported to have inhibition effect on HepG2 cells, with an IC50 value of 3.13 %*v*/*v* [68–72]. In this experiment, ethanol concentrations of TE-NPs samples were in the range of 0.001–1.518 %*v*/*v* (for CTEV and CTEVHA) and 0.001–0.983 %*v*/*v* (for CTOP and CTOPHA), depending on the equivalent curcumin concentration in each formulation. These maximum levels of ethanol contained in TE-NPs samples were 2–3 times lower than IC50. Therefore, it can be assumed that ethanol has a negligible inhibition effect on HepG2 cells. In addition, it was shown that the treatment using the developed optimal formulations (CTOP, and CTOPHA) at equivalent curcumin concentration of 50–100 µM inhibited the proliferation of HepG2 better than CTEV and CTEVHA (*p* value < 0.05). To investigate the enhancing effect of sodium hyaluronate, TE-NPs formulations with sodium hyaluronate coating were developed and tested for their inhibitory effect to HepG2 cells. The results show that for CTEVHA and CTOPHA formulations, at 50 µM equivalent curcumin concentration, the sodium hyaluronate coated TE-NPs formulations showed significantly higher inhibitory effects than the uncoated formulations (*p* value < 0.05). This indicates that sodium hyaluronate coating may enhance the inhibition effect of TE-NPs toward HepG2 cells. This enhancing effect is dose-dependent. **Table 4.** Cytotoxicity study of selected turmeric rhizome extract nanoparticles by MTT assay (Mean

hibited the proliferation of HepG2 better than CTEV and CTEVHA (*p* value < 0.05).

effect of TE-NPs might be due to other compositions contained in the turmeric rhizome in addition to curcumin. These compositions were DCM, BDCM, and turmerone compounds [6,7,9]. Ethanol was also reported to have inhibition effect on HepG2 cells, with an IC50 value of 3.13 %*v*/*v* [68–72]. In this experiment, ethanol concentrations of TE-NPs samples were in the range of 0.001–1.518 %*v*/*v* (for CTEV and CTEVHA) and 0.001–0.983 %*v*/*v* (for CTOP and CTOPHA), depending on the equivalent curcumin concentration in each formulation. These maximum levels of ethanol contained in TE-NPs samples were 2–3 times lower than IC50. Therefore, it can be assumed that ethanol has a negligible inhibition effect on HepG2 cells. In addition, it was shown that the treatment using the developed optimal formulations (CTOP, and CTOPHA) at equivalent curcumin concentration of 50–100 µM in-

**Table 4.** Cytotoxicity study of selected turmeric rhizome extract nanoparticles by MTT assay (Mean ± SD, *n* = 3). **% Cell Viability** 


± SD, *n* = 3).

*Molecules* **2022**, *27*, x FOR PEER REVIEW 8 of 16

**Figure 5.** Cytotoxicity of turmeric rhizome extract nanoparticles quantified by MTT assay. The data are expressed as the percentage relative to the vehicle (control) and are shown as mean ± SD (*n* = 3). CTEVHA and CTEV = TE-NPs prepared from ethanol extract (EV) with and without sodium hyaluronate (NaHA), respectively. CTOPHA and CTOP = TE-NPs prepared from optimal formulation with and without sodium hyaluronate (NaHA), respectively. \*, *p* < 0.05 versus free curcumin. **Figure 5.** Cytotoxicity of turmeric rhizome extract nanoparticles quantified by MTT assay. The data are expressed as the percentage relative to the vehicle (control) and are shown as mean ± SD (*n* = 3). CTEVHA and CTEV = TE-NPs prepared from ethanol extract (EV) with and without sodium hyaluronate (NaHA), respectively. CTOPHA and CTOP = TE-NPs prepared from optimal formulation with and without sodium hyaluronate (NaHA), respectively. \*, *p* < 0.05 versus free curcumin.

To investigate the enhancing effect of sodium hyaluronate, TE-NPs formulations with sodium hyaluronate coating were developed and tested for their inhibitory effect to HepG2 cells. The results show that for CTEVHA and CTOPHA formulations, at 50 µM equivalent curcumin concentration, the sodium hyaluronate coated TE-NPs formulations showed significantly higher inhibitory effects than the uncoated formulations (*p* value < 0.05). This indicates that sodium hyaluronate coating may enhance the inhibition effect of TE-NPs toward HepG2 cells. This enhancing effect is dose-dependent.

#### **3. Materials and Methods**

#### *3.1. Materials*

Turmeric (*Curcuma longa* Linn., Zingiberaceae) rhizome powder was purchased from a medicinal herb store in Bangkok, Thailand. Plant sample was identified by Dr. Pongtip Sithisarn, Department of Pharmacognosy, Faculty of Pharmacy, Mahidol University, Bangkok, Thailand. Curcumin 98% was purchased from AK Scientific, Union city, CA, USA. Turmeric oil was purchased Thai-China Flavours and Fragrances Industry Co. Ltd., Phra Nakhon Si Ayutthaya, Thailand. Sodium hyaluronate was purchased from Bloomage Freda Biopharm, Jinan, China. Curcumin, desmethoxycurcumin, and bisdesmethoxycurcumin standards were purchased from USP, Rockville, MD, USA. Analytical HPLC grade acetonitrile was purchased from Scharlab S.L., Barcelona, Spain. Absolute ethanol, hexane, chloroform, and n-butanol were purchased from RCI Labscan Limited, Bangkok, Thailand. Ethyl acetate was purchased from J.T. Baker, Phillipsburg, NJ, USA. The 95% ethanol was purchased from The Liquor Distillery Organization, Chachoengsao, Thailand. Water for Injection was purchased from A.N.B. Laboratories Co., Ltd., Bangkok, Thailand. Benzene was purchased from Panreac quimica SA, Barcelona, Spain. Dimethyl sulfoxide (DMSO) ≥99.5%, 3-[4,5 dimethylthiazole-2-yl]-2,5-diphenyl tetrazolium bromide (MTT) dye, phosphomolybdic acid hydrate, and Dulbecco's Modified Eagle Medium (DMEM) were purchased from Sigma-Aldrich, St. Louis, MO, USA. Phosphate-buffered solution pH 7.4 and fetal bovine serum (FBS) were purchased from JR Scientific, Inc., Woodland, CA, USA. Penicillin streptomycin solution was purchased from Life Technologies, Carlsbad, CA, USA.

#### *3.2. Preparation of Turmeric Rhizome Fraction*

Turmeric powder (200 g) was mixed with 95% ethanol (600 g). After being kept at room temperature for 48 h, the mixture was filtered through filter paper (Whatman no. 2) and nylon filter pore size 0.45 µm, consecutively. The filtrate was dried at 50 ◦C. The ethanol extract (EV) was dispersed in water with weight to volume ratio of EV/water of 1:10. The mixture was sonicated for 10 min. Solvent–solvent extraction process was conducted using four solvents including hexane, chloroform, ethyl acetate, and n-butanol. The ratio of solvent/EV aqueous dispersion used was 1:1 by volume. First, hexane was added into EV aqueous dispersion, stirred for 30 min, and left at room temperature until the aqueous phase was completely separated from the hexane phase. Then the hexane phase was withdrawn. Hexane extraction was repeated with the remaining turmeric aqueous dispersion for two times. Three collected parts of hexane phase were combined and dried at 50 ◦C by using a rotary evaporator model Buchi Rota vapor R200 (BÜCHI Labortechnik AG, Flawil, Switzerland). The remaining turmeric aqueous dispersion was further extracted using chloroform with the extraction procedure exactly the same as that described above for hexane. The remaining turmeric aqueous dispersion was further extracted by using ethyl acetate and then n-butanol, consecutively. The dried turmeric rhizome fractions obtained from solvent extraction—i.e., ethanol extract (EV), hexane fraction (HE), chloroform fraction (CF), ethyl acetate fraction (EA), n-butanol fraction (BU), and the remaining aqueous fraction (AQ)—were weighed and dissolved in 95% ethanol to obtain the final concentration of 2.5 %*w*/*w*. The fractions in ethanol were prepared and kept in a glass bottle with screw cap, stored at 5 ◦C, and protected from light until use.

#### *3.3. Characterization and Curcuminoids Content Analysis of Turmeric Rhizome Extracts*

Phytochemical analysis of turmeric rhizome fractions was carried out by using thin layer chromatography (TLC). Five microliters of turmeric rhizome fractions and 0.05 %*w*/*v* standard curcumin, desmethoxycurcumin, and bisdesmethoxycurcumin were separately spotted on a TLC plate (silica gel GF254). Benzene/chloroform/ethanol 49:49:2 by volume was used as a solvent system. The solvent front was 8 cm. After development, the TLC plate was examined under UV at the wavelengths of 254 nm and 366 nm in a UV chamber. Then, the TLC plate was sprayed with 10% phosphomolybdic acid spray reagent and heated at 105 ◦C for 5 min. The hRf values of the samples were calculated and compared

with hRf values of the standards and hRf values specified in Thai Herbal Pharmacopoeia 2020, Volume 1 [56].

Curcuminoid content of turmeric rhizome fractions (curcumin, desmethoxycurcumin, and bisdesmethoxycurcumin) were quantitatively analyzed using a modified HPLC method developed by Wichitnithad W and coworkers (2009) [57]. A Shimadzu-VP system equipped with a SCL-10A VP controller, a LC-10AD VP pump, a SIL-10AD VP auto- injector, a DGU-14 degasser, an SPD-10A VP UV-VIS detector, and Shimadzu CLASS-VP software (Shimadzo corporation, Kyoto, Japan) were used together with Hypersil GOLD C18 column (250 × 4.6 mm i.d.; 5 µm, Thermo Fisher Scientific Inc., Waltham, MA, USA). A reversephase HPLC analysis was carried out by using an isocratic system with 1% acetic acid and acetonitrile at the volume ratio of 61:39 as a mobile phase at a flow rate of 1.2 mL/min. The injection volume was 10 µL and analytical time was 25 min. A detection wavelength of 425 nm was used. The method that was used was validated for accuracy, precision, specificity, linearity, and sensitivity.

#### *3.4. Preparation of Turmeric Rhizome Extract Nanoparticles*

Turmeric rhizome extract nanoparticles (TE-NPs) were spontaneously formed by the solvent displacement method called the Ouzo effect, which was first named and described by Vitale and Katz in 2003 [47]. The formulations consisted of 2.5 %*w*/*w* turmeric rhizome fraction in ethanol solution, Water for Injection (WFI), external curcumin (extCM), and/or sodium hyaluronate (NaHA). By using a syringe with a 25G needle, turmeric rhizome fraction in ethanol solution at certain quantity by weight was dropped into WFI at the rate of 60 drops per minute. The mixture was continuously stirred at 400 rpm for 10 min. For cases where extCM was used, it was completely dissolved in turmeric rhizome fraction in ethanol solution by sonication for 10 min before subsequent nanoparticles formation. For cases where NaHA was added, 0.1 %*w*/*w* NaHA aqueous solution was dropped into the TE-NPs dispersion. The TE-NPs dispersion was continuously stirred at 400 rpm for 10 min.

### *3.5. Characterization and Curcuminoids Content Analysis of Turmeric Rhizome Extract Nanoparticles*

The particle characteristics (z-average, d50, d90, derived count rate, and PDI) and zeta potential values of TE-NPs were studied by dynamic light scattering technique (DLS) using Zetasizer Nano ZS (Malvern Instruments, Worcestershire, UK). The measurement was set at equilibrated time of 2 min, 173◦ detection optics backscatter detection, numbers of run 10 times, run duration 10 s, and the measurement was carried out in triplicates. The refractive index required for size measurement was determined by Abbe Refractometer NAR-3T (Atago, Tokyo, Japan) set at 25 ◦C and the wavelength of 589 nm. The refractive index of turmeric rhizome fraction of 1.38 was used. Curcuminoids content in TE-NPs were quantitatively analyzed using a validated modified HPLC method, described above.

#### *3.6. Determination of the Optimal Turmeric Rhizome Extract Nanoparticles Formulation*

To determine the optimal TE-NPs formulation, the regression model was constructed by using the mixture–process variables experimental (MPV) design [73]. The quantity of 2.5 %*w*/*w* HE, CF, and EA ethanol solutions were selected as mixture components. The quantity of external curcumin (extCM) and 0.1 %*w*/*w* sodium hyaluronate (NaHA) aqueous solution were selected as process variables. Physicochemical properties of TE-NPs—namely, curcuminoid contents, %LA of curcuminoids, and particle characteristics—were dependent variables. The ranges of actual and coded mixture components and process variables are shown in Table 5 The 47-run MPV design was generated by Design-Expert 9 (Stat-Ease, Inc., Minneapolis, MN, USA) using the best optimal design algorithm with D-optimality criterion. The basis for 47 runs was the 36 terms in the MPV model, 5 extra-points to assess model lack-of-fit, 5 replicated points, and 1 additional center point. TE-NPs were prepared by the method described above.


**Table 5.** Variables used in MPV design.

HE = hexane fraction; CF = chloroform fraction; EA = ethyl acetate fraction; CM = curcumin; and NaHA = sodium hyaluronate.

The physicochemical properties of TE-NPs were measured, and the results were used to construct regression models by ANOVA with backward elimination regression at alpha of 0.05. The predictability of regression models was validated by an external data set of 10 formulations that were not included in the MPV design data set. To demonstrate the predictability, the predictive root mean square error (predictive RMSE) and predictive r<sup>2</sup> were calculated according to the following equations [74]:

$$\text{predicted RMSE} = \sqrt{\frac{\sum \left( \mathbf{y}\_{\text{experimental}} - \mathbf{y}\_{\text{predicted}} \right)^{2}}{\mathbf{N}}} \tag{6}$$

$$\text{predicted } \mathbf{r}^2 = 1 - \frac{\sum \left( \mathbf{y}\_{\text{experimental}} - \mathbf{y}\_{\text{predicted}} \right)^2}{\sum \left( \mathbf{y}\_{\text{experimental}} - \mathbf{y}\_{\text{mean}} \right)^2} \tag{7}$$

where yexperimental is the dependent variable value obtained from the experiment, ymean is an average dependent variable of the results obtained from the experiment, ypredicted is the dependent variable value obtained from the regression model, and N is total number of experimental points.

To determine the optimal TE-NPs formulation, the contour plots were constructed with the acceptance ranges of the desired dependent variables and overlayed by Design Expert 9. The overlapping area was an optimal region. Each individual point in this optimal region represented an appropriate formulation. The optimal formulation could be selected from this region.

#### *3.7. Preparation, Characterization, and Cytotoxicity Test of the Optimal Turmeric Rhizome Extract Nanoparticles*

The optimal turmeric rhizome extract nanoparticles formulation (CTOP) obtained from MPV design was prepared by the method described above with aseptic techniques. The physicochemical properties and particle morphology of CTOP were determined. The particle morphology was observed by using a transmission electron microscope (TEM) model Hitachi HT7700 (Hitachi High-Technologies Corporation, Tokyo, Japan) with accelerated voltage preset of 80 kV. Two techniques were used to prepare the sample. First, the TE-NPs dispersion was dropped on a paraffin film. A Formvar-coated copper grid was gently placed on the drop for 2 min, removed, stained by 2% uranyl acetate solution, and left overnight in desiccator before study. Second, the grid was placed on the drop of the TE-NPs dispersion on the paraffin film, removed, and left overnight in desiccator saturated with vapor of osmium tetroxide before study.

The cytotoxicity test using MTT assay was carried out for the optimal TE-NPs formulation in comparison with free curcumin and the selected TE-NPs formulations.

#### 3.7.1. Cell Culture

Human hepatoma (HepG2) cells (catalog number HB-8065 from American Type Culture Collection; ATCC) were maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1:100 penicillin/streptomycin (10,000 units/mL) at 37 ◦C and 5% CO2.

### 3.7.2. MTT Assay (Cytotoxicity Assay)

HepG2 cells (1 <sup>×</sup> <sup>10</sup><sup>4</sup> cells/well) were cultured in 96-well plates, in 200 µL of DMEM supplemented with 10% FBS and 1% streptomycin/penicillin solution (Gibco) and incubated at 37 ◦C and 5% CO<sup>2</sup> humidified atmosphere for overnight as previously described [75]. Cells were treated with solubilized free curcumin in dimethyl sulfoxide (DMSO) and TE-NPs at the final equivalent curcumin concentrations of 0.1, 1, 10, 25, 50, and 100 µM. DMSO 1 %*v*/*v* was used as a control. This study was performed in triplicate, with two replicate wells. The relative number of viable cells was determined after 24 h incubation by adding 1 mg/mL of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT) and incubating the cells for a further 4 h. The formazan crystals formed were dissolved with DMSO. The absorbance values of the solution at wavelength of 570 nm, which was directly relative to the viable cells, were determined using an Infinite M200 microplate reader (Tecan Sales Austria GmbH, Grödig, Austria). The percentage of cell viability was calculated as follows:

$$\% \text{Cell viability} = \frac{\text{Absorbance of treated cells}}{\text{Absorbance of control cells}} \times 100\tag{8}$$

#### *3.8. Statistical Analysis*

The data obtained are expressed as mean ± standard deviation (SD) of triplicates. Unpaired t-test or one-way ANOVA was used to compare means (α = 0.05). All analyses were performed using PASW Statistics for Windows, version 18.0 (SPSS Inc., Chicago, IL, USA).

#### **4. Conclusions**

Nanoparticles were successfully prepared from turmeric rhizome fractions in this study. By applying 47-run D-optimal mixture–process variables experimental design, the appropriate formulation of stable nanoparticles was obtained. The optimal TE-NPs formulation had physicochemical properties within the acceptance limits after a 3-month storage period. In addition, regression models with good predictability for three desired dependent variables including z-average, d50, and d<sup>90</sup> could be determined. The results from a cytotoxicity study using human hepatoma HepG2 cells show that the optimal TE-NPs had stronger cytotoxic effects than free curcumin. Thus, optimal TE-NPs could be successfully developed by using the mixture–process variables experimental design. It was found that the aqueous solubility of curcumin was increased in the optimal formulation system. Moreover, the addition of curcumin in the TE-NPs formulation might be applicable for cases where there is a biological variation of curcuminoids content. It should be noted that an inhibition effect of TE-NPs was found in an in vitro experiment using HepG2 cells. Further clinical study should be performed to confirm this result.

**Supplementary Materials:** The following supporting information can be downloaded online, Table S1: Physicochemical properties of turmeric rhizome extract nanoparticles formulation obtained from mixture—process variables experimental design.

**Author Contributions:** Conceptualization, P.P.L. and S.A.; methodology, P.P.L., S.P., P.S., S.M. and S.A.; software, S.A.; formal analysis, S.A.; investigation, S.A.; resources, P.S., S.M. and S.A.; data curation, S.A.; writing—original draft preparation, S.A.; writing—review and editing, P.P.L., S.P., P.S. and S.M.; visualization, S.A.; supervision, P.P.L.; project administration, P.P.L. and S.A.; funding acquisition, P.P.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available in the article.

**Acknowledgments:** A special acknowledgement is extended to the Faculty of Pharmacy, Mahidol University, Thailand, for providing research facilities. It is also extended to Silom Medical Co., LTD., Thailand, for Ph.D. scholarship provided to Sakchai Auychaipornlert.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples of the compounds are not available from the authors.

#### **References**


## *Article* **Pharmaceutical Development and Design of Thermosensitive Liposomes Based on the QbD Approach**

**Dorina Gabriella Dobó 1,\* , Zsófia Németh <sup>1</sup> , Bence Sipos <sup>1</sup> , Martin Cseh <sup>1</sup> , Edina Pallagi <sup>1</sup> , Dániel Berkesi <sup>2</sup> , Gábor Kozma <sup>2</sup> , Zoltán Kónya <sup>2</sup> and Ildikó Csóka <sup>1</sup>**


**Abstract:** This study aimed to produce thermosensitive liposomes (TSL) by applying the quality by design (QbD) concept. In this paper, our research group collected and studied the parameters that significantly impact the quality of the liposomal product. Thermosensitive liposomes are vesicles used as drug delivery systems that release the active pharmaceutical ingredient in a targeted way at ~40–42 ◦C, i.e., in local hyperthermia. This study aimed to manufacture thermosensitive liposomes with a diameter of approximately 100 nm. The first TSLs were made from DPPC (1,2-dipalmitoyl-snglycerol-3-phosphocholine) and DSPC (1,2-dioctadecanoyl-sn-glycero-3-phosphocholine) phospholipids. Studies showed that the application of different types and ratios of lipids influences the thermal properties of liposomes. In this research, we made thermosensitive liposomes using a PEGylated lipid besides the previously mentioned phospholipids with the thin-film hydration method.

**Keywords:** quality by design; quality planning; initial risk assessment; critical factors; thermosensitive liposomes; thin-film hydration method

### **1. Introduction**

Liposomes are spherical formations, vesicles located by a membrane bilayer, in the nanometre size range [1]. The first liposomes were made by Bangham et al. in 1965 [2]. The inner phase of liposomes is separated from the dispersed media by a double lipid layer. The size of these vesicles can range from 20 nm to some µm. The thickness of the membrane is around 4 nm [1]. Phospholipids are important components of liposomes that, due to their hydrophobic tails and hydrophilic heads, are able to form a double membrane layer. One of their crucial pharmaceutical advantages is that these vesicles can incorporate hydrophobic and hydrophilic active pharmaceutical ingredients (APIs) as well. The hydrophilic drug is found in the hydrophilic core of the liposomes, while the lipophilic API is in the wall of the vesicles [1]. Liposomes may contain more than just a conventional drug. Recently, there has been an increase in the use of gene therapies, specifically in tumour therapy and the treatment of congenital genetic diseases [3–7]. Liposomes are also very useful in this area, as they can deliver genes to cells. In addition to the pharmaceutical industry, they are also widely researched and applied by the cosmetics and food industries. Liposomes can be produced in numerous ways. The thin-film hydration technique, also known as the Bangham method—named after Alec D. Bangham, the developer of the practice [2]—is one of the most frequently used preparation methods. This method is still widely used for the typical production of non-phospholipid/phospholipid vesicles located in a membrane bilayer and formed by supramolecular assembly [8–11]. Its first step is to solve the lipids of the membrane and the hydrophobic API in an organic solvent, which is later removed.

**Citation:** Dobó, D.G.; Németh, Z.; Sipos, B.; Cseh, M.; Pallagi, E.; Berkesi, D.; Kozma, G.; Kónya, Z.; Csóka, I. Pharmaceutical Development and Design of Thermosensitive Liposomes Based on the QbD Approach. *Molecules* **2022**, *27*, 1536. https://doi.org/10.3390/ molecules27051536

Academic Editor: Faiyaz Shakeel

Received: 28 January 2022 Accepted: 23 February 2022 Published: 24 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

This elimination can happen via lyophilisation or vacuum evaporation. The hydration of the film follows this step. The most often used hydration media are aqueous solutions: buffers, physiological saline solution, or solutions of carbohydrates, etc. This hydration step always has to be performed above the phase transition temperature (Tm) of the lipids [12]. Liposomes are used broadly; their application in tumour therapy started at the beginning of the 1990s [13]. Numerous types of liposomes are known. One of them is grouped according to the properties of the ligands bound to their surface by the liposomes. Examples are conventional, PEGylated, immunoliposomes and bioresponsive liposomes. Bioresponsive liposomes include thermosensitive liposomes. Thermosensitive liposomes (TSL) contain phospholipids with very different Tm values. The membrane of the liposome transforms from its solid, gel-like form to a highly permeable form at Tm, whose state is typical for hydrophilic materials [14,15]. A more effective way to target the active ingredient of liposomes is to attach signalling and targeting molecules to the liposome (antibodies, peptides, oligosaccharides, etc.). They bind to the recognised specific target molecule located either on the cancer cells themselves or on the surface of the endothelium of the altered vascular system that encloses the tumour. The only drawback is the heterogeneity of the tumours. A variety of antigens can be expressed by a particular tumour, and it is not certain that a particular antibody will target what would be required in a given therapy. Thermosensitive liposomes release their active components in the appropriate area under the influence of energy transfer or a biological signal. Examples of such signals include temperature change, pH change, light exposure, and ultrasound [14]. Thermosensitive liposomes are vesicles used for drug delivery that release their API in a targeted way at ~40–42 ◦C, i.e., in local hyperthermia [14]. TSLs can be further grouped as low-temperature-sensitive liposomes (LTSL) and high-temperature-sensitive liposomes (HTSL). LTSLs release the incorporated API above the physiological body temperature (37 ◦C), while in the case of the HTSLs, it happens at even higher temperatures. The API release from LTSLs is induced by the hyperthermic area formed due to the extra heat originating from the increased metabolic processes of the tumour tissue. In the case of HTSLs, this release happens at a place which is heated by an external heat source. The energy arising from the external source may increase the temperature even higher than the tumour's original value [16]. TSLs are more selective and stable than traditional (natural phospholipid-containing) liposomes. With the help of TSLs, 20–30 times greater API concentrations can be achieved in the targeted tissues than with the free drugs [17], and 5–10 times greater than with traditional liposomes [18]. Thus, for various diseases, local therapy can be performed (i.e., tumour therapy), which can decrease the toxicity of the procedure in parallel. It is favourable for the patient if the therapy is targeted because, in this way, the applied API does not damage the healthy areas. The first TSLs were made from DPPC (1,2-dipalmitoylsn-glycero-3-phosphocholine) and DSPC (1,2-dioctadecanoyl-sn-glycero-3-phosphocholine) phospholipids (DPPC:DSPC = 7:3) [19]. Studies showed that the application of different types and ratios of lipids influences the thermal properties of liposomes [14]. Later research demonstrated that the addition of PEGylated phospholipids to the lipid mixture influences the size [20] and the thermosensitivity of the liposomes [12,13].

Since the beginning of the 2000s, the quality by design (QbD) method has been taking the place of the 'in-process study'-based process and quality control system (quality by testing = QbT) in the pharmaceutical industry. The QbD concept is a knowledge- and risk- assessment-based quality management approach, applied principally in the industry [21–24], but it can also be extended and applied in the early pharmaceutical research and development (R&D) phase [24]. The essential elements in a QbD approach are the following: determining the quality target product profile (QTPP), selecting the critical quality attributes (CQAs) and critical process parameters (CPPs), performing the risk assessment (RA), design of experiments (DoE), developing a design space (DS) with a proper control strategy, and finally managing the product lifecycle based on aspects of continuous improvement. These steps and definitions are presented in the relevant documents of the International Council for Harmonisation of Technical Requirements for Pharmaceuticals

for Human Use (ICH) [17–19,22]. According to the ICH descriptions, CQAs are related to the quality, safety, and efficacy profile of the product. In contrast, critical material attributes (CMAs) and CPPs are connected to the selected production method. By performing RA using a risk estimation matrix, we can obtain the ranking of CQAs, CMAs, and CPPs by the degree of their impact on the targeted product quality. The factors that have the highest critical impact on the final product have to be the focus of the development process. These should be the key elements in the factorial design of the experimental work. The QbD-guided procedure in development and industrial manufacturing provides more information and knowledge about the final product and the manufacturing process and has advantages in marketing authorisation procedures [22,23]. Risk factors regarding the development of liposomes have already been analysed from different viewpoints by researchers worldwide. A general overview of the QbD approach for liposome formulation without a defined preparation method completed with characterisation techniques was provided by Porfire et al. [25]. Xu et al. analysed the size, encapsulation efficiency, and stability-affecting factors regarding the formulation, process, analytical method, and instrumentation reliability for liposomes prepared using the thin-film hydration method [26]. Although examples for quality analysis can be found in the literature [27], there have been no cases concerning application of the R&D QbD model on TSLs. Due to the emerging need for novel drug delivery systems capable of responding to the variety of pH, ionic, enzymatic, and thermal changes in the human body, quality management is a highly recommended factor in the development process [28–30]. Thus, the novelty of this work is the extension of the QbD approach to thermosensitive liposomes based on the previous statements of our research group in a risk-assessment-based development proposal for liposomes [31].

Our research team has been working on lipid-based nanoparticles for years. The aim of this study was to produce stable TSL liposomes with an average diameter of 100–200 nm using a QbD-based experimental design. Concerning the QbD assay, quality management tools such as editing an Ishikawa diagram and performing a risk assessment process were applied.

During the work, various wall-forming components (DPPC, DSPC, and DSPE-PEG3000 (1,2-distearoyl-sn-glycero-3-phosphoethanolamine-*N*-[amino(polyethylene glycol)-3000]), solvents (ethanol, methanol, and chloroform), and hydration media (phosphate-buffered saline solutions of different pHs and isotonic saline solution) were used to influence thermosensitive properties and particle size, characterised by material analysis methods using differential scanning calorimetry (DSC), thermogravimetry (TGA), dynamic light scattering (DLS), transmission electron microscopy (TEM), Fourier-transform infrared spectroscopy (FT-IR), and atomic force microscopy (AFM) techniques.

#### **2. Materials and Methods**

#### *2.1. Materials*

Two different compositions and two hydration media were used in this study to form thermosensitive liposomes.

The following phospholipids were used: DPPC-1,2-dipalmitoyl-sn-glycero-3-phosphocholine (Avanti Polar Lipids, Alabaster, AL, USA), DSPC-1,2-dioctadecanoyl-sn-glycero-3-phosphocholine (Avanti Polar Lipids, Alabaster, AL, USA), and DSPE-PEG3000-1,2-distearoyl-sn-glycero-3 phosphoethanolamine-*N*-[methoxy(polyethylene glycol)-3000] (Avanti Polar Lipids, Alabaster, AL, USA), solved in ethanol 96% (Molar Chemicals Kft., Budapest, Hungary). The excipients were used in different ratios (Table 1).

Phosphate buffer solution pH 7.4 (PBS pH 7.4) and sodium chloride physiological solution (saline solution) pH 5.5 [32] were used as hydration media. The composition of these solutions was as follows: PBS pH 7.4: 8.0 g/L NaCl, 0.20 g/L KCl, 1.44 g/L Na2HPO<sup>4</sup> × 2 H2O, 0.12 g/L KH2PO4; saline solution: 0.9 g/L NaCl dissolved in distilled water. The materials used to make these hydration media were the following: sodium chloride (NaCl) (Molar Chemicals Ltd., Budapest, Hungary), potassium chloride (KCl) (Molar Chemicals Ltd., Budapest, Hungary), disodium hydrogen phosphate dihydrate

(Na2HPO<sup>4</sup> × 2 H2O) (Spektrum-3D Ltd., Debrecen, Hungary), and dipotassium phosphate (K2HPO4) (Spektrum-3D Ltd., Debrecen, Hungary).

None of the formulations contained active pharmaceutical ingredients (API).



#### *2.2. Methods*

2.2.1. Elements of the QbD Design

Development of the Knowledge Space and Determination of the QTPP, CQAs, CMAs, and CPPs

In order to perform the initial RA, the first step was the collection of all the relevant influencing factors of the desired liposomal formulation. This collection and systemic evaluation are parts of the knowledge space development [33]. To evaluate the cause and effect relationships between the factors, an Ishikawa diagram [34] was set up, which helps determine the QTPP elements and identify the critical factors. QTPP is related to the quality, safety, and efficacy of the product, considering, e.g., the route of administration, dosage form, bioavailability, strength, stability, etc. [33,35]. In this study, the following was chosen as the QTPP: a nanosized, stable, thermo-responsive (thermosensitive) liposomal formulation for nasal administration capable of reaching the CNS.

CQAs are the physical, chemical, biological, or microbiological properties or characteristics that should be within an appropriate limit, range, or distribution to ensure the desired product quality, derived from the QTPPs and/or prior knowledge, and always dependent upon predefined goals. Other critical factors can be linked to the materials used, these are the CMAs, and the process factors whose variability has an impact on the quality of the final product are the CPPs [35]. In this experiment, the following CQAs were identified: the zeta potential, particle size, and morphology of the liposomes and their phase transition temperature (Tm). The group of the CMAs/CPPs in this study includes the preparation of the mixture, dissolution of the lipids, hydration medium, and the method of the stabilisation (drying/lyophilisation) process.

#### 2.2.2. Risk Assessment

The RA procedure was performed using the LeanQbD® software (QbD Works LLC, Fremont, CA, USA, www.qbdworks.com accessed on 18 January 2022). The first element of this procedure was the interdependence rating between the QTPPs and the CQAs, followed by the same procedure between the CQAs and the CMAs/CPPs. A three-level scale was used to describe the relationship between the parameters: 'high' (H), 'medium' (M), or 'low' (L). Then, a risk occurrence rating of the CMAs/CPPs (or probability rating step) was made, and the same three-grade scale (H/M/L) was applied for the analysis. As the output of the initial RA evaluation, Pareto diagrams [36] were generated by the software, which presented the numeric data and the ranking of the CQAs and the CMAs/CPPs according to their potential impact on the desired final product (QTPP). The Pareto charts not only show the differences of the CQAs, CMAs, and CPPs by their effect but they also help to select the factors of a potential experimental design. A relative occurrence–relative severity chart was set up to visualise the severity of the CPP/CMA elements compared to each other.

#### *2.3. Design of Experiments (DoE)*

The DoE was made according to the results of the initial RA. A "two-level 3-factor full factorial design" model was applied for the design of the experimental studies. 2 <sup>3</sup> = 8 experiments were performed with this model to determine the relationship between the selected factors and the quality of the final liposomal product. The factors applied in the factorial design were those which showed the highest critical effect on the desired product according to the RA. These are: the ratio of the phospholipids building up the liposomes: DPPC:DSPC:DSPE-PEG3000 (+1 value: 80:15:5 n/n%; −1 value: 70:25:5 n/n%); the type of the hydration medium (+1 value: physiological saline solution; −1 value: pH = 7.4 phosphate-buffered saline); and the method of physical stabilisation (+1 value: lyophilisation; −1 value: vacuum drying). The effect of these factors was measured by testing the character of the CQAs (zeta potential, particle size, morphology of the liposomes, and Tm) as a response.

#### *2.4. Preparation of Thermosensitive Liposomes*

The liposomes were prepared from DPPC, DSPC, and DSPE-PEG3000 phospholipids with the thin-film hydration method. As the first step, 2.5 mg/mL stock solutions were prepared from the lipids with ethanol (96% EtOH-Molar Chemicals). The prepared stock solutions were mixed according to the defined mole ratios of the lipids (Table 1). After that, the solvent was removed via vacuum evaporation (Büchi Rotavapor: 60 ◦C, 25 rpm) of the ethanol. This process resulted in the formation of a lipid film layer on the wall of the round-bottom flask. During the hydration step, the appropriate medium (pH = 7.4 PBS (phosphate-buffered saline) (self-prepared) or 0.9% NaCl solution-physiological saline (selfprepared)) was poured on the top of the film; then the flask was placed into a thermostated ultrasonic bath (60 ◦C, 30 min) to form the liposomes. To improve the size and size distribution values of the liposomes, the ultrasonic treatment was followed by membrane filtration. The shaping of the liposomes was performed in two steps via vacuum membrane filtration using a 0.45 µm (nylon membrane disk filter 47 mm, Labsystem Ltd., Budapest, Hungary), then a 0.22 µm membrane filter (Ultipor® N66 nylon 6.6 membrane disk filter 47 mm, Pall Corporation, New York, NY, USA), while the vacuum was created by a vacuum pump (Rocker 400 oil-free vacuum pump, Rocker Scientific Co., Ltd. New Taipei City, Taiwan). The prepared liposome samples were immediately investigated for vesicle size, polydispersity, and zeta potential and then lyophilised for stability purposes. The samples were conserved by lyophilisation with a SanVac CoolSafe freeze dryer (LaboGeneTM Lillerod, Lillerød, Denmark). First, 1 or 2 mL of the samples was frozen at normal atmospheric pressure, gradually decreasing the temperature from +25 ◦C to −40 ◦C. The vacuum was created when the temperature of the samples reached the desired value, reducing the pressure to 0.01 atmosphere, where the samples were stored for 8–10 h. After this period, the temperature of the tray was increased manually, step by step, from −40 ◦C to +30 ◦C until the pressure reached the normal atmosphere.

The applied lipid compositions (mole ration of lipids) and hydration media are shown in Table 1.

#### *2.5. Characterisation of the Liposomes*

#### 2.5.1. Vesicle Size and Zeta Potential of the Liposomes

The dynamic light scattering (DLS) technique, which describes the size distribution of the vesicles, was used to determine the size and the polydispersity index of the liposomes. For a measurement, 1 mL of liposome suspension was used and then diluted. Zeta potential is the potential difference between the dispersion medium and the liquid layer adsorbed on the surface of the particles, which can be used—among other things—to investigate the stability of a suspension. Low absolute zeta potential values predict the aggregation of the dispersed particles, while high values show major repulsion between the particles. Suspensions are considered stable in the latter case. The measurements were taken three

times for each sample. Our measurements were performed with a Malvern Zetasizer Nano ZS device.

### 2.5.2. Differential Scanning Calorimetry (DSC) and Thermogravimetric Analysis (TGA) Investigations

During the thermogravimetric measurements, the samples were heated to the defined temperature in the given atmosphere and investigated for mass changes. Our measurements were carried out in an oxygen atmosphere in the temperature range of 25–70 ◦C with a Mettler Toledo TGA/DSC1 STARe System apparatus at a heating rate of 10 ◦C/min. The measurements were carried out on ~5–10 mg of the samples in all cases.

To determine the thermosensitivity of the liposomes, their size was checked, the samples were kept above the expectable phase transition temperature for a while, and then cooled back to room temperature and tested for size changes. There is a further, simpler, and faster method to determine the thermosensitive characteristic—differential scanning calorimetry (DSC) measurement. These investigations were performed in the temperature range of 25–70 ◦C on ~5 mg of the samples at a heating rate of 10 ◦C/min. The measurements were performed three times for each sample.

#### 2.5.3. Transmission Electron Microscopy (TEM) Measurements

The size, structure, and morphology of the liposomes were characterised by transmission electron microscopy (TEM) measurements. The TEM images were made with an FEI Tecnai G2 X-Twin HRTEM microscope (FEI, Hillsboro, OR, USA) with an accelerating voltage of 200 kV. Suspensions were prepared from the formulations with ethanol and then spread onto a copper grid coated with a 3 mm diameter carbon film. For particle size and distribution analysis, public domain image analyser software—ImageJ—was used (https://imagej.nih.gov/ij/index.html accessed on 20 January 2022).

#### 2.5.4. Fourier-Transform Infrared (FT-IR) Spectroscopy Measurements

Mid-infrared (MIR) spectroscopy provides information about the chemical bonds of the materials and, in the case of crystalline compounds, the rearrangements in the crystal structures. The results of functionalisation made on the liposomes were investigated with an Avatar 330 FT-IR Thermo Nicolet spectrometer equipped with an infrared light source and optics in absorbance mode. The measurements were made from powder samples in the 3500–400 cm−<sup>1</sup> wavelength range with a spectral resolution of 4 cm−<sup>1</sup> . The samples were mixed and pulverised with KBr and then pressed to form pellets. Pure KBr pellets were used as references. The measurements were performed three times for each sample.

#### 2.5.5. Atomic Force Microscopy (AFM)

One drop of solution was pipetted onto freshly cleaved mica (Muscovite mica, V-1 quality, Electron Microscopy Sciences, Washington, DC, USA) for the experiment. The AFM images were obtained using the tapping mode on an NT-MDT SolverPro Scanning Probe Microscope (NT-MDT, Spectrum Instruments, Moscow, Russia) under ambient conditions. AFM tips type PPP-NCHAuD-10 manufactured by NANOSENSORS (Neuchâtel, Switzerland) was applied with a nominal radius of curvature of 2 nm and 15 µm length. The non-contact silicon cantilevers had a typical force constant of 42 N/m and a resonance frequency of 330 kHz. Further information on the tip—thickness: 4.0 µm, length: 125 µm, width: 30 µm.

#### *2.6. Statistical Analysis*

Data analysis, statistics, and graphs were made from the experimental data with the Microsoft® Excel® (Microsoft Office Professional Plus 2013, Microsoft Excel 15.0.5023.100, Microsoft Corporation, Washington, WA, USA), OriginPro® 8.6 software (OriginLab® Corporation, Northampton, MA, USA), and the JMP® 13 Software (SAS Institute, Cary, NC, USA). One-way ANOVA with post-hoc Tukey test was applied using GraphPad

Prism (GraphPad Software Inc., San Diego, CA, USA). *p* < 0.05 was considered statistically significant, comparing the size, PDI, and zeta potential of the formulations. significant, comparing the size, PDI, and zeta potential of the formulations. **3. Results and Discussion** 

châtel, Switzerland) was applied with a nominal radius of curvature of 2 nm and 15 μm length. The non-contact silicon cantilevers had a typical force constant of 42 N/m and a resonance frequency of 330 kHz. Further information on the tip—thickness: 4.0 μm,

Data analysis, statistics, and graphs were made from the experimental data with the Microsoft® Excel® (Microsoft Office Professional Plus 2013, Microsoft Excel 15.0.5023.100, Microsoft Corporation, Washington, WA, USA), OriginPro® 8.6 software (OriginLab® Corporation, Northampton, MA, USA), and the JMP® 13 Software (SAS Institute, Cary, NC, USA). One-way ANOVA with post-hoc Tukey test was applied using GraphPad Prism (GraphPad Software Inc., California, USA). *p* < 0.05 was considered statistically

#### **3. Results and Discussion** *3.1. Initial Knowledge Space Development*

length: 125 μm, width: 30 μm.

*2.6. Statistical Analysis* 

#### *3.1. Initial Knowledge Space Development* According to the QbD methodology, the first step of this work was to systematically

*Molecules* **2022**, *27*, x FOR PEER REVIEW 7 of 18

According to the QbD methodology, the first step of this work was to systematically evaluate the literature and the accurate collection of all the relevant information. Nanosystems, such as liposomes, also require special attention for the formulation to deliver the expected product quality. The properties of the starting materials—each step of the manufacturing process, possible investigations, desired product quality, and possible therapeutic uses—must also be thoroughly examined. The development of the knowledge space was visualised by interpreting the classic 4M Ishikawa diagram (Figure 1). evaluate the literature and the accurate collection of all the relevant information. Nanosystems, such as liposomes, also require special attention for the formulation to deliver the expected product quality. The properties of the starting materials—each step of the manufacturing process, possible investigations, desired product quality, and possible therapeutic uses—must also be thoroughly examined. The development of the knowledge space was visualised by interpreting the classic 4M Ishikawa diagram (Figure 1).

**Figure 1.** Ishikawa diagram of the target product and the related factors. Abbreviations not used before: temp., temperature; BA, bioavailability*.* **Figure 1.** Ishikawa diagram of the target product and the related factors. Abbreviations not used before: temp., temperature; BA, bioavailability.

The Ishikawa diagram later contributed to the risk assessment process and identified the range of possible QTPP, CQA, CMA, and CPP elements. Based on the previous The Ishikawa diagram later contributed to the risk assessment process and identified the range of possible QTPP, CQA, CMA, and CPP elements. Based on the previous work of our research group, our goal was to develop a carrier system that is suitable for human use due to its favourable properties with increased impact. The selected CQAs can be seen in Table 2.

Based on the defined QTPPs, our goal was to produce thermosensitive liposomes with a monodisperse size distribution around 100 nm, which can serve as API-loadable nanocarriers, provide higher drug release due to the thermal change, and are easily transported across biological membranes. The product must be physically stable, which is a critical parameter for both particle size and shelf-life. The possibility of targeting via alternative administration routes must always be considered as these routes allow for improved pharmacokinetics and bioavailability.

Dosage form Liposomal colloid solution


**Table 2.** QTPP elements, their target, and the justification of LTSLs. Physical stability Stabilisable in solid form and retain the vesicle size after the dissolution achieved by the appropriate formulation, which results in stable

*Molecules* **2022**, *27*, x FOR PEER REVIEW 8 of 18

**Table 2.** QTPP elements, their target, and the justification of LTSLs

work of our research group, our goal was to develop a carrier system that is suitable for human use due to its favourable properties with increased impact. The selected CQAs

tients' adherence to therapy.

lution and absorption profiles.

Liquid forms can be used in multiple dosage forms—such as nasal drops, eye drops, and nasal sprays—which improve pa-

They can pass more efficiently across biological membranes in the appropriate nanosize range, resulting from improved disso-

The thermodynamic stability of colloidal solutions can be

#### *3.2. Risk Assessment 3.2. Risk Assessment*

can be seen in Table 2.

**QTPP Element Target Justification** 

The QbD-based risk assessment consists of two steps: first, an interdependence rating must be established between the QTPP-CQA and the CMA/CPP-CQA elements. Figure 2 shows the three-level interdependence rating between QTPPs and CQAs, where relationships were characterised as high, medium, and low risk. The QbD-based risk assessment consists of two steps: first, an interdependence rating must be established between the QTPP-CQA and the CMA/CPP-CQA elements. Figure 2 shows the three-level interdependence rating between QTPPs and CQAs, where relationships were characterised as high, medium, and low risk.


**Figure 2. Figure 2.**  Result of the interdependence rating between the QTTPs and the CQAs. Result of the interdependence rating between the QTTPs and the CQAs.

The CQAs chosen appear to be at high risk in at least one relation with the QTPP elements. A definite requirement for liquid colloidal dosage forms is the nanosize range as well as its distribution. Zeta potential has been assigned critical value because, under appropriate surface charge conditions, nanoparticles with a given morphology repel each other and keep the solution in a colloidal dispersed state. The phase transition temperature plays a major role in the site of the activity of the biological response, which determines through which administrative routes the system can release the drug from the potential drug-loaded liposome in an effective concentration, thereby achieving an enhanced thera-


peutic effect. The relationships between the CMA/CPP and CQA elements are shown in Figure 3. hanced therapeutic effect. The relationships between the CMA/CPP and CQA elements are shown in Figure 3.

The CQAs chosen appear to be at high risk in at least one relation with the QTPP elements. A definite requirement for liquid colloidal dosage forms is the nanosize range as well as its distribution. Zeta potential has been assigned critical value because, under appropriate surface charge conditions, nanoparticles with a given morphology repel each other and keep the solution in a colloidal dispersed state. The phase transition temperature plays a major role in the site of the activity of the biological response, which determines through which administrative routes the system can release the drug from the potential drug-loaded liposome in an effective concentration, thereby achieving an en-

*Molecules* **2022**, *27*, x FOR PEER REVIEW 9 of 18

**Figure 3.** Estimation of the interrelated impacts of the critical quality (CQAs) and material attributes (CMAs), and the critical process parameters (CPPs). **Figure 3.** Estimation of the interrelated impacts of the critical quality (CQAs) and material attributes (CMAs), and the critical process parameters (CPPs).

The film hydration method is a complex, multi-step sequence of operations that involves risks not only in its entirety, but also in its sub-steps. The CMA/CPP elements always include the composition, which in this case is the amount of lipids that make up the membrane. In the process, several solvents are used, such as dissolution medium, effluent solvent, hydration medium, and solvent for dissolving the final product. Of course, the material properties of these fluids also influence the manufacturing sub-processes, which are directly related to the product characteristics. For freeze-drying as a drying method that increases physical stability, the previous method of our research group was used, which—although it can represent many critical parameters alone—is excellent for increasing the shelf-life of the formulations. Vacuum drying and filtration through the membrane determine the shape in addition to the particle size. However, since a fixed, nanosized pore size membrane was used, deformation is adequately successful with its application. The probability rating—i.e., the quantification of the risk impacts—is shown in Figure 4 in Pareto diagrams. The results obtained were calculated by the software, which presented the numeric data and the ranking of the CQAs and the The film hydration method is a complex, multi-step sequence of operations that involves risks not only in its entirety, but also in its sub-steps. The CMA/CPP elements always include the composition, which in this case is the amount of lipids that make up the membrane. In the process, several solvents are used, such as dissolution medium, effluent solvent, hydration medium, and solvent for dissolving the final product. Of course, the material properties of these fluids also influence the manufacturing sub-processes, which are directly related to the product characteristics. For freeze-drying as a drying method that increases physical stability, the previous method of our research group was used, which—although it can represent many critical parameters alone—is excellent for increasing the shelf-life of the formulations. Vacuum drying and filtration through the membrane determine the shape in addition to the particle size. However, since a fixed, nanosized pore size membrane was used, deformation is adequately successful with its application. The probability rating—i.e., the quantification of the risk impacts—is shown in Figure 4 in Pareto diagrams. The results obtained were calculated by the software, which presented the numeric data and the ranking of the CQAs and the CMAs/CPPs according to their potential impact on the desired final product (QTPP). *Molecules* **2022**, *27*, x FOR PEER REVIEW 10 of 18

**Figure 4.** Pareto diagrams presenting the severity scores based on probability rating. **Figure 4.** Pareto diagrams presenting the severity scores based on probability rating.

Based on the probability rating, the three membrane-forming lipids were assigned the highest severity score. Besides hydration time, the cryoprotectant also showed a high risk, so fixed concentration and time were chosen to optimise the formulation around

tribution. As we chose to investigate only five CQAs, it is not a problem as the investigation methods allow the exact determination of these parameters. The relative occurrence– relative severity graph in Figure 5 supports the severity scores and gives a more visual

**Figure 5.** Relative severity–relative occurrence chart representing the CMA/CPP factors.

representation of the influencing CMA/CPP factors.

Based on the probability rating, the three membrane-forming lipids were assigned the highest severity score. Besides hydration time, the cryoprotectant also showed a high risk, so fixed concentration and time were chosen to optimise the formulation around these parameters. As for CQAs, no significant severity score difference was observed between particle size, zeta potential, phase transition temperature, and particle size distribution. As we chose to investigate only five CQAs, it is not a problem as the investigation methods allow the exact determination of these parameters. The relative occurrence–relative severity graph in Figure 5 supports the severity scores and gives a more visual representation of the influencing CMA/CPP factors. the highest severity score. Besides hydration time, the cryoprotectant also showed a high risk, so fixed concentration and time were chosen to optimise the formulation around these parameters. As for CQAs, no significant severity score difference was observed between particle size, zeta potential, phase transition temperature, and particle size distribution. As we chose to investigate only five CQAs, it is not a problem as the investigation methods allow the exact determination of these parameters. The relative occurrence– relative severity graph in Figure 5 supports the severity scores and gives a more visual representation of the influencing CMA/CPP factors.

Based on the probability rating, the three membrane-forming lipids were assigned

**Figure 4.** Pareto diagrams presenting the severity scores based on probability rating.

*Molecules* **2022**, *27*, x FOR PEER REVIEW 10 of 18

**Figure 5.** Relative severity–relative occurrence chart representing the CMA/CPP factors. **Figure 5.** Relative severity–relative occurrence chart representing the CMA/CPP factors.

#### *3.3. Design of Experiments (DoE)*

These results of the RA gave the basis of setting up a factorial experimental design. The selected factors came from the RA (the ratio of the phospholipids, the type of the hydration media, and the physical stabilisation method) and were applied on two levels. The pattern of the experiments and the factors are presented schematically in Figure 5. The variables are the CMAs, and the CPPs found to have the highest critical effect on the desired product mentioned above, and each of them was used on a minimum and a maximum level, as shown in Figure 6. The characteristics of the prepared liposomes (zeta potential, particle size, morphology of the liposomes, and Tm) were analysed (as a response) after each experiment.

These results of the RA gave the basis of setting up a factorial experimental design. The selected factors came from the RA (the ratio of the phospholipids, the type of the hydration media, and the physical stabilisation method) and were applied on two levels. The pattern of the experiments and the factors are presented schematically in Figure 5. The variables are the CMAs, and the CPPs found to have the highest critical effect on the desired product mentioned above, and each of them was used on a minimum and a maximum level, as shown in Figure 6. The characteristics of the prepared liposomes (zeta potential, particle size, morphology of the liposomes, and Tm) were analysed (as a re-

**Figure 6.** Schematic illustration of the factorial experimental design. **Figure 6.** Schematic illustration of the factorial experimental design.

#### *3.4. Results of the Vesicle Size and the Zeta Potential Analysis 3.4. Results of the Vesicle Size and the Zeta Potential Analysis*

*3.3. Design of Experiments (DoE)* 

sponse) after each experiment.

The samples were prepared after the QbD-based planning. The following table (Table 3) presents the size, zeta potentials, and PDI values of the liposomes measured after synthesis and two weeks later (in the meantime, the samples were stored at 4 °C) after heat treatment at 46 °C. The samples were prepared after the QbD-based planning. The following table (Table 3) presents the size, zeta potentials, and PDI values of the liposomes measured after synthesis and two weeks later (in the meantime, the samples were stored at 4 ◦C) after heat treatment at 46 ◦C.

**Table 3.** Size, zeta potential, and PDI values of the samples made from different compositions measured after synthesis, two weeks later, and after heat treatment **Table 3.** Size, zeta potential, and PDI values of the samples made from different compositions measured after synthesis, two weeks later, and after heat treatment.


70:25:5\_PSS 154 ± 5 −2.98 ± 0.93 0.34 ± 0.02 75 ± 1 −4.53 ± 1.44 0.22 ± 0.006 78 ± 1 −2.54 ± 1.21 0.22 ± 0.004

70:25:5\_PBS 165 ± 1 −3.56 ± 0.59 0.34 ± 0.005 97 ± 1 −6.00 ± 2.36 0.53 ± 0.04 71 ± 1 −3.79 ± 1.19 0.360 ± 0.36 The mean vesicle size of the 80:15:5\_PSS sample was 75 nm, and its zeta potential value was −2.74 mV; the mean vesicle size of the 80:15:5\_PBS sample was 130 nm, and its zeta potential value was −3.46 mV; the mean vesicle size of the 70:25:5\_PSS sample was 154 nm, and its zeta potential value was −2.98 mV; the mean vesicle size of the 70:25:5\_PBS sample was 165 nm, and its zeta potential was −3.56 mV. According to the literature, the best circulation time can be obtained for vesicles with a size of 100 nm; thus, we will try to optimise the production process in our future experiments [37]. The measured zeta potential values met expectations; the applied lipids had no charges, so high values were not expected. The results were strengthened by one-way ANOVA with post-hoc Tukey HSD statistical analysis. After synthesis, the 80:15:5\_PSS sample showed more significant size reduction and smaller PdI values compared to the other three for-The mean vesicle size of the 80:15:5\_PSS sample was 75 nm, and its zeta potential value was −2.74 mV; the mean vesicle size of the 80:15:5\_PBS sample was 130 nm, and its zeta potential value was −3.46 mV; the mean vesicle size of the 70:25:5\_PSS sample was 154 nm, and its zeta potential value was −2.98 mV; the mean vesicle size of the 70:25:5\_PBS sample was 165 nm, and its zeta potential was −3.56 mV. According to the literature, the best circulation time can be obtained for vesicles with a size of 100 nm; thus, we will try to optimise the production process in our future experiments [37]. The measured zeta potential values met expectations; the applied lipids had no charges, so high values were not expected. The results were strengthened by one-way ANOVA with post-hoc Tukey HSD statistical analysis. After synthesis, the 80:15:5\_PSS sample showed more significant size reduction and smaller PdI values compared to the other three formulations (80:15:5\_PSS vs. 80:15:5\_PBS/70:25:5\_PSS/70:25:5\_PBS \*\*, *p* < 0.01). Concerning the zeta potential value, there was no significance found between the samples.

> It is essential to note the correlation between the size values measured after synthesis when investigating the effect of different hydration media on the same compositions. Applying PSS, the size of the liposomes was smaller than that obtained with the same compositions when PBS was used. The size of the vesicles in the case of the samples made from the 80:15:5 lipid composition was: PSS (75 ± 2 nm) < PBS pH 7.4 (154 ± 5 nm), and the same tendency could be observed for the samples made from the 70:25:5 lipid compositions as well. However, the zeta potential values were slightly more negative when PBS solution was used. The ionic strength of the hydration medium influences the value of the zeta potential; the higher the ionic strength, the more compact the ion layer formed around the vesicles and, due to this fact, the higher the zeta potential [38]. In the presented case, the ionic strength of the hydration medium was: saline solution (0.15 M) < PBS pH 7.4 (0.16 M).

The samples were investigated in a dispersed state after two weeks of storage in a refrigerator (~4 ◦C); furthermore, these samples were kept above their phase transition temperature (46 ◦C) for 30 min as DLS measurements were rerun. 46 ◦C was chosen as the temperature of the treatment because the liposomes disintegrate above 40 ◦C. The change in the size of the liposomes proves their instability in a dispersed state that highlights the importance of stabilisation by lyophilisation. Based on the PDI values, the 80:15:5\_PSS samples can be considered homogeneous after synthesis. After two weeks, all the samples were homogeneous except for the 70:25:5\_PBS sample; however, after heat treatment, these data are not relevant. It can be assumed that the stability of the liposomes decreases without treatment because their size was reduced; however, this size reduction resulted in a homogeneous size distribution based on the PDI values. In the case of the 80:15:5 samples, a notable size increase was observed after heat treatment. The 70:25:5 samples hydrated with PBS behaved differently because there was a decrease in their size instead of growth, but the change in the size of the vesicles was recognisable. Based on the obtained results, it was concluded that the liposomes destabilise when kept above their phase transition temperature, which may later lead to API release. 80:15:5\_PSS was significantly smaller than the others (\*\*, *p* < 0.01). Regarding PDI of the samples, 80:15:5\_PSS vs. 80:15:5\_PBS showed no significance after storage, whilst 80:15:5\_PSS vs. 70:25:5\_PSS/70:25:5\_PBS showed significant difference (\*\*, *p* < 0.01). Although the PDI of 70:25:5\_PSS was smaller, its size was larger, so the 80:15:5\_PSS formulation still proved to be favourable even after two weeks of storage. The zeta potential value of 80:15:5\_PSS also showed significant difference compared to 80:15:5\_PBS and 70:15:5\_PBS formulations (\*\*, *p* < 0.01) but insignificance can be found vs. 70:25:5\_PSS. The results of the DLS measurements led to the conclusion that the 80:15:5\_PSS sample shows the most homogenous size distribution measured after synthesis, and two weeks later, thus this sample proved to be the most stable compared to the others. After heat treatment, concerning size, all formulations showed significant differences compared to each other (\*\*, *p* < 0.01) whilst no significance was observed for the zeta potential values. The PDI values were significantly higher (80:15:5\_PSS vs. 80:15:5\_PBS \*\*, *p* < 0.01) or lower (80:15:5\_PSS vs. 70:25:5\_PSS, \*\* *p* < 0.01) but were insignificant compared to 70:25:5\_PBS.

DLS and zeta potential measurements were repeated after the synthesis of the 80:15:5\_PPS sample (Table 4), before lyophilisation, but now at different temperatures. These temperatures were 10, 20, 30, 35, 37, and 40 ◦C.


**Table 4.** Size, zeta potential, and PDI values of the 80:15:5\_PSS sample at different temperatures.

Based on the measurement results, it can be stated that the size of liposomes is almost constant at the storage temperature (10–20 ◦C) and then increases continuously with increasing temperature. The results conclude that liposomes are destabilised by keeping them above their phase transition temperature, which will later lead to the release of their active ingredient.

3.4.1. Results of the Differential Scanning Calorimetry (DSC) and the Thermogravimetric Analysis (TGA) Investigations

Figure 7 demonstrates the results of the TGA and DSC measurements for the 80:15:5\_PSS sample; however, the same curve was obtained in every case. Based on the TGA curve, there was no significant change in the mass of the sample. The slight increase in weight can

be considered a measurement error because, due to its sensitivity, the instrument can show a weight increase when such a small amount of sample is measured. strument can show a weight increase when such a small amount of sample is measured.

*Molecules* **2022**, *27*, x FOR PEER REVIEW 13 of 18

**Table 4.** Size, zeta potential, and PDI values of the 80:15:5\_PSS sample at different temperatures

**T (°C) Size (nm) PDI Zeta Potential (mV)**  10 64 ± 1.2 0.29 ± 0.04 −2.23 ± 0.1 20 69 ± 3.2 0.24 ± 0.01 −1.42 ± 0.16 30 80 ± 1.8 0.26 ± 0.01 −1.92 ± 0.41 35 78 ± 1.1 0.26 ± 0.01 −1.54 ± 0.43 37 80 ± 0.4 0.26 ± 0.01 −1.98 ± 0.41 40 85 ± 2.8 0.27 ± 0.01 −1.39 ± 0.5

Based on the measurement results, it can be stated that the size of liposomes is almost constant at the storage temperature (10–20 °C) and then increases continuously with increasing temperature. The results conclude that liposomes are destabilised by keeping them above their phase transition temperature, which will later lead to the release of their

3.4.1. Results of the Differential Scanning Calorimetry (DSC) and the Thermogravimetric

80:15:5\_PSS sample; however, the same curve was obtained in every case. Based on the TGA curve, there was no significant change in the mass of the sample. The slight increase in weight can be considered a measurement error because, due to its sensitivity, the in-

Figure 7 demonstrates the results of the TGA and DSC measurements for the

active ingredient.

Analysis (TGA) Investigations

**Figure 7.** TGA and DSC curves of the 80:15:5\_PSS sample. **Figure 7.** TGA and DSC curves of the 80:15:5\_PSS sample.

Figure 7 demonstrates the TGA and DSC results for the 80:15:5\_PSS. Similar curves were obtained for all samples. Based on the TGA curve, there was no significant change in the mass of the sample. The slight increase in weight is a measurement error due to the Figure 7 demonstrates the TGA and DSC results for the 80:15:5\_PSS. Similar curves were obtained for all samples. Based on the TGA curve, there was no significant change in the mass of the sample. The slight increase in weight is a measurement error due to the small sample amount.

small sample amount. The stability of the freeze-dried samples can be characterized by the glass transition temperature (Tg) [39]. Glass transition and melting temperatures of 26 °C and 57 °C were found in the DSC curve in Figure 7, respectively. Based on the literature references, DPPC melting transition is around 38 °C [39,40], however, it is well known that Tm can increase in lipid mixtures [41–43]*.* The thermodynamic parameters (ΔHm and Tm) provide im-The stability of the freeze-dried samples can be characterized by the glass transition temperature (Tg) [39]. Glass transition and melting temperatures of 26 ◦C and 57 ◦C were found in the DSC curve in Figure 7, respectively. Based on the literature references, DPPC melting transition is around 38 ◦C [39,40], however, it is well known that T<sup>m</sup> can increase in lipid mixtures [41–43]. The thermodynamic parameters (∆H<sup>m</sup> and Tm) provide important information on liposome stability and the pharmacokinetics of the API. The higher the ∆Hm, the stiffer the bilayers are [39]. In all of our samples an enthalpy change (∆Hm) of 7.8 J/g were found.

portant information on liposome stability and the pharmacokinetics of the API. The

#### higher the ΔHm, the stiffer the bilayers are [39]. In all of our samples an enthalpy change 3.4.2. Results of the Fourier-Transform Infrared (FT-IR) Spectroscopy Measurements

(ΔHm) of 7.8 J/g were found. The FT-IR spectra (Figure 8) show the results of the samples made from different compositions with PSS (A) and PBS (B) media. The FT-IR spectra of the samples made with different hydration media are different. In every case, the spectra include two different regions. The high wavenumber part of the spectrum (3100–2800 cm−<sup>1</sup> ) contains a contribution from C-H stretching vibrations only [44]. However, it mostly originates from the hydrocarbon chains. The low wavenumber region of the spectrum (below 1800 cm−<sup>1</sup> ) is essentially related to the polar head groups of the lipids, as indicated. The ester ν(C=O) is usually the strongest peak near 1735 cm−<sup>1</sup> and 1797 cm−<sup>1</sup> , followed by the phosphate contributions near 1243 cm−<sup>1</sup> (νas(PO2)) and 1103 cm−<sup>1</sup> (A), and 1065 cm−<sup>1</sup> (B) (νs(PO2)) [45]. The hydrocarbon chains do contribute near 1465 cm−<sup>1</sup> , but all-trans conformations rather absorb near 1468 cm−<sup>1</sup> .

The contribution of the lipid hydrocarbon chains is present in various spectral regions. The most prominent ones appear between 3050 and 2800 cm−<sup>1</sup> . This region essentially contains C-H stretching bands from different vibrational modes: νas(CH2) near 2917 cm−<sup>1</sup> , νs(CH2) near 2850 cm−<sup>1</sup> . These vibrations present several interesting features: (1) the fact that there is little overlap with other vibrations, including complex systems such as cells and tissues; (2) these vibrational modes are largely uncoupled from other modes, i.e., they do not depend on the lipid head group; and (3) they are sensitive to the structure (disordering) of the chains [44].

formations rather absorb near 1468 cm−1.

3.4.2. Results of the Fourier-Transform Infrared (FT-IR) Spectroscopy Measurements

The FT-IR spectra (Figure 8) show the results of the samples made from different compositions with PSS (A) and PBS (B) media. The FT-IR spectra of the samples made with different hydration media are different. In every case, the spectra include two different regions. The high wavenumber part of the spectrum (3100–2800 cm−1) contains a contribution from C-H stretching vibrations only [44]. However, it mostly originates from the hydrocarbon chains. The low wavenumber region of the spectrum (below 1800 cm−1) is essentially related to the polar head groups of the lipids, as indicated. The ester ν(C=O) is usually the strongest peak near 1735 cm−1 and 1797 cm−1, followed by the phosphate contributions near 1243 cm−1 (νas(PO2)) and 1103 cm−1(A), and 1065 cm−1 (B) (νs(PO2)) [45]. The hydrocarbon chains do contribute near 1465 cm−1, but all-trans con-

**Figure 8.** FT-IR spectra of samples made from different compositions: (**A**) Samples made with PSS medium from the compositions of 80:15:5 (blue) and 70:25:5 (red) lipid ratios. (**B**) Samples made with PBS medium from the compositions of 80:15:5 (blue) and 70:25:5 (red) lipid ratios. **Figure 8.** FT-IR spectra of samples made from different compositions: (**A**) Samples made with PSS medium from the compositions of 80:15:5 (blue) and 70:25:5 (red) lipid ratios. (**B**) Samples made with PBS medium from the compositions of 80:15:5 (blue) and 70:25:5 (red) lipid ratios.

The contribution of the lipid hydrocarbon chains is present in various spectral re-3.4.3. Results of the Transmission Electron Microscopy (TEM) and the Atomic Force Microscopy (AFM) Spectroscopy Measurements

gions. The most prominent ones appear between 3050 and 2800 cm−1. This region essentially contains C-H stretching bands from different vibrational modes: νas(CH2) near 2917 cm−1, νs(CH2) near 2850 cm−1. These vibrations present several interesting features: (1) the fact that there is little overlap with other vibrations, including complex systems such as cells and tissues; (2) these vibrational modes are largely uncoupled from other modes, i.e., they do not depend on the lipid head group; and (3) they are sensitive to the structure The liposomes are visible in the Transmission Electron Microscopy image (Figure 9A) of the 80:15:5\_PPS sample. The scale in the picture is not validated for precise size determination; however, it provides a good indication of the size of the liposomes. It can be seen from the picture that the average size of the liposomes is ~100 nm. This result correlates well with the accurately defined size values obtained from the DLS measurements. *Molecules* **2022**, *27*, x FOR PEER REVIEW 15 of 18

**Figure 9.** TEM (**A**) and AFM (**B**) image of the 80:15:5\_PPS sample. **Figure 9.** TEM (**A**) and AFM (**B**) image of the 80:15:5\_PPS sample.

AFM measurements were also performed, giving a proper three-dimensional surface profile of the 80:15:5\_PPS samples. These recordings are illustrated in the following figure (Figure 9B). AFM measurements were also performed, giving a proper three-dimensional surface profile of the 80:15:5\_PPS samples. These recordings are illustrated in the following figure (Figure 9B).

The AFM images show that liposomes with a homogeneous size distribution at nearly 100 nm were prepared. The measurement results are the same as the particle sizes obtained during the DLS and TEM measurements. The AFM images show that liposomes with a homogeneous size distribution at nearly 100 nm were prepared. The measurement results are the same as the particle sizes obtained during the DLS and TEM measurements.

#### **4. Conclusions 4. Conclusions**

chosen.

In conclusion, the characteristics of the desired liposomal formulation and the factors that can influence these features were defined by following the steps of the quality by design method. After performing the risk assessment, the key element of the QbD—a factorial experimental design—was developed based on the RA results. Therefore, the QbD-based product and experimental design and the liposome preparation were carried In conclusion, the characteristics of the desired liposomal formulation and the factors that can influence these features were defined by following the steps of the quality by design method. After performing the risk assessment, the key element of the QbD—a factorial experimental design—was developed based on the RA results. Therefore, the QbD-based product and experimental design and the liposome preparation were carried out to obtain

out to obtain the nanosized delivery system. Furthermore, thermosensitivity was proved, and the most stable sample with the ideal composition—the 80:15:5\_PSS formula—was

**Author Contributions:** Conceptualisation, D.G.D.; Data curation, B.S., D.B. and G.K.; Formal analysis, D.G.D. and Z.N.; Funding acquisition, Z.K. and I.C.; Investigation, D.G.D., D.B. and G.K.; Methodology, D.G.D., Z.N., M.C. and E.P.; Project administration, D.G.D.; Resources, Z.K. and I.C.; Software, Z.N., B.S. and E.P.; Supervision, Z.K. and I.C.; Validation, Z.K. and I.C.; Visualisation, M.C.; Writing—original draft, D.G.D., Z.N., B.S., M.C. and E.P.; Writing—review and editing, Z.N.

**Funding:** This work was supported by the Gedeon Richter's Talentum Foundation, the Ministry of Human Capacities, Hungary grant, TKP-2020 (Interdisciplinary Excellence Centre), the construction EFOP 3.6.3-VEKOP-16-2017-00009, and the GINOP project (2.2.1-15-2016-00007). Project no. TKP2021-EGA-32 has been implemented with the support provided by the Ministry of Innovation and Technology of Hungary from the National Research, Development and Innovation Fund, fi-

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study, in the collection, analyses or interpretation of data, in the writing of the man-

**Sample Availability:** Samples of the compounds are not available from the authors.

nanced under the TKP2021-EGA funding scheme.

**Informed Consent Statement:** Not applicable.

uscript, or in the decision to publish the results.

**Institutional Review Board Statement:** Not applicable.

All authors have read and agreed to the published version of the manuscript.

the nanosized delivery system. Furthermore, thermosensitivity was proved, and the most stable sample with the ideal composition—the 80:15:5\_PSS formula—was chosen.

**Author Contributions:** Conceptualisation, D.G.D.; Data curation, B.S., D.B. and G.K.; Formal analysis, D.G.D. and Z.N.; Funding acquisition, Z.K. and I.C.; Investigation, D.G.D., D.B. and G.K.; Methodology, D.G.D., Z.N., M.C. and E.P.; Project administration, D.G.D.; Resources, Z.K. and I.C.; Software, Z.N., B.S. and E.P.; Supervision, Z.K. and I.C.; Validation, Z.K. and I.C.; Visualisation, M.C.; Writing—original draft, D.G.D., Z.N., B.S., M.C. and E.P.; Writing—review and editing, Z.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Gedeon Richter's Talentum Foundation, the Ministry of Human Capacities, Hungary grant, TKP-2020 (Interdisciplinary Excellence Centre), the construction EFOP 3.6.3-VEKOP-16-2017-00009, and the GINOP project (2.2.1-15-2016-00007). Project no. TKP2021-EGA-32 has been implemented with the support provided by the Ministry of Innovation and Technology of Hungary from the National Research, Development and Innovation Fund, financed under the TKP2021-EGA funding scheme.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study, in the collection, analyses or interpretation of data, in the writing of the manuscript, or in the decision to publish the results.

**Sample Availability:** Samples of the compounds are not available from the authors.

#### **Abbreviations**


## **References**


## *Article* **Development and Evaluation of Chitosan Nanoparticles for Ocular Delivery of Tedizolid Phosphate**

**Mohd Abul Kalam 1,2 , Muzaffar Iqbal 3,4 , Abdullah Alshememry 1,2 , Musaed Alkholief 1,2 and Aws Alshamsan 1,2,\***


**Abstract:** This study investigates the development of topically applied non-invasive chitosannanoparticles (CSNPs) for ocular delivery of tedizolid phosphate (TZP) for the treatment of MRSArelated ocular and orbital infections. An ionic-gelation method was used to prepare TZP-encapsulated CSNPs using tripolyphosphate-sodium (TPP) as cross-linker. Particle characterization was performed by the DLS technique (Zeta-Sizer), structural morphology was observed by SEM. The drug encapsulation and loading were determined by the indirect method. In-vitro release was conducted through dialysis bags in simulated tear fluid (pH 7) with 0.25% Tween-80. Physicochemical characterizations were performed for ocular suitability of CSNPS. An antimicrobial assay was conducted on different strains of Gram-positive bacteria. Eye-irritation from CSNPs was checked in rabbits. Transcorneal flux and apparent permeability of TZP from CSNPs was estimated through excised rabbit cornea. Ionic interaction between the anionic and cationic functional groups of TPP and CS, respectively, resulted in the formation of CSNPs at varying weight ratios of CS/TPP with magnetic stirring (700 rpm) for 4 h. The CS/TPP weight ratio of 3.11:1 with 10 mg of TZP resulted in optimal-sized CSNPs (129.13 nm) with high encapsulation (82%) and better drug loading (7%). Release profiles indicated 82% of the drug was released from the TZP aqueous suspension (TZP-AqS) within 1 h, while it took 12 h from F2 to release 78% of the drug. Sustained release of TZP from F2 was confirmed by applying different release kinetics models. Linearity in the profile (suggested by Higuchi's model) indicated the sustained release property CSNPs. F2 has shown significantly increased (*p* < 0.05) antibacterial activity against some Gram-positive strains including one MRSA strain (SA-6538). F2 exhibited a 2.4-fold increased transcorneal flux and apparent permeation of TZP as compared to TZP-AqS, indicating the better corneal retention. No sign or symptoms of discomfort in the rabbits' eyes were noted during the irritation test with F2 and blank CSNPs, indicating the non-irritant property of the TZP-CSNPs. Thus, the TZP-loaded CSNPs have strong potential for topical use in the treatment of ocular MRSA infections and related inflammatory conditions.

**Keywords:** tedizolid-phosphate; chitosan; nanoparticles; antibacterial; eye-irritation transcornealpermeation

## **1. Introduction**

Among ocular infections, methicillin-resistant *Staphylococcus aureus* (MRSA) infections in the eyes and orbits are the most important. Such infections are often treated inappropriately [1]. The most common presentations of ocular infections by MRSA are keratitis (36%), eyelid problems (24%), conjunctivitis, cellulitis, and dacryocystitis (20%) and around nearly half (48%) of the infections are found to be vision-threatening [2]. As per the Kaiser

**Citation:** Kalam, M.A.; Iqbal, M.; Alshememry, A.; Alkholief, M.; Alshamsan, A. Development and Evaluation of Chitosan Nanoparticles for Ocular Delivery of Tedizolid Phosphate. *Molecules* **2022**, *27*, 2326. https://doi.org/10.3390/ molecules27072326

Academic Editor: Ildiko Badea

Received: 27 February 2022 Accepted: 31 March 2022 Published: 4 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Permanente study, roughly 13% of ocular MRSA infections were found in infants, where conjunctivitis was the main sign [3]. Due to many serious infections caused by MRSA, it has become a significant clinical challenge and economic burden [4]. Tedizolid phosphate (TZP) is a novel oxazolidinone antibiotic to treat the infections caused by MRSA that has become a new defense weapon [5]. It is also used against the vancomycin-resistant *enterococci* [6] and some linezolid-resistant strains [7]. It was approved by the US-FDA in June 2014 for acute bacterial skin and skin structure infections [8]. The chemical structure and structural activity relationships of tedizolid (TDZ) are illustrated in Figure 1. TZP is a prodrug which is rapidly converted in vivo to its active form TDZ by acid and alkaline phosphatases [9,10]. Therefore, either TDZ or TZP can be used in eye preparations. It differs from other members of the oxazolidinone class as it has a modified side chain at the C5 position of the oxazolidinone nucleus which instructs the activity against some linezolid-resistant microorganisms and has an optimized C- and D-ring system that improve its potency through additional binding site interactions [8]. The antibacterial activity of TZP/TDZ is facilitated by inhibiting the bacterial protein synthesis. Linezolid is also an oxazolidinone antibiotic approved by the FDA in 2000; however, it induces peripheral and optic neuropathy in humans, so its clinical use is limited for prolonged therapy, while TZP/TDZ has no such adverse effects [9,11]. the Kaiser Permanente study, roughly 13% of ocular MRSA infections were found in infants, where conjunctivitis was the main sign [3]. Due to many serious infections caused by MRSA, it has become a significant clinical challenge and economic burden [4]. Tedizolid phosphate (TZP) is a novel oxazolidinone antibiotic to treat the infections caused by MRSA that has become a new defense weapon [5]. It is also used against the vancomycinresistant *enterococci* [6] and some linezolid-resistant strains [7]. It was approved by the US-FDA in June 2014 for acute bacterial skin and skin structure infections [8]. The chemical structure and structural activity relationships of tedizolid (TDZ) are illustrated in Figure 1. TZP is a prodrug which is rapidly converted in vivo to its active form TDZ by acid and alkaline phosphatases [9,10]. Therefore, either TDZ or TZP can be used in eye preparations. It differs from other members of the oxazolidinone class as it has a modified side chain at the C5 position of the oxazolidinone nucleus which instructs the activity against some linezolid-resistant microorganisms and has an optimized C- and D-ring system that improve its potency through additional binding site interactions [8]. The antibacterial activity of TZP/TDZ is facilitated by inhibiting the bacterial protein synthesis. Linezolid is also an oxazolidinone antibiotic approved by the FDA in 2000; however, it induces peripheral and optic neuropathy in humans, so its clinical use is limited for prolonged therapy, while TZP/TDZ has no such adverse effects [9,11].

*Molecules* **2022**, *27*, x FOR PEER REVIEW 2 of 22

**Figure 1.** Structure–activity relationships for tedizolid. Where, A–D symbolize the different aromatic rings in the molecular structure of Tedizolid. The Ring-A = Oxazolidinone ring, Ring-B = Aryl group, Ring-C = meta-fluorine and para-oriented electron withdrawing or unsaturated ring and Ring-D = para-oriented ring structure, provides additional sites for H-bonding. **Figure 1.** Structure–activity relationships for tedizolid. Where, A–D symbolize the different aromatic rings in the molecular structure of Tedizolid. The Ring-A = Oxazolidinone ring, Ring-B = Aryl group, Ring-C = meta-fluorine and para-oriented electron withdrawing or unsaturated ring and Ring-D = para-oriented ring structure, provides additional sites for H-bonding.

Although vancomycin is the choice of antibiotic for the treatment of MRSA-infections, its efficacy has been compromised due to emergence of resistant strains of *S. aureus* Although vancomycin is the choice of antibiotic for the treatment of MRSA-infections, its efficacy has been compromised due to emergence of resistant strains of *S. aureus* [12].

[12]. These finding encouraged us to develop a topically applied non-invasive nano-carrier for ocular delivery of TZP to treat MRSA-related eye and orbital infections. We presumed that TZP would stand a better chance of accomplishing the critical prerequisite for new antibiotics in this era of increasing multi-drug resistance, including MRSA and other resistant strain eye infections. After topical administration, the ocular availability of drugs is limited due to strong self-protective and defensive ocular barriers. The nasolacrimal drainage, noncorneal absorption, and robust corneal impenetrability [13] limit the ocular availability (5–7%) of topically applied drugs [14–16]. The availability of drugs can be improved by prolonging the precorneal retention of the dosage forms and enhancing the These finding encouraged us to develop a topically applied non-invasive nano-carrier for ocular delivery of TZP to treat MRSA-related eye and orbital infections. We presumed that TZP would stand a better chance of accomplishing the critical prerequisite for new antibiotics in this era of increasing multi-drug resistance, including MRSA and other resistant strain eye infections. After topical administration, the ocular availability of drugs is limited due to strong self-protective and defensive ocular barriers. The nasolacrimal drainage, noncorneal absorption, and robust corneal impenetrability [13] limit the ocular availability (5–7%) of topically applied drugs [14–16]. The availability of drugs can be improved by prolonging the precorneal retention of the dosage forms and enhancing the corneal and conjunctival transport of the drugs. In some conditions repeated application of

dosage forms into eyes is needed which may cause corneal pigmentation, mechanical injury, or sensitivities to the eyes [17]. To avoid the frequent application of eye preparations and to attain an effective and prolonged drug concentration into ocular tissues, the development of an appropriate dosage form is needed. Drug encapsulation into nano-carriers is one of the best approaches to overcome the shortfalls of conventional ophthalmic dosage forms [17–19]. Such carriers extend the ocular retention of the drug which can improve its transcorneal flux and intraocular availability [20,21].

Chitosan (CS) is hydrophilic, mucoadhesive, non-toxic, biodegradable polysaccharide [22,23], which also stabilizes tear fluids and increases the precorneal/corneal contact time of CSNPs [24]. Due to high viscosity and sufficient adhesion with the ocular surfaces CSNPs may reduce nasolacrimal drainage [25,26] and consequently, improve the ocular bioavailability of encapsulated TZP [27] which will augment its activity against Grampositive and MRSA infections with reduced dosing frequency and easy topical instillation with good patient compliance.

Thus, we developed and characterized CS based nanoparticles (NPs) to prolong ocular retention and achieve an effective drug concentration. For the in vitro release of TZP, physicochemical characterization of a TZP-CSNP suspension for ocular suitability was performed. Antibacterial activity of TZP from NPs was determined against *B. subtilis* and *S. aureus* strains including one MRSA strain (SA-6538). Transcorneal permeation of TZP from CSNPs was tested in excised rabbit corneas and eye irritation from CSNPs was tested in rabbit eyes [28,29]. In vivo efficacy of TZP-CSNPs was estimated by analyzing the aqueous humor concentration of tedizolid (active form of TZP), which was reported in our previous publication [30].

#### **2. Materials and Methods**

#### *2.1. Materials*

Tedizolid phosphate (C17H15FN6O6P; MW 450.318 Da) was of ≥98% purity, purchased from "Beijing Mesochem Technology Co. Ltd. (Beijing, China)". Low MW Chitosan (50–190 kDa) based on viscosity 20–300 cP, at 1 wt.% in acetic acid (1%) at 25 ◦C and 75–85% de-acetylated, Tripolyphosphate-sodium (TPP) and sodium dihydrogen phosphate (KH2PO4) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Glacial acetic acid, HPLC grade methanol and acetonitrile were purchased from BDH, Ltd. (Poole, UK). RC-dialysis membrane (MWCO: 12–14 kDa) was purchased from Spectra Por, Spectrum Laboratories Inc., (Rancho Dominguez, CA, USA). Mannitol was purchased from Qualikems Fine Chem Pvt. Ltd. (Vadodara, India). Purified water was obtained using a Milli-Q® water purifier (Millipore, Molsheim, France). All other chemicals used were of analytical grade and solvents of HPLC grade.

#### *2.2. Chromatographic Analysis of TZP*

Reverse-phase (RP) high-performance liquid chromatography (HPLC) with UV-detection (at 251 nm) was used for the quantification of TZP following the reported HPLC-UV method [31,32]. In brief, an HPLC system (Waters® 1500-series controller, Milford, MA, USA) was used, which was equipped with a UV-detector (Waters® 2489, dual absorbance detector, Milford, MA, USA), a binary pump (Waters® 1525, Milford, MA, USA), and an automated sampling system (Waters® 2707 Autosampler, Milford, MA, USA). The HPLC system was monitored by Breeze software. An RP C<sup>18</sup> analytical column (Macherey-Nagel 250 × 4.6 mm, 5 µm) at 40 ◦C was used for this analysis. The mobile phase consisted of 65:35 *v*/*v* of 0.02 M sodium acetate buffer (the pH was adjusted to 3.5 by hydrochloric acid) and acetonitrile was pumped isocratically at 1 mL/min of flow rate. The total run time was 10 min. The injection volume was 30 µL. The standard stock solution of TZP was prepared in methanol (100 <sup>µ</sup>g·mL−<sup>1</sup> ) and working standard solutions (0.25–50 <sup>µ</sup>g·mL−<sup>1</sup> ) were prepared by serial dilution of the stock solution with 65:35, *v*/*v* mixture of the mobile phase.

#### *2.3. Formulation Development*

The TZP-loaded CSNPs were prepared by ionic-gelation of chitosan (CS) with a crosslinker of tripolyphosphate-sodium (TPP) [33], with slight modification for highly lipophilic drugs [22,23,34,35]. Briefly, 10 mg of TZP was dissolved in 200 µL of DMSO in triplicate. The drug solution was added slowly (with magnetic stirring at 500 rpm) into a previously prepared 13.5 mL 0.4, 0.6, and 0.8% *w*/*v*, solutions of CS in 1%, *v*/*v* glacial acetic acid (pH 3.0). Simultaneously, the TPP solutions in Milli-Q water (at 0.2, 0.4 and 0.6%, *w*/*v*) were prepared and pH of these solutions was maintained at 7.2 with 100 mM potassium dihydrogen phosphate buffer. Thereafter, 6.5 mL of TPP solution was added dropwise (at the rate of 1.5 mL·min−<sup>1</sup> ) to 13.5 mL of CS solution containing TZP with continuous magnetic stirring at 700 rpm for 4 h at 10 ◦C [36]. The details of the constituents used to prepare three optimal TZP-loaded CSNPs are summarized in Table 1. The excess drug (possibly un-encapsulated) was washed by centrifugation (13,500 rpm) for 15 min at 10 ◦C. Finally, collection of TZP-loaded CSNPs was performed by washing with Milli-Q® water through ultracentrifugation (30,000 rpm) for 30 min at 4 ◦C. Around 10 mL of CSNP suspension was filtered through a 450 µ filtration unit, frozen at −80 ◦C, freeze-dried (at −50 ◦C and 0.01 mbar pressure for 24 h) in a FreeZone-4.5 freeze dry system (Labconco Corporation, MO, USA), and stored at −20 ◦C for further studies. Mannitol (1%, *w*/*v*) was added into the suspension as cryoprotectant before freeze-drying [37].

**Table 1.** Formulation of tedizolid phosphate (TZP) loaded-CSNPs.


\* In all cases the drug (TZP) was dissolved in 200 µL DMSO prior to its addition into CS solution. Low-molecularweight chitosan (CS), tripolyphosphate sodium (TPP), nanoparticles (NPs).

#### *2.4. Characterization of the CSNPs*

#### 2.4.1. Particle Size, Polydispersity-Index (PDI) and Zeta-Potential Measurements

The hydrodynamic diameter as particle size, polydispersity-index (PDI) and zeta potentials of the developed CSNPs were evaluated by dynamic light scattering (DLS) analysis using a Zetasizer Nano Series (Nano-ZS, Malvern Instruments Ltd., Worcestershire, UK) [38]. The DLS also known as photon correlation spectroscopy, measures the Brownian movement and relates this to particle's size by enlightening the particles with the laser and analyzing the fluctuations in the intensities of the scattered light, then utilizes this to calculate the particle's size. DLS was performed at a fixed detection arrangement of 90◦ angle to the laser light and the center of the cuvette area. The suspensions of CSNPs were further diluted with Milli-Q® water for the above measurements, because low a concentration of samples is beneficial for maximizing the amount of scattering from the measurement sample. For zeta potential, by considering the dielectric constant of water (≈78.5) at 25 ◦C, the electrophoretic mobility was determined and then the Henry equation was applied (these processes were performed by the software, DTS V-4.1, Malvern, UK). The magnitude of zeta potential (mV) gives an indication of the potential stability of any colloidal system. All the measurements were performed in triplicate.

#### 2.4.2. Transmission Electron Microscopy (TEM)

The morphology and structural characterization of the optimal formulation (TZP-CSNPs, F2) was carried out using transmission electron microscopy (TEM), JEOL TEM (JEM-1010). The TEM analysis was performed under light microscopy, operated at 80 kV with point-to-point resolution [39]. The magnification of images was 50–80 K (X). A combination of bright-field imaging at increasing magnification and diffraction modes was

used to expose the structure and size of the NPs. The suspension of F2 was further diluted with Milli-Q water prior to the analysis. Dilution was performed to overcome certain challenges including the images overlapping, difficulty in detection of small particles, and obscured signals during observation due to the presence of the surrounding matrix and background noise. In order for the electron beams to transmit through a very thin specimen and interact with it, a drop of the nanosuspension was put on the carbon coated copper grids and stained with Phosphotungstic acid (2% solution). The grids were air dried overnight and then the particle morphology was observed at ambient temperature.

### 2.4.3. X-ray Diffraction Study

The X-ray diffraction study on powdered samples was performed using an Ultima-IV Goniometer (Rigaku, Inc., Tokyo, Japan) over a 5.0◦ to 70.0◦ 2*θ* range at a scan speed of 1.0◦ per min to examine the crystalline nature of the encapsulated drug into the CSNPs as compared to the pure drug. The X-ray tube anode material was Cu with Ka2 elimination, the Ka2/Ka1 intensity ratio was 0.10 nm, and it was monochromatized with graphite crystal. The diffractograms were obtained at 40 kV tube voltage and 40 mA, and the generator was in step scan mode (step size 0.02◦ and counting time was 1 s per step).

#### 2.4.4. Encapsulation Efficiency and Drug Loading Capacity

The encapsulation and loading of TZP into the CSNPs were determined by indirect methods (i.e., quantification of unencapsulated drugs). The amount of TZP encapsulated into NPs and the percentage drug loading were calculated by the difference between the total (initial) amounts of drug used for the preparation of the NPs and the drug analyzed in the supernatant after centrifugation of the suspension of CSNPs [10]. Briefly, 4 mg of CSNPs was suspended in methanol, vortexed and centrifuged at 13,500 rpm for 15 min. Supernatant was collected and the concentration of drug in the supernatant was analyzed by HPLC-UV [31,32]. The percentages of encapsulation efficiency (%EE) and drug loading (%DL) were calculated by Equations (1) and (2)*:*

$$\%EE = \left(\frac{\text{Initial amount of TZP used (mg)} - Amount of TZP in supernantant (mg)}{\text{Initial amount of TZP used (mg)}}\right) \times 100\tag{1}$$
 
$$\text{// Initial amount of TTP used (ma)} - Amount of TTP in supernantant (ma)}$$

$$\% \text{DL} = \left(\frac{\text{Initial amount of TZP used (mg)} - \text{Amount of TZP in supernantant (mg)}}{\text{Total amount of CSNPs (mg)}}\right) \times 100\tag{2}$$

#### 2.4.5. Physicochemical Characterization

The physicochemical characterization of TZP-loaded CSNPs was performed to ensure its suitability for ocular use. The characterization parameters included the transparency of the nanosuspension of TZP-CSNPs by visual observation under light alternatively against black and white background at 25 ◦C and pH 7.2. The drug content in the TZP-CSNPs was estimated by the HPLC-UV method as described above. The pH of the CSNP suspension was measured using a calibrated pH meter (Mettler Toledo MP-220, Schweiz, Switzerland) and osmolarity was checked using an Osmometer (Fiske Associates, Waterford, PA, USA). The viscosity of the CSNP suspension was determined at ocular physiological (≈35 ± 0.5 ◦C) and non-physiological (≈25 ± 0.5 ◦C) temperatures [40] using a sine-wave vibro viscometer (Model SV-10, A & D Co., Ltd., Tokyo, Japan). The viscosity of simulated tear fluid (STF) was also measured as a control for comparative analysis.

#### *2.5. In Vitro Drug Release and Release Kinetics*

The suspension of optimal formulation (F2) was made isotonic with mannitol solution and subjected to in vitro drug release study. Simulated tear fluid (STF) with 0.25%, *w*/*v* of Tween-80 was used as a release medium for this experiment. The STF was prepared by dissolving NaCl (3.4 g), NaHCO<sup>3</sup> (1.1 g), KCl (0.7 g), and CaCl2·2H2O (0.04 g) in 500 mL of Milli-Q® water. A dialysis bag was used as a release barrier [41]. Around 1 mL of F2 suspension (~821.5 µg of TZP) was put into the dialysis bags, and both ends of the bags were tied with threads. The bags filled with formulation were put into beakers containing 50 mL of STF. All the beakers were put into a shaking water bath (100 strokes per min) at 37 ± 1 ◦C. At different elapsed times, 1 mL aliquots were taken out from the beakers and an equal volume of fresh release medium was put into the beakers after each sampling. The collected aliquots were centrifuged at 13,500 rpm (10 min at 10 ◦C). The supernatants were collected and 30 µL was injected into the HPLC-UV system to analyze the TZP concentration. The drug release from TZP aqueous suspension (TZP-AqS) was also checked as a control. TZP-AqS was prepared by suspending TZP (~8.22 mg) in 10 mL of Polysorbate-20 solution (0.5%, *w*/*v*) in Milli-Q® water [42,43]. All the experiments were performed in triplicate. Cumulative amount of TZP released as %DR was calculated using Equation (3).

%*DR* = *Conc*. <sup>µ</sup>g·mL−<sup>1</sup> × *Dilution Factor* × *Volume o f release medium* (mL) *Initial dose o f TZP used f or the experiment* (µg) × 100 (3)

In vitro release data were fitted into release kinetic model equations including zeroorder, first-order, Higuchi matrix square-root, Hixson–Crowell cube-root and Korsmeyer– Peppas. The best-fit model for the release of TZP from CSNPs was classified on the basis of highest correlation coefficient (*R* 2 ) value. From the slope and intercept of the plots of the kinetic models, two specific release kinetic parameters, i.e., *n* and *k* were calculated [44]. The *n*-value is also known as release/diffusion exponent, suggesting the mechanism of drug release from the CSNPs and *k* denotes the rate constant [19,45,46].

### *2.6. Antimicrobial Study*

Testing of the antimicrobial activity of the F2 and TZP AqS was performed by the agar diffusion method [47,48]. Bacterial strains for the assessment were obtained from the Department of Pharmaceutics, College of Pharmacy, King Saud University. The strains were chosen from the Global Priority Pathogens List. Three Gram-positive American type culture collections (ATCC) of *Bacillus subtilis*, *Staphylococcus aureus*, and MRSA (SA-6538) were used for their TZP susceptibility (F2). The Mueller–Hinton agar (MHA) plates were prepared and each strain was spread on to the separate plates. Wells of 6 mm diameter were created by a sterile borer. In the first well, 40 µL of TZP-AqS (32.86 µg of TZP) was placed, into the second well 40 µL of F2 (~32.86 µg of TZP), and in the third well, the same volume of blank CSNPs (without TZP) was inoculated. After 1 h, the plates were incubated at 37 ◦C for 24 h, and after 24 h the zone of inhibition for each product was measured. The entire assessment was performed in triplicate.

#### *2.7. In Vivo Animal Study*

New Zealand albino rabbits weighing 2.5–3.5 kg were made available by the College of Pharmacy, Animal care and use center, King Saud University, Riyadh, Saudi Arabia, for the in vivo eye irritation experiment. The protocol for the animal use was approved by the King Saud University Research Ethics Committee with approval number KSU-SE-18–25 (amended). Animals were housed in light-controlled air-conditioned areas at 75 ± 5% RH according to the Guide for the Care and Use of Laboratory Animals recommended by the center. All the animals were healthy (free from any ocular clinical defects), were kept on a pellet diet (standard for rabbits) with water ad-libitum and fasted overnight before starting the experiment.

### 2.7.1. Ocular Irritation Study

Based on the performance of physical and physicochemical characteristics, in vitro drug release, only the optimal formulation (F2) was chosen for the eye irritation test, which was compared with the blank formulation. The irritation study was performed by following Draize's test in healthy rabbits [29]. The study was performed following the guidelines of the Association for Research in Vision and Ophthalmology (ARVO) for animal use in

ophthalmic and vision research. According to these guidelines, only one eye (the right eye) of all rabbits was chosen for the test formulations and 0.9% NaCl was put into the left eyes (as negative control) to assess the ocular safety of the products. Normally, for one test formulation, a maximum of six rabbits is used. In the present study, we used only three rabbits for one test formulation, as we expected there might be some severe eye irritation and ocular damage, as suggested in a previous report [49]. Thus, six rabbits were divided in two groups for the irritation test of F2 and blank CSNPs (without TZP). Around 40 µL of each product was put into the lower conjunctival sac of each animal of the respective groups. All the rabbits received three consecutive doses in the conjunctival sac of right eyes at intervals of 10 min for the acute eye irritation test. After 1 h of exposure, the treated eyes were periodically examined for any injuries or signs and symptoms in the iris, cornea, and conjunctiva, or any alteration in the treated eyes as compared to the normal eyes. The photographs were captured by slit lamp microscope (Model-4ZL, Takagi, Japan) for irritation scoring purposes. The level of eye irritation was evaluated according to the guideline for scoring [28] on the basis of discomfort to the animals as well as the signs and symptoms such as swelling, redness, edema, or chemosis in the cornea, conjunctiva, and iris or any watery/mucoidal discharge [50]. The scoring was performed and the irritation potential of the tested formulations was categorized according to the described systems [51,52].

### 2.7.2. Transcorneal Permeation

In vitro transcorneal permeation of TZP from CSNPs (F2) across the rabbit cornea, was performed using double-jacketed transdermal diffusion cells assembled with the automated sampling system SFDC 6, LOGAN, New Jersey, NJ, USA [50]. The rabbits used in the irritation test were kept on a washout period for three weeks. After injecting an overdose of a mixture of Ketamine, HCl, and Xylazine, the animals were sacrificed. Eyes were taken out and the corneas were separated. The freshly excised cornea (permeation barrier) was fitted between the donor and receptor compartments in such a way that the epithelial layer of the cornea faced towards the donor compartment of the cell. The receptor compartment of the diffusion cells was filled with STF (pH 7.4) containing Tween-80 (0.25%, *w*/*v*). A small magnetic bar was also put into the receptor compartment. The cells were placed on the LOGAN instrument and water at 37 ± 1 ◦C, was run through the outer jacket. For each group (in triplicate), 500 µL of suspensions of F2 (~410.8 µg of TZP) and TZP-AqS (410.8 µg of TZP) were placed in the donor compartments and the instrument was started with magnetic stirring. Samples from the receptor compartment were collected at different time points up to 4 h. The continuous magnetic stirring could remove air bubbles (if generated during sampling) from the receptor compartment. The concentration of the drug (µg·mL−<sup>1</sup> ) that had passed through the cornea and present in the collected samples was analyzed by the HPLC-UV method [31,32]. The amount of drug that had permeated across the cornea was calculated by considering the volume of receptor compartment (5.2 mL), the cross sectional area (0.5024 cm<sup>2</sup> ) and the initial concentration of TZP (*C<sup>0</sup>* = 821.6 <sup>µ</sup>g·mL−<sup>1</sup> ) using Equation (4) and plotted against time.

$$\text{Amount of drug permuted} \left(\text{ựg} \cdot \text{cm}^{-2}\right) = \frac{\text{Conc.} \left(\text{ựg} \cdot \text{mL}^{-1}\right) \times \text{DF} \times \text{Volume of receptor compartment } \left(\text{mL}\right)}{\text{Area of corona involved} \left(\text{cm}^2\right)}\tag{4}$$

The slope of this plot was used to determine the permeation parameters (steadystate flux, *J*, and apparent permeability, *Papp*). The *Papp* is also known as the permeation coefficient. These permeation parameters were calculated using Equations (5) and (6).

$$J\left(\mu \text{gcm}^{-2} \cdot \text{h}^{-1}\right) = \frac{dQ}{dt} \tag{5}$$

$$P\_{app} \left(\text{cm} \cdot \text{h}^{-1}\right) = \frac{J}{\mathcal{C}\_0} \tag{6}$$

where *Q* is the amount of TZP crossed through the cornea, (*dQ/dt*) is the linear ascent of the slope, *t* is the contact time of the product with the epithelial layer of corneal, and *C*<sup>0</sup> is the initial drug concentration present in the donor compartment of the diffusion cell.

#### *2.8. Statistical Analysis of the Data*

The data are presented as mean with standard deviation (±SD) unless otherwise indicated. Statistical analysis was performed using GraphPad Prism: Version 5 (GraphPad Software, Inc., San Diego, CA, USA). The parameters were compared by *t*-test with *p* values less than 0.05 (*p* < 0.05) considered statistically significant.

#### **3. Results and Discussion**

#### *3.1. Formulation Development*

The ionic-gelation method was used for preparation of the CSNPs where TPP sodium acted as cross-linker [33]. The TZP-CSNPs were optimized by considering the excipients (CS and TPP) concentrations and keeping 120 min of stirring time. The optimization of TZP-CSNPs was performed following our previous publication wherein we optimized indomethacin-loaded CSNPs using a three-factor three-level Box–Behnken experimental design [34]. Thus, in the present study, optimal concentrations of CS and TPP (0.6 and 0.4 mg/mL, respectively) with 120 min stirring time and 10 mg of TZP, resulted in CSNPs with the desired features. Constraints, including the minimum particle size with maximum encapsulation efficiency (%EE), drug loading (%DL) and zeta potential (ZP), were applied for optimization of the TZP-CSNPs. Based on the obtained responses (parameters mentioned in Table 2), the F2 CSNPs were found to be the best one among the three formulations tried (F1–F3). Thus, this formula was selected for further study.

**Table 2.** Physical characteristics of the TZP-CSNPs (Mean ± SD, *n* = 3).


F1–F3 (Formulations 1 to Formulation 3) and PDI = Polydispersity index.

The ionic interaction between the high charge density (six ionic groups) of negatively charged functional groups of TPP and the positively charged quaternary amine groups (NH4) of CS resulted in optimal CSNP formation at particular weight ratios of CS/TPP, with magnetic stirring at 700 rpm at room temperature. Among the three (F1–F3) formulations, F2 was chosen for further studies, based on its smallest particle size with maximum encapsulation efficiency and comparatively better loading capacity. Briefly, at low weight ratio of CS/TPP (81:26 mg with 10 mg of TZP) and at magnetic stirring rate of 700 rpm for 3–4 h was found suitable to obtain optimum-sized particles (129.13 nm) with high encapsulation (82.15%) and better drug loading capacity (7.02%), as shown in Table 2. Before putting the drug into CS solution, it was dissolved in 200 µL of DMSO, due to the highly lipophilic and poorly soluble nature of TZP. It was only 1%, (*v*/*v*) of total volume of the formulation, which is permissible because even for in vitro cytotoxicity studies. In general, by increasing the CS concentration, particle size increases; however, in the case of F2, the size was smaller than F1. This might be due to the fact that the CS was exactly 3.12-fold higher than TPP in F2, while it was 4.15-fold in case of F1. The size of F3 was unexpectedly high, which might be due to very low weight ratio of CS/TPP (CS was 2.77-fold TPP), therefore, due to lack of proper weight ratio of CS/TPP, the ionic interaction between them could not occur properly. Contrary to F1, in F2, the anionic functional groups of TPP showed better ionic interaction with the positively charged amine groups of CS due to their proper weight ratio, which might be the reason for its improved physical performance.

Chitosan (CS) was chosen as main excipient to develop the TZP-loaded CSNPs, because of its natural hydrophilic, biodegradable, and mucoadhesive properties with a non-toxic and non-irritant (to eyes) safety profile. It was expected that CS would maintain and stabilize tear fluids on ocular surfaces, hence would reduce the drainage and prolong the ocular contact time of the nanocarrier [14,19,53]. It has shown minor broad-spectrum antibacterial activity [54] against some Gram-positive and Gram-negative bacteria and also has some antifungal activity [55–57].

Moreover, CS was found to sustain the intraocular penetration of loaded drugs by binding with corneal epithelium and causing reversible loosening of tight junctions of corneal epithelium. Hence, it was determined to be one of the best natural polymers (of biological origin) for ophthalmic purposes [58]. It has been extensively utilized for the development of numerous products for ocular use including nanoemulsions [59], indomethacin-loaded nanocapsules [33], cyclosporine-A-loaded CSNPs [60], ofloxacin [27], and acyclovir-loaded microspheres [61]. Moreover, due to electrostatic interaction with the negatively charged mucin layers, the corneal and conjunctival epithelial penetration of CS-NP/liposome-CSNP complexes were achieved [28,61].

#### *3.2. Particle Characterization and Morphology of CSNPs*

Dynamic light scattering (DLS) analysis by Zetasizer was used for the characterization of the developed CSNPs including the size, polydispersity, and zeta potential. In the case of F1, at CS/TPP (54 mg and 13 mg each) and 10 mg TZP, the obtained particle size (227.23 nm) was larger, with a higher PDI value (0.833). The zeta potential was +20.6 mV and encapsulation efficiency was comparatively lower (61.4%), but the loading capacity was similar (7.97%), as compared to F2. F2, with CS/TPP weight ratio of 81/39 mg and 10 mg of drug, the obtained particle size was largest (129.13 nm), with a PDI of 0.373. For F3, a CS/TPP weight ratio of 108/39 mg and 10 mg of drug, the obtained particle size was the largest of the three formulations (472.06 nm) with a slightly higher PDI (0.576), as compared to F2 (0.373). The resultant low particle size of the developed CSNPs in this investigation could be suitable for ocular application as human eyes can tolerate the particulate materials with sizes ≤10 µ without any potential ocular irritation or corneal abrasion while the larger particles may cause scratching of ocular surfaces and discomfort to eyes [62]. Thus, reduction in nanoparticle size would improve patient compliance and provide comfort during the dose administration.

For F3, the zeta potential was +36.6 mV which is excellent for stable dispersion of the CSNPs but the encapsulation efficiency (69.92%) and drug loading (4.45%) were the lowest among the three (F1–F3) developed formulations. The results of physical characterization, including the particle size, PDI, zeta potential, encapsulation efficiency (%EE), and drug loading (%DL), are summarized in Table 2. The high positive zeta potential values of CSNPs (+20.6 to +36.6 mV) obtained in the present investigation, predict good physical stability of the developed colloidal nanocarriers (CSNPs). The same surface charges (positive) have strong electrostatic repulsion among the NPs to prevent self-aggregation. The polydispersity index measures the NPs' size distribution, where small values are indicative of the unimodal distribution and stable dispersion of the CSNPs in the medium. The particle size and zeta potential distribution curves of the optimized CSNPs are represented in Figure 2a,b, respectively.

Therefore, based on the above findings (Table 2), F2 was selected as the best formulation among the three developed formulations (F1–F3). To substantiate its suitability for ocular application, F2 was chosen for morphological characterization by TEM imaging. The TEM imaging of F2 revealed discrete spherical particles, well separated from each other (i.e., without potential aggregation) with solid, densely structured NPs. TEM images at two separate magnifications (80 K and 50 K) are shown in Figure 3.

**Figure 2.** Particle size distribution (**a**) and zeta potential distribution (**b**) of the optimized TZPloaded CSNPs (F2). **Figure 2.** Particle size distribution (**a**) and zeta potential distribution (**b**) of the optimized TZP-loaded CSNPs (F2). *Molecules* **2022**, *27*, x FOR PEER REVIEW 11 of 22

the vacuum to be free of any interference [64]. Thus, vacuum and the voltage of the electron-beam irradiation are important conditions because the highly dispersed NPs remain mobile under the electron-beam irradiation, which may interfere with the imaging. There-**Figure 3.** TEM images of TPZ-loaded CSNPs (F2): Performed at 80,000 magnification and 200 nm scale (**a**) and at 50,000 magnification and 500 nm scale (**b**). **Figure 3.** TEM images of TPZ-loaded CSNPs (F2): Performed at 80,000 magnification and 200 nm scale (**a**) and at 50,000 magnification and 500 nm scale (**b**).

fore, the TEM analysis was performed under light microscopy operated at 80 kV accelerating voltage to provide high resolution and prevent any damage caused by higher-energy electron irradiation. The low accelerating voltage (80 kV), as compared to higher energy (200–1000 kV) electrons used for metallic particles and intermediate voltage (200–400 kV) for high resolution electron microscopy of non-metallic and biological specimens. *3.3. X-ray Diffraction Analysis*  The X-ray diffractogram spectra of TZP, pure TZP, low molecular weight chitosan (CS), Tripolyphosphate sodium (TPP), mannitol, and TZP-loaded CSNPs (F2) are illustrated in Figure 4. The diffractogram of pure TZP (Figure 4a) has characteristic sharp and intense peaks at 2*θ* values of 14.4°, 23.8°, 38.1°, and 44.3°, with intensities of 3490 cps (with The characterization of NPs involves the exploration of the structures at the nano scale. The size, shape, and any surface layers/absorbents on NPs is a crucial first step to understand the relationships between NPs, performance, quality, and safety/toxicity. It is also important whether any changes have occurred as a result of sample preparation, e.g., oxidation/reduction, during the process of checking the morphology of NPs by TEM [39,63].

I/I0 of 100 and Bragg's or *d*-value 6.145), 2526 cps (with I/I0 of 73 and *d*-value 3.735), 2492 cps (with I/I0 of 72 and *d*-value 2.36) and 1036 cps (with I/I0 of 30 and *d*-value 2.04), respec-

intensities of 1524 cps (I/I0 of 100 and *d*-value 2.366) and 593 cps (I/I0 of 39 and *d*-value 2.047), while the presence of a less intense (237 cps) broad peak at 2*θ* of 19.9° with a *d*value of 4.457 and I/I0 of only 16.0, suggests the less crystalline, or somewhat amorphous, characteristics of CS. Figure 4c (for TPP), shows intense peaks at 2*θ*s of 19.8°, 29.1°, 32.5°, and 36.6°, with intensities of 696 cps (I/I0 of 79 and *d*-value 4.48), 564 cps (I/I0 of 64 and *d*value 3.066), 884 cps (I/I0 of 100 and *d*-value 2.753), and 468 cps (I/I0 of 53 and *d*-value 2.453), respectively, suggesting the crystallinity of TPP. Figure 4d (for mannitol) has intense peaks at 2*θ* values of 15.0°, 19.1°, 21.4°, and 23.8° with intensities of 1814 cps (I/I0 of 36 and *d*-value 5.901), 4026 cps (I/I0 of 80 and *d*-value 4.642), 1861 cps (I/I0 of 37 and *d*-value 4.148), and 5092 cps (I/I0 of 100 and *d*-value 3.736), indicating the crystalline character of

The diffractogram of TZP-encapsulated CSNPs (F2) lyophilized with mannitol (Figure 4e) has low intensity characteristic peaks of TZP at 2*θ* values of 14.6°, 23.4°, 38.7°, and 44.3°, with intensities of 366 cps (I/I0 of 36 and *d*-value 6.062), 1042 cps (I/I0 of 100 and *d*-value 3.798), 213 cps (I/I0 of 21 and *d*-value 2.324), and 203 cps (I/I0 of 20 and *d*-value 2.043), indicating that the TZP was entrapped in the core of the NPs or in the matrix of the polymer in an amorphous state and there was no any degradation interaction with the mannitol. Similarly, in Figure 4f, almost diminished or very low intensity characteristic peaks of TZP can be seen. However, the characteristic peaks of CS at 2*θ*s of 19.3° and 38.2° with intensities of 543 cps (I/I0 of 77 and *d*-value 4.595) and 169 cps (I/I0 of 169 and *d*-value 2.453) can be seen in Figure 4f. Moreover, the characteristic crystalline peaks of TPP at 2*θ*s of 32.5° and 36.6° with intensities of 713 cps (I/I0 of 100 and *d*-value 2.752) and 342 cps (I/I0 of 48 and *d*-value 2.453), indicate that the TZP was well encapsulated in amorphous form

into the core of CSNPs rather than adsorbed onto the surfaces of the NPs.

mannitol.

Due to some challenges, such as image overlapping, difficulty in detection of small particles, and obscured signals during observation due to the presence of the surrounding matrix and background noise, the samples were diluted with Milli-Q water before the analysis. This enabled the electron beams to transmit through the highly diluted specimens and interact with them for surface imaging, when the NPs should be present around the vacuum to be free of any interference [64]. Thus, vacuum and the voltage of the electron-beam irradiation are important conditions because the highly dispersed NPs remain mobile under the electron-beam irradiation, which may interfere with the imaging. Therefore, the TEM analysis was performed under light microscopy operated at 80 kV accelerating voltage to provide high resolution and prevent any damage caused by higher-energy electron irradiation. The low accelerating voltage (80 kV), as compared to higher energy (200–1000 kV) electrons used for metallic particles and intermediate voltage (200–400 kV) for high resolution electron microscopy of non-metallic and biological specimens.

#### *3.3. X-ray Diffraction Analysis*

The X-ray diffractogram spectra of TZP, pure TZP, low molecular weight chitosan (CS), Tripolyphosphate sodium (TPP), mannitol, and TZP-loaded CSNPs (F2) are illustrated in Figure 4. The diffractogram of pure TZP (Figure 4a) has characteristic sharp and intense peaks at 2*θ* values of 14.4◦ , 23.8◦ , 38.1◦ , and 44.3◦ , with intensities of 3490 cps (with I/I<sup>0</sup> of 100 and Bragg's or *d*-value 6.145), 2526 cps (with I/I<sup>0</sup> of 73 and *d*-value 3.735), 2492 cps (with I/I<sup>0</sup> of 72 and *d*-value 2.36) and 1036 cps (with I/I<sup>0</sup> of 30 and *d*-value 2.04), respectively, indicating the crystallinity of pure TZP. The diffractogram of low-molecular-weight CS (Figure 4b) has only two intense peaks at 2*θ* values of 38.0◦ and 44.2◦ with intensities of 1524 cps (I/I<sup>0</sup> of 100 and *d*-value 2.366) and 593 cps (I/I<sup>0</sup> of 39 and *d*-value 2.047), while the presence of a less intense (237 cps) broad peak at 2*θ* of 19.9◦ with a *d*-value of 4.457 and I/I<sup>0</sup> of only 16.0, suggests the less crystalline, or somewhat amorphous, characteristics of CS. Figure 4c (for TPP), shows intense peaks at 2*θ*s of 19.8◦ , 29.1◦ , 32.5◦ , and 36.6◦ , with intensities of 696 cps (I/I<sup>0</sup> of 79 and *d*-value 4.48), 564 cps (I/I<sup>0</sup> of 64 and *d*-value 3.066), 884 cps (I/I<sup>0</sup> of 100 and *d*-value 2.753), and 468 cps (I/I<sup>0</sup> of 53 and *d*-value 2.453), respectively, suggesting the crystallinity of TPP. Figure 4d (for mannitol) has intense peaks at 2*θ* values of 15.0◦ , 19.1◦ , 21.4◦ , and 23.8◦ with intensities of 1814 cps (I/I<sup>0</sup> of 36 and *d*-value 5.901), 4026 cps (I/I<sup>0</sup> of 80 and *d*-value 4.642), 1861 cps (I/I<sup>0</sup> of 37 and *d*-value 4.148), and 5092 cps (I/I<sup>0</sup> of 100 and *d*-value 3.736), indicating the crystalline character of mannitol.

The diffractogram of TZP-encapsulated CSNPs (F2) lyophilized with mannitol (Figure 4e) has low intensity characteristic peaks of TZP at 2*θ* values of 14.6◦ , 23.4◦ , 38.7◦ , and 44.3◦ , with intensities of 366 cps (I/I<sup>0</sup> of 36 and *d*-value 6.062), 1042 cps (I/I<sup>0</sup> of 100 and *d*-value 3.798), 213 cps (I/I<sup>0</sup> of 21 and *d*-value 2.324), and 203 cps (I/I<sup>0</sup> of 20 and *d*-value 2.043), indicating that the TZP was entrapped in the core of the NPs or in the matrix of the polymer in an amorphous state and there was no any degradation interaction with the mannitol. Similarly, in Figure 4f, almost diminished or very low intensity characteristic peaks of TZP can be seen. However, the characteristic peaks of CS at 2*θ*s of 19.3◦ and 38.2◦ with intensities of 543 cps (I/I<sup>0</sup> of 77 and *d*-value 4.595) and 169 cps (I/I<sup>0</sup> of 169 and *d*-value 2.453) can be seen in Figure 4f. Moreover, the characteristic crystalline peaks of TPP at 2*θ*s of 32.5◦ and 36.6◦ with intensities of 713 cps (I/I<sup>0</sup> of 100 and *d*-value 2.752) and 342 cps (I/I<sup>0</sup> of 48 and *d*-value 2.453), indicate that the TZP was well encapsulated in amorphous form into the core of CSNPs rather than adsorbed onto the surfaces of the NPs.

**Figure 4.** XRD patterns of pure TZP (**a**), low MW CS (**b**), TPP (**c**), mannitol (**d**), TZP-CSNPs (F2) lyophilized with mannitol (**e**), and TZP-CSNPs (F2) (**f**). **Figure 4.** XRD patterns of pure TZP (**a**), low MW CS (**b**), TPP (**c**), mannitol (**d**), TZP-CSNPs (F2) lyophilized with mannitol (**e**), and TZP-CSNPs (F2) (**f**).

#### *3.4. Physicochemical Characterization*

The transparency, drug content, osmolarity, pH, and viscosity of the TPZ-CSNPS were tested and are summarized in Table 3. Osmolarity of the CSNPs was measured in the range of 302–306 mOsmol·L −1 , which is almost equal to the osmolarity of tear fluid (302 mOsmol·L −1 ) in normal eye conditions [65]. The viscosity of F2 (20.85 cPs at 35 ◦C, normal ocular surface temperature) was almost equal to the optimum viscosity (20 cPs) that the human eye can easily tolerate without any blurring of vision.


**Table 3.** Physicochemical characteristics of TZP-CSNPs (mean ± SD, *n* = 3).

\* Simulated tear fluid (STF) was prepared by dissolving 0.68 g NaCl, 0.22 g NaHCO3, 0.008 g CaCl2·2H2O, and 0.14 g KCl in 100 mL of Milli-Q® water, cPs (Centipoises, 1 cP = 1 mPa·s) and " . . . " indicates that the drug content was not measured.

#### *3.5. In Vitro Drug Release and Kinetics*

The in vitro release of the drug through dialysis bags in simulated tear fluid (pH 7) with 0.25% *w*/*v* of Tween-80 was found to be suitable for the release of TZP from NPs and the aqueous suspension. Tween-80 was added to increase the solubility of the highly lipophilic and poorly soluble nature of TZP into aqueous environment. The in vitro drug release profile (Figure 5a,b) shows that around 82% of the drug was released from the TZP-AqS within 1 h, while it took 12 h to release 78% of the drug from the NPs in a sustained manner. From assessment of the release profiles, TZP-AqS showed that almost all the drug was released from the suspension within 3 h, suggesting that the optimized TZP-loaded CSNPs (F2) could be an important tool for prolonged and sustained release of TZP for topical ocular application.

The sustained release property of the NPs was further confirmed by applying the release kinetics models [17]. In general, the CSNPs show a two-step release pattern—an initial burst release phase followed by a slow-release pattern. In the present investigation only sustained release of the drug occurred from the CSNPs, which might be due to the low aqueous solubility of the drug. This is also beneficial to maintain the therapeutic index of the drug for prolonged effect with reduced dosing frequency.

Applying the different kinetic models, it was observed that the in vitro release of TZP from F2 could be better explained by two models (Higuchi's square root and first order release models). The curve between the square root of time and the fraction of drug released was almost linear (Higuchi's square root model) and its extrapolation crossed through the origin (Figure 5c). The linearity in the release profile (as suggested in Higuchi's square root model) indicated the sustained release property of the optimized CSNPs (F2). Among the applied models, the highest value of the correlation coefficient (*R* 2 ), 0.9976, was found with the Higuchi's square root model (Table 4). Considering the *R* <sup>2</sup> values and slope of different kinetic equations, the diffusion or release exponent (*n*-value) was calculated. The obtained *n*-value (0.109) according to Higuchi's square root model (for F2) indicated that the mechanism of drug release from F2 followed Fickian diffusion. Apart from Higuchi's square root model, the second-best fit model for the release of TZP from the optimized CSNPs was the first-order model (with *R* <sup>2</sup> = 0.9936) (Figure 5d). The values of the correlation coefficient and release-exponents are presented in Table 4.

**Figure 5.** In-vitro release profile of TZP in STF from CSNPs as compared to TZP-AqS (**a**); from the aqueous suspension only (**b**); release kinetics of TZP from CSNPs (F2) that followed Higuchi's square root of time plot (**c**); and second-best fit was the first-order model (**d**). "\*" *p* < 0.05; TZP-CSNPs (F2) vs. TZP-AqS. **Figure 5.** In-vitro release profile of TZP in STF from CSNPs as compared to TZP-AqS (**a**); from the aqueous suspension only (**b**); release kinetics of TZP from CSNPs (F2) that followed Higuchi's square root of time plot (**c**); and second-best fit was the first-order model (**d**). "\*" *p* < 0.05; TZP-CSNPs (F2) vs. TZP-AqS.

**Table 4.** Release kinetics model equations. **Table 4.** Release kinetics model equations.


of time) 0.9976 0.2525 0.10964 *R* <sup>2</sup> = Coefficient of correlation and *n* = Release/diffusion exponent.

*R*2 = Coefficient of correlation and *n* = Release/diffusion exponent. In general, the sustained release of drugs from the biodegradable polymeric matrix (CS-matrix in the present study) is assumed to occur by three different mechanisms—(a) release of drug from the polymer matrix due to the erosion of the matrix, (b) diffusion of drug molecules through the polymer matrix, or (c) a combination of diffusion of drug molecules and degradation of polymer matrix [19,66,67]. The pattern of drug release from CSNPs in the present investigation is indicative of the mechanism of degradation and ero-In general, the sustained release of drugs from the biodegradable polymeric matrix (CS-matrix in the present study) is assumed to occur by three different mechanisms—(a) release of drug from the polymer matrix due to the erosion of the matrix, (b) diffusion of drug molecules through the polymer matrix, or (c) a combination of diffusion of drug molecules and degradation of polymer matrix [19,66,67]. The pattern of drug release from CSNPs in the present investigation is indicative of the mechanism of degradation and erosion of chitosan molecules, which was the reason for the continuous, sustained release of TZP from F2 and control of the release pattern for up to 12 h.

#### sion of chitosan molecules, which was the reason for the continuous, sustained release of *3.6. Antimicrobial Activity of TZP-CSNPs (F2)*

TZP from F2 and control of the release pattern for up to 12 h. *3.6. Antimicrobial Activity of TZP-CSNPs (F2)*  The results of an antimicrobial susceptibility test by the agar diffusion method are summarized in Table 5A. The TZP-loaded CSNPs (F2) showed significantly (*p* < 0.05) im-The results of an antimicrobial susceptibility test by the agar diffusion method are summarized in Table 5A. The TZP-loaded CSNPs (F2) showed significantly (*p* < 0.05) improved activity against Gram-positive bacteria such as *B. subtilis* and *S. aureus*, including one MRSA strain (SA 6538), as compared to TZP-AqS (Figure 6). Relatively little activity was noted for the blank CSNPs, as compared to the two tested formulations.

proved activity against Gram-positive bacteria such as *B. subtilis* and *S. aureus*, including


**Table 5.** Zone of inhibitions obtained in agar diffusion test by F2 as compared to TZP-AqS. Blank CSNPs were used as control.

\* SED = Standard error of the difference and D = Difference between two means, SD = Standard deviation, n = times repeated the experiment, p = probability (for significance), q = Studentized range statistic and F2 = Formulation 2 (TZP-CSNPs). *Molecules* **2022**, *27*, x FOR PEER REVIEW 15 of 22

**Figure 6.** Antimicrobial activity of TZP-containing formulations as compared to blank CSNPs against some Gram-positive bacteria, including one MRSA strain. Results are presented as mean ± SD, *n* = 3. "\*" *p* < 0.05; F2 vs. other formulations (for *B. subtilis*), "#" *p* < 0.05; F2 vs. other formulations (for *S. aureus*). "\$" *p* < 0.05; F2 vs. other formulations (for MRSA SA-6538). **Table 5.** Zone of inhibitions obtained in agar diffusion test by F2 as compared to TZP-AqS. Blank **Figure 6.** Antimicrobial activity of TZP-containing formulations as compared to blank CSNPs against some Gram-positive bacteria, including one MRSA strain. Results are presented as mean ± SD, *n* = 3. "\*" *p* < 0.05; F2 vs. other formulations (for *B. subtilis*), "#" *p* < 0.05; F2 vs. other formulations (for *S. aureus*). "\$" *p* < 0.05; F2 vs. other formulations (for MRSA SA-6538).

CSNPs were used as control. **(A) Microorganisms Zone Diameters (mm), Mean ± SD,** *n* **= 3 By TPZ-AqS By TPZ-CSNPs (F2) By Blank CSNPs**  *B. subtilis* 25.77 ± 3.23 34.83 ± 2.78 7.83 ± 1.59 *S. aureus* 23.63 ± 2.28 36.93 ± 2.65 8.36 ± 1.47 MRSA (*SA* 6538) 23.46 ± 1.27 32.46 ± 1.18 5.66 ± 0.98 **(B) Statistical Analysis by One-Way Analysis of Variance Tukey's Multiple Comparison Test Mean Difference q = Sq. Root \* (D/SED)** *p* **< 0.05**  TZP-AqS vs. TZP-CSNPs One-way analysis of variance followed by Tukey's multiple comparison test using GraphPad Prism V-5.0 were used to check the level of significance between the two formulations, as compared to blank CSNPs (against the tested microorganism). *p* < 0.05 was considered as the level of significance. The data obtained are presented in Table 5B. The improved antimicrobial activity of the TZP formulations indicates that the formulation processes did not alter the intrinsic or inherent antimicrobial property of TZP. Moreover, the processes did not alter the structure–activity relationship of TZP. Therefore, we can conclude that the encapsulation of TZP into the CSNPs not only increases the bioavailability of the drug but could also increase its antimicrobial potency against the tested microorganisms.

#### (F2) −10.46 10.64 Yes TZP-AqS vs. TZP-CSNPs *3.7. Ocular Irritation Study*

croorganisms.

(F2) 17.00 17.31 Yes TZP-AqS vs. Blank CSNPs 27.46 27.95 Yes The scores and signs of discomfort during the eye irritation study for CSNPs (blank and F2) are shown in Table 6. No obvious symptoms of discomfort were noted in the

range statistic and F2 = Formulation 2 (TZP-CSNPs).

\* SED = Standard error of the difference and D = Difference between two means, SD = Standard deviation, n = times repeated the experiment, p = probability (for significance), q = Studentized

One-way analysis of variance followed by Tukey's multiple comparison test using GraphPad Prism V-5.0 were used to check the level of significance between the two formulations, as compared to blank CSNPs (against the tested microorganism). *p* < 0.05 was considered as the level of significance. The data obtained are presented in Table 5B. The improved antimicrobial activity of the TZP formulations indicates that the formulation processes did not alter the intrinsic or inherent antimicrobial property of TZP. Moreover, the processes did not alter the structure–activity relationship of TZP. Therefore, we can conclude that the encapsulation of TZP into the CSNPs not only increases the bioavailability of the drug but could also increase its antimicrobial potency against the tested mi-

rabbits treated with the two products. Figure 7a,f, are pictures of NaCl-, F2-, and blank CSNP- treated left eyes of rabbits, respectively. Figure 7b shows mild redness (red arrow) without inflammation of conjunctiva but with mild abnormal discharge (black arrow), 1 h after dosing with blank CSNPs. The redness and slight mucoidal discharge continued until 3 h (Figure 7c). These symptoms disappeared at 6 h (Figure 7d) and the eye regained its normal condition (green arrow) at 24 h (Figure 7e). In contrast, no such findings were noted in the F2 treated eyes even at 1 h (Figure 7g). In fact, the F2 treated eyes did not show any symptoms of irritation at any time-point (Figure 7g–j). The normal recovery in the blank CSNP-treated animals was due to the strong natural defensive mechanism of the eyes themselves. Moreover, this might be attributable to non-irritant properties of the biocompatible excipients (CS and TPP) in the formulation. *3.7. Ocular Irritation Study*  The scores and signs of discomfort during the eye irritation study for CSNPs (blank and F2) are shown in Table 6. No obvious symptoms of discomfort were noted in the rabbits treated with the two products. Figure 7a,f, are pictures of NaCl-, F2-, and blank CSNPtreated left eyes of rabbits, respectively. Figure 7b shows mild redness (red arrow) without inflammation of conjunctiva but with mild abnormal discharge (black arrow), 1 h after dosing with blank CSNPs. The redness and slight mucoidal discharge continued until 3 h (Figure 7c). These symptoms disappeared at 6 h (Figure 7d) and the eye regained its normal condition (green arrow) at 24 h (Figure 7e). In contrast, no such findings were noted in the F2 treated eyes even at 1 h (Figure 7g). In fact, the F2 treated eyes did not show any symptoms of irritation at any time-point (Figure 7g–j). The normal recovery in the blank CSNP-treated animals was due to the strong natural defensive mechanism of the eyes themselves. Moreover, this might be attributable to non-irritant properties of the biocom-

*Molecules* **2022**, *27*, x FOR PEER REVIEW 16 of 22


**Table 6.** Weighted irritation scores during the testing of F2 and blank CSNPs in rabbit eyes. patible excipients (CS and TPP) in the formulation.

F2 = Formulation 2 (TZP-CSNPs). F2 = Formulation 2 (TZP-CSNPs).

**Figure 7.** Treated rabbit eyes during irritation experiments. Figures (**a**,**f**) showing the NaCl-treated eye of two groups. After topical application of blank CSNPs at 1 h, exhibiting mild redness (red **Figure 7.** Treated rabbit eyes during irritation experiments. (**a**,**f**) showing the NaCl-treated eye of two groups. After topical application of blank CSNPs at 1 h, exhibiting mild redness (red arrow) without inflammation of conjunctiva but with mild abnormal discharge (black arrow) (**b**); at 3 h (**c**); at 6 h (**d**) and at 24 h (**e**). After topical application of TZP-CSNPs (F2) at 1 h (**g**); at 3 h (**h**); at 6 h (**i**) and at 24 h (**j**). Other images show no redness or abnormal discharge, with green arrows indicating normal features.

As a result of application of blank CSNPs, a slight irritation was found in one animal with some mucoidal discharge, which was given a score of 1. No opacity in the treated eyes was found. Therefore, the cornea, iris, and conjunctiva scored 0 for both the formulations. Adopting the scoring classification system for ocular irritation [52], the maximum mean total scores (MMTS) were calculated. The MMTS after 24 h, for the blank CSNPs was 19.33 (>15.1 but <25), while it was only 9.00 for F2 (>2.6 but <15) (Table 7). Therefore, the blank-CSNP formulation was judged to be "mildly irritating" while F2 was "minimally irritating" to the rabbit eyes. The low MMTS value for F2 indicates the merits of the product for ocular use.


**Table 7.** Maximum mean total score (MMTS) calculations based on the scores represented in Table 5.

All animals remained active and healthy throughout the study, this demonstrated that the TZP-CSNPs were non-irritant to the rabbits' eyes. No traces of formulation were found on visual observation after 24 h, signifying the complete disposition and degradation of the treatments. Overall, the "minimally irritating" nature of TZP-CSNPs (F2) in the present investigation was demonstrated, in agreement with previous reports where chitosan based nanocarriers were applied for topical ocular delivery of dexamethasone [50], forskolin [68], and clarithromycin [69]. Thus, we conclude that TZP-loaded CSNPs were tolerated well by rabbit eyes.

#### *3.8. Transcorneal Permeation of TZP*

For this study, we used Tween-80 at 0.25%, (*w*/*v*) added to the STF to enhance the solubility of TZP into the release medium, because TZP is a highly lipophilic drug. The study was performed for 4 h only, because we did not supply any nutrients to the corneal tissue during the experiment. From the graphs in Figure 8 and the values of permeation parameters (Table 8), the CSNPs (F2) demonstrated linearity in the permeation of encapsulated drugs as compared to the conventional formulation (TZP-AqS). However, the cumulative amounts of permeated TZP at 4 h were 51.74 and 58.05 <sup>µ</sup>g·cm−<sup>2</sup> for TZP-AqS and F2, respectively. The pattern of drug permeation was completely different, as a significantly (*<sup>p</sup>* < 0.05) higher quantity of the drug (33.41 <sup>µ</sup>g·cm−<sup>2</sup> ) permeated from TZP-AqS within 1 h, compared to F2 (only 16.05 <sup>µ</sup>g·cm−<sup>2</sup> 1 h). Similarly, 49.81 <sup>µ</sup>g·cm−<sup>2</sup> of drug permeated from TZP-AqS at 2 h, while a similar quantity took 3.5 h from F2 (50.41 <sup>µ</sup>g·cm−<sup>2</sup> ). Around a 1.6-fold increase in flux (*J*) and *Papp* of the drug was achieved by F2 as compared to AqS, as represented in Table 7. Finally, from the pattern of permeation profiles, we conclude that the developed nano-carriers (F2) could provide sustained delivery of the encapsulated TZP, compared to the conventional suspension of the drug. Moreover, we expected that the developed TZP-encapsulated CSNPs would enhance the prolonged and sustained release of the drug into eyes, hence would improve its ocular bioavailability.

**Figure 8.** Transcorneal permeation of TZP from F2 and TZP-AqS (Mean ± SD, *n* = 3). **Figure 8.** Transcorneal permeation of TZP from F2 and TZP-AqS (Mean ± SD, *n* = 3).

**4. Conclusions Table 8.** Parameters of transcorneal permeation for F2 and TZP-AqS (Mean ± SD, *n* = 3).


kinetics investigation on in vitro data revealed the release of TZP from F2 primarily fol-

conclude that the developed nano-carriers (F2) could provide sustained delivery of the encapsulated TZP, compared to the conventional suspension of the drug. Moreover, we expected that the developed TZP-encapsulated CSNPs would enhance the prolonged and sustained release of the drug into eyes, hence would improve its ocular bioavailability.

**Parameters TZP-AqS CSNPs (F2)** 

(μg·cm−2) at 4 h 51.74 ± 2.31 58.05 ± 2.44

Steady-state flux, *J* (μg·cm−2·h−1) 17.50 ± 3.32 28.12 ± 1.41 Permeability coefficient, *Papp* (cmh−1) (2.13 ± 0.41) × 10−2 (3.42 ± 0.17) × 10−<sup>2</sup>

**Table 8.** Parameters of transcorneal permeation for F2 and TZP-AqS (Mean ± SD, *n* = 3).

Cumulative amount of drug permeated

#### lowed the Higuchi square root model (*R*2 = 0.9976 and release exponent, *n* = 0.1096) indi-**4. Conclusions**

cating the mechanism was Fickian diffusion. The optimized CSNPs (F2) showed a 1.35– 1.56-fold increase in the antibacterial activity of TZP against some Gram-positive microorganisms with highest value of zone of inhibition (36.9 mm) against *B. subtilis*. No sign of discomfort in the eyes of rabbits during the irritation test indicated excellent ocular tolerance, around 2–4-fold increased flux (≈28.1 μg/cm2/h), and apparent permeability (≈3.4 × 10−2 cm/h) with the highest amount of drug permeated (≈58.05 μg/cm2 at 4 h), indicating its higher transcorneal permeation compared to AqS. Though it might be out of scope of the present communication, further investigation has been performed in rabbit eyes to determine the ocular bioavailability of tedizolid (the active form of TZP). Approximately 2.6 to 5.8-fold improved pharmacokinetic parameters were obtained with F2, as compared to its counter formulation (TZP-AqS). Outcomes of the investigation were reported in our previous publication during the application of a The results of particle characterization, physicochemical, morphological, and in vitro release properties showed an efficient encapsulation (≈61.4–82.2%) of TZP into CSNPs by ionic gelation of CS and TPP. The reported HPLC-UV method was successful for the analysis of TZP. In vitro release profiling suggests a sustained release of TZP from optimal CSNPs (F2) for up to 12 h (81.6 ± 5.84%) in STF (pH 7) with Tween-80 (0.25% *w/v*). Release kinetics investigation on in vitro data revealed the release of TZP from F2 primarily followed the Higuchi square root model (*R* <sup>2</sup> = 0.9976 and release exponent, *n* = 0.1096) indicating the mechanism was Fickian diffusion. The optimized CSNPs (F2) showed a 1.35–1.56-fold increase in the antibacterial activity of TZP against some Gram-positive microorganisms with highest value of zone of inhibition (36.9 mm) against *B. subtilis*. No sign of discomfortin the eyes of rabbits during the irritation test indicated excellent ocular tolerance, around 2–4-fold increased flux (≈28.1 <sup>µ</sup>g/cm2/h), and apparent permeability (≈3.4 <sup>×</sup> <sup>10</sup>−<sup>2</sup> cm/h) with the highest amount of drug permeated (≈58.05 <sup>µ</sup>g/cm<sup>2</sup> at 4 h), indicating its higher transcorneal permeation compared to AqS.

Though it might be out of scope of the present communication, further investigation has been performed in rabbit eyes to determine the ocular bioavailability of tedizolid (the active form of TZP). Approximately 2.6 to 5.8-fold improved pharmacokinetic parameters were obtained with F2, as compared to its counter formulation (TZP-AqS). Outcomes of the investigation were reported in our previous publication during the application of a developed and validated UPLC-MS/MS method for the quantification of tedizolid in rabbit aqueous humor [30]. The CSNP-based controlled delivery of TZP would have potential ocular and other topical or oral applications. The delivery system might serve as an optimal model to encapsulate therapeutic agents including drugs, peptides, vitamins, enzymes, fatty acids, etc. Moreover, the CSNPs as carriers for TZP have strong potential for topical use treating ocular MRSA infections and associated inflammatory conditions. Further investigations are needed to validate the developed carrier system for its clinical applications to authenticate the safety and efficacy for human trials.

#### **5. Future Prospects**

The developed nanocarrier system (TZP-CSNPs) would be a fruitful exploration for the treatment of MRSA and other Gram-positive microbial ocular infections. This research was expected to give an excellent product at lower cost in the pharmaceutical field. This may utilize the nation's inherent potential to provide a better platform between research (product development) and industrial collaboration. Such a collaborative approach will have all the means to achieve the ambitions, dreams, and visions of any nation. The successful achievement of the goal of this study, i.e., the encapsulation and topical ocular delivery of TZP could improve quality of life and benefit the healthcare system as follows: (a) A focus on promoting preventive care could help clinicians to reduce infectious diseases, which would encourage patients to make use of such an efficient drug delivery system as a primary step to target multiple diseases. (b) This study may help in corporatization with efficient and high-quality healthcare services and service providers that would promote competition among manufacturers and providers. This in turn would improve the capability, efficiency, and productivity of healthcare and treatment. Thus, effectively increasing the number of options available to patients. (c) To achieve the goal of corporatization, the responsibility of health care provision can be transferred to the public sector that will compete against the private sector, which will offer citizens high-quality health care facilities and allow the government to focus on its legislative, regulatory, and supervisory roles.

**Author Contributions:** Conceptualization, M.A.K. and M.A.; methodology, M.A.K.; software, M.I.; validation, M.A.K., A.A. (Abdullah Alshememry), A.A. (Aws Alshamsan) and M.I.; formal analysis, M.A.K.; investigation, M.I.; resources, M.A.K and M.A.; data curation, M.A.K. and M.I.; writing original draft preparation, M.A.K.; writing—review and editing, A.A. (Aws Alshamsan) and A.A. (Abdullah Alshememry); visualization, M.A.K.; supervision, A.A. (Alshamsan. A); project administration, M.A.K., A.A. (Aws Alshamsan) and M.A.; funding acquisition, M.A.K., M.A. and A.A. (Aws Alshamsan). All authors have read and agreed to the published version of the manuscript.

**Funding:** This project was funded by the National Plan for Science, Technology and Innovation (MAARIFAH), King Abdulaziz City for Science and Technology, Kingdom of Saudi Arabia, Award Number (2-17-03-001-0035).

**Institutional Review Board Statement:** The animal study protocol was approved by King Saud University Research Ethics Committee with approval number KSU-SE-18–25 (amended on 8 May 2020) for studies involving animals.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** This study did not report any data.

**Acknowledgments:** Authors are thankful to the National Plan for Science, Technology and Innovation (MAARIFAH), King Abdulaziz City for Science and Technology, Kingdom of Saudi Arabia, for the grant with Award Number 2-17-03-001-0035.

**Conflicts of Interest:** We declare that we have no known competing financial interest or personal relationships that could influence the work reported in this paper.

**Sample Availability:** Samples of the compounds are available from the authors.

#### **References**


## *Article* **Levonorgestrel Microneedle Array Patch for Sustained Release Contraception: Formulation, Optimization and In Vivo Characterization**

**Amarjitsing Rajput 1,2,\* , Riyaz Ali M. Osmani 1,3,\* , Achyut Khire <sup>1</sup> , Sanket Jaiswal <sup>1</sup> and Rinti Banerjee <sup>1</sup>**


**Abstract:** Background: The goal of this work was to develop a levonorgestrel liposome-loaded microneedle array patch for contraception. Methods: Levonorgestrel-loaded liposome was formulated by a solvent injection technique, characterized, and studied. Results: The formulated liposomes were characterized for particle size (147 ± 8 nm), polydispersity index (0.207 ± 0.03), zeta potential (−23 ± 4.25 mV), drug loading (18 ± 3.22%) and entrapment efficiency (85 ± 4.34%). A cryo high-resolution transmission electron microscopy and cryo field emission gun scanning electron microscopy study showed spherical shaped particles with a smooth surface. The in vitro drug release and in vivo pharmacokinetic study showed sustained behaviour of Levonorgestrel for 28 days. Conclusion: The levonorgestrel liposome-loaded microneedle array patch showed better contraception than the drug-loaded microneedle array patch.

**Keywords:** drug delivery; nanotechnology; contraception; levonorgestrel; liposomes; microneedle; sustained-release

#### **1. Introduction**

Despite advances in contraception methods available in the market, 67 million pregnancies were unplanned throughout the world in 2017 [1]. The number of pregnancies accounts for around 56 million abortions every year [2]. Over 40% of births are unwanted worldwide, with approximately 20% of those ending in abortion [3]. This high prevalence of unwanted pregnancy leads to high economic and social burdens for women. The reasons for unintended pregnancy include limited access to contraception, younger populations, poverty, unmarried people, side effects associated with contraceptive methods, cultural or religious conflicts, poor-quality contraceptive agents, preferences of users and providers, and gender-associated barriers. In the world, 62% of married women (age 15 to 49) use family planning procedures and modern methods are used by 56% of the population [4]. Many methods are currently available, such as combined oral contraceptives (COCs) or the pill, progestogen-only pills (POPs), implants, progestogen-only injectable devices, monthly injectable devices or mixed injectable contraceptives (CIC), mixed contraceptive patches, mixed contraceptive vaginal rings (CVRs), intrauterine devices containing copper (IUDs), intrauterine device loaded with levonorgestrel (IUDs), male and female condoms, male sterilization (vasectomy), female sterilization (tubal ligation), lactational amenorrhea method (LAM), contraception pills for emergency use (composition: ulipristal acetate 30 mg or levonorgestrel 1.5 mg), the standard days m (SDM), basal body temperature (BBT) method, two day method, sympto-thermal method, etc. Traditional methods are also used for

**Citation:** Rajput, A.; Osmani, R.A.M.; Khire, A.; Jaiswal, S.; Banerjee, R. Levonorgestrel Microneedle Array Patch for Sustained Release Contraception: Formulation, Optimization and In Vivo Characterization. *Molecules* **2022**, *27*, 2349. https://doi.org/10.3390/ molecules27072349

Academic Editors: Iola F. Duarte Ciceco and Abdelwahab Omri

Received: 17 January 2022 Accepted: 31 March 2022 Published: 6 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

contraception, such as the calendar method or rhythm method and withdrawal (coitus interruptus) [3]. These are effective contraception methods, but some require frequent dosing; as a result, they suffer from poor patient compliance [5]. Modern contraception methods need healthcare professionals for administration, hence are not suitable for people in low-income countries [6]. Thus, it is essential to formulate a safe, cost-effective, longacting, self-administrable and patient-acceptable contraceptive-based system for global use [7–9].

Liposomes are vesicular structures that consist of a lipid bilayer enclosed by an aqueous membrane. Liposomes are widely used as lipid-based nanocarrier systems for both hydrophilic and hydrophobic compounds [10]. They have shown significant enhancement in the delivery and therapeutic potency of different classes of active pharmaceutical ingredients. The effective delivery of liposomes improves various drugs' pharmacodynamic and pharmacokinetic properties [11,12].

Microneedles are micrometer-size structures that penetrate through the skin layer, not to the dermis, to promote the passage of the drug across the skin layers without causing pain [13–16]. Microneedles are an economical and suitable dosage form for selfadministration by the patient. They are made of biocompatible and biodegradable materials and release the drugs slowly. An incredible, extremely unique, effective and minimally invasive development has been achieved in the era of microneedle technology to improve intradermal and transdermal delivery of different medications [17–22]. Microneedles are considered as a viable approach for the immediate and long-acting delivery of drugs across the skin [23]. Therefore, microneedles are drawing more attention from researchers and clinicians [24].

In this study, we developed and characterized a contraceptive hormone (levonorgestrel, LNG) liposome-based microneedle array patch in terms of different parameters such as microscopy, scanning electron microscopy (SEM), in vitro skin piercing, in vitro skin penetration, mechanical strength, etc. Finally, we investigated in vitro drug release and in vivo pharmacokinetics in rats for 28 days. The aims of the study were to (i) reduce the dosing frequency, (ii) avoid side effects associated with frequent administration of other contraceptive agents, and (iii) provide sustained drug release for 28 days.

#### **2. Materials and Methods**

#### *2.1. Materials*

Levonorgestrel was purchased from Cayman Chemicals, Bangalore, India. Soya phosphatidylcholine (SPC) and oleic acid were gifted and purchased from VAV life sciences, Mumbai, India, and Merck India Ltd., Mumbai, India, respectively. Gelatin and polyvinyl alcohol were purchased from MP Biomedical, Mumbai, India. Polydimethylsiloxane (PDMS) and elastomer were purchased from DuPont, Mumbai, India. All chemicals and reagents used for the study were analytical grade. All solvents used for analytical study were HPLC-grade.

#### *2.2. Methods*

### 2.2.1. Preparation of Liposomes

Levonorgestrel liposomes were formulated by the solvent injection method, as reported by Goudo et al. [25]. Weighed quantities of soya phosphatidylcholine and oleic acid were dissolved in ethanol. Simultaneously, a weighed amount of phosphate buffer solution (PBS, pH 7.4) was subjected to stirring on a magnetic stirrer (Remi, MLV, Mumbai, India). After that, the ethanolic lipid solution was injected into the phosphate buffer solution (pH 7.4) using a 20-gauge needle (dimensions: 0.9 mm × 25 mm) dropwise to form multilamellar vesicles and allowed to stir for 1 h at 1000 rpm. The formed multilamellar vesicles (MLVs) were further processed using high-speed homogenization (IKA, T50 Digital Ultra Turrax, Ahmedabad, India) at 10,000 rpm for 2 min. Finally, large unilamellar vesicles (LUVs) were subjected to stirring for 24 h to remove the solvent.

#### 2.2.2. Characterization of Liposomes

Measurement of Particle Diameter and Zeta Potential

The prepared liposomes were characterized for particle size, polydispersity index, and zeta potential using a particle size measurement instrument based on the dynamic light scattering (DLS) principle (Malvern Instruments, Zetasizer Nano ZS Ultra, Malvern, UK). The experiments were conducted in triplicate using 2 mL of the formulation at 25 ± 2 ◦C with a 90◦ scattering angle [26,27].

#### Determination of Drug Loading and Encapsulation Efficiency

Using the centrifugation technique, the formulated liposomes were further characterized for drug loading and encapsulation efficiency. A definite quantity of liposome (2 mL) was centrifuged (Beckman Coulter Inc., Optima XPN-100 Ultracentrifuge, Indianapolis, IN, USA) at 32,000 rpm with average g-force of 152535 (rotor type-swinging bucket SW 32 Ti, tube-polypropylene quick seal with 10 mL capacity) for 6 h at 4 ◦C to separate the free drug from the liposome. Then, the upper layer was separated, filtered through a 0.22 µ filter, and quantified by HPLC (Jasco, 4000, Tokyo, Japan) with a photodiode array (PDA) detector (Jasco, MD 4015, Tokyo, Japan). The drug loading and encapsulation efficiency were calculated using Equations (1) and (2), respectively [28,29].

$$\text{Drug loading} = \frac{\text{Total drug amount} - \text{Free drug amount}}{\text{Total liquid amount}} \times 100\tag{1}$$

$$\text{Entropyment efficiency} = \frac{\text{Total drug amount} - \text{Free drug amount}}{\text{Total drug amount}} \times 100\tag{2}$$

Cryo High-Resolution Transmission Electron Microscopy (Cryo HR TEM)

The shape and surface of the prepared liposomes were determined using a cryo highresolution transmission electron microscope (Cryo HR TEM) (JEOL, JEM 2100, Tokyo, Japan). A single drop of liposome dispersion (10 µL) was kept on a copper grid coated with carbon and stained negatively using a phosphotungstic acid solution (1% (*w*/*v*)) and observed at 100 kV voltage [30].

#### Cryo Field Emission Gun Scanning Electron Microscopy (Cryo FEG SEM)

The surface structure of developed liposomes was determined using cryo field emission gun scanning electron microscopy (cryo FEG SEM) (JEOL, JSM-7600F, Tokyo, Japan). The liposome sample (5–10 µL) was placed on a carbon tape and lyophilized using liquid nitrogen. The images were captured by a cryo FEG SEM microscope, operated at 30 kV voltage with suitable magnification [31].

#### 2.2.3. Fabrication of Microneedle Array Patch

A master mold with metallic microneedles formed as an array was fabricated using the EDM facility at IDEMI, Mumbai. The diameter of the mold was 35 mm, and the needle dimensions were 900 µm height, 300 µm base diameter, and pitch of 1.5 mm, for an 18 × 18 array. To obtain a working mold consisting of needle cavities, polydimethylsiloxane (PDMS) compound (a mixture of elastomer and curing agent 10:1, mixed manually and degassed) was poured on the master mold and heated at 90 ◦C in the oven for 50 min. The cured compound was separated from the master mold and was ready for microneedle array patch fabrication. The PDMS molds were filled with PVA–gelatin mixture and LNG-loaded liposomes, all in defined proportions. After that, the molds were put onto the centrifuge's plate rotor and centrifuged (Thermo Scientific, Fiberlite™ H3-LV Large Volume Swinging Bucker rotor, New York, NY, USA) for 45 min at 3000 rpm at g-force of 1840× *g*. After centrifugation, the mold was kept in the isolation chamber (at room temperature and 45% RH) for 24 h for drying. After drying, the microneedles were separated from the mold and characterized [16,32].

## 2.2.4. Characterization of LNG Liposome-Loaded Microneedle Array Patch Microscopy Study

Microneedle array patch images were captured from different directions and angles using a stereomicroscope (Nikon, SZX2, Tokyo, Japan) to study the needles' shapes, sizes, and arrangements in the array.

#### Field Emission Gun Scanning Electron Microscopy (FEG SEM)

The drug-loaded microneedle array patch was placed on an ultra-microtome, and a section of suitable size was used for the study. The microneedle array patch was placed on two-sided tape with a metal stub and sputter-coated (10 nm thick) with Au/Pd and studied with a field emission gun scanning electron microscope (JEOL, JSM-7600F, Japan). The microneedle was examined for the drug in the microneedle [33].

#### In Vitro Drug Release

A USP type II dissolution apparatus was used to conduct an in vitro release evaluation of a liposome (containing LNG) loaded microneedle array patch. (Electrolab, TDT-08L, Navi Mumbai, India). For the release study, one microneedle array patch was suspended in 75 mL of phosphate buffer pH 7.4 (release medium) containing 0.1% *w*/*w* sodium azide at 37 ◦C at a speed of 50 rpm. Samples (2 mL) were removed and replaced with an equivalent volume of the release medium at various time intervals (1, 2, 4, 6, 8, 12, and every 24 h up to 28 days) to maintain sink condition. The amount of drug release was determined using an HPLC with a PDA detector [34].

#### Mechanical Strength Study

A universal testing machine (UTM) (Tinius Olsen, Inc., H5KS, Buskerud, Norway) was used to determine the mechanical strength of the LNG liposome-loaded microneedle array patch. Then, the patch was placed on a bottom circular plate workbench. Initially, the spacing between the microneedle array patch and the top circular plate was maintained at 3 mm. The upper workbench then slid down to the microneedle array patch at a 1 mm·min−<sup>1</sup> pace. After that, the upper workbench moved down to the microneedle array patch at a speed of 1 mm·min−<sup>1</sup> . The testing machine determined the load and displacement values, and the weight-displacement curve was plotted to calculate the compression strength [33].

#### Skin Piercing Study

The strength desired for a liposome-loaded microneedle array patch for piercing into the skin was studied using properly shaved and excised skin of rat. The skin was subjected to 37 ◦C for 2 h prior to the study. The skin was kept on Styrofoam block, flattened, and maintained with the help of pins. The LNG liposome-loaded microneedle array was attached to a universal testing machine (UTM) probe with two-side tape. The probe was brought down onto the skin, allowed to penetrate the skin up to a certain distance (0.5 mm), and remained in site for 30 s. After that, the skin was treated with rhodamine dye solution for 30 s and imaged under an optical microscope [35,36].

#### Skin Irritancy Study

Before and after microneedle array patch insertion, skin integrity was studied on Wistar female rats weighing 200–250 g. The blank microneedle and microneedle array patch loaded with liposomes was applied to the dorsal part of the rat. The rat skin was observed for any reactions (i.e., redness, inflammation, or swelling) 24 h post application of the microneedle array patch [37,38].

#### 2.2.5. Statistical Analysis

The data were presented as means with standard deviation (n = 6). Microsoft Excel and GraphPad Prism (GraphPad Software Inc. La Jolla, CA, USA, trial version 5.0) were used to perform the statistical calculations [39].

#### **3. Results and Discussion**

#### *3.1. Preparation of LNG Loaded Liposomes*

Various lipids such as hydrogenated soy phosphatidylcholine (HSPC), distearoyl phosphatidylcholine (DSPC), dioleoyl phosphatidylcholine (DOPC), dipalmitoylphosphatidylcholine, egg phosphatidylcholine, and soy phosphatidylcholine were screened for the preparation of LNG loaded liposomes. The liposomes were prepared using the widely used ethanol injection method. The ethanol injection method is very useful for developing liposomes on a large scale. It is also an easy and continuous method [40]. The various parameters, such as syringe gauge, stirring speed, homogenization time, homogenization speed, and solvent quantity, were studied during preliminary trials (data not shown). It was found that 21 G syringe gauge, 1000 rpm stirring speed, 2 min sonication time, 10,000 rpm homogenization speed, and 2.5 mL of solvent were the optimal parameters for the formulation of liposomes with desired particle size and entrapment efficiency. During the preparation of liposomes, we used a perfusion automatic device or pump. The optimization was carried out using different injection speeds. The injection speed of 0.5 mL/min was considered suitable for preparation of uniform-size liposomes.

Finally, the liposomes prepared using soy phosphatidylcholine as a lipid showed the desired characteristics. Along with soy PC, oleic acid was also added to the formulation to promote the permeation of microneedles through the skin. Oleic acid was used as a permeation enhancer, which acts by modulating the extracellular lipids of the stratum corneum layer (the major barrier to skin permeation) of the skin [41]. It also decreases the resistance of the skin to diffusion by networking with the lipid matrix, leading to enhanced fluidity of the lipid [42,43]. The composition of all batches is shown in Table 1.


**Table 1.** Composition of liposome batches.

#### *3.2. Characterization of Liposomes*

3.2.1. Particle Size, Polydispersity Index, and Zeta Potential

The particle size, polydispersity index, and zeta potential were determined by a method based on the zeta sizer's dynamic laser light scattering mechanism. The average particle size and polydispersity index were measured by the equilibration time of 120 s for each sample using triplicate measurement. The particle sizes of batches L1, L2, and L3 increased with soy PC concentration, from 207 ± 4.21 to 245 ± 7.15 nm. The addition of oleic acid into the liposome formulation resulted in flexible liposomes with smaller particle sizes (Batch L4). The particle sizes of the L5 and L6 batches were found to be 169 nm and 157 nm, respectively, due to variation in the ratio of soy PC to oleic acid. The particle graph of optimized batch L6 is shown in Supplementary Materials Figure S1.

The polydispersity index shows how the particles in a formulation are distributed in size. The narrow particle size distribution suggested that the formulation had a homogeneous particle size distribution. The particle size distribution was <0.350 in all batches [44].

The colloidal dispersion stability was determined by measuring the zeta potential of the formulations. When the value of zeta potential increases, the attraction between the particles decreases, resulting in the enhancement of the stability of the formulation [45]. The impact of soy PC on the zeta potential was also studied. It suggested that as the concentration of soy PC increased, the zeta potential value also increased, as observed in batch L6 shown in Table 2. Batch L6 showed the highest zeta potential, of around −19 ± 4 mV.


**Table 2.** Characterization of liposome batches.

The results are expressed as mean ± SD (n = 3).

#### 3.2.2. Encapsulation Efficiency

The encapsulation efficiency results for all batches are shown in Table 2. The batches were prepared using different drug: lipid ratios such as 1:5, 1:7, and 1:10 to determine their impact on entrapment efficiency. It was observed that as the drug: lipid ratio increased from 1:5 to 1:10, the encapsulation efficiency was enhanced, from 65.34 ± 5.44 to 85.24 ± 6.15. The increase in the quantity of lipids provided additional space for accumulating the drug and preventing the loss of the drug into the external phase. This may be due to an excess of lipid forming a layer around the particle. The superior imperfections in the crystal lattice of lipid provide adequate space to encapsulate drug particles [46–48].

The batches were also prepared with different ratios of soya PC: oleic acid, such as 5:5, 6:4, and 7:3. The increase in the quantity of lipids relative to the oleic acid resulted in higher entrapment efficiency, of around 85.24 ± 6.15 (Batch L6).

#### 3.2.3. Cryo High-Resolution Transmission Electron Microscopy (Cryo HR TEM)

The cryo HR TEM image of LNG-loaded liposomes showed a particle size of around 100 nm, as shown in Figure 1. The particle size obtained using dynamic light scattering was found to be 157 nm. This difference in the particle size obtained using both techniques was due to their different analysis mechanisms. The DLS technique determines the particle size by measuring the hydrodynamic diameter of the particle, whereas TEM determines the particle size with the particle mounted on the grid. The Cryo HR TEM image of liposomes showed a monodispersed unilamellar vesicle with a spherical shape [49]. *Molecules* **2022**, *27*, x FOR PEER REVIEW 7 of 13

ing DLS.

**Figure 1.** Cryo HR TEM image of optimized liposome (Batch L6). microneedle. The needle-shaped microneedle had an average length and a base diameter **Figure 1.** Cryo HR TEM image of optimized liposome (Batch L6).

**Figure 2.** Cryo FEG SEM image of optimized liposome (Batch L6).

**4. Fabrication of Microneedle Array Patch** 

was selected for further formulation development.

*5.2. Scanning Electron Microscopy (SEM) Study* 

needle-shaped needles with their arrangement with the base.

*5.1. Microscopy Study* 

**5. Characterization of LNG Liposome-Loaded Microneedle Array Patch** 

3.2.4. Cryo Field Emission Gun Scanning Electron Microscopy (Cryo FEG SEM)

The cryo FEG SEM imaging study found liposomes containing LNG to be around

The microneedle array patch was prepared using a combination of the polymer matrix (PVA: Gelatin). PVA and gelatin are considered safe, biocompatible, and biodegradable materials [50–52]. A combination of PVA: gelatin solution was utilized prepare the microneedles and increase their mechanical strength. The various microneedle array patches were designed using different ratios of PVA: gelation, such as 1:0.25, 1:0.5, and 1:1. The microneedles were characterized for mechanical strength and prepared using a 1:0.5 ratio showed the desired mechanical strength. Hence, a 1:0.5 ratio of PVA: gelatin

Microscopy-derived images of the microneedles (Supplementary Figure S2) showed

Figure 3 shows a scanning electron microscopy image of the LNG liposome-loaded

#### 3.2.4. Cryo Field Emission Gun Scanning Electron Microscopy (Cryo FEG SEM) 3.2.4. Cryo Field Emission Gun Scanning Electron Microscopy (Cryo FEG SEM)

*Molecules* **2022**, *27*, x FOR PEER REVIEW 7 of 13

**Figure 1.** Cryo HR TEM image of optimized liposome (Batch L6).

The cryo FEG SEM imaging study found liposomes containing LNG to be around 100–200 nm in size and spherically shaped with a smooth surface (Figure 2). The results were the same as those obtained by TEM analysis and the particle size measurements using DLS. The cryo FEG SEM imaging study found liposomes containing LNG to be around 100–200 nm in size and spherically shaped with a smooth surface (Figure 2). The results were the same as those obtained by TEM analysis and the particle size measurements using DLS.

**Figure 2.** Cryo FEG SEM image of optimized liposome (Batch L6). **Figure 2.** Cryo FEG SEM image of optimized liposome (Batch L6).

#### **4. Fabrication of Microneedle Array Patch 4. Fabrication of Microneedle Array Patch**

The microneedle array patch was prepared using a combination of the polymer matrix (PVA: Gelatin). PVA and gelatin are considered safe, biocompatible, and biodegradable materials [50–52]. A combination of PVA: gelatin solution was utilized prepare the microneedles and increase their mechanical strength. The various microneedle array patches were designed using different ratios of PVA: gelation, such as 1:0.25, 1:0.5, and 1:1. The microneedles were characterized for mechanical strength and prepared using a 1:0.5 ratio showed the desired mechanical strength. Hence, a 1:0.5 ratio of PVA: gelatin was selected for further formulation development. The microneedle array patch was prepared using a combination of the polymer matrix (PVA: Gelatin). PVA and gelatin are considered safe, biocompatible, and biodegradable materials [50–52]. A combination of PVA: gelatin solution was utilized prepare the microneedles and increase their mechanical strength. The various microneedle array patches were designed using different ratios of PVA: gelation, such as 1:0.25, 1:0.5, and 1:1. The microneedles were characterized for mechanical strength and prepared using a 1:0.5 ratio showed the desired mechanical strength. Hence, a 1:0.5 ratio of PVA: gelatin was selected for further formulation development.

#### **5. Characterization of LNG Liposome-Loaded Microneedle Array Patch**  *5.1. Microscopy Study*  **5. Characterization of LNG Liposome-Loaded Microneedle Array Patch**

#### *5.1. Microscopy Study*

Microscopy-derived images of the microneedles (Supplementary Figure S2) showed needle-shaped needles with their arrangement with the base. Microscopy-derived images of the microneedles (Supplementary Figure S2) showed needle-shaped needles with their arrangement with the base.

#### *5.2. Scanning Electron Microscopy (SEM) Study*  Figure 3 shows a scanning electron microscopy image of the LNG liposome-loaded *5.2. Scanning Electron Microscopy (SEM) Study Molecules* **2022**, *27*, x FOR PEER REVIEW 8 of 13

microneedle. The needle-shaped microneedle had an average length and a base diameter Figure 3 shows a scanning electron microscopy image of the LNG liposome-loaded microneedle. The needle-shaped microneedle had an average length and a base diameter of 900 µm and 300 µm, respectively. The LNG liposome-loaded microneedle had a smooth surface without any cracks or fractures. of 900 µm and 300 µm, respectively. The LNG liposome-loaded microneedle had a smooth surface without any cracks or fractures.

In the case of the dissolution of the microneedle array patch, the needle starts to dis-

Figure 4 shows the release profile of liposomes and LNG in the microneedle (LP). In the liposome-loaded microneedle array patch, no initial burst release was observed within 24 h. This may be because of the slow dissolution of LNG, which is sparingly water-soluble, and the release of the LNG from the lipid bilayer. To be released, the LNG has to come out from the lipid bilayer to the surface of the liposomes, and then it will be released. Once the LNG comes into contact with the dissolution medium, the release of the drug takes place. LNG also showed resistance to the highly water-soluble PVA and gelatin matrix. The liposome-loaded microneedle array patch showed 57% drug release in 4 weeks.

**Figure 4.** In vitro drug release study of optimized (Batch L6) versus levonorgestrel (API) loaded

The pure LNG (API) drug release profile showed an initial burst release within 24 h. That may be due to the amount of drug that directly underwent dissolution in the dissolution medium, as there was no barrier to the release of the drug. Almost 98% of drug release was observed within 4 weeks, which was similar to results reported earlier in the

*5.3. In Vitro Release Study* 

within 35 ± 10 min.

microneedle.

**Figure 3.** SEM image of levonorgestrel liposome-loaded microneedle. literature [53]. **Figure 3.** SEM image of levonorgestrel liposome-loaded microneedle.

#### *5.3. In Vitro Release Study 5.3. In Vitro Release Study*

In the case of the dissolution of the microneedle array patch, the needle starts to dissolve in 15 ± 5 min, and complete dissolution of the microneedle array patch was observed within 35 ± 10 min. In the case of the dissolution of the microneedle array patch, the needle starts to dissolve in 15 ± 5 min, and complete dissolution of the microneedle array patch was observed within 35 ± 10 min.

of 900 µm and 300 µm, respectively. The LNG liposome-loaded microneedle had a smooth

Figure 4 shows the release profile of liposomes and LNG in the microneedle (LP). In the liposome-loaded microneedle array patch, no initial burst release was observed within 24 h. This may be because of the slow dissolution of LNG, which is sparingly water-soluble, and the release of the LNG from the lipid bilayer. To be released, the LNG has to come out from the lipid bilayer to the surface of the liposomes, and then it will be released. Once the LNG comes into contact with the dissolution medium, the release of the drug takes place. LNG also showed resistance to the highly water-soluble PVA and gelatin matrix. The liposome-loaded microneedle array patch showed 57% drug release in 4 weeks. Figure 4 shows the release profile of liposomes and LNG in the microneedle (LP). In the liposome-loaded microneedle array patch, no initial burst release was observed within 24 h. This may be because of the slow dissolution of LNG, which is sparingly water-soluble, and the release of the LNG from the lipid bilayer. To be released, the LNG has to come out from the lipid bilayer to the surface of the liposomes, and then it will be released. Once the LNG comes into contact with the dissolution medium, the release of the drug takes place. LNG also showed resistance to the highly water-soluble PVA and gelatin matrix. The liposome-loaded microneedle array patch showed 57% drug release in 4 weeks.

*Molecules* **2022**, *27*, x FOR PEER REVIEW 8 of 13

**Figure 3.** SEM image of levonorgestrel liposome-loaded microneedle.

surface without any cracks or fractures.

**Figure 4.** In vitro drug release study of optimized (Batch L6) versus levonorgestrel (API) loaded microneedle. **Figure 4.** In vitro drug release study of optimized (Batch L6) versus levonorgestrel (API) loaded microneedle.

The pure LNG (API) drug release profile showed an initial burst release within 24 h. That may be due to the amount of drug that directly underwent dissolution in the dissolution medium, as there was no barrier to the release of the drug. Almost 98% of drug release was observed within 4 weeks, which was similar to results reported earlier in the literature [53]. The pure LNG (API) drug release profile showed an initial burst release within 24 h. That may be due to the amount of drug that directly underwent dissolution in the dissolution medium, as there was no barrier to the release of the drug. Almost 98% of drug release was observed within 4 weeks, which was similar to results reported earlier in the literature [53].

The patch could be designed for weekly, monthly, quarterly, and annual application based on users' needs. For this purpose, the dose of the drug, patch size, and the number of microneedles per patch would need to be modified.

#### *5.4. In Vivo Study in Rats*

For further investigation, the behavior of the liposome-loaded microneedle array patch was studied in vivo in rats. After administering the liposome-loaded microneedle and pure drug via a transdermal route, the in vivo behavior was studied in rats (Figure 5).

The liposome-loaded microneedle array patch achieved a concentration above the human therapeutic level (200 pg/mL) in 8 days [54]. It showed an initial burst release within 2 days after application of the patch to the rats and maintained an LNG concentration above the human therapeutic level for more than 1 month. The average amount of drug release was >30 µg/day.

Conversely, pure LNG required 10 days to achieve the human therapeutic level. Similarly to the liposome-loaded microneedle patch, it showed an initial burst release within 2 days. It released more drug and released it faster than the liposome-loaded microneedle.

of microneedles per patch would need to be modified.

**Figure 5.** In vivo profile of levonorgestrel liposome-loaded microneedle versus levonorgestrel (API) loaded microneedle. **Figure 5.** In vivo profile of levonorgestrel liposome-loaded microneedle versus levonorgestrel (API) loaded microneedle.

The patch could be designed for weekly, monthly, quarterly, and annual application based on users' needs. For this purpose, the dose of the drug, patch size, and the number

For further investigation, the behavior of the liposome-loaded microneedle array patch was studied in vivo in rats. After administering the liposome-loaded microneedle and pure drug via a transdermal route, the in vivo behavior was studied in rats (Figure

#### The liposome-loaded microneedle array patch achieved a concentration above the *5.5. Mechanical Strength Study*

*5.4. In Vivo Study in Rats* 

5).

human therapeutic level (200 pg/mL) in 8 days [54]. It showed an initial burst release within 2 days after application of the patch to the rats and maintained an LNG concentration above the human therapeutic level for more than 1 month. The average amount of drug release was >30 µg/day. Conversely, pure LNG required 10 days to achieve the human therapeutic level. Similarly to the liposome-loaded microneedle patch, it showed an initial burst release within The mechanical strength is an indicator of the skin penetration of the microneedle. Various materials were studied to obtain sufficient mechanical strength. The materials included PEG, PVA, gelatin alone, as well as PVA: PEG (1:1), PEG: gelatin (1:1), and PVA: gelatin (1:1). Finally, PVA: gelatin (1:0.5) resulted in intact needle-shaped microneedles without any bends. It was observed that 1 N strength was sufficient to penetrate through the skin. The graph of mechanical strength is shown in Supplementary Figure S3.

2 days. It released more drug and released it faster than the liposome-loaded microneedle.

#### *5.6. Skin Piercing Study*

*5.5. Mechanical Strength Study*  The mechanical strength is an indicator of the skin penetration of the microneedle. Various materials were studied to obtain sufficient mechanical strength. The materials included PEG, PVA, gelatin alone, as well as PVA: PEG (1:1), PEG: gelatin (1:1), and PVA: gelatin (1:1). Finally, PVA: gelatin (1:0.5) resulted in intact needle-shaped microneedles without any bends. It was observed that 1 N strength was sufficient to penetrate through Piercing microneedles through the skin is essential for the transdermal drug delivery system. To achieve this, the skin-piercing microneedles should have sufficient mechanical strength. The microneedle array patch consisted of an array (18 × 18) consisting of 324 microneedles that were applied onto the rat skin at a speed of 0.1 mm/s. Rat skin is elastic in nature, but as the force increased rapidly, it lost its elasticity. Almost 100% of the microneedles penetrated the skin, as shown in Supplementary Figure S4.

#### the skin. The graph of mechanical strength is shown in Supplementary Figure S3. *5.7. Skin Irritation Study*

*5.6. Skin Piercing Study*  Piercing microneedles through the skin is essential for the transdermal drug delivery system. To achieve this, the skin-piercing microneedles should have sufficient mechanical strength. The microneedle array patch consisted of an array (18 × 18) consisting of 324 microneedles that were applied onto the rat skin at a speed of 0.1 mm/s. Rat skin is elastic in nature, but as the force increased rapidly, it lost its elasticity. Almost 100% of the microneedles penetrated the skin, as shown in Supplementary Figure S4. The microneedle patches were pressed into the dorsal part of the skin for 10 s and then left for 60 s. Then, images of the skin were captured using a camera (Nikon D3500, Japan) using the same conditions (e.g., for light, exposure time, zoom) at 0 and 24 h after application of the patch (Figure S5). The rats tolerated the LNG-loaded microneedle array patch well, with no evidence of redness, inflammation, or swelling to the skin 24 h post patch application. The pain score was calculated using a scale ranging from 0 (no pain) to 10 (hypodermic pain). The pain score was found to be 0, i.e., no pain.

#### **6. Conclusions**

In the present investigation, liposomes were prepared with a particle size of 157 ± 4.54 nm, polydispersity index of 0.231 ± 0.007, zeta potential of −19 ± 4, and entrapment efficiency of 85%. The cryo HR TEM and cryo FEG SEM analyses showed smooth, spherically shaped particles that correlated with DLS data. The microneedle array patch was developed with a backing membrane, and the microneedles remained implanted under the skin surface after removing the patch. The microneedles were fabricated using a biodegradable polymer that released LNG for >30 days, as shown by in vitro and in vivo study in rats, indicating the promising ability to serve in long-acting contraceptive application. Thus, we concluded that the microneedle array patch would be easy to administer, non-invasive, biodegradable, and suitable for more than 1 month of delivery of the contraceptive agent (LNG).

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/molecules27072349/s1, Figure S1: Particle size graph of optimized (Batch L6); Figure S2: Microscopy image of microneedle array patch loaded with rhodamine dye; Figure S3: Mechanical behavior of microneedle array under a compression force; Figure S4: Skin penetration study of microneedle array patch after insertion into the rat skin; Figure S5: Skin irritancy study after administration of microneedle.

**Author Contributions:** A.R.: Performing experiments, writing and editing, supervision; R.A.M.O.: Performing experiments, conceptualization, investigation, writing—original draft; A.K.: Data analysis, investigation on animal study, review and discussion; S.J.: Performing experiments; R.B.: Conceptualisation, study design, supervision, funding acquisition, data analysis, writing and draft review. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors are thankful to the Bill & Melinda Gates Foundation, USA, for providing financial assistance for this research work through research grant OPP1184017 IITB-Self-administered microneedle array patch for contraceptive delivery [IITB Project Code: RD/0118-GATES00-001].

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors are thankful to the IRCC and SAIF Department of Indian Institute of Technology Bombay, Mumbai, Maharashtra, for providing needed help and instrumentation facilities, and are also grateful to VAV Lipids, Mumbai, for providing a gift sample of lipid for research work.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


## *Article* **Self-Nanoemulsifying Drug Delivery System (SNEDDS) of Apremilast: In Vitro Evaluation and Pharmacokinetics Studies**

**Ahad S. Abushal <sup>1</sup> , Fadilah S. Aleanizy <sup>1</sup> , Fulwah Y. Alqahtani <sup>1</sup> , Faiyaz Shakeel <sup>1</sup> , Muzaffar Iqbal 2,3 , Nazrul Haq <sup>1</sup> and Ibrahim A. Alsarra 1,\***


**Abstract:** Psoriatic arthritis is an autoimmune disease of the joints that can lead to persistent inflammation, irreversible joint damage and disability. The current treatments are of limited efficacy and inconvenient. Apremilast (APR) immediate release tablets Otezla® have 20–33% bioavailability compared to the APR absolute bioavailability of 73%. As a result, self-nanoemulsifying drug delivery systems (SNEDDS) of APR were formulated to enhance APR's solubility, dissolution, and oral bioavailability. The drug assay was carried out using a developed and validated HPLC method. Various thermodynamic tests were carried out on APR-SNEDDS. Stable SNEDDS were characterized then subjected to in vitro drug release studies via dialysis membrane. The optimum formulation was F9, which showed the maximum in vitro drug release (94.9%) over 24 h, and this was further investigated in in vivo studies. F9 was composed of 15% oil, 60% Smix, and 25% water and had the lowest droplet size (17.505 ± 0.247 nm), low PDI (0.147 ± 0.014), low ZP (−13.35 mV), highest %T (99.15 ± 0.131) and optimum increases in the relative bioavailability (703.66%) compared to APR suspension (100%) over 24 h. These findings showed that APR-SNEDDS is a possible alternative delivery system for APR. Further studies are warranted to evaluate the major factors that influence the encapsulation efficiency and stability of APR-containing SNEDDS.

**Keywords:** apremilast; psoriatic arthritis; pharmacokinetics studies; SNEDDS; solubility; dissolution; oral bioavailability

## **1. Introduction**

Psoriasis is a well-known chronic inflammatory autoimmune disease of the skin that occurs in 2–4% of the world's population. Both psoriasis and psoriatic arthritis are remitting and relapsing diseases [1,2]. The co-existence of environmental factors or stress factors can trigger the onset of psoriatic arthritis. The treatment of psoriatic arthritis is based on the initial assessment of the disease severity, which is determined by the degree of inflammation, pain, the number of joints involved, and the degree of disability. The treatment can be achieved using single or multiple drug therapies depending on the disease stage and considering the patients' preference (route of administration, frequency and side effects tolerability) [3]. The currently used treatment regimens involve non-steroidal antiinflammatory drugs (NSAIDs), intra-articular corticosteroid injections, disease modifying antirheumatic drugs (DMARDs), and biologics [4,5].

NSAIDs are mostly used for mild psoriatic arthritis alone or combined with other agents like intra-articular corticosteroid injections. This combination is used for symptomatic relief only; it creates synergetic anti-inflammatory action since they act on different inflammatory pathways [6,7]. They do not alter the disease progression course and

**Citation:** Abushal, A.S.; Aleanizy, F.S.; Alqahtani, F.Y.; Shakeel, F.; Iqbal, M.; Haq, N.; Alsarra, I.A. Self-Nanoemulsifying Drug Delivery System (SNEDDS) of Apremilast: In Vitro Evaluation and Pharmacokinetics Studies. *Molecules* **2022**, *27*, 3085. https://doi.org/ 10.3390/molecules27103085

Academic Editor: Jessica Rosenholm

Received: 27 April 2022 Accepted: 11 May 2022 Published: 11 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

their side effects are not well tolerated by most patients, especially their gastric adverse events [5,6]. DMARDs are more effective in treating psoriatic arthritis than NSAIDs or intra-articular corticosteroid injections since they can not only improve the inflammatory symptoms, but also reduce the progressiveness of the disease, which improves patient's health-related quality of life [4]. Biologics are the most expensive treatment option; despite that, they are still used because many studies proved that they are the most effective in treating psoriatic arthritis and are superior in overcoming the inflammation symptoms and pain, minimizing the progression of the diseased joints, and enhancing the quality of life of patients when compared to the other remedies. Unfortunately, they have a major drawback as they lose their efficacy during the treatment course as the body produces antibodies against them [5].

Apremilast (APR) is the first orally administered drug approved for the treatment of active psoriatic arthritis in adults. It was also assigned for the treatment of dermatologic psoriasis and many other diseases such as rheumatoid arthritis, atopic dermatitis and Beçhet's syndrome [5]. It belongs to a group of drugs known as cyclic nucleotide phosphodiesterase type-4 (PDE-4) inhibitors. In comparison to the common treatment regimens, treatment with APR was tolerated by patients and associated with better overall disease-related improvements. APR is a class IV drug, which means it has poor water solubility and permeability, hence it has poor rate of dissolution and consequently poor oral bioavailability [8,9]. Very limited formulation approaches were found in the literature to enhance APR solubility and oral bioavailability [10,11]. The marketed film coated immediate release tablets of APR Otezla® have 20–33% bioavailability compared to APR absolute bioavailability of 73% [5]. On the other hand, many studies represented SNEDDS as promising delivery systems for pharmaceutical drugs due to their tremendous advantages in enhancing solubility, spontaneously occurring emulsification, and thermodynamic and kinetic stabilities [12–18]. The aim of the present study was to create and optimize SNEDDS of APR to increase its solubility and dissolution rate, which sequentially will upgrade the extent of the oral bioavailability and therapeutic efficacy of the drug.

#### **2. Results**

#### *2.1. Solubility Studies*

The results of APR equilibrium solubility are illustrated in the table below (Table 1). The results of APR equilibrium solubility in water, oils, surfactants, and cosurfactants were greatly variable. The maximum equilibrium solubility of APR was observed in Transcutol-HP with a value of 55.01 ± 3.19 mg/mL followed by Tween-80 with a value of 48.54 ± 3.76 mg/mL; thus, these two components were further used as a surfactant and cosurfactant in APR-SNEDDS. The maximum oil solubility of APR was observed in Lauraglycol-FCC (36.54 ± 2.78 mg/mL) compared to the other oils; Lauroglycol-90 (28.21 ± 1.45 mg/mL), Capryol-PGMC (21.41 ± 1.32 mg/mL), Capryol-90 (18.15 ± 1.03 mg/mL), and Triacetin (11.42 ± 0.95 mg/mL); thus, Lauraglycol-FCC was selected as the oil phase in APR-SNEDDS. The least equilibrium solubilites were observed with water (0.01 ± 0.00 mg/mL), ethanol (0.66 ± 0.01 mg/mL) and IPA (2.07 ± 0.10 mg/mL) apparently due to poor APR hydrophilicity. However, water was preferred to be used as the aqueous phase in APR-SNEDDS due to its inert nature, high miscibility with the formulation component, high formulation compatibility and its frequent use in the literature [19,20].

**Table 1.** Equilibrium solubility values of apremilast (APR) in different oils, surfactants, cosurfactants, and water at 25 ◦C (mean ± SD, *n* = 3).




#### *2.2. Pseudo-Ternary Phase Diagrams for APR SNEDDS*

A total of six phase diagrams were developed (Figure 1A–F); each contained different ratios of aqueous phase, oil phase and Smix. Depending on the Smix ratios mainly, the first phase diagram (A) with (1:0) Smix ratio showed the least emulsification areas. Next to it, was phase diagram (B) with (1:2) Smix ratio, which showed very small emulsifications areas too. For phase diagrams (C) and (D) with (1:1) and (2:1) Smix ratios respectively, the maximum emulsification areas were observed, but phase diagram (C) was superior to phase diagram (D) with slightly bigger emulsification areas. The last two phase diagrams (E) with (3:1) Smix ratio and (F) with (4:1) Smix ratio, showed moderate emulsification areas when compared to the least emulsification areas (A,B) and maximum emulsification areas (C,D) observed. From the above findings, phase diagram (C) with (1:1) Smix ratio and the largest emulsification areas was chosen for APR-SNEDDS formulation development. *Molecules* **2022**, *27*, x FOR PEER REVIEW 4 of 19

**Figure 1.** Pseudo-ternary phase diagrams showing SNEDDS zones for oil phase (Lauroglycol-90)), aqueous phase (water), surfactant (Tween-80) and cosurfactant (Transcutol-HP) at Smix ratios of (**A**) 1:0, (**B**) 1:2, (**C**) 1:1, (**D**) 2:1, (**E**) 3:1, and (**F**) 4:1. **Figure 1.** Pseudo-ternary phase diagrams showing SNEDDS zones for oil phase (Lauroglycol-90), aqueous phase (water), surfactant (Tween-80) and cosurfactant (Transcutol-HP) at Smix ratios of (**A**) 1:0, (**B**) 1:2, (**C**) 1:1, (**D**) 2:1, (**E**) 3:1, and (**F**) 4:1.

were developed. Each SNEDDS contained 5 mg of the drug in a total of 1 mL formulation. The formulations were prepared considering almost an entire range of SNEDDS zones in phase diagram with various Lauraglycol-FCC (oil phase) concentrations (10, 15, 20, 25% *v*/*v*), Smix concentrations (30, 40, 45, 50, 55, 60% *v*/*v*) and de-ionized water (aqueous phase) concentrations (25, 30, 35, 40, 50, 55% *v*/*v*). The drug was dissolved completely in Lauraglycol-FCC before the addition of Tween 80: Transcutol-HP followed by titration

with de-ionized water. The composition of APR SNEDDS is included in Table 2.

**Table 2.** Composition of 1 mL APR-SNEDDS each containing 5 mg of the drug.

After choosing 1:1 Smix ratio of Tween 80 and Transcutol-HP (Phase diagram 1C),

**Code Oil (%) Smix (%) Water (%) Total (mL)** F1 10 40 50 1 mL F2 15 40 50 1 mL F3 20 40 40 1 mL F4 25 40 35 1 mL F5 15 30 55 1 mL F6 15 45 40 1 mL F7 15 50 35 1 mL F8 15 55 30 1 mL F9 15 60 250 1 mL

The formulations F3, F4, F7, F8 and F9 withstood the testing and showed no lack or loss of stability in terms of phase separation (flocculation, coalescence, phase inversion)

*2.3. Formulation Development*

*2.4. Thermodynamic Stability Testing*

#### *2.3. Formulation Development*

After choosing 1:1 Smix ratio of Tween 80 and Transcutol-HP (Phase diagram 1C), which gave the maximum nano-emulsification areas, nine APR-SNEDDS, namely (F1–F9), were developed. Each SNEDDS contained 5 mg of the drug in a total of 1 mL formulation. The formulations were prepared considering almost an entire range of SNEDDS zones in phase diagram with various Lauraglycol-FCC (oil phase) concentrations (10, 15, 20, 25% *v*/*v*), Smix concentrations (30, 40, 45, 50, 55, 60% *v*/*v*) and de-ionized water (aqueous phase) concentrations (25, 30, 35, 40, 50, 55% *v*/*v*). The drug was dissolved completely in Lauraglycol-FCC before the addition of Tween 80: Transcutol-HP followed by titration with de-ionized water. The composition of APR SNEDDS is included in Table 2.


**Table 2.** Composition of 1 mL APR-SNEDDS each containing 5 mg of the drug.

#### *2.4. Thermodynamic Stability Testing*

The formulations F3, F4, F7, F8 and F9 withstood the testing and showed no lack or loss of stability in terms of phase separation (flocculation, coalescence, phase inversion) or drug precipitation. The rest of the formulations F1, F2, and F6 were metastable and F5 was unstable (Table 3).

**Table 3.** Results for self-nanoemulsication and thermodynamic tests.


CENT.: centrifugation, H&C: heating-cooling cycle, FPT: freeze-pump thaw cycle, M: metastable, S: stable, Un.: unstable, √ : passed the test.

#### *2.5. Self-Nanoemulsification Efficiency Test*

The results for the self-nanoemulsification test are shown in Table 3. All APR-SNEDDS (F1–F9) were subjected to self-nanoemulsification efficiency test to assess their ability to maintain their stability upon dilution with aqueous phase at different pH values (neutral, acidic and basic). The efficiency of each formulation was evaluated and graded via visual inspection and the use of a system for grading. All the formulations (F1–F9) were graded as grade (A) as they rapidly formed clear NEs within 1 min and maintained their physical and thermodynamic stabilities.

#### *2.6. Physicochemical Characterization*

The physicochemical characterization of APR-SNEDDS was carried out on the most stable formulations (F3, F4, F7, F8 and F9) by testing the following parameters: droplet size, polydispersity index (PDI), zeta potential (ZP), refractive index (RI), percentage of transmittance (%T) and surface morphology by transmission electron microscopy (TEM); but the later one was for the optimized APR-SNEDDS only. The results for the physicochemical characterization of APR-SNEDDS are included in Table 4. In terms of droplet size, all the results recorded were below 25 nm, which indicated high formulations uniformity and stability. In general, the droplet size was found to be reduced as the percentage of oil phase decreased in the formulation (15% oil phase in formulation F7, F8 and F9 that had the lowest droplet size). The mean droplet size was lowest in formulation F9 (17.505 ± 0.247 nm), which might be due to the presence of the highest percentage of Smix ratio in the formulation (60%) that provided relatively high solubilizing capacity. The lowest PDI was 0.109 ± 0.019 in formulation F3 and the highest PDI was 0.278 ± 0.014 in formulation F8. The ZP values were negative for all formulations, F3 = −11.2, F4 = −17.4, F7 = −20.55, F8 = −17.65 and F9 = −13.35 mV, which was due to the composition of (o/w) APR-SNEEDS that presented the negatively charged molecules at the surface, due to the presence of the fatty acid esters in Lauraglycol-FCC (oil phase). The negative charges created repulsive forces between the nanoemulsion droplets, which reflected on the physical stability of the formulations in terms of the absence of droplets combination or phase separation that resulted in the clear and transparent appearances of the formulations [18]. The mean RI of the formulations was 1.340, which in the case of SNEDDS formulation meant that the formulation is of isotropic nature. The %T of the formulations was measured to determine their clarity/transparency translated into their ability to transmit the light rather than absorbing or blocking it. All the results recorded were ≥ 95%; formulation F3 had the lowest %T = 95.94% and formulation F9 had the highest %T = 99.15%. From the above findings, formulation F9 was selected as the optimum APR SNEDDS, based on its lowest droplet size (17.505 ± 0.247), relatively low value of PDI (0.147 ± 0.014) and ZP (−13.35 mV), average RI (1.337) and highest %T (99.15 ± 0.131). Therefore, the TEM analysis of its surface morphology was carried out and the results are presented below in Figure 2. The shape of the droplets was spherical and their size was ≤50 nm. *Molecules* **2022**, *27*, x FOR PEER REVIEW 6 of 19 had the lowest %T = 95.94% and formulation F9 had the highest %T = 99.15%. From the above findings, formulation F9 was selected as the optimum APR SNEDDS, based on its lowest droplet size (17.505 ± 0.247), relatively low value of PDI (0.147 ± 0.014) and ZP (−13.35 mV), average RI (1.337) and highest %T (99.15 ± 0.131). Therefore, the TEM analysis of its surface morphology was carried out and the results are presented below in Figure 2. The shape of the droplets was spherical and their size was ≤50 nm. **Table 4.** Physicochemical characterization of APR-SNEDDS. **SNEDDS Characterization Parameter ± SD Droplet Size (nm) PDI ZP (mV) RI %T** F3 24.95 ± 0.169 0.109 ± 0.019 −11.2 1.343 ± 0.001 95.94 ± 0.221 F4 37.07 ± 2.234 0.237 ± 0.070 −17.4 1.341 ± 0.000 96.6 ± 0.222 F7 18.725 ± 0.275 0.139 ± 0.022 −20.55 1.341 ± 0.001 96.67 ± 0.128 F8 19.335 ± 0.021 0.278 ± 0.014 −17.65 1.339 ± 0.001 97.25 ± 0.022 F9 17.505 ± 0.247 0.147 ± 0.014 −13.35 1.337 ± 0.001 99.15 ± 0.131 SD: standard deviation, PDI: polydispersity index, ZP: zeta potential, mV: millivolts RI: refractive index; %T: percentage of transmittance.

*2.7. In Vitro Drug Release Studies*

**Figure 2.** Transmission electron microscopy (TEM) for optimized SNEDDS F9 showing the droplets shape and size in submicron range. **Figure 2.** Transmission electron microscopy (TEM) for optimized SNEDDS F9 showing the droplets shape and size in submicron range.

F9 was >80% but for formulation F4 it was 76.69% and for APR suspension, it was 31% only. As the study continued, the cumulative drug release from APR-SNEDDS and APR suspension continued to increase steadily until the end of the study (24 h). At 24 h, the cumulative drug release from F7, F8 and F9 exceeded 92% with F9 having the highest percentage of cumulative drug release (94.919% ≈ 95%). While, F4 had the least drug release compared to the other SNEDDS, reaching 81.36% cumulative drug release, followed by the APR suspension that had its maximum observed cumulative drug release throughout the entire study with a value of 40%. The ascending order for the cumulative drug release for the formulations at the 24 h time point was as follows: APR suspension = 40.3%,

The results of in vitro drug release studies are presented in Figure 3. The drug release pattern from APR-SNEDDS and APR-suspension was immediate and rapid but with APR-SNEDDS having a greater percentage of drug release during the first hours of the study compared to APR suspension. During the first 3 h of the study, APR-SNEDDS released more than 30% drug compared to the APR-suspension that released only 19.49%.


**Table 4.** Physicochemical characterization of APR-SNEDDS.

SD: standard deviation, PDI: polydispersity index, ZP: zeta potential, mV: millivolts, RI: refractive index; %T: percentage of transmittance.

#### *2.7. In Vitro Drug Release Studies Molecules* **2022**, *27*, x FOR PEER REVIEW 7 of 19

The results of in vitro drug release studies are presented in Figure 3. The drug release pattern from APR-SNEDDS and APR-suspension was immediate and rapid but with APR-SNEDDS having a greater percentage of drug release during the first hours of the study compared to APR suspension. During the first 3 h of the study, APR-SNEDDS released more than 30% drug compared to the APR-suspension that released only 19.49%. Both formulations continued to release the drug gradually until the steady state was reached at 8 h. By that time, the cumulative drug release for formulations F3, F7, F8 and F9 was >80% but for formulation F4 it was 76.69% and for APR suspension, it was 31% only. As the study continued, the cumulative drug release from APR-SNEDDS and APR suspension continued to increase steadily until the end of the study (24 h). At 24 h, the cumulative drug release from F7, F8 and F9 exceeded 92% with F9 having the highest percentage of cumulative drug release (94.919% ≈ 95%). While, F4 had the least drug release compared to the other SNEDDS, reaching 81.36% cumulative drug release, followed by the APR suspension that had its maximum observed cumulative drug release throughout the entire study with a value of 40%. The ascending order for the cumulative drug release for the formulations at the 24 h time point was as follows: APR suspension = 40.3%, F4 = 81.36%, F3 = 88.11%, F7 = 91.80% and F8 = 93.20% and F9 = 95%. From these results, APR-SNEDDS F9 was selected for optimization and further investigation. F4 = 81.36%, F3 = 88.11%, F7 = 91.80% and F8 = 93.20% and F9 = 95%. From these results, APR-SNEDDS F9 was selected for optimization and further investigation. *2.8. Kinetic Analysis of Drug Release Data* Different kinetic models regarding the mechanism of drug release were studied, including zero-order model, first-order model, Higuchi model, the Hixon–Crowell model and the Korsemeyer–Peppas model [21–23]. The correlation coefficients (R<sup>2</sup> ) and kinetic of drug release from APR-SNEDDS (F3, F4, F7, F8 and F9) and APR suspension are shown in Table 5. The different values of R<sup>2</sup> indicated the best model fit the drug release pattern from each formulation. The F3 formulation had R<sup>2</sup> = 0.999 following zero order release kinetics and Hixon–Crowell kinetics too. The best model fit for F4 was zero order kinetics with R<sup>2</sup> = 0.999. For formulations F7, F8 and the optimum formulation F9, the highest R<sup>2</sup> values were 0.997, 0.996 and 0.9995, respectively, fitting Hixon–Crowell drug release kinetics. For APR-suspension, the best model fitting its drug release kinetic was the Higuchi model with R<sup>2</sup> = 0.996.

**Figure 3.** Cumulative in vitro release of APR from prepared APR-SNEDDS and APR suspension via dialysis membrane over 24 h. **Figure 3.** Cumulative in vitro release of APR from prepared APR-SNEDDS and APR suspension via dialysis membrane over 24 h.

The analysis of APR in male rat plasma samples was performed using a UHPLC-MS*/*MS method as reported in the literature [24]. The concentration of APR in rat plasma

**Crowell**

**Korsemeyer-Peppas**

**Table 5.** The correlation coefficients and kinetics of APR release from SNEDDS and suspension.

F4 0.127 0.999 9.936 0.978 0.983 0.994 0.987 1.547 F7 0.113 0.992 10.353 0.951 0.995 0.997 0.995 1.671 F8 0.112 0.994 10.783 0.954 0.995 0.998 0.996 1.737 F9 0.112 0.993 11.217 0.955 0.995 0.997 0.995 1.803 Suspension 0.262 0.975 7.747 0.897 0.996 0.987 0.993 1.324

*2.9. Bioavailability (In Vivo) Study and Pharmacokinetic Evaluation* 

**Formulation Zero Order First Order Higuchi Hixon-**

#### *2.8. Kinetic Analysis of Drug Release Data*

Different kinetic models regarding the mechanism of drug release were studied, including zero-order model, first-order model, Higuchi model, the Hixon–Crowell model and the Korsemeyer–Peppas model [21–23]. The correlation coefficients (R<sup>2</sup> ) and kinetic of drug release from APR-SNEDDS (F3, F4, F7, F8 and F9) and APR suspension are shown in Table 5. The different values of R<sup>2</sup> indicated the best model fit the drug release pattern from each formulation. The F3 formulation had R<sup>2</sup> = 0.999 following zero order release kinetics and Hixon–Crowell kinetics too. The best model fit for F4 was zero order kinetics with R <sup>2</sup> = 0.999. For formulations F7, F8 and the optimum formulation F9, the highest R<sup>2</sup> values were 0.997, 0.996 and 0.9995, respectively, fitting Hixon–Crowell drug release kinetics. For APR-suspension, the best model fitting its drug release kinetic was the Higuchi model with R <sup>2</sup> = 0.996.

**Table 5.** The correlation coefficients and kinetics of APR release from SNEDDS and suspension.


#### *2.9. Bioavailability (In Vivo) Study and Pharmacokinetic Evaluation*

The analysis of APR in male rat plasma samples was performed using a UHPLC-MS/MS method as reported in the literature [24]. The concentration of APR in rat plasma samples was obtained using a calibration curve plotted between the concentration of APR and area ratio of APR to an internal standard (IR). The calibration curve of APR was found to be linear in the concentration range of 1.47–350 ng/mL. Both formulations showed immediate and rapid drug release during the first two hours of the study with the suspension reaching its maximum concentration of 20 ng/mL, but the F9 formulation continued to release drug sharply with a concentration of 103 ng/mL by that time. After two hours, the drug release from the suspension decreased gradually, reaching 0 ng/mL concentration at the end of the study (24 h). APR-optimized SNEDDS continued to increase readily after two hours until the maximum concentration of 119 ng/mL was reached at 5 h. After that, it decreased steeply, reaching 90 ng/mL concentration at 6 h. The next hours showed a gradual decrease in the drug plasma concentration until the end time point was reached (24 h) with a concentration of 22 ng/mL. Overall, the release profile of APR from optimized SNEDDS F9 was significant compared to the drug suspension (*p* < 0.05). The comparative in vivo APR release after oral administration of optimized SNEDDS and APR suspension are shown in Figure 4.

The results of each pharmacokinetic parameter (mean ± SD) of APR after an oral administration of optimized formulation (F9) and APR-suspension (3 mg/kg) are given in Table 6.

samples was obtained using a calibration curve plotted between the concentration of APR and area ratio of APR to an internal standard (IR). The calibration curve of APR was found to be linear in the concentration range of 1.47–350 ng*/*mL. Both formulations showed immediate and rapid drug release during the first two hours of the study with the suspension reaching its maximum concentration of 20 ng*/*mL, but the F9 formulation continued to release drug sharply with a concentration of 103 ng*/*mL by that time. After two hours, the drug release from the suspension decreased gradually, reaching 0 ng*/*mL concentration at the end of the study (24 h). APR-optimized SNEDDS continued to increase readily after two hours until the maximum concentration of 119 ng*/*mL was reached at 5 h. After that, it decreased steeply, reaching 90 ng*/*mL concentration at 6 h. The next hours showed a gradual decrease in the drug plasma concentration until the end time point was reached (24 h) with a concentration of 22 ng*/*mL. Overall, the release profile of APR from optimized SNEDDS F9 was significant compared to the drug suspension (*p* < 0.05). The comparative in vivo APR release after oral administration of optimized SNEDDS and APR suspension

**Figure 4.** Plasma concentration–time plots of APR after oral administration of optimized SNEDDS and APR suspension in rats (Mean ± SD; *n* = 6; dose 3 mg*/*kg). **Figure 4.** Plasma concentration–time plots of APR after oral administration of optimized SNEDDS and APR suspension in rats (Mean ± SD; *n* = 6; dose 3 mg/kg).

The results of each pharmacokinetic parameter (mean ± SD) of APR after an oral ad-**Table 6.** Pharmacokinetic parameters of APR after an oral administration of optimized SNEDDS and APR suspension (3 mg/kg) in rats.


\* *p* < 0.05 significant compared to APR suspension.

#### AUC0-t (ng.h*/*mL) 462.83 ± 52.25 3256.76 ± 212.50 \* **3. Discussion**

are shown in Figure 4.

Psoriatic arthritis is a progressive inflammatory disease that can lead to persistent inflammation, irreversible joint damage and disability. Current treatment options for psoriatic arthritis are limited because they lack optimal efficacy, and are mostly inconvenient for patients as they are quite expensive, involve injections, and are associated with serious adverse events. APR is the most recently approved oral anti-psoriatic arthritis drug and has been found superior to conventional treatment choices in adult patients with active disease. APR is classified as class IV according to the biopharmaceutics classification system (BCS). Class IV drugs are characterized by their low solubility and poor permeability, which affects their dissolution and absorption [8,9]. The lipid formulation approach appears as a promising approach that can be utilized for improving the solubility, dissolution properties and oral bioavailability of poorly soluble drugs [25–27]. SNEDDS were reported in many studies as the best formulation to solve the problems associated with class II, III and IV

drugs, which have poor solubility or/and poor permeability, which affects their overall oral bioavailability [12–18]. This anhydrous formulation can rapidly form fine oil-in-water nanoemulsions upon dispersion in the gastrointestinal fluids under mild agitation imparted by the gastric motility [18]. Formation of submicron droplets upon dilution produce a large interfacial surface area for transfer of the drug, which may result in increased rate and extent of absorption and hence, improved bioavailability [12]. These formulations maintain the drug in a dissolved state throughout the GI tract and therefore, may enhance the bioavailability of poorly soluble drugs, for which absorption is dissolution rate limited [17,18]. To our knowledge, very limited approaches have been reported in the literature to enhance APR solubility, dissolution, permeability and oral bioavailability [10,11]. Therefore, the aim of this study was to enhance APR's drawbacks using a less complicated and more reproducible method, which was perfectly achieved through the application of SNEDDS as a potential drug delivery system for APR. The current research was meant to enhance APR's therapeutic efficacy by improving its in vitro rate of dissolution, solubility, and bioavailability. APR SNEEDDS were developed using a spontaneous emulsification method via the construction of pseudo-ternary phase diagrams, while different thermodynamic tests were carried out on the developed SNEDDS based on centrifugation, heating and cooling cycles, and freeze-pump thaw cycles. Subsequently, the thermodynamically stable SNEDDS were characterized by self-nanoemulsification efficiency, droplet size, PDI, ZP, RI, %T and surface morphology. The optimized SNEDDS of APR were then used for in vivo evaluation followed by statistical analysis.

The equilibrium solubility data of solutes in different components at room temperature or physiological body temperature was the technique that was applied for selecting the components to develop the suitable SNEDDS [18]. Screening of components by carrying out equilibrium solubility studies using the shake flask method [28] was the very first and the most important step in APR-SNEDDS fabrication, and determined the most suitable components for SNEDDS formulation. Their selection was also made upon their safety; they fall under GRAS category and their acceptability for oral pharmaceutical use. APR is a class IV drug; it is poorly soluble in water and its water solubility as a mole fraction at room temperature and atmospheric pressure is 2.74 <sup>×</sup> <sup>10</sup>−<sup>7</sup> as per EMA and USFDA [8,9]. The equilibrium solubility of APR in different SNEDDS components was found to vary significantly with the maximum equilibrium solubility being observed in Transcutol-HP, followed by Tween-80. These two components were used as surfactant and cosurfactant in APR-SNEDDS. The combination of a surfactant with a cosurfactant in the formation of o/w nanoemulsions with improved levels of solubilization greatly reduced interfacial tension and decreased interface fluidity [13]. The process of selection of the surfactant and cosurfactant in the further study was governed by the efficiency of emulsification and the solubilization ability of APR. In general terms, the surfactant was selected with emphasis being placed on the continuous phase of the nanoemulsion with the hydrophilic surfactant for nanoemulsion with the aqueous phase as the phase of dispersion and vice-versa [29,30]. The maximum oil solubility of APR was found in Lauraglycol-FCC and hence selected as the oil phase for formulation development. The oil solubility of the drug is crucial for its stability in the formulation and throughout the ingestion process in the GIT. It maintains the drug in the solubilized form, which prevents its precipitation that hinders its desirable solubility and absorption. The solubility studies were meant to identify the most preferred oil phase and surfactant to cosurfactant ratio for the development of the APR SNEDDS formulation. It was also observed in the literature that creating a distinction between the most suitable oil and surfactant to cosurfactant ratio that has the maximal solubilizing potential for the drug under investigation was essential, as it would lead to the improvement of the drug loading [18]. The results for solubility studies also showed that the least equilibrium solubility was observable with water, ethanol and IPA. The variation exhibited was significant and this can be explained based on the poor hydrophilicity of APR. This implies that APR is a molecule whose interactions with water and other polar substances are more favorable thermodynamically as compared to the interactions with

either ethanol or IPA. The rule of thumb here is that the solubility of APR molecules in water is more than 1 mass percentage as long as the condition of having at least one neutral hydrophilic group for each 5 carbons is met or at least a single electrically charged hydrophilic group for each of the 7 carbons is met. As such, APR seems to attract water out of air. However, water was the most preferred solvent in the aqueous phase in APR SNEDDS considering that it is inert in nature, while it has other properties like high miscibility with formulation components and high formulation compatibility [19,20]. The self-emulsion formulations made up of oil, surfactant, cosurfactant and the drug should be clear and monophasic liquid at ambient temperature upon addition to the aqueous phase, while the solution should have good solvent properties [13]. The solubility studies were assessed in further pseudo-ternary phase diagrams interpretation and formulation development. Pseudo-ternary phase diagrams showed that lipophilic drugs like APR are preferably solubilized in the o/w nanoemulsions, while the w/o systems seem to be the better option for the hydrophilic drugs [29,30]. The loading of the drug at each formulation was found to be the most critical design factor in developing the nanoemulsion systems for the drug, considering that it is poorly soluble, which depends on the drug solubility at different formulation components. The formulation volume was minimized to the highest possible values to allow for the delivery of the therapeutic dose of the drug in the encapsulated form. In the case of oral formulation development, the drug solubility in the oil phase is of particular importance. The reasoning here is that the ability of the nanoemulsion to maintain APR in solubilized form was greatly dependent on the solubility of the drug in oil phase (Lauraglycol-FCC). Whenever the surfactant or cosurfactant was found to contribute to the drug solubilization, then it was easy to conclude that there is a high risk of precipitation since the dilution of the NEs in the gastrointestinal tract can contribute to the lowered solvent capacity of either the surfactant or cosurfactant [15]. Another suggestion is that it is essential to have a sound understanding of the factors that influence the capacity of drug loading while ensuring that the capability of the system is maintained to undergo the monophasic dilution with water, while also ensuring that the tendency for drug precipitation and crystallization is minimal [16]. Large amounts of surfactants were found to cause gastrointestinal and skin irritation upon oral and topical administration. This implies that the proper selection of the surfactants was essential, where it was essential to determine the surfactant concentration properly and use the minimum concentration in the formulation. Nonionic surfactants were also found to be less toxic as compared to the ionic counterparts, where they were also observed from the literature to have lower CMCs [31,32]. Furthermore, o/w nanoemulsion dosage forms for oral and parenteral use based on the nonionic surfactants were more likely to offer better in vivo stability [31]. The cosurfactant is a characteristic component in naoemulsions and an essential entity that is meant to maintain nanoemulsion systems at low surfactant concentrations [18]. The cosurfactant was observed to increase the mobility of the hydrocarbon tail, while it allowed greater penetration of the oil into the region. Alcohols (O-H) are reported to increase the miscibility of the aqueous and oil phases, bearing in mind that they tend to partition between the phases. Therefore, since the maximal solubility of APR was observed in Lauraglycol-FCC as compared to other oils, the nanoemulsion area was applied as the criteria for assessment and evaluation of the cosurfactants [18]. The pseudo-ternary phase diagrams were developed for APR-SNEDDS based on the spontaneous emulsification or aqueous phase titration method [31,32]. These phase diagrams were developed in order to optimize the APR SNEDDS. The size of the nanoemulsion regions in the phase diagrams was compared at an interchangeable Smix ratio with the major measures including keeping the surfactant at the same levels while altering the cosurfactant and vice versa. There were six diagrams that were constructed, each consisting of three plots with each plot representing the different phases of the formulation. In the first plot, it is observed as the oil phase comprising of Lauraglycol-FCC, while the other plot was for the aqueous phase comprising of the water. The third phase was for the Smix ratio comprising Tween-80: Transcutol-HP. In the course of formation of the Smix,

Tween-80 and Transcutol-HP were mixed together in several mass ratios from 1:0 to 4:1. On the other hand, the oil phase Lauraglycol-FCC was mixed in ratios of 1:9 to 9:1 with the Smix ratios. After this, there was a drop-wise titration for the oil-Smix ratios done by the water. The aim of creation of phase diagrams was to examine the maximum nano-emulsification. The research was based on the observation that the larger the size of the field of nanoemulsification, the greater the nano-emulsification efficiency of the system. It was also observed that whenever the length of chain was increased, there was an increase in the area of existence of the nanoemulsion. As such, the nanoemulsion formation was a function of the composition of the system, where the existence of the nanoemulsion formation was illustrated with the help of the pseudo-ternary phase diagram. In as much as the order of the mixing of the various components did little in terms of influencing the formation of the nanoemulsion, the system was kept at a thermodynamically stable condition that was pathindependent. The other major observation was that there was no distinct conversion from the w/o to the o/w NEs. The rest of the region of the phase diagram was representative of the turbid and conventional emulsions. There was also careful observation of the formulations to ensure that the metastable systems were not selected, even though the free energy that was consumed in the formation of the NEs was very low, while the formation was thermodynamically spontaneous. For further optimization of the system, the effect of the surfactant and cosurfactant ratio on nanoemulsion formation was determined. A total of six phase diagrams were developed with each containing different ratios of the aqueous phase, oil phase, and the Smix. Based on the Smix ratios, the first phase diagram with the (1:0) Smix ratio portrayed the least nano-emulsification areas, while the phase diagram with the (1:2) Smix ratio portrayed very small nano-emulsification areas too. For the phase diagrams with (1:1) and (2:1) Smix ratios, the maximal nano-emulsification areas were observed with the diagram with the Smix ratio of (1:1) portraying a superior phase diagram. The other phase diagrams with (3:1) and (4:1) Smix ratios showed moderate nano-emulsification areas as compared to the least emulsification areas (1:0) and (1:2). From the results obtained, the phase diagram with the (1:1) Smix ratio and the largest nano-emulsification areas was selected for the APR-SNEDDS formulation development. It was concluded that whenever the cosurfactant is absent or present at lower concentrations, the surfactant cannot have the potential of sufficiently reducing the o/w interfacial tension. An o/w NEs region was found towards the rich apex of the phase diagram. The maximum concentration of oil that could be solubilized as shown in the phase diagram was at 66% of Smix. Whenever the cosurfactant was added to the surfactant within equivalent amounts, a higher nanoemulsion region was exhibited. The increase in the nanoemulsion region relative to the addition of the surfactant is attributed to the reduction in the interfacial tension and increased fluidity of the interface at Smix [18]. The selection of the phase diagram with the (1:1) Smix ratio for the APR-SNEDDS formulation development was based on the observation that the higher the nanoemulsion field is, the greater the nanomulsification efficiency of the system. In the current research, the (1:1) Smix ratio of Tween 80 and Transcutol-HP gave the maximal nano-emulsification area, which was selected for the development of the nine APR-SNEDDS (F1-F9). Each of the mixtures contained 5 mg of APR in a total volume of 1 mL. The preparation of the formulations was based on the entire range of SNEDDS zones in the phase diagram with varied levels of the oil phase concentrations with Lauraglycol-FCC being given a preference and the deionized water or aqueous phase concentrations being applied. APR as the drug under investigation was allowed to dissolve completely in Lauraglycol-FCC before the Smix (Tween 80 and Transcutol-HP) was added to the mixture, followed by titration with deionized water. For the purpose of exclusion of the possibility of metastable formulations, thermodynamic stability tests were carried out. Most of the representative formulations were extracted from the o/w nanoemulsion region of the phase diagram, which was constructed at an Smix ratio of (1:1) as it was observed to show the largest nano-emulsification areas for the APR-SNEDDS formulations development. Thermodynamic stability tests were carried out on the nine formulated APR-SNEDDS for the exclusion of the metastable and unstable formulations by the application

of various external conditions that could impact on the stability. Formulations including F3, F4, F7, F8, and F9 were found to withstand the tests, as they portrayed no lack or loss of stability in terms of phase separation and drug precipitation. Thermodynamic stability was a measure that could aid in conferring the long shelf life to the nanoemulsion as compared to the ordinary emulsion [16]. The formulations were prepared based on the nature of the entire range of the SNEDDS zones in the phase diagram with various Lauraglycol-FCC (oil phase) concentrations of 10, 15, 20, 25% *v*/*v*, the Smix concentrations of 30, 40, 45, 50, 55, 65% *v*/*v*, and aqueous phase concentrations of 25, 30, 35, 40, 50, 55 *v*/*v*. The most stable formulations (F3, F4, F7, F8, and F9) were subjected to the physicochemical characterization with tests aiming at understanding the effect of the droplet size, PDI, ZP, RI, %T and surface morphology. In the first test of the droplet size, all the results were recorded to be below 25 nm, which was an indicator that there was high formulation uniformity and stability. The droplet size was heavily dependent on the oil phase formulation, where the droplet size was reduced related to the decrease in the percentage of the oil phase formulation with 15% oil phase formulation F7, F8, and F9 having the lowest droplet size readings. The mean droplet size was approximately 23.517 nm with the lowest result being recorded in formulation F9 that recorded 17.505 nm. The reduced droplet size can be explained in terms of the presence of the highest percentage of Smix ratio in the formulation (60%), which provided relatively high solubilizing capacity. Studies have shown that the droplet size distribution is one of the most essential characteristics that affect the in vivo fate of NEs as it influences the bioactive release rate and absorption. The production of NEs with smaller droplet sizes is highly recommended as it provides extremely low surface tension for the entire system and the interfacial tension of the o/w droplets [17,18]. The mean PDI for APR-SNEDDS was reported as 0.182, which is an indication of the narrow size distribution and more uniformity of the droplets within the formulations. Smaller particles tend to resist gravity separation, flocculation, coalescence, and creaming. The ZP values were also found to be negative for all formulations F3 (−11.2 mV), F4 (−17.4 mV), F7 (−20.55 mV), F8 (−17.65 mV), while F9 was −13.35 mV. The charge difference among different formulations was possible due to different compositions of different formulations. The significant charge differences between the formulations F3 and F7 could be possible due to the high concentration of Smix in formulation F7 compared to formulation F3. The explanation behind the observation is attributed to the composition of the o/w APR-SNEDDS that presented negatively charged molecules at the surface due to the presence of fatty acid esters in the Lauraglycol-FCC (oil phase). The negative charges were also found to create repulsive forces between the nanoemulsion droplets, which reflected the physical stability of the formulations in terms of the absence of droplet combinations or phase separation that resulted in the clear and transparent appearances of the formulations. The %T of the formulations was also considered as an essential component that would aid in the determination of their clarity/transparency translated into their ability to transmit the light as opposed to absorbing or blocking it. All the results recorded were less than or equal to 95% with formulation F3 having the lowest %T (95.94%) and formulation F9 having the highest %T (99.15%). From the physicochemical characterization of APR-SNEDDS, formulation F9 was selected as the optimum APR SNEDDS, considering that it has the lowest droplet size, and relatively low value of PDI and ZP, while it had the highest %T.

The in vitro drug release studies were carried out to investigate the release profile of APR from the APR SNEDDS that were stable including F3, F4, F7, F8, and F9 and the APR-suspension over a period of 24 h though a dialysis membrane. The results show that within the first three hours of the study, APR-SNEDDS released more than 30% of the drug as compared to the APR-suspension that released only 19.49%. From the cumulative in vitro release of APR from prepared APR-SNEDDS and APR suspension over a period of 24 h, it can be said that there are variations because of the changes in the suspension agent that affected the drug release pattern from the suspension formulation compared to the superior APR-SNEDDS (F9) formulation. The results show that the formulation and process parameters in the preparation of NE containing APR is critical in obtaining the desirable

attitudes for effective drug delivery. Different kinetic models regarding the mechanism of drug release were studied, including the zero-order model, first-order, Higuchi model, Hixon–Crowell model and Korsemeyer–Peppas model [21–23]. The different values of R 2 indicated the best model fit the drug release pattern from each formulation. The F3 formulation had R<sup>2</sup> = 0.999 following zero order release kinetics and Hixon–Crowell kinetics too. The best model fit for F4 was zero order kinetics with R<sup>2</sup> = 0.999. For formulations F7, F8 and the optimum formulation F9, the highest R<sup>2</sup> values were = 0.997, 0.996 and 0.9995, respectively, fitting Hixon–Crowell drug release kinetics. For APR-suspension, the best model fit its drug release kinetic was the Higuchi model with R<sup>2</sup> = 0.996. The different pattern of the drug release model in different formulations could be possible due to the presence of different concentrations of oil phase and Smix. For the in vivo drug release, comparisons were made on rat plasma concentrations of the optimized APR-SNEDDS (F9) compared to that of the APR suspension. In male rats, the plasma concentration of APR after oral administration is too low [11]. The HPLC method is not able to detect the low concentration of APR in plasma. The UPLC-MS/MS method is a very sensitive method, which is able to detect the low concentration of APR in rat plasma. Hence, the UPLC-MS/MS method was used to determine APR in rat plasma [24]. Both formulations were found to portray immediate and rapid drug release during the first two hours of the study with the suspension reaching its maximum concentration at 20 ng/mL, while the F9 formulation continued to release drug sharply with a concentration of 103 ng/mL by that time. After a period of two hours, the drug release from the suspension decreased gradually, reaching 0 ng/mL concentration at the end of the study (24 h). While APRoptimized SNEDDS continued to increase readily after two hours until the maximum concentration of 119 ng/mL was reached at 5 h. After that, it decreased steeply, reaching 90 ng/mL concentration at 6 h. The next hours showed a gradual decrease in the drug plasma concentration until the end time point was reached (24 h) with a concentration of 22 ng/mL. Overall, the release profile of APR from optimized SNEDDS F9 was significant compared to the drug suspension (*p* < 0.05). The noncompartmental pharmacokinetic model was used to calculate different pharmacokinetic parameters of APR including Cmax, AUC0–t, AUC0–inf, λz, T<sup>1</sup> 2 , Tmax and relative bioavailability [33–35]. The most significant parameters compared to the APR suspension were the Tmax = 4.00 ± 0.96 h, which was 2.0 ± 1.70 h for the APR suspension, AUC0–t = 3256.76 ± 212.50 ng.h/mL compared to APR suspension 462.83 ± 52.25 ng.h/mL, AUC0–<sup>∞</sup> = 3481.04 ± 235.51 ng.h/mL compared to APR suspension 488.13 ± 61.31 ng.h/mL and the relative bioavailability = 703.66% compared to APR suspension = 100%, which indicated a seven-fold increase in APR bioavailability (*p* < 0.05).

#### **4. Materials and Methods**

### *4.1. Materials*

APR was purchased from Beijing Mesochem Technology Pvt. Ltd. (Beijing, China). From Gattefossé (Lyon, France), Lauroglycol-90, Capryol-90, Labrasol, Capryol-PGMC, Transcutol-HP, Labrafil-M1944CS, Lauroglycol-FCC, Labrafac-PG and Peceol were purchased. Ethanol, isopropyl alcohol (IPA), polyethylene glycol-400 (PEG-400), ethylene glycol (EG), propylene glycol (PG), Tween-80, Triton-X100 and Tween-85 were acquired from Sigma Aldrich (St. Louis, MO, USA). Cremophor-EL was acquired from BASF (Cheshire, UK). From Nikko Chemicals (Tokyo, Japan), Sefsol-218 and HCO-60 were obtained. HPLC-grade solvents, Ethanol Chromasolv® absolute for HPLC was purchased from Sigma Aldrich (St. Louis, MO, USA) and methanol HPLC-grade was purchased from Fischer Scientific (Waltham, MA, USA). Lastly, from ELGA water purification system (Wycombe, UK), the deionized water was procured.

#### *4.2. Screening of Components*

The equilibrium solubility of APR was examined in different oils (Triacetin, Lauroglycol-90, Lauroglycol-FCC, Capryol-90 and Capryol-PGMC), surfactants (Tween-80, Labrasol, Cremophor-EL and Triton-X100), cosurfactants (Transcutol-HP, PEG-400, ethanol, PG, EG, IPA and water). The water is frequently used as an aqueous phase, as found in the literature [19,20]. The method used to confirm the saturated solubility of APR was the equilibrium method [28]. The solubility of APR in each component was determined at 25 ◦C. The excess amount of solid APR was added in known amounts of each component in triplicates. Each mixture was vortexed for about 5 min and transferred to the "OLS 200 Grant Scientific Biological Shaker (Grant Scientific, Cambridge, UK)" at the shaking speed of 100 rpm for the period of 72 h [9]. After 72 h, each mixture was removed from the biological shaker, and filtered and centrifuged at 5000 rpm. The supernatants were taken, diluted suitably with mobile phase (wherever applicable) and subjected for the analysis of APR content using RP-HPLC method at 254 nm. The concentration of APR in solubility samples was determined by a calibration curve of APR.

### *4.3. Construction of Pseudo-Ternary Phase Diagrams for APR SNEDDS*

Nanoemulsions are multicomponent systems and therefore, pseudo-ternary phase diagrams are most suitably constructed for them. After choosing the SNEDDS components from the solubility studies, the pseudo ternary phase diagrams are constructed. Generally, phase diagrams are graphical plots that are used to examine different thermodynamic parameters of a given system. They show the relationship between the different system phases at equilibrium or even various conditions. They identify the factors that could affect the equilibrium such as temperature, pressure, concentration and pH. The number of plots is related to the number of the components in a system. In the case of SNEDDS, phase diagrams identify the emulsification areas of the nanoemulsion and the number the plots on the phase diagram is three. One plot is for the oil phase, the second plot is for the aqueous phase, and the third one represents the surfactants mixture ratio (Smix ratio). In APR SNEDDS, the aqueous phase used was de-ionized water, the oil phase was Lauroglycol-90, the surfactant was Tween-80, and the cosurfactant was Transcutol-HP. To form the Smix, Tween-80 and Transcutol-HP were mixed together in several mass ratios from 1:4 and 4:1 ratios. Then, the oil phase Lauroglycol-90 was mixed from 1:9 to 9:1 ratios with the Smix ratios. After that, gradual or drop wise titration for the oil-Smix ratios by the de-ionized water was carried out, refereeing this step as phase titration. The appearance of each mixture was observed in terms of clarity and turbidity during the titration. A clear transparent appearance stood for nanoemulsion and a turbid milky appearance stood for regular emulsion. Finally, the physical observations were marked on the phase diagram to note the optimum ratios for the further APR SNDDS formulation [30–32].

#### *4.4. Formulation Development*

Using the aqueous phase titration method/spontaneous emulsification method to create the phase diagrams, the maximum SNEDDS zones for APR-SNEDDS were identified [31,32]. The maximum SNEDDS zones were observed with 1:1 mass ratio of Tween-80 and Transutol-HP. Nine APR-SNEDDS in a total of 1 mL each were utilized using the 1:1 Smix ratio, namely F1, F2, F3, F4, F5, F6, F7, F8 and F9, considering almost an entire range of SNEDDS zones in the phase diagram with various Lauroglycol-FCC concentrations (10, 15, 20, 25% *v*/*v*), Smix concentrations (30, 40, 45, 50, 55, 60% *v*/*v*), and aqueous phase concentrations (25, 30, 35, 40, 50, 55% *v*/*v*). Each formulation contained 5 mg of the drug dissolved completely in Lauroglycol-FCC before the addition of Tween-80 and Transcutol-HP mixture followed by vortex shaking until a clear and transparent mixture was obtained. The deionized water was added gradually by a drop wise pattern while vortexing with different concentrations used for each formulation to produce particulate free and clear formulations. The composition of each formulation is illustrated in Table 2.

#### *4.5. Thermodynamic Stability Testing*

The purpose of thermodynamic stability testing is to exclude metastable APR-SNEDDS (stable under certain conditions and/or reform slowly) and unstable APR-SNEDDS (unstable under standard conditions and/or do not reform) through applying various external conditions that might affect their stability, such as centrifugation, heating–cooling cycles and freeze–pump–thaw cycles. The centrifugation of APR-SNEDDS (F1–F9) was carried out at 5000 rpm, 25 ◦C for 30 min, the heating–cooling cycles were carried out between 4 ◦C (refrigerator) and 50 ◦C (oven) for 48 h for 3 cycles, and the freeze–pump–thaw cycles were carried out between −21 (freeze) and +25 ◦C (thaw) for 24 h for 3 cycles [36,37].

#### *4.6. Self-Nanoemulsification Efficiency Test*

The APR-SNEDDS that withstood the thermodynamic stability testing were subsequently subjected to self-nanoemulsification efficiency testing. The aim of this test is to examine the SNEDDS stability regarding the occurrence of phase separation or precipitation upon dilution with water. In order to conduct this test, the dilution of 1 mL from each APR SNEDDS was done 500 times with different diluents, such as 0.1 N HCl, deionized water and phosphate buffer (pH 6.8). The efficiency of each SNEDDS was evaluated and graded via inspection and the use of a system for grading [12,16]:

Grade A: Rapidly forming clear/transparent nanoemulsion (emulsify within 1 min) Grade B: Rapidly forming bluish white nanoemulsion (emulsify within 2 min) Grade C: Milky emulsions (take more than 2 min to emulsify)

Grade D: Dull, grayish milky emulsions (take more than 3 min to emulsify)

Grade E: Emulsions with oil globules at the surface (take more than 5 min to emulsify).

#### *4.7. Physicochemical Characterization*

Developed APR SNEDDs were physicochemically characterized by testing a number of variables such as droplet size, PDI, ZP, RI, %T, and surface morphology [37–39]. For the droplet size measurement, 1 drop of APR-SNEDDS was diluted with water at 25 ◦C with a scattering angle of 90◦ using Malvern Particle Size Analyzer (Malvern Instruments Ltd., Holtsville, NY, USA). PDI was measured using the same dilution, temperature, scattering angle and instrument as the droplet size measurement. Additionally, ZP of diluted APR-SNEDDS with water was measured using glass electrodes at pH 7.0. The RI was measured by Abbes Refractometer without any sample dilution. The %T was estimated for 1 drop APR-SNEDDS diluted with methanol and a blank of methanol using a spectrophotometer at 550 nm detection wavelength. Lastly, the surface morphology evaluation of optimized APR-SNEDDS was performed using the dilution by TEM at 100–200 Kv.

#### *4.8. In Vitro Dissolution Studies*

The purpose of this study was to develop a comparison between in vitro APR release from the developed APR-SNEDDS (F1-F9) and APR suspension. The investigation was carried out using a dialysis membrane from Spectrum Medical Industries (Mumbai, India; MWCO 12,000 Da) and dissolution apparatus in accordance with United States Pharmacopoeia (USP) XXIV method [18] with the following conditions: rotational speed fixed at 100 rpm in 500 mL dissolution media of pH controlled 6.8 phosphate buffer, at 37 ± 0.5 ◦C temperature. An amount of 1 mL from both the APR SNEDDS and APR suspension was transferred to the dialysis bags. From each formulation, a 3 mL sample was withdrawn at regular time intervals and replaced at the same time with 3 mL of drug free dissolution media (phosphate buffer pH 6.8). The amount of APR in each sample was determined using the reported RP-HPLC method [8]. The drug release mechanism from the SNEDDS formulation was studied via the application of different kinetic models such as zero order, first order, Higuchi, Hixson–Crowell, and Korsemeyer–Peppas models [21–23].

#### *4.9. Bioavailability Study and Pharmacokinetic Evaluation*

A single oral dose parallel-built study was performed for a bioavailability and pharmacokinetic study on twelve male Wistar Albino rats weighing around 200–250 kg, provided from the Animal Care and Use Centre, College of Pharmacy, King Saud University, Riyadh, Saudi Arabia. The entire study was performed in accordance with King Saud University Animal Care and Use Committee guidelines and were approved by the Animal Ethical Committee of King Saud University (Approval number: SE-19-123). Before starting the experiment, the rats were acclimatized in plastic cages under common lab conditions, maintaining controlled temperature and humidity of 25 ± 2 ◦C and 55 ± 5% RH, respectively, with a light/dark cycle (12 h), drinking water, and feeding on rats' pellet diet ad libitum. The rats were randomly divided into two groups (*n* = 6 in each group), which served as APR suspension (sodium carboxymethyl cellulose, 0.5% *w*/*v*) and optimized APR SNEDDS (F9) treatment groups, respectively. The rats were fasted overnight before the experiments. Blood samples (approximately 500 µL) were taken from the retro-orbital plexus into heparinized microfuge tubes at 0, 1, 2, 4, 6 and 24 h after oral administration of APR (3 mg/kg, oral) in both groups. Plasma samples were harvested by centrifuging the blood at 5000× *g* for 8 min. Plasma samples were mixed with acetate buffer (pH 4.6) in the ratio of 1:10 (buffer: plasma) and stored in a deep freezer at −80 ± 10 ◦C until further analysis.

The drug analysis was carried out using a reported UPLC-MS/MS method [24]. A validated and reported UPLC-MS/MS (UPLC, Waters Acquity, Milford, MA, USA) was employed to determine the concentration of APR in rat plasma [24]. The chromatographic conditions involved the use of a Acquity BEH C18 column (100 mm × 2.1 mm, 1.7 µm), mobile phase mixture (85:15, *v*/*v*) of acetonitrile and 10 mM ammonium acetate and flow rate of 0.30 mL/min. The eluted compounds (APR and IS) were detected by tandem mass spectrometry using TQ detector (Waters Corp., Milford, MA, USA) attached to an electrospray ionization (ESI) source operating in negative ionization mode. A protein precipitation method with the use of ethyl acetate as a solvent were carried out for the extraction of the drug from rat plasma. In this study, celecoxib was used as the IS. About 20 µL of IS combined with 200 µL of rat plasma (2.0 µg/mL) and 2.0 mL of ethyl acetate. The mixture was vortexed for 2.0 min. The samples were centrifuged at 50,000 rpm for about 5 min and 500 µL of the supernatant was removed and placed in a sample vial for further analysis in the UPLC-MS/MS system. The analysis of APR in rat plasma was carried out via the injection of about 5 µL sample into UPLC-MS/MS.

The plasma concentration values of APR at different time intervals were used to evaluate its pharmacokinetic profiles by plotting drug concentration–time curves. The software used for calculation of pharmacokinetic parameters of APR was WinNonlin software (Pharsight Co., Mountain View, CA, USA) [35]. The noncompartmental pharmacokinetic model was used to calculate the Cmax, Tmax, AUC0–t, AUC0–inf, λz and T<sup>1</sup> [40,41].

2

#### *4.10. Statistical Analysis*

Diverse physicochemical variables, drug delivery and biological data was analyzed using GraphPad InStat® software (San Diego, CA, USA) applying unpaired Dunnett's test. Differences between each two related parameters were considered statistically significant for a *p*-value of ≤0.05.

#### **5. Conclusions**

APR is the first orally administered drug approved for the treatment of active psoriatic arthritis, which is a painful and inconvenient disease that can lead to other diseases and disability. In comparison to the common treatment regimens, the treatment with APR was well tolerated by the patients and associated with better overall disease-related improvements. In terms of side effects, it was found that APR has minimal adverse events but only upon the initiation of therapy, and they can be resolved or controlled during the treatment course. However, the marketed film-coated immediate release tablets of

APR Otezla® have 20–33% bioavailability compared to an APR absolute bioavailability of 73%. To our knowledge, very limited approaches to enhance APR solubility, dissolution, permeability and oral bioavailability have been reported in the literature. Therefore, the aim of this study was to enhance APR's drawbacks using a less complicated and more reproducible method. This was perfectly achieved through the application of SNEDDS as a potential drug delivery system for APR. SNEDDS were reported in many studies working as solvers for the problems associated with class II, III and IV drugs, which have poor solubility or/and poor permeability, which affects their overall oral bioavailability. In conclusion, the optimum formulation was F9, composed of 15% oil, 60% Smix, and 25% aqueous phase with the lowest droplet size (17.505 ± 0.247 nm), low PDI (0.147 ± 0.014), low ZP (−13.35 mV), highest %T (99.15 ± 0.131), maximum in vitro drug release (94.9%) over 24 h and optimum relative bioavailability (703.66%). Following the promising results of the current study, future studies should be carried out to evaluate the major factors that influence the encapsulation efficiency and stability of APR-containing NEs and the application of the formulations for the oral delivery of APR.

**Author Contributions:** Conceptualization, F.S. and I.A.A.; methodology, A.S.A., F.S.A. and F.Y.A.; software, N.H.; validation, F.S., F.S.A. and I.A.A.; formal analysis, N.H.; investigation, A.S.A., M.I. and N.H.; resources, I.A.A.; data curation, F.S.; writing—original draft preparation, F.S.; writing review and editing, I.A.A. and F.S.A.; visualization, I.A.A.; supervision, I.A.A. and F.S.; project administration, I.A.A.; funding acquisition, F.S.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research project was supported by Researchers Supporting Project number (RSP-2021/340), King Saud University, Riyadh, Saudi Arabia and APC was supported by the RSP.

**Institutional Review Board Statement:** The animal study protocol was approved by the Animal Ethical Committee of King Saud University (Approval number: SE-19-123).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** This study did not report any data.

**Acknowledgments:** Authors are thankful to the Researchers Supporting Project number (RSP-2021/340), King Saud University, Riyadh, Saudi Arabia for supporting this work.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples of the compounds APR are available from the authors.

#### **References**


## *Article* **Novel C<sup>60</sup> Fullerenol-Gentamicin Conjugate–Physicochemical Characterization and Evaluation of Antibacterial and Cytotoxic Properties**

**Aleksandra Nurzynska <sup>1</sup> , Piotr Piotrowski 2,\* , Katarzyna Klimek 1,\* , Julia Król 2 , Andrzej Kaim <sup>2</sup> and Grazyna Ginalska <sup>1</sup>**


**Abstract:** This study aimed to develop, characterize, and evaluate antibacterial and cytotoxic properties of novel fullerene derivative composed of C<sup>60</sup> fullerenol and standard aminoglycoside antibiotic– gentamicin (C<sup>60</sup> fullerenol-gentamicin conjugate). The successful introduction of gentamicin to fullerenol was confirmed by X-ray photoelectron spectroscopy which together with thermogravimetric and spectroscopic analysis revealing the formula of the composition as C60(OH)12(GLYMO)11(Gentamicin)0.8. The dynamic light scattering (DLS) revealed that conjugate possessed ability to form agglomerates in water (size around 115 nm), while Zeta potential measurements demonstrated that such agglomerates possessed neutral character. In vitro biological assays indicated that obtained C<sup>60</sup> fullerenol-gentamicin conjugate possessed the same antibacterial activity as standard gentamicin against *Staphylococcus aureus*, *Staphylococcus epidermidis*, *Pseudomonas aeruginosa*, and *Escherichia coli*, which proves that combination of fullerenol with gentamicin does not cause the loss of antibacterial activity of antibiotic. Moreover, cytotoxicity assessment demonstrated that obtained fullerenolgentamicin derivative did not decrease viability of normal human fibroblasts (model eukaryotic cells) compared to control fibroblasts. Thus, taking into account all of the results, it can be stated that this research presents effective method to fabricate C<sup>60</sup> fullerenol-gentamicin conjugate and proves that such derivative possesses desired antibacterial properties without unfavorable cytotoxic effects towards eukaryotic cells in vitro. These promising preliminary results indicate that obtained C<sup>60</sup> fullerenol-gentamicin conjugate could have biomedical potential. It may be presumed that obtained fullerenol may be used as an effective carrier for antibiotic, and developed fullerenol-gentamicin conjugate may be apply locally (i.e., at the wound site). Moreover, in future we will evaluate possibility of its applications in *inter alia* tissue engineering, namely as a component of wound dressings and implantable biomaterials.

**Keywords:** functionalized fullerenes; fullerenols; antibiotics; gentamicin; antibacterial properties; cytotoxicity; skin fibroblasts; nanomedicine

### **1. Introduction**

The fullerenes, a family of carbon allotropes, represent very promising group of chemical molecules in the context of biomedical applications [1–5]. In this case, their hydrophilic derivatives have been gained the special scientific attention, thanks to water solubility [2] and biological activities, such as anti-oxidative and radical-quenching [6,7], antibacterial [8,9], antiviral [10,11], DNA photocleavage [12,13] as well as enzyme-inhibiting [14–16] or neuroprotective [17,18] properties. Thus, thanks to aforementioned features, it was suggested that water-soluble fullerene derivatives could be potentially used as antioxidant

**Citation:** Nurzynska, A.; Piotrowski, P.; Klimek, K.; Król, J.; Kaim, A.; Ginalska, G. Novel C<sup>60</sup> Fullerenol-Gentamicin Conjugate–Physicochemical Characterization and Evaluation of Antibacterial and Cytotoxic Properties. *Molecules* **2022**, *27*, 4366. https://doi.org/10.3390/ molecules27144366

Academic Editor: Faiyaz Shakeel

Received: 27 May 2022 Accepted: 6 July 2022 Published: 7 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

agents, therapeutics for the treatment of HIV infection, and also as photosensitizers or drug carriers [1,3,19,20].

The applications of water-soluble fullerenes as drug delivery vehicles are especially focused on anticancer therapy [1,3,21,22]. For instance, Chaudhuri et al. [23] demonstrated that a very popular chemotherapeutic agent–doxorubicin (DOX) together with polyethylene glycol (PEG) may be conjugated with polyhydroxylated fullerene (PHF) C60, which allows to obtain nanocomplex with high therapeutics efficacy and lower cytotoxicity compared to free doxorubicin. PHFs were also linked with other anti-cancer drug, namely methotrexate (MTX). Bahuguna et al. [24] received FLU-MTX nanocomplex, which was characterized by increased availability of drug to the biological system and increased cytotoxicity toward cancer cells compared to free MTX. In turn, Prylutska et al. [25] developed anti-cancer nanocomplex consisted of C<sup>60</sup> fullerene aqueous colloid solution (C60FAS) and cisplatin (Cis). The authors showed that C<sup>60</sup> + Cis nanocomplex exhibited higher toxicity towards human leukemia cells compared to cisplatin alone. Moreover, in vivo study with Lewis lung carcinoma (LLC) C57BL/6J male mice demonstrated that C<sup>60</sup> + Cis conjugate inhibited tumor growth more potently than free cisplatin.

Nevertheless, to the best of our knowledge, there are no research reports which present adducts composed of fullerenols and antimicrobials. Such composition seems to be promising thanks to biological activities of fullerenols and antimicrobial agents. In other words, we assumed that C<sup>60</sup> fullerenol will be not only a nanocarrier for gentamicin, but also may exhibit beneficial biological properties (for instance antioxidant or free-radical scavenging). It is especially important as an application of antibiotics most often is associated with unfavorable side effects. It is known that aminoglycoside antibiotics (such as gentamicin) may cause nephrotoxicity (damage of kidneys) and ototoxicity (damage of hearing organ) [26,27]. Thus, our preliminary study aimed to develop novel fullerene derivative composed of C<sup>60</sup> fullerenol and standard aminoglycoside antibiotic–gentamicin. Desired product was synthesized by binding (3-glycidyloxypropyl)trimethoxysilane linker to fullerenol surface using part of its hydroxyl groups and further reaction of introduced oxirane moieties with gentamicin –NH<sup>2</sup> groups. Synthesized nanomaterial was evaluated using X-ray photoelectron spectroscopy (XPS), Fourier-transform infrared spectroscopy (FT-IR), thermogravimetry (TGA), and dynamic light scattering (DLS). This nanoconjugate was also subjected to initial evaluation of biological properties. The antibacterial activity of fullerenol-gentamicin nanocarrier was assessed towards Gram-positive and Gram-negative bacterial strains (i.e., *Staphylococcus aureus*, *Staphylococcus epidermidis*, *Pseudomonas aeruginosa*, and *Escherichia coli*, respectively). Moreover, its cytotoxicity was estimated using model eukaryotic cells–normal human skin fibroblasts (BJ cell line). To our best knowledge, it is first research, which demonstrates effective method allowing to conjugate fullerenol with aminoglycoside antibiotic.

#### **2. Results and Discussion**

## *2.1. Characterization of C<sup>60</sup> Fullerenol-Gentamicin Derivative*

Composition of synthesized C<sup>60</sup> fullerenol-gentamicin conjugate was investigated using X-ray photoelectron spectroscopy (XPS). Survey XPS spectrum obtained for the newly synthesized fullerene derivative revealed the presence of carbon, oxygen, nitrogen, and silicon atoms, namely elements expected for gentamicin functionalized fullerenol (see Supplementary Material, Figure S1).

C1s spectrum obtained for C<sup>60</sup> fullerenol-gentamicin derivative (**60FGG**) (Figure 1a) was deconvoluted with very good correlation into four signals [28]. The lowest binding energy peak, centered at 284.6 eV is assigned to sp2 carbon from fullerene cage. Second signal was registered at 285.6 eV and can be attributed to the sp3 carbon. Next contribution (at 286.6eV) was ascribed to C-O carbon atoms from fullerenol, linker, and gentamicin moieties. The highest binding energy peak, arising from C = O carbon atoms, was registered at 288.0 eV and suggests that traces of hydroxyl groups present at the C<sup>60</sup> surface were oxidized during the synthesis process.

**Figure 1.** The XPS spectra obtained for C<sup>60</sup> fullerenol-gentamicin derivative (**60FGG**)–C1s spectrum (**a**), N1s core level region (**b**), and Si2p region (**c**).

Single peak observed in the N1s core level region (Figure 1b) with maximum located at 399.9 eV is assigned to the nitrogen atoms from amino groups of gentamicin bound to the fullerenol derivative surface. This result confirms successful introduction of aforementioned aminoglycoside antibiotic.

Another strong evidence of the modification of fullerenol surface was provided by Si2p region of XPS spectrum (Figure 1c), where double peak is observed with signals centered at 103.6 and 104.2 eV assigned to 2p3/2 and 2p1/2 silicon atoms from the GLYMO linker [29]. Those signals show expected area ratio of 2:1 and splitting of 1.2 eV.

Additionally, lack of signals associated with sulfur atoms implies that introduced gentamicin molecules were neutralized during the final step of synthesis. Released sulfate ions were removed during the dialysis process.

The XPS results along with thermogravimetric measurements [30] (Figure 2) allowed to calculate the approximate composition of fullerenol, its GLYMO derivative, and synthesized fullerenol-gentamicin conjugate. Single step thermal decomposition of **60F** resulted in weight loss of 35%, which corresponds to 23 hydroxyl groups. GLYMO functionalized fullerenol and its gentamicin adduct revealed more complex TGA curves, both with shape close to two step decomposition. Total weight loss was determined to be 77% in case of **60FG** and 80% for **60FGG**. Thus, final composition of synthesized nanomaterial was estimated to be C60(OH)12(GLYMO)11(Gentamicin)0.8.

**Figure 2.** Thermogravimetric analysis of fullerenol (**60F**), GLYMO modified fullerenol (**60FG**), and gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**) at 5 K/min under N<sup>2</sup> atmosphere.

Functionalization of C<sup>60</sup> fullerenol (**60F**) with GLYMO allowed to form **60FG** and then to introduce gentamicin molecules in order to obtain desired conjugate, i.e., **60FGG** that was also confirmed by means of FT-IR spectroscopy (Figure 3). First of all, **60FGG** spectrum shows broad band from 3700 to 2400 cm−<sup>1</sup> , characteristic to gentamicin substrate. Despite relatively strong intensity of this signal, small peak at 2842 cm−<sup>1</sup> , also observed in **60FG** spectrum is still present, indicating that final product has **60FG** contributions associated with its alkane C-H stretching modes. Additionally, amine N-H bending mode is present at around 1635 cm−<sup>1</sup> in both gentamicin functionalized fullerene and unmodified gentamicin samples. This band overlaps with signals attributed to C=C stretching vibrations, which are present in obtained fullerene derivatives at around 1600 cm−<sup>1</sup> . Presence of signal centered around 1420 cm−<sup>1</sup> in all fullerene related samples: **60F**, **60FG** and **60FGG** confirms presence of C<sup>60</sup> core in those products [31]. Broad band at around 1130 cm−<sup>1</sup> present in both **60FGG** and **G** samples can be associated with C-O stretching vibrations coming from secondary and tertiary alcohol groups present in the gentamicin structure. Spectrum of **60FG** shows band at 826 cm−<sup>1</sup> , which can be assigned to asymmetric ring deformation of epoxide groups [32–34], coming from introduced GLYMO linkers. Intensity of this signal is significantly lowered in the sample after reaction with gentamicin, which is in good agreement with expected reaction pathway, leading to epoxide ring opening due to reaction with amino groups from antibiotic molecules. Another very important signal, which is associated with gentamicin moiety can be observed in **60FGG** spectrum at 614 cm−<sup>1</sup> , additionally confirming presence of the aminoglycoside antibiotic in final product.

**Figure 3.** FT-IR spectra recorded in KBr disk for: gentamicin (**G**), fullerenol (**60F**), GLYMO modified fullerenol (**60FG**), and gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**).

The dynamic light scattering (DLS) measurements allowed to determine the hydrodynamic diameter of the synthesized fullerene derivatives in water (Figure 4). Size distribution for **60F** revealed single peak at approximately 92 nm, which is value typical for fullerenol associations [35–37]. Introduction of gentamicin onto fullerenol surface results in slight increase of the size of agglomerates, which was found to be around 115 nm.

DLS analysis was accompanied by Zeta potential measurements. Both fullerenol (**60F**) and gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**) revealed to form neutral agglomerates [38]. Corresponding Zeta potential values were determined to be −0.1 mV and 0.1 V for **60F** and **60FGG**, respectively (see Supplementary Material, Figures S2 and S3). Results obtained for C<sup>60</sup> fullerenol-gentamicin conjugate are in good agreement with XPS data, which suggested that gentamicin was neutralized during **60FGG** synthesis and purification process.

**Figure 4.** Size distributions obtained from DLS measurements for: fullerenol **60F** (**a**) and C<sup>60</sup> fullerene gentamicin conjugate **60FGG** (**b**).

## *2.2. Antibacterial Properties of Gentamicin Functionalized C<sup>60</sup> Fullerene Derivative*

Firstly, antibacterial tests showed that hydroxylated C<sup>60</sup> fullerene (**60F**) did not exhibit antibacterial activity towards all tested bacterial strain. Interestingly, some authors indicated that water-soluble C<sup>60</sup> fullerene derivatives possessed antibacterial activities [8,9]. For instance, Deryabin et al. [8] developed water-soluble C<sup>60</sup> fullerene derivatives bearing amine (AF) and carboxylic (CF) groups. The authors indicated that both derivatives had ability to form nanoclusters in aqueous solutions–diameters ranged from 2 to 200 nm (AF) and from 70 to 100 nm (CF) (based on DLS measurements). Nevertheless, only AF derivatives possessed antibacterial properties. The authors suggested that this phenomenon was associated with electrostatic interactions between fullerene derivatives and cells. Fullerene derivative (AF) with positive charge adhered to negatively charged bacterial cells, while negatively charged CF derivatives did not possess ability to interact with bacterial cells. Thus, it seems that charge of fullerene derivative has greater influence on antibacterial activity than particle size. In our study, we showed (DLS measurements, Section 2.1) that size distribution of **60F** in water was 92 nm, while introduction of gentamicin onto fullerenol surface resulted in slight increase of the size of agglomerates, which was found to be around 115 nm. Thus, the particle size of both products (**60F** and **60FGG**) was comparable. In turn, Zeta potential measurements showed that both possessed neutral charges. Thus, these results may explain the lack of bacterial properties of **60F**–neutrally charged derivative rather should not have ability to adhere to negatively charged bacterial cells. Our study also demonstrated that both gentamicin sulfate salt (**G**) and conjugate composed of C<sup>60</sup> fullerenol and gentamicin (**60FGG**) possessed the same activity–values of MIC and MBC for both compounds were identical and ranged from 0.125 to 1 µg/mL, depending on bacterial strain (Tables 1 and 2). Thus, these results indicated, that conjugate composed of **G** and **60F** did not exhibit greater antibacterial activity than **G** alone, but allowed to maintain the antibacterial properties of **G**. In other words, these antibacterial assays confirmed that newly developed method for fabrication of **60FGG** allows to obtain effective, antibacterial agent. It is worth noting that the aim of this study is not to develop C<sup>60</sup> fullerenol with antibacterial properties or C<sup>60</sup> fullerenol-gentamicin conjugate with better activity than gentamicin alone. We aimed to develop nanoconjugate which will maintain antibacterial activity of gentamicin and will not exhibit cytotoxic properties towards eukaryotic cells.


**Table 1.** MIC values for *S. aureus*, *S. epidermidis*, *P. aeruginosa*, and *E. coli* after 24 h incubation with aqueous solutions of gentamicin sulfate salt (**G**), C<sup>60</sup> fullerenol (**60F**), and gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**).

<sup>a</sup> Minimum Inhibitory Concentration (MIC)–the lowest compound concentration that inhibits more than 90% of bacterial growth. <sup>b</sup> ND (not determined). In tested concentrations, no antibacterial activity was observed.

**Table 2.** MBC values for *S. aureus*, *S. epidermidis*, *P. aeruginosa*, and *E. coli* after treatment with aqueous solutions of gentamicin sulfate salt (**G**), C<sup>60</sup> fullerenol (**60F**), and gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**).


<sup>a</sup> Minimum Bactericidal Concentration (MBC)–the lowest compound concentration that decreased the number of colonies by <sup>≥</sup>99.99% compared to the control growth. <sup>b</sup> NT (not tested). Lack of antibacterial activity was determined already during MIC evaluation.

## *2.3. Cytotoxic Properties of Gentamicin Functionalized C<sup>60</sup> Fullerene*

The MTT assay indicated that all tested solutions prepared from gentamicin sulfate salt (**G**), C<sup>60</sup> fullerenol (**60F**), and gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**) were non-cytotoxic towards normal human skin fibroblasts (Figure 5). The cell viability after treatment with tested solutions of **G**, **60F**, and **60FGG** was comparable with viability of control cells (cultured without investigated compounds) and close to 100%. These results indicated that introduction of gentamicin to C<sup>60</sup> fullerenol did not have unfavorable influence on cell behavior. Thus, obtained **60FGG** conjugate possessed desired antibacterial activity (please see Section 2.2) and at the same time was safe for model eukaryotic cells (human fibroblasts) in vitro.

**Figure 5.** Viability of normal human skin fibroblasts (BJ cell line, CRL-2522TM) after 24 h treatment with aqueous solutions of gentamicin sulfate salt (**G**), C<sup>60</sup> fullerenol (**60F**), and gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**). Gentamicin (**G**) was evaluated at concentration of 4–0.008 µg/mL, C<sup>60</sup> fullerenol (**60F**) at concentration of 29.25–0.058 µg/mL, while concentrations of gentamicin and C<sup>60</sup> fullerenol in obtained conjugate (**60FGG**) ranged from 4–0.008 µg/mL and 29.25–0.058 µg/mL, respectively. The cell viability was assessed by MTT assay. The results were not statistically significant (*p* > 0.05) compared to control, namely cell incubated with culture medium without tested compounds; one-Way ANOVA test, followed by a Tukey's multiple comparison test.

In general, water-soluble fullerene derivatives exhibit no or minimal cytotoxicity [19]. Nevertheless, it is also known that cytotoxicity of these derivatives may be controlled by *inter alia* derivatization of their surfaces. Thus, the increase in cytotoxicity of fullerene

derivatives is suggested as a promising approach in antibacterial or cancer therapies [39]. Aforementioned positive-charged amine derivatives of fullerene (AF) (please see Section 2.2) possessed antibacterial activity thanks to their ability to adhere to negatively charged bacterial cells. Nevertheless, cell membrane of eukaryotic cells are also negative-charged, which suggests that such derivative will also adhere to these cells. Indeed, the authors of cited manuscript proved that AF derivatives adhered also to human erythrocytes, which suggests that they may have cytotoxic activity towards eukaryotic cells [8]. On the other hand, some authors demonstrated that water-soluble fullerene derivatives linked with anticancer agents exhibited increased cytotoxicity towards cancer cells compared to free anticancer agents [24,25]. However, they evaluated the influence of obtained derivatives only towards cancer cells. Thus, it is possible that such amine derivatives of fullerene may also have unfavorable influence on normal cells.

It is known that cytocompatibility is a mandatory feature of potential medicinal products [40,41]. Thus, non-cytotoxic **60FGG** derivative seems to be promising candidate for further analysis.

## **3. Materials and Methods**

The general plan of performed experiments was introduced in Figure 6.

**Figure 6.** Schematic diagram presenting the experiments performed within this study.

### *3.1. Materials*

Dimethylformamide (DMF), gentamicin sulfate salt, (3-glycidyloxypropyl)trimethoxysilane (GLYMO), methanol, penicillin-streptomycin solution, phosphate buffered saline (PBS), sodium dodecyl sulfate (SDS), sodium hydroxide, tetrabutylammonium hydroxide (TBAH), thiazolyl blue tetrazolium bromide (MTT), toluene, and trypsin-EDTA solution (0.25%) were obtained from Merck, Warsaw, Poland. Fetal bovine serum (FBS) was purchased from Pan-Biotech, Aidenbach, Germany. Eagle's Minimum Essential Medium (EMEM), normal human skin fibroblasts (BJ cell line, CRL-2522TM), Staphylococcus aureus (ATCC 25923), Staphylococcus epidermidis (ATCC 12228), Pseudomonas aeruginosa (ATCC 27859), and Escherichia coli (ATCC 25922) were supplied by ATCC, Manassas, USA, while Mueller-Hinton agar (MHA) and Mueller-Hinton broth (MHB) by Oxoid, Hampshire, UK.

## *3.2. Synthesis and Characterization of C<sup>60</sup> Fullerenol-Gentamicin Derivative*

Gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**) was obtained in three-step synthesis. In the first step we obtained hydroxylated C<sup>60</sup> using modified method reported by Li and Takeuchi [42]. Briefly, C<sup>60</sup> (200 mg) was dissolved in 150 mL of toluene, then 10 mL of saturated NaOH solution was added, followed by addition of catalytic amount of TBAH. The solution was then stirred for an hour, until the organic layer became colorless and

brown precipitate appeared. Toluene was decanted from the solid and 25 mL of deionized water was added. After the solution was stirred for additional 10 h, another 50 mL of water were added and the resultant mixture was filtered through a fluted filter. Obtained orange solution was concentrated on rotary evaporator to c.a. 20 mL and 125 mL of methanol was added to precipitate product. Afterwards, crude product was re-dissolved in deionized water and dialyzed (1 kDa cut off). Resultant solution was precipitated using methanol, centrifuged and dried in a vacuum oven. TGA measurements allowed to estimate the composition of synthesized fullerenol to be C60(OH)24.

The second step was concentrated on the addition of (3-glycidyloxypropyl)trimethoxysilane (GLYMO) to the hydroxyl groups previously introduced onto fullerene surface (Figure 7). 150 mg of synthesized fullerenol was dispersed in 12 mL of anhydrous toluene using ultrasound. Then 1.2 mL of GLYMO was added and resultant mixture was refluxed in argon atmosphere for 24 h. Then, methanol was added and mixture was centrifuged. The liquids were decanted from the solid, which was washed with methanol and then dried in vacuo.

**Figure 7.** Scheme for synthesis of gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**) using fullerenol as a substrate.

The final step was performed using the modified method reported by Lizza et al. [43], where epoxide ring was opened using primary amino groups from gentamicin molecules. Crude GLYMO functionalized fullerene **60FG** (65 mg, 0.02 mmoL) was sonicated in 8 mL of DMF, followed by addition of gentamicin (94 mg, 0.2 mmol) dissolved in small amount of DMF. The reaction was carried out at 60 ◦C for 24 h. The water (18 mL, 1 mmoL) was added and the mixture was centrifuged to separate insoluble residue. Obtained brown liquid was decanted and dialyzed (1 kDa cut off). Then, it was concentrated on rotary evaporator and methanol was added in order to form a precipitate of the final product, which composition was estimated to be C60(OH)12(GLYMO)11(Gentamicin)0.8.

Obtained products were characterized using X-ray photoelectron spectroscopy (XPS), Fourier-transform infrared spectroscopy (FT-IR), and thermogravimetry (TGA). XPS measurements were carried out using a VG ESCALAB 210 electron spectrometer (Thermo Fisher Sceintific, Bothell, WA, USA) equipped with an Al-Kα source (1486.6 eV). XPS data were calibrated using the binding energy of C1s = 284.6 eV as the internal standard. The infrared experiments were performed on the Shimadzu FTIR-8400S (Shimadzu, Kioto, Japan). The samples were prepared as KBr disks. Thermogravimetric measurements were performed using Q50 TGA (TA Instruments, New Castle, DE, USA). The analyzed sample was previously dried under vacuum at 60 ◦C and the measurement was carried out under flow of nitrogen with heating rate of 5 K/min. Size of synthesized fullerene derivatives their Zeta potential were measured in water using Zetasizer Nano (Malvern, UK).

## *3.3. Biological Properties of PEG Functionalized C<sup>60</sup> Fullerenol-Gentamicin Conjugate*

For biological assays, the following samples were used: C<sup>60</sup> fullerenol (**60F**), gentamicin sulfate salt (**G**), and C<sup>60</sup> fullerenol-gentamicin conjugate (**60FGG**). The samples were put into 1.5 mL eppendorf tubes and sterilized by ethylene oxide. During preparation of aqueous solutions of samples (in sterile deionized water), estimated structure of synthesized fullerene derivative (C60(OH)12(GLYMO)11(Gentamicin)0.8) was taken into account. It was assumed that **60FGG** contains 88 wt.% of modified C<sup>60</sup> fullerenol and 12 wt.% of gentamicin. Thus, tested concentrations of **60F** and **G** in solutions of **60FGG** ranged 29.25–0.057 µg/mL and 4–0.007 µg/mL, respectively. In parallel, **60F** was tested at concentrations ranged 29.25–0.057 µg/mL, while **G** at concentrations ranged 4–0.007 µg/mL.

#### 3.3.1. Antibacterial Properties In Vitro

In order to evaluate antibacterial activity of tested samples, minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) were determined. The MIC was estimated using microdilution method according to the CLSI M7A7 standard [44]. Briefly, the two-fold dilutions of tested samples were prepared in MHB broth. Then, 100 µL of solutions were added to 96-well plate. 100 µL of MHB was used as a control solution. The bacteria inoculum at 0.5 McFarland standard density (1 <sup>×</sup> <sup>10</sup><sup>8</sup> CFU/mL) was prepared in sterile 0.85% saline using nephelometer (BD PhoenixSpecTM Nephelometer, Thermo Fisher Scientific, Waltham, WA, USA). Then, inoculum was 200-times diluted in order to obtain <sup>5</sup> <sup>×</sup> <sup>10</sup><sup>5</sup> CFU/mL. 100 <sup>µ</sup>L of diluted inoculum was added to 100 <sup>µ</sup>L of tested sample solutions or MHB (control), which were placed in 96-well plate. The plates were incubated in air at 37 ◦C without agitation for 24 h. Then, the absorbance was read at 600 nm (Synergy H4 automatic plate reader, BioTek, Winooski, VT, USA). The experiment was performed in three separate measurements in triplicate. The MIC was defined as the lowest compound concentration that inhibits more than 90% of bacterial growth. Then, in order to determine MBC, 100 µL of each well without bacterial growth was seeded on Petri dish containing MHA. After 24 h incubation at 37 ◦C without agitation, the colonies were count. MBC was defined as the lowest compound concentration that decreased the number of colonies by ≥99.99% compared to the control growth. The experiment was performed in three separate replicates.

#### 3.3.2. Cytotoxic Properties In Vitro

Cytotoxicity of compounds was evaluated using normal skin fibroblasts (BJ cell lines) as described in details previously [45]. Briefly, BJ cells were seeded in 96-well plate at concentration of 2 <sup>×</sup> <sup>10</sup><sup>4</sup> cell/well and incubated at 37 ◦C for 24 h. On the next day, the two-fold dilutions of tested samples were prepared in EMEM medium. Then, cell medium from above the cells was gently removed and 100 µL of prepared solutions were added. 100 µL of fresh EMEM medium was used as a control solution. After 24 h incubation at 37 ◦C, fibroblast viability was assessed using MTT assay. The experiment was performed in three separate measurements in octuplicate.

#### **4. Conclusions**

In this study, (3-glycidyloxypropyl)trismethoxysilane functionalized fullerene (**60FG**) was used as a vehicle for a standard aminoglycoside antibiotic–gentamicin (**G**). It is worth underlining that this is the first study which shows effective method to conjugate fullerenol with aminoglycoside antibiotic. We believe that such conjugate may allow to link biological properties of fullerenol (for instance antioxidant or free-radical scavenging) and antibacterial properties of gentamicin. It seems to be very important since the application of antibiotics (such as gentamicin) is associated with unfavorable generation of free radicals, which may lead to nephrotoxicity. Our preliminary results demonstrated that applied three-step fabrication procedure allowed to obtain gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**), as proven by XPS analysis, termogravimetric evaluation, and FT-IR measurements. Obtained **60FGG** derivative exhibited beneficial antibacterial activity against Gram-positive and Gram-negative bacterial strains, indicating that such conjugate enabled the maintaining of inhibition activity of gentamicin. Moreover, cell culture experiments proved that **60FGG** derivative did not possess cytotoxic properties towards normal human fibroblasts (model eukaryotic cells). From biomedical point of view, these preliminary results are very promising and allow to consider further application of obtained **60FGG** derivative. For instance, obtained fullerenol-gentamicin conjugate could be potentially used

as injection solution with antibacterial properties for the treatment of chronic wounds or osteochondral defects. It may be also applied as a component of bioactive biomaterials, such as wound dressings or implantable biomaterials, which are prone to bacterial infections. Presumably, addition of fullerenol-gentamicin conjugate to biomaterial, instead of gentamicin alone, allow to obtain construct possessing ability to release antibiotic in controllable manner and hopefully with antioxidant properties, thanks to presence of C<sup>60</sup> fullerenol. Nevertheless, in order to verify these hypotheses, additional experiments are needed. Thus, we plan to expand our research involving (e.g., evaluation of antioxidant properties of 60FGG derivative and possibility to use it as a component of bioactive biomaterials).

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27144366/s1, Figure S1: Survey XPS spectrum of gentamicine modified C<sup>60</sup> fullerene (**60FGG**), Figure S2: Zeta potential measurements of C<sup>60</sup> fullerenol (**60F**), Figure S3: Zeta potential measurements of gentamicin functionalized C<sup>60</sup> fullerene (**60FGG**).

**Author Contributions:** Conceptualization, A.N., A.K. and G.G.; methodology, P.P. and K.K.; software, A.N.; validation, A.N., P.P. and K.K.; formal analysis, K.K.; investigation, A.N., P.P., K.K. and J.K.; resources, A.K. and G.G.; data curation, K.K.; writing—original draft preparation, A.N. and P.P.; writing—review and editing, A.N., K.K., A.K. and G.G.; visualization, A.N.; supervision, A.K. and G.G.; project administration, G.G.; funding acquisition, A.K. and G.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by DS2 project of Medical University of Lublin, Poland. This paper was developed using the equipment purchased within agreement No. POPW.01.03.00-06- 010/09-00 Operational Program Development of Eastern Poland 2007–2013, Priority Axis I, Modern Economy, Operations 1.3. Innovations Promotion.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples of C<sup>60</sup> fullerenol-gentamicin derivative are available from the authors.

#### **References**


## *Article* **Perspectives of Positively Charged Nanocrystals of Tedizolid Phosphate as a Topical Ocular Application in Rabbits**

**Abdullah Alshememry 1,† , Musaed Alkholief 1,†, Mohd Abul Kalam <sup>1</sup> , Mohammad Raish <sup>1</sup> , Raisuddin Ali <sup>1</sup> , Sulaiman S. Alhudaithi <sup>1</sup> , Muzaffar Iqbal <sup>2</sup> and Aws Alshamsan 1,\***


**Abstract:** The aim of this study was the successful utilization of the positively charged nanocrystals (NCs) of Tedizolid Phosphate (TZP) (0.1% *w*/*v*) for topical ocular applications. TZP belongs to the 1, 3-oxazolidine-2-one class of antibiotics and has therapeutic potential for the treatment of many drug-resistant bacterial infections, including eye infections caused by MRSA, penicillin-resistant *Streptococcus pneumonia* and vancomycin-resistant *Enterococcus faecium*. However, its therapeutic usage is restricted due to its poor aqueous solubility and limited ocular availability. It is a prodrug and gets converted to Tedizolid (TDZ) by phosphatases in vivo. The sterilized NC<sup>1</sup> was subjected to antimicrobial testing on Gram-positive bacteria. Ocular irritation and pharmacokinetics were performed in rabbits. Around a 1.29 to 1.53-fold increase in antibacterial activity was noted for NC<sup>1</sup> against the *B. subtilis*, *S. pneumonia, S. aureus* and MRSA (SA-6538) as compared to the TZP-pure. The NC<sup>1</sup> -AqS was "practically non-irritating" to rabbit eyes. There was around a 1.67- and 1.43 fold increase in t1/2 (h) and Cmax (ngmL−<sup>1</sup> ) while there were 1.96-, 1.91-, 2.69- and 1.41-times increases in AUC0–24h,AUC0–∞,AUMC0–<sup>∞</sup> and MRT0–∞, respectively, which were found by NC<sup>1</sup> as compared to TZP-AqS in the ocular pharmacokinetic study. The clearance of TDZ was faster (11.43 mLh−<sup>1</sup> ) from TZP-AqS as compared to NC<sup>1</sup> (5.88 mLh−<sup>1</sup> ). Relatively, an extended half-life (t1/2; 4.45 h) of TDZ and the prolonged ocular retention (MRT0–∞; 7.13 h) of NC<sup>1</sup> was found, while a shorter half-life (t1/2; 2.66 h) of TDZ and MRT0–∞(t1/2; 5.05 h)was noted for TZP-AqS, respectively. Cationic TZP-NC<sup>1</sup> could offer increased transcorneal permeation, which could mimic the improved ocular bioavailability of the drug in vivo. Conclusively, NC<sup>1</sup> of TZP was identified as a promising substitute for the ocular delivery of TZP, with better performance as compared to its conventional AqS.

**Keywords:** tedizolid; antimicrobial; nanocrystals; eyeirritation; ocular pharmacokinetics; transcorneal permeation

## **1. Introduction**

Nanotechnology-based drug delivery systems (DDS) have overcome some of the pitfalls associated with conventional ophthalmic products (solutions, eye drops, suspensions, emulsions, etc.) such as improving the aqueous solubility and stability of poorly soluble/ lipophilic drugs [1–3]. In general, the frequent application of a topical ophthalmic dose (one–two drops) of any conventional eye drops of an antibiotic is needed in the affected eyes and only ~1–5% of the applied drug becomes available to the internal eye tissues. The poor ocular availability of conventional eye preparations have encouraged the development of novel nanocarriers-based ocular DDS, which would prolong the ocular retention of the applied dosage forms, permeate the drug(s) across the corneal and conjunctival area and improve the ocular (corneal and conjunctival) absorption and hence the bioavailability of

**Citation:** Alshememry, A.; Alkholief, M.; Abul Kalam, M.; Raish, M.; Ali, R.; Alhudaithi, S.S.; Iqbal, M.; Alshamsan, A. Perspectives of Positively Charged Nanocrystals of Tedizolid Phosphate as a Topical Ocular Application in Rabbits. *Molecules* **2022**, *27*, 4619. https:// doi.org/10.3390/molecules27144619

Academic Editor: Ildiko Badea

Received: 13 June 2022 Accepted: 18 July 2022 Published: 20 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

the drug(s) together with minimizing eyeirritation/toxicity and visual interruption, as is associated with ocular gels [4,5].

In the present study, positively charged nanocrystals (NCs) of Tedizolid Phosphate (TZP) were used for ocular delivery. The NCs were prepared using asmall amount of stabilizer(s) with a drug [6–8], representing a good alternative to the existing colloidal nanocarriers such asnanoemulsions [9,10], microemulsion [11,12], liposomes [13,14], niosomes [15,16], polymeric nanoparticles (NPs) [17,18], dendrimer nanoparticles [19,20], solid lipid nanoparticles [3,21] and polymeric micelles [22–24], etc.

Despite some drawbacks associated with nanocarriers, their potential in ocular delivery for numerous drugs has been explored well, as these carrier systems have improved the ocular availability of many poorly soluble drugs while reducing the dosing frequency of the applied dose and hence any toxicity [17,23,25,26]. Moreover, the potential of NCs in ocular applications has remained relatively unnoticed due to the availability of numerous proven bioadhesive polymeric-NPs [26–28].

TZP is a phosphate monoester and a prodrug that gets converted to its active form Tedizolid (TDZ) by phosphatase enzymes during its in vivo fate [29,30]. TDZ is a 1,3-oxazolidin-2-one class of antibiotic, frequently used in the infections caused by drug-resistant bacteria, including the methicillin-resistant *Staphylococcus aureus* (MRSA), penicillin-resistant *Streptococcus pneumonia* and vancomycin-resistant *Enterococcus faecium*, etc. [31,32]. TDZ differs from the other members of 1, 3-oxazolidin-2-one by having a modified side chain at the C5 site of the1, 3-oxazolidin-2-one nucleus, which advises its action against some linezolid-resistant pathogenic microbes [33,34]. TDZ inhibits the bacterial protein synthesis by binding to the 23S rRNA of the 50S subunit of the ribosome, as is done by other oxazolidinone antibiotics [35]. The frequency of the occurrence of the resistance to TDZ is very low and it is 4–8-times more potent than linezolid against the species mentioned above [30]. The details of the structure–activity relationship (SAR) and the mechanism of action of TDZ have been explained well in our previous reports [26,36].

Due to the above reasons, we supposed that TZP might be a good choice of antibiotic in the present scenario of growing multidrug-resistant eye infections due to MRSA and many other resistant strains. In the present study, we investigated the in vitro antimicrobial efficacy of TZP-NCs against certain strains, the ocular irritation potential (if any) of NCs, the ocular pharmacokinetics of TDZ in rabbit eyes andthe ex vivo transcorneal permeation (through excised rabbit cornea) of TZP-NCs as compared to the conventional TZP-aqueous suspension (TZP-AqS). The developed TZP-NCs were characterized well andan in vitro release of TZP through the dialysis membrane was performed and reported in the previous part of this article [37]. The previously reported LC-MS/MS method was successfully utilized for the quantitative determination of TDZ in rabbit aqueous humor samples.

#### **2. Materials and Methods**

#### *2.1. Materials*

Tedizolid and Tedizolid Phosphate (C17H15FN6O6P; MW 450.32 Da) with more than 98% purity were purchased from "Beijing Mesochem Technology Co., Ltd. (Beijing, China)". Ketamine. HCl(TEKAM®, 50 mgmL−<sup>1</sup> ) was purchased from HIKMA Pharmaceuticals (Amman, Jordan). Stearylamine and mannitol were purchased from Alpha Chemika, Mumbai, India and Qualikems Fine Chem Pvt. Ltd. (Vadodara, India), respectively. The HPLC grade methanol and acetonitrile were purchased from "BDH Ltd. (Poole, England)". Polyvinyl alcohol (Mw 16,000), Poloxamer-188 (Pluronic-F68), Sodium Lauryl Sulfate and Benzalkonium chloride were purchased from Sigma Aldrich (St. Louis, MO, USA). Milli-Q® water was obtained by a Millipore filter unit (Millipore, Molsheim, France). All the other chemicals and solvents were of analytical grade and HPLC grade, respectively.

#### *2.2. Methods*

2.2.1. Nanocrystals of Tedizolid Phosphate

The NCs of TZP were formulated by the antisolvent precipitation technique, using homogenization and probe sonication steps. The optimal formulation (TZP-NC1) was well characterized. The characterization parameters included the size, polydispersityindex, zetapotential, structural morphology by scanning electron microscopy, FTIR for any interaction with the excipients, crystallinity by differential scanning calorimetry and Xray diffraction studies, physicochemical characterization of the NCs for ocular suitability, saturation solubility, in vitro drug release in simulated tear fluid and storage stability at three different temperatures for 6 months. The data regarding these experiments have been published as a separate article in another journal [37]. For the ease of the reader, here we have included the composition of the formulations as mentioned in Table 1. Therefore, here only was the optimized formulation further subjected to the following studies for its in vivo ocular suitability in rabbits.


**Table 1.** Composition of TZP-containing formulations.

#### 2.2.2. Sterilization and Sterility Testing

The final formulations (aqueous suspensions of TZP-NC1 and TZP-pure) were prepared in an aseptic area as per the guidelines available concerning the aseptic filling method for the ophthalmic dosage forms because aseptic processing is highly regulated with considerable guidance in the US Code of Federal Regulations (CFR 21), FDA documents and the EU-GMPS "Rules and Guidance for Pharmaceutical Manufacturers and Distributors" [38,39]. Although the final products were prepared in an aseptic area, we still performed the sterilization of the products because these were intended for in vivo studies in rabbit eyes. Terminal sterilization by autoclaving in the final container is possible for the products if the stability of the drugs/products is not adversely affected by the moist heat (121 ◦C). Considering these facts, therefore, the AqS of TZP-NC<sup>1</sup> and TZP-pure were aseptically filled in the HDPE container. Before the aseptic filling, the bulk preparation TZP-NC1 was sterilized by filtration. TZP-NC1 was filtered through a 0.22 µm membrane filter into the final sterile 10 mL capacity HDPE container. Such membrane filters can remove most of the bioburden including bacteria and fungi [40]. TZP-AqS was not terminally sterilized; rather, it was prepared in the aseptic area using freshly autoclaved Milli-Q water.

The sterility testing of the sterilized TZP-NC<sup>1</sup> was performed according to the USP method [41]. Briefly, two containers of TZP-NC<sup>1</sup> were tested for sterility. TheTZP-NC1-AqS (2 mL) from the two containers of sterilized products was pooled out in the aseptic condition. The pooled samples were further diluted with 8 mL of autoclaved double distilled water. The sterile syringe filter (0.22 µm pore size) is a type of membrane filter (Corning Inc., New York, NY 14831, USA) that was fixed in a membrane-filter funnel unit. The filter was moistened with Fluid-A (1 g of peptic digest of animal tissues in 1000 mL of distilled water). The diluted pooled TZP-NC<sup>1</sup> suspension was then passed through the membrane filter in an aseptic condition. As the product contained an antimicrobial agent (TZP), the membrane was washed repeatedly (4 times) with 100 mL of sterilized Fluid-A. Thereafter, the membrane was then divided into two parts; one part was transferred to Soybean Casein Digest Media (for molds/fungi and lower bacteria) and was incubated at 20–25 ◦C for

10 days, and the other portion of the membrane was put into Fluid Thioglycollate Media (for aerobic and/or anaerobic bacteria) and was incubated at 30–35 ◦C for 10 days.

#### 2.2.3. Antimicrobial Study

The antimicrobial activity of the TZP-NCs-AqS and conventional TZP-AqS was accomplished through the agar diffusion method [26,42]. The bacterial strains for this testing were chosen from the "Global Priority Pathogens List" and are available at "Department of Pharmaceutics, College of Pharmacy, King Saud University". A total of four Gram-positive American Type Culture Collections (ATCC) of *Bacillus subtilis*, *Streptococcus pneumonia, Staphylococcus aureus* and MRSA (SA-6538) were used for their susceptibility toward TZP. The MHA plates were aseptically prepared and the chosen strains were spread out on the separate MHA-containing plates. Using a sterile borer, three wells of ~6 mm diameter were made. In the first well, 30 µL of conventional TZP-AqS (30.0 µg of TZP) was inoculated, in the second well, 30 µL of TZP-NC1-AqS (30.0 µg of TZP)was inoculated, and in the third well, the same volume of blank AqS without TZP was transferred. All the plates were left untouched for 1 h for the proper diffusion of the products into the medium and the plates were incubated at 37 ◦C for 24 h. Thereafter, the zones of inhibition created by the test products on the plates were measured. The antimicrobial assessment was accomplished in triplicate. The results are represented as the mean ± SD of the three measurements. Statistical analysis was performed using GraphPad Prism: Version 5 (GraphPad Software, Inc., San Diego, CA, USA). A oneway analysis of variance followed by Tukey's multiple comparison test was conducted by considering *p* < 0.05 as statistically significant.

#### 2.2.4. In Vivo Animal Study

New Zealand white rabbits weighing 2.0–3.0 kg were used for thein vivo studies. The protocol for animal use was approved by the Research Ethics Committee at King Saud University (approval No. KSU-SE-18-25, amended). The animals were housed in airconditioned rooms with 75 ± 5% relative humidity, as per the "Guide for the Care and Use of Laboratory Animals". All the animals were healthy (free from ocular problems). "The animals were kept on a standard pellet diet and watere *ad libitum* and "fasted overnight before starting the experiment.

#### Eye Irritation

This study was performed on healthy rabbits by following Draize's test [43]. We followed the guidelines of "The Association for Research in Vision and Ophthalmology (ARVO)" for animal use in "Ophthalmic and Vision Research". So, only the left eyes of the animals were selected for the test samples and the right eyes were left untreated. Based on the characterizations to obtainan optimized formulation, the nanocrystals (NC1) were considered for eye irritation tests as compared to conventional TZP-AqS.

Generally, six rabbits are taken for one test product; in the present investigation, we used three animals for one test product, as there might have been a chance of severe ocular irritation and damage [44]. Additionally, we had constraints with the number of animals used. Six rabbits were divided into two groups, three for NC<sup>1</sup> and three for TZP-AqS (conventional). For acute irritation, three consecutive doses (at 10 min intervals) of TZP-AqS and the suspension of NC<sup>1</sup> (40 µL) were instilled in the right eyes of each animal of the respective groups. After one hour of dosing, the eyes were visually observed periodically for 24 h for any injuries or signs and symptoms in the conjunctiva, iris and cornea or for any changes in the treated eyes other than that of the NaCl treated. The photographs of the eyes were clicked for scoring. Additionally, the level of irritation was assessed [45] based on the discomfort to the animals and signs and symptoms including redness, swelling (edema), chemosis in the conjunctiva, cornea and iris, or mucoidal/non-mucoidal discharge [28]. The scoring was performed and the irritation (if any) due to NC<sup>1</sup> was characterized as per the designated system [46,47].

#### Ocular Pharmacokinetics (PK)

The TDZ concentration in the aqueous humor (AqH) was determined to check the ocular bioavailability of the active form of the drug (active form) after the topical ocular application of the TZP (prodrug)-containing formulations in the healthy rabbits. Six rabbits were divided into two groups (one for TZP-NC<sup>1</sup> and the second for TZP-AqS). Forty microliters (40 µL, equivalent to 40 µg of TZP) of the sterilized formulations were applied to the left eyes of the rabbits of the respective groups [17,28]. Half an hour post dosing, the rabbits were desensitized with an intravenous injection of a Ketamine. HCl and Xylazine mixture [17,28,48]. Subsequently, around 40 µL of the AqH was aspirated by a 29-gauge needle attached to an insulin syringe at stipulated times. The collected samples were prepared and analyzed by liquid chromatography and the mass spectrometric (LC-MS/MS) method [36].

#### Chromatography of TDZ and Mass Spectrometric Conditions (LC-MS/MS)

The chromatographic and mass spectrometric conditions for the analysis of TDZ were previously reported by our group in detail [36]. Briefly, the "UPLC system (Acquity™) connected with a triple-quadruple Tandem Mass-Spectrometer Detector (TQD) (Waters®, Milford, MA, USA)" was used. The chromatographic separation of TDZ (active moiety) and Linezolid as the internal standard (IS) was accomplished on "Acquity™ HILIC column (2.1 <sup>×</sup> 100 mm, 1.7 <sup>µ</sup>m)", fitted with 0.22 <sup>µ</sup> of a stainless-steel fritfilter (Waters®, Milford, MA, USA). The column temperature was maintained at 40 ◦C. The mobile phase was composed of acetonitrile and 20 mM of ammonium acetate at an 85:15 (*v*/*v*) ratio and was pumped at a 0.3 mLmin−<sup>1</sup> flow rate. The injection volume was 3 µL and the total runtime was 3 min for the elution of the drugs. The Tedizolid (TDZ) and the IS were eluted with retention times (*R*t) of 1.12 and 1.32 min, respectively. The TQD fitted with the electrospray ionization interface was operated in positive mode for the detection of the two elutes. The optimal "TQD parameters were: the source temperature (150 ◦C), capillary voltage (3.7 kV), dwell time (0.161 s), desolvation temperature (350 ◦C), desolvation gas (N2) flow rate (600 L.h−<sup>1</sup> ), cone gas flow rate (50 L.h−<sup>1</sup> ) and collision gas (Argon) flow rate (0.13 mL.min−<sup>1</sup> )". The optimal MS/ MS conditions including the cone voltages were 32 V and 34 V (for TDZ and IS, respectively) whereas the collision energy was 18 eV (for both elutes). The "Multiple reactions monitoring (MRM) was used for the quantification of TDZ and IS with the parent to daughter ion transitions (*m*/*z*) of 371.15→343.17 and 338.18→296.22, respectively". "The UPLC-MS/MS system was operated by Mass-Lynx Software (V-4.1, SCN-714)" while the obtained chromatograms were processed by the "Target LynxTM program" as reported [36,49].

#### 2.2.5. Transcorneal Permeation

The transcorneal permeation of TZP from NC<sup>1</sup> across the excised rabbit cornea was performed using "fabricated double-jacketed transdermal diffusion cells assembled with the automated sampling system-SFDC 6, LOGAN, Somerset, NJ, USA"(a schematic representation of the Franz diffusion cell is shown in Figure S1, appeared in Supplementary materials) [28]. After three weeks (the washout period) of the irritation study, the same rabbits were sacrificed by an overdose intravenous injection of a Ketamine. HCl and Xylazine mixture (15 and 3 mgkg−<sup>1</sup> b. wt., respectively). The left eyes (used as the control in Draize's test) were taken out and the corneas were excised and fitted between the donor and receptor components of the diffusion cells, where the epithelial layer of the cornea was towards the donor component. The STF with SLS (0.5% *w*/*v*) was filled in the receptor component and a small magnetic bead was also put into it. The filled cells were placed on different stations of the LOGAN instrument and water (at 37 ± 1 ◦C) was allowed to flow into the outer jacket of the cells. For each formulation (in triplicate), 500 µL (0.1%, *w*/*v*) of the suspension of NC<sup>1</sup> and the drug-aqueous suspension (TZP-AqS) was put into the donor components and the instrument was switched on with magnetic stirring. Sampling was conducted from the receptor component at different time points until 4 h and the drug

(µgmL−<sup>1</sup> ) was analyzed by the HPLC-UV method as mentioned above. The permeated amount of the drug (µgcm−<sup>2</sup> ) through the cornea was calculated. The calculation was performed by considering the volume of the receptor compartment (5.2 mL), where DF stands for the dilution factor, as well as the involved corneal cross-section area (0.5024 cm<sup>2</sup> ) and the initial drug concentration (1000 µgmL−<sup>1</sup> ), using Equation(1).

$$\text{Permeated amount of drug} \left(\text{\textquotesingle gcm}^{-2}\right) = \frac{\text{Conc.} \left(\text{\textquotesingle gcm}^{-1}\right) \times \text{DF} \times \text{Volume of receptor compartment} \left(\text{mL}\right)}{\text{Area ofcome involved} \left(\text{cm}^2\right)}\tag{1}$$

The slope of the time versus permeated amount plot was applied to determine the permeation parameters (flux and apparent permeability/permeability coefficient) using the following Equations (Equations (2) and (3)):

$$\text{Steady state flux i.e., } J \left( \mu \text{gcm}^{-2}.\text{h}^{-1} \right) = \frac{dQ}{dt} \tag{2}$$

$$\text{Apparent permeability i.e., } P\_{app} \left( \text{cm.h}^{-1} \right) = \frac{f}{\mathcal{L}\_0} \tag{3}$$

where "*Q*" = amount of drug passed through the excised cornea, (*dQ/dt*) = linear ascent of the slope, "*t*" = contact time of formulation with corneal epithelial layer and "*C*0" = initial concentration of TZP.

Moreover, after finishing the transcorneal permeation studies, the used corneas were weighed, dipped into 1.0 mL of methanol, left overnight to be dried at around 75–80 ◦C and then reweighed. From the weight differences, the corneal hydration level was estimated [3,50].

#### 2.2.6. Statistical Analysis

The results are represented as the mean with standard deviation (±SD) unless otherwise indicated (as ± SEM was used for the PK parameters). Statistical analysis was performed using GraphPad Prism: Version5 (GraphPad Software, Inc., San Diego, CA, USA). A non-compartmental approach was used for the estimation of the PK parameters by "*PK*-Solver Software, Nanjing, China using MS-Excel-2013" [51]. The comparative analysis of the data was accomplished by the Student's *t*-test and *p* < 0.05 was considered statistically significant.

#### **3. Results and Discussion**

#### *3.1. Formulation and Characterization of the Optimized Formulation*

The optimized nanocrystal (TZP-NC1) was suitable for ocular use, having a size range of 154.3 ± 17.9 nm with good crystalline morphology, a good polydispersityindex (0.243 ± 0.009) and a zetapotential of +31.6 ± 3.8 mV. The smaller particle size and larger surface area of the nanocrystals helps them to cross the mucus layer of the tear film, which increases the residence time of formulation in the eye by keeping them in contact with the corneal tissues. The increased contact time with cornea may increase the absorption, which further translates into an improved bioavailability. The nanocrystals are also responsible for the increased corneal permeation, which will help in the treatment of intraocular diseases. The positive zeta potential values suggest an enhanced electrostatic interaction of the nanocrystals with the negatively charged mucin layer, which helps with increasing the residence time of the drug in the eye. The freeze drying of NC<sup>1</sup> with mannitol (1%, *w*/*v*) provided good stabilization to NC1, prevented crystal growth and provided isoosmolarity to the NC1-suspension after redispersion in dextrose (5%, *w*/*v*), where the drug content was 96.4%. The FTIR spectroscopy indicated no alteration in the basic molecular structure of TZP after nanocrystallization, and the DSC and X-ray diffraction validated the reduced crystallinity of TZP-NC. The solubility of NC<sup>1</sup> in the simulated tear fluid (STF) with sodium lauryl sulfate (SLS, 0.5%, *w*/*v*) resulted in a 1.6-fold increase as compared to the pure TZP due to its nanosizing. The redispersion of freeze-dried NC<sup>1</sup> produced

a clear transparent aqueous suspension of NC<sup>1</sup> with osmolarity (≈298 mOsm.L−<sup>1</sup> ) and viscosity (≈21.07 cps at 35 ◦C). A relatively higher (≈78.8%) release of the drug from NC<sup>1</sup> was obtained as compared to the conventional TZP-aqueous suspension (≈43.4%) at 12 h in STF with SLS (0.5%, *w*/*v*). The NC<sup>1</sup> was found to be physically (size, PDI, ZP) and chemically (drug content) stable at 4 ◦C, 25 ◦C and 40 ◦C for 6 months. The above findings encourages us that the topical ocular application of TZP-NC<sup>1</sup> in rabbits is one of the best alternatives to the conventional aqueous suspension of the poorly soluble TZP with an improved performance.

#### *3.2. Interpretation of Sterility Testing*

During the incubation, both media were visually observed every day for 10 days to see the appearance of any turbidity due to microbial growth. The media should be clear and transparent against a light source. The appearance of turbidity or cloudiness in the media is indicative of microbial growth. In the present study, no turbidity/cloudiness was found in any of the two culture media. Thus, the tested sample of TZP-NC1-AqS passed the sterility test and could be suitable for ophthalmic purposes.

#### *3.3. Antimicrobial Activity*

The results of theantimicrobial activity experiment using the "agar diffusion method" aresummarized in Table 2. The TZP-NC<sup>1</sup> showeda significant improvement (*p* < 0.05) in the antimicrobial action against all the tested Gram-positive microbes (*Bacillus subtilis*, *Streptococcus pneumonia, Staphylococcus aureus* and MRSA (SA-6538)) as compared to the conventional TZP-AqS, as illustrated in Figure 1. Relatively very little antimicrobial activity was illustrious for the blank-AqS than those of the tested TZP-containing products. Such little activity by the blank-AqS was due to the presence of some antibacterial excipients (those added in the AqS except TZP), such as the quaternary ammonium benzalkonium chloride (0.01%) which has broad-spectrum antibacterial activity and acts by interacting with the negatively charged bacterial membrane [52] and polyvinyl alcohol [53]. The antibacterial activities in the present investigation were further substantiated by the previous findings, where an improved activity of TZP-loaded chitosan nanoparticles was reported against the conventional TZP-AqS [26].


**Table 2.** Zones of inhibition attained by TZP-NC<sup>1</sup> -AqS and conventional TZP-AqS by agar diffusion test method; the blank-AqS was used as control. Data are the mean of three measurements with SD.

\* CI = Confidence interval.

The level of significance between the two TZP preparations in comparison to the blank AqS against the used microbes was performed using a one-way analysis of variance (a one-way ANOVA) followed by Tukey's multiple comparison test using GraphPad Prism V-5.0 by considering the *p* < 0.05 as statistically significant; the data obtained arerepresented (Table 1).The antimicrobial activity of TZP-NC<sup>1</sup> was enhanced as compared to the conventional TZP-AqS. These findings also pointed out that the process of nanocrystallization could not adversely affect the fundamental (antimicrobial) property as well as the structure– activity relationship (SAR) of the oxazolidinone antibiotic (TZP) in the present investigation. Both of the TZP preparations showed significantly (*p* < 0.05) increased antimicrobial activity as compared to the blank AqS (no activity).Thus, we could assume that the size reduction following the nanocrystallization increased the antimicrobial effectiveness of TZP. This might be attributed to the fact that the size reduction in the crystals could increase the aqueous solubility of the highly lipophilic drug (TZP), which increased the drug intake into the bacteria and inhibited their protein synthesis by binding to the 23S rRNA of the 50S subunit of the ribosome [35,54]. compared to the conventional TZP-AqS. These findings also pointed out that the process of nanocrystallization could not adversely affect the fundamental (antimicrobial) property as well as the structure–activity relationship (SAR) of the oxazolidinone antibiotic (TZP) in the present investigation. Both of the TZP preparations showed significantly (*p* < 0.05) increased antimicrobial activity as compared to the blank AqS (no activity).Thus, we could assume that the size reduction following the nanocrystallization increased the antimicrobial effectiveness of TZP. This might be attributed to the fact that the size reduction in the crystals could increase the aqueous solubility of the highly lipophilic drug (TZP), which increased the drug intake into the bacteria and inhibited their protein synthesis by binding to the 23S rRNA of the 50S subunit of the ribosome [35,54].

The level of significance between the two TZP preparations in comparison to the blank AqS against the used microbes was performed using a one-way analysis of variance (a one-way ANOVA) followed by Tukey's multiple comparison test using GraphPad Prism V-5.0 by considering the *p* < 0.05 as statistically significant; the data obtained arerepresented (Table 1).The antimicrobial activity of TZP-NC1 was enhanced as

*Molecules* **2022**, *27*, x FOR PEER REVIEW 8 of 17

TPZ-NC1 vs. Blank-AqS Yes 27.70 to 32.77 TZP-AqS vs. Blank-AqS Yes 16.70 to 21.77

\* CI = Confidence interval.

**Figure 1.** Antimicrobial activity of TZP-containing products as compared to the blank aqueous suspension (AqS) against *Bacillus subtilis*, *Streptococcus pneumonia*, *Staphylococcus aureus* and MRSA (SA-6538). Results are represented as mean with SD of three measurements. "a" *p* < 0.05, TZP-NC1 versus other test substances (for *B.subtilis*); "b" *p* < 0.05, TZP-NC1 versus other test substances (for *S. pneumonia*); "c" *p* < 0.05, TZP-NC1 versus other test substances (for *S.aureus*); "d" *p* < 0.05, TZP-NC1 versus other test substances (for MRSA SA-6538). **Figure 1.** Antimicrobial activity of TZP-containing products as compared to the blank aqueous suspension (AqS) against *Bacillus subtilis*, *Streptococcus pneumonia*, *Staphylococcus aureus* and MRSA (SA-6538). Results are represented as mean with SD of three measurements. "a" *p* < 0.05, TZP-NC<sup>1</sup> versus other test substances (for *B.subtilis*); "b" *p* < 0.05, TZP-NC<sup>1</sup> versus other test substances (for *S. pneumonia*); "c" *p* < 0.05, TZP-NC<sup>1</sup> versus other test substances (for *S.aureus*); "d" *p* < 0.05, TZP-NC<sup>1</sup> versus other test substances (for MRSA SA-6538).

#### *3.4. Eye Irritation 3.4. Eye Irritation*

The ocular irritation (if any) caused by the application of NC1 as compared to TZP-AqS (conventional) was investigated for 24 h by considering the NaCl-treated eyes The ocular irritation (if any) caused by the application of NC<sup>1</sup> as compared to TZP-AqS (conventional) was investigated for 24 h by considering the NaCl-treated eyes as normal. Any alterations in the cornea, conjunctiva and iris were visually observed [55]. Based on the signs and symptoms of eye irritation, the scoring for irritation was performed by following the grading and scoring systems (Table S1, appeared in Supplementary materials). The signs and symptoms included redness, swelling, hemorrhage, chemosis, cloudiness (mucoidal) and edema, etc., which could have possibly occurred in the treated eyes [46]. The type of irritation was categorized according to the ocular irritation classification [47] mentioned in Table S2 (appeared in Supplementary materials). The obtained scores during the experiment for the test samples are summarized in Table 3.


**Table 3.** Weighted scores for the eye irritation test of TZP-NC<sup>1</sup> -AqS as compared to conational TZP-AqS.

No clear signs of ocular discomfort were noted in the treated rabbits during the irritation testing of NC<sup>1</sup> as compared to TZP-AqS. Figure 2a,a' are the representative images of the normal saline (NaCl, 0.9%)-treated eyes for the TZP-AqS- and NC1-treated animals, respectively. Figure 2b,b' show the redness of the conjunctiva with mild mucoidal discharge (red arrow) after 1 h post application of TZP-AqS and NC1, respectively. Among the three rabbits treated with TZP-AqS, one showed mild redness (less intense) and mucoidal discharge even at 3 h (Figure 2c, red arrow), while the NC1-treated rabbits did not show any such abnormal ocular discharge at 3 h (Figure 2c', green arrow).The less intense redness and mucoidal discharge by the TZP-AqS-treated rabbits at 3 h was probably due to the larger size of the suspended particles and PVA (which was added as a suspending agent in AqS), which caused some corneal abrasion. Hence, to overcome such unwanted phenomena, the ocular physiological secretions (mucoidal discharge) occurred and such secretions remained until 6 h (Figure 2d, black arrow), while no such signs and symptoms were noted at 6 h in the eyes of the NC1-treated rabbits. The redness in the treated eye was much reduced or almost recovered and clear, as denoted by the green arrow (Figure 2d'). The redness of the conjunctiva and ocular inflammation completely disappeared from their normal state (green arrows) at 24 h post topical application of TZP-AqS and NC1, as illustrated in Figure 2e and 2e', respectively. The redness of the conjunctiva and eye inflammation was gone and the eyes regained their normal conditions after 24 h postapplication of the test products. The disappearance of such symptoms was due to the natural defense system of the eyes and the use of the Generally Recognized as Safe (GRAS) excipients in the formulations [22,26].

The ocular application of TZP-AqS caused minimal irritation in one rabbit with redness of the eye and mucoidal discharge, which was given a score of one. No corneal lesions or opacity were observed; hence, the cornea, conjunctiva and iris were given a score of 0 (Table 3). A reported classification system for irritation scoring [47] was followed to calculate the maximum mean total score (MMTS). The MMTS for TZP-AqS and TZP- NC1- AqS after 24 h of their application was 8.00 (>2.6 and <15.1, minimally) and 1.33 (>0.6 and <2.6, practically none), as mentioned in Table 4. Thus, the conventional TZP-AqS was "minimally irritating", while the TZP-NC1-AqS was "practically non-irritating" to the rabbit eyes; thus, there is hope for its ocular application. All the involved animals remained healthy and active without any odd signs of ocular irritation during the experiment, except

a few as stated above. Thus, we concluded that the conventional TZP-AqS, as well as the developed TZP-NC1-AqS, were well tolerated by the rabbit eyes. toms was due to the natural defense system of the eyes and the use of the Generally Recognized as Safe (GRAS) excipients in the formulations [22,26].

as normal. Any alterations in the cornea, conjunctiva and iris were visually observed [55]. Based on the signs and symptoms of eye irritation, the scoring for irritation was performed by following the grading and scoring systems (Table S1, appeared in Supplementary materials). The signs and symptoms included redness, swelling, hemorrhage, chemosis, cloudiness (mucoidal) and edema, etc., which could have possibly occurred in the treated eyes [46]. The type of irritation was categorized according to the ocular irritation classification [47] mentioned in Table S2 (appeared in Supplementary materials). The obtained scores during the experiment for the test samples are summarized in Table 3.

No clear signs of ocular discomfort were noted in the treated rabbits during the irritation testing of NC1 as compared to TZP-AqS. Figure 2a,a' are the representative images of the normal saline (NaCl, 0.9%)-treated eyes for the TZP-AqS- and NC1-treated animals, respectively. Figure 2b,b' show the redness of the conjunctiva with mild mucoidal discharge (red arrow) after 1 h post application of TZP-AqS and NC1, respectively. Among the three rabbits treated with TZP-AqS, one showed mild redness (less intense) and mucoidal discharge even at 3 h (Figure 2c, red arrow), while the NC1-treated rabbits did not show any such abnormal ocular discharge at 3 h (Figure 2c', green arrow).The less intense redness and mucoidal discharge by the TZP-AqS-treated rabbits at 3 h was probably due to the larger size of the suspended particles and PVA (which was added as a suspending agent in AqS), which caused some corneal abrasion. Hence, to overcome such unwanted phenomena, the ocular physiological secretions (mucoidal discharge) occurred and such secretions remained until 6 h (Figure 2d, black arrow), while no such signs and symptoms were noted at 6 h in the eyes of the NC1-treated rabbits. The redness in the treated eye was much reduced or almost recovered and clear, as denoted by the green arrow (Figure 2d'). The redness of the conjunctiva and ocular inflammation completely disappeared from their normal state (green arrows) at 24 h post topical application of TZP-AqS and NC1, as illustrated in Figure 2e and 2e', respectively. The redness of the conjunctiva and eye inflammation was gone and the eyes regained their normal conditions after 24 h postapplication of the test products. The disappearance of such symp-

*Molecules* **2022**, *27*, x FOR PEER REVIEW 9 of 17

**Figure 2.** Eye images captured during irritation study. Representative images of 0.9% NaCl-treated eyes (**a**) and (**a'**). Post topical application of conventional TZP-AqS at 1 h (**b**) (red arrow); at 3 h (**c**) (red arrow); at 6 h (**d**) (black arrow); and at 24 h (**e**) (green arrow). Post application of suspension of NC1 at 1 h (**b'**) (red arrow); at 3 h (**c'**) (green arrow); at 6 h (**d'**) (green arrow); and at 24 h (**e'**) (green arrow).Images are not showing any abnormal watery discharge or intense redness, indicating the normal features of rabbit eyes, represented by green arrows. **Figure 2.** Eye images captured during irritation study. Representative images of 0.9% NaCl-treated eyes (**a**) and (**a'**). Post topical application of conventional TZP-AqS at 1 h (**b**) (red arrow); at 3 h (**c**) (red arrow); at 6 h (**d**) (black arrow); and at 24 h (**e**) (green arrow). Post application of suspension of NC<sup>1</sup> at 1 h (**b'**) (red arrow); at 3 h (**c'**) (green arrow); at 6 h (**d'**) (green arrow); and at 24 h (**e'**) (green arrow).Images are not showing any abnormal watery discharge or intense redness, indicating the normal features of rabbit eyes, represented by green arrows.

**Table 4.** Calculation of maximum mean total score (MMTS) by considering the obtained scores.


#### *3.5. Ocular Pharmacokinetics*

The previously developed LC-MS/MS method by our group was effectively used for the quantification of TDZ in the aqueous humor (AqH) obtained from the rabbit eyes [36]. The level of TDZ in the AqH versus time plots and the calculated pharmacokinetic parameters for the two TZP formulations are, respectively, illustrated in Figure 3 and Table 5. After the topical application of the two formulations, a fast release of the drug (TZP) was found from NC1-AqS as compared to the conventional TZP-AqS, indicating a faster absorption of TDZ and an attained a maximum concentration (Cmax) of 829.21 <sup>±</sup> 38.27 ngmL−<sup>1</sup> and 580.92 <sup>±</sup> 45.48 ngmL−<sup>1</sup> , respectively, at 2 h of Tmax. Thereafter, the concentration of the active form of the drug in the AqH decreased in a log-linear fashion, with the average elimination half-lives of 4.45 h and 2.66 h for NC<sup>1</sup> and TZP-AqS, respectively, just after the

second sampling at 2 h, which was indicative of the fast absorption (up to 2 h) of TDZ from NC<sup>1</sup> as compared to its counter formulation.

tion (up to 2 h) of TDZ from NC1 as compared to its counter formulation.

Table 5. After the topical application of the two formulations, a fast release of the drug (TZP) was found from NC1-AqS as compared to the conventional TZP-AqS, indicating a faster absorption of TDZ and an attained a maximum concentration (Cmax) of 829.21 ± 38.27 ngmL−1 and 580.92 ± 45.48 ngmL−1, respectively, at 2 h of Tmax. Thereafter, the concentration of the active form of the drug in the AqH decreased in a log-linear fashion, with the average elimination half-lives of 4.45 h and 2.66 h for NC1 and TZP-AqS, respectively, just after the second sampling at 2 h, which was indicative of the fast absorp-

*Molecules* **2022**, *27*, x FOR PEER REVIEW 11 of 17

**Figure 3.** The drug concentration–time profile of AqH samples after topical application of conventional TZP−AqS and TZP-NC1−AqS in rabbit eyes. Results are the mean of three measurements (three animals per group) with SEM. # (*p* < 0.05) represents the significant difference between NC1 as compared to conventional AqS. **Figure 3.** The drug concentration–time profile of AqH samples after topical application of conventional TZP−AqS and TZP-NC1−AqS in rabbit eyes. Results are the mean of three measurements (three animals per group) with SEM. # (*p* < 0.05) represents the significant difference between NC<sup>1</sup> as compared to conventional AqS.

**Table 5.** Ocular pharmacokinetics of TZP-containing formulations. The data are represented as mean with ± SEM of three readings, where # (*p* < 0.05) represents the significant difference between NC1as compared to conventional AqS. **Table 5.** Ocular pharmacokinetics of TZP-containing formulations. The data are represented as mean with <sup>±</sup> SEM of three readings, where # (*p* < 0.05) represents the significant difference between NC<sup>1</sup> as compared to conventional AqS.


Other than Tmax, the differences were statistically significant (*p* < 0.05) in the rest of the pharmacokinetic parameters for the two formulations. Around a 1.67- and 1.43-fold increase in t1/2 (h) and Cmax (ngmL−1), respectively, was found from NC1 as compared to the pure drug suspension (TZP-AqS). Approximately 1.96-, 1.91-, 2.69- and 1.41-times increases in AUC0–24h, AUC0–∞, AUMC0–∞ and MRT0–∞, respectively, were obtained for the Other than Tmax, the differences were statistically significant (*p* < 0.05) in the rest of the pharmacokinetic parameters for the two formulations. Around a 1.67- and 1.43-fold increase in t1/2 (h) and Cmax (ngmL−<sup>1</sup> ), respectively, was found from NC<sup>1</sup> as compared to the pure drug suspension (TZP-AqS). Approximately 1.96-, 1.91-, 2.69- and 1.41-times increases in AUC0–24h, AUC0–∞, AUMC0–<sup>∞</sup> and MRT0–∞, respectively, were obtained for the active form of TZP from NC<sup>1</sup> than that of TZP-AqS. The clearance (CL/F) of TDZ was faster from TZP-AqS (11.43 mLh−<sup>1</sup> ) than that of NC<sup>1</sup> (5.88 mLh−<sup>1</sup> ). The faster clearance of the drug from TZP-AqS could be the primary reason for the relatively low ocular bioavailability, which was further justified by the extended half-life (t1/2; 4.45 h) of TDZ and the prolonged ocular retention (MRT0–∞; 7.13 h) of NC<sup>1</sup> compared to the shorter half-life (t1/2; 2.66) of TDZ and MRT0–<sup>∞</sup> and of TZP-AqS (t1/2; 5.05 h), which was further confirmed by the fast elimination of TDZ as the drug concentration was not detectable at 24 h from AqS.

Cl/F (mLh−1) 11.43 ± 1.25# 5.88 ± 0.22 1.95

Overall, the comparative pharmacokinetic profiling illustrated an improved ocular bioavailability of TDZ from NC<sup>1</sup> as compared to TZP-AqS. This might be due to the high positive ZP (+29.4 mV) of NC<sup>1</sup> (due to the presence of Benzalkonium chloride and stearylamine), which could interact electrostatically with the negatively charged mucin layer on ocular surfaces and increase the contact time of NC1. This interaction could help improve the penetration of NC<sup>1</sup> across the cornea, which in turn could improve its cellular uptake and hence the bioavailability of TDZ [6,26,36]. Additionally, the nanosize range of NC<sup>1</sup> could also be a reason for its increased transcorneal permeation. Similarly, the significantly increased ocular bioavailability of hydrocortisone and some other poorly soluble glucocorticoids were reported from the nanosuspension more than their microrange formulations [56]. Conclusively, the nanocrystallization of TZP could have the potential to enhance the ocular bioavailability of TDZ at a relatively low dose and could have the reduced dosing frequency of the TZP-NC1-AqS as an ophthalmic formulation.

#### *3.6. Transcorneal Permeation*

The cumulative amounts of the drug that were permeated (µgcm−<sup>2</sup> ) were plotted against time (h), as shown in Figure 4, and from these plots the permeation parameters were calculated and summarized in Table 6. The TZP-NC1-AqS showed a linear permeation of TZP as compared to the conventional TZP-AqS up to 4 h. Overall, the cumulative amounts of the permeated drug were 44.32 <sup>±</sup> 1.74 and 70.43 <sup>±</sup> 3.52 <sup>µ</sup>gcm−<sup>2</sup> (at 4 h) from the conventional TZP-AqSand TZP-NC1-AqS, respectively. *Molecules* **2022**, *27*, x FOR PEER REVIEW 13 of 17

**Figure 4.** Transcorneal permeation of TZP from conventional TZP−AqSand TZP−NC1-AqS (mean ± SD, *n* = 3). \* (*p* < 0.05) represents the significant difference between NC1as compared to conventional AqS. **Figure 4.** Transcorneal permeation of TZP from conventional TZP−AqSand TZP−NC<sup>1</sup> -AqS (mean ± SD, *n* = 3). \* (*p* < 0.05) represents the significant difference between NC<sup>1</sup> as compared to conventional AqS.

The corneal hydration levels were found to be 77.29 ± 1.23% and 78.05 ± 1.27% for the conventional TZP−AqS- and TZP−NCs−AqS-treated excised rabbit corneas, respec-**Table 6.** Parameters of transcorneal permeation from conventional TZP−AqS and TZP−NC1−AqS (mean ± SD, *n* = 3).

**Parameters TZP-AqS (Conventional) TZP-NC1-AqS** 

pH 6.18 ± 0.46 7.03 ± 0.35

(μgcm−2) at 4th h 44.32 ± 1.74 70.43 ± 3.52

Steady-state flux, *J* (μgcm−2h−1) 19.18 ± 1.03 31.65 ± 2.39

Corneal hydration level (%) 77.29 ± 1.23 78.05 ± 1.27

Around a 1.29 to 1.53-fold increase in antibacterial activity was noted against *B. subtilis*, *S. pneumonia, S. aureus* and MRSA (SA-6538) as compared to the pure TZP. The ocular irritation study indicated that the conventional TZP−AqS was "minimally irritating" and NC1-AqS was "practically non-irritating" to the rabbit eyes; thus, there is hope for its ocular application. Around a 1.67- and 1.43-fold increase in t1/2 (h) and Cmax (ngmL−1) occurred, while 1.96-, 1.91-, 2.69- and 1.41-times increases in AUC0–24h, AUC0–∞, AUMC0–∞ and MRT0–∞, respectively, were found for TDZ (active of TZP) by NC1 as com-


Cumulative amount of drug permeated

**4. Conclusions** 

The transcorneal permeation of TZP was higher from AqS-NC<sup>1</sup> throughout the experiment as compared to its counter formulation. In the case of NC1, around 34.94 <sup>±</sup> 3.58 <sup>µ</sup>gcm−<sup>2</sup> of TZP was traversed at 1.5 h, while roughly around same amount of the drug (35.92 <sup>±</sup> 1.94 <sup>µ</sup>gcm−<sup>2</sup> ) was passed through the cornea at 2 h from the conventional TZP-AqS. Around 42.1 µgcm−<sup>2</sup> of the drug was crossed at 2.5 h from the TZP-AqS; after that, the amount of crossed drug through the cornea was not increased significantly (*p* < 0.05) and it was comparable (44.32 µgcm−<sup>2</sup> ) until 4 h. However, an increased permeation of TZP (51.15 µgcm−<sup>2</sup> ) was noted from the NC<sup>1</sup> at 2.5 h and there was a linear progression until 4 h (70.42 µgcm−<sup>2</sup> ). Overall, the permeated amount of drug from the NC<sup>1</sup> was significantly (*p* < 0.05) increased as compared to the TZP-AqS. The enhanced permeation of TZP from the NC<sup>1</sup> form indicated that the nanocrystallization of the drug improved its solubility in the aqueous media. The reduced size of the TZP-NC<sup>1</sup> could easily cross the cornea as compared to the particle size (574.5 nm) of TZP-AqS. The pH of the formulations (Table 6) were appropriate for the transcorneal permeation of TZP. Additionally, the partitioning of a neutral drug species between *n*-octanol/water (LogP) of TZP is around 4.89 at a neutral pH (7.0). Therefore, it was assumed that, as the pH of the TZP-NC1-AqS was closer to the neutral pH or the pH of the tear fluids, a larger fraction of the TZP remained unionized in the NCs of TZP as compared to its conventional AqS [57], which might be attributed to the enhanced transcorneal permeation of TZP from NC1. Conclusively, from the permeation profiles, the NC<sup>1</sup> form of TZP could offer a linearly increased permeation of the drug, which could mimic the improved ocular bioavailability of the drug during in vivo application of TZP-NC1-AqS as compared to its counter formulation (conventional TZP-AqS).

The corneal hydration levels were found to be 77.29 ± 1.23% and 78.05 ± 1.27% for the conventional TZP−AqS- and TZP−NCs−AqS-treated excised rabbit corneas, respectively. The obtained values were in the range of a normal hydration level (between 75% to 80%) [58] at pH 6.18 (for conventional TZP-AqS) and 7.03 (TZP-NCs−AqS) as summarized in Table 6. The corneal hydration levels as mentioned in Table 6 were below 80%; therefore, the damage that appeared during the ex vivo transcorneal permeation experiment to the corneal was considered reversible and non-damaging [58].

#### **4. Conclusions**

Around a 1.29 to 1.53-fold increase in antibacterial activity was noted against *B. subtilis*, *S. pneumonia, S. aureus* and MRSA (SA-6538) as compared to the pure TZP. The ocular irritation study indicated that the conventional TZP−AqS was "minimally irritating" and NC1-AqS was "practically non-irritating" to the rabbit eyes; thus, there is hope for its ocular application. Around a 1.67- and 1.43-fold increase in t1/2 (h) and Cmax (ngmL−<sup>1</sup> ) occurred, while 1.96-, 1.91-, 2.69- and 1.41-times increases in AUC0–24h, AUC0–∞, AUMC0–<sup>∞</sup> and MRT0–∞, respectively, were found for TDZ (active of TZP) by NC<sup>1</sup> as compared to TZP−AqS. The clearance of TDZ was slower (5.88 mLh−<sup>1</sup> ) from NC<sup>1</sup> as compared to TZP−AqS (11.43 mLh−<sup>1</sup> ).This was further substantiated by the extended half-life (t1/2; 4.45 h) of TDZ and the prolonged ocular retention (MRT0–∞; 7.13 h) of NC<sup>1</sup> as compared to the shorter half-life (t1/2; 2.66) of TDZ and MRT0–∞, as well as of TZP−AqS(t1/2; 5.05 h), and due to fast elimination rate of the conventional AqS, the concentration of TDZ was not detected in the 24 h AqH samples. The cationic TZP−NC<sup>1</sup> could offer an increased transcorneal permeation of the drug, which could mimic the improved ocular bioavailability of the drug in vivo. Summarily, the cationic NC<sup>1</sup> of TZP is a promising alternative for the ocular delivery of TZP, with an amplified performance comparatively to the conventional TZP−AqS. Further, in vivo studies (rabbit uveitis models) are warned to check the antiinflammatory activity of the drug during bacterial eye infections, including the different ocular anterior and posterior segment inflammatory conditions and some retinal ailments following the topical application of the developed TZP−NC1.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27144619/s1, Table S1: Grading system for ocular irritation test; Table S2: Classification of eye irritation scoring system; Figure S1: Schematic representation of Franz Diffusion cell.

**Author Contributions:** Conceptualization, M.A.K., R.A. and M.A.; methodology, M.A.K., M.R; software, M.I.; validation, M.A.K., R.A., A.A. (Abdullah Alshememry), A.A. (Aws Alshamsan) and M.I.; formal analysis, M.A.K. and M.R.; investigation, M.I.; resources, M.A.K. and M.A.; data curation, M.A.K. and M.I.; writing-original draft preparation, M.A.K. and S.S.A.; writing-review and editing, S.S.A., A.A. (Aws Alshamsan) and A.A. (Abdullah Alshememry); visualization, M.A.K.; supervision, A.A. (Aws Alshamsan); project administration, M.A.K., A.A. (Aws Alshamsan) and M.A.; funding acquisition, M.A.K., M.A. and A.A. (Aws Alshamsan). All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Researchers Supporting Project (RSP-2022-R490), King Saud University, Riyadh, Saudi Arabia.

**Institutional Review Board Statement:** The protocol for animal use was approved by the Research Ethics Committee at King Saud University (approval No. KSU-SE-18-25, amended).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The authors are thankful to the Researchers Supporting Project (RSP-2022-R490), King Saud University, Riyadh, Saudi Arabia.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


AqS Aqueous Suspension

#### **References**


## *Article* **Erlotinib-Loaded Dendrimer Nanocomposites as a Targeted Lung Cancer Chemotherapy**

**Wafa K. Fatani <sup>1</sup> , Fadilah S. Aleanizy <sup>1</sup> , Fulwah Y. Alqahtani <sup>1</sup> , Mohammed M. Alanazi <sup>2</sup> , Abdullah A. Aldossari <sup>2</sup> , Faiyaz Shakeel <sup>1</sup> , Nazrul Haq <sup>1</sup> , Hosam Abdelhady <sup>3</sup> , Hamad M. Alkahtani <sup>4</sup> and Ibrahim A. Alsarra 1,\***


**Abstract:** Lung cancer is the main cause of cancer-related mortality globally. Erlotinib is a tyrosine kinase inhibitor, affecting both cancerous cell proliferation and survival. The emergence of oncological nanotechnology has provided a novel drug delivery system for erlotinib. The aims of this current investigation were to formulate two different polyamidoamine (PAMAM) dendrimer generations—generation 4 (G4) and generation 5 (G5) PAMAM dendrimer—to study the impact of two different PAMAM dendrimer formulations on entrapment by drug loading and encapsulation efficiency tests; to assess various characterizations, including particle size distribution, polydispersity index, and zeta potential; and to evaluate in vitro drug release along with assessing in situ human lung adenocarcinoma cell culture. The results showed that the average particle size of G4 and G5 nanocomposites were 200 nm and 224.8 nm, with polydispersity index values of 0.05 and 0.300, zeta potential values of 11.54 and 4.26 mV of G4 and G5 PAMAM dendrimer, respectively. Comparative in situ study showed that cationic G4 erlotinib-loaded dendrimer was more selective and had higher antiproliferation activity against A549 lung cells compared to neutral G5 erlotinib-loaded dendrimers and erlotinib alone. These conclusions highlight the potential effect of cationic G4 dendrimer as a targeting-sustained-release carrier for erlotinib.

**Keywords:** polyamidoamine dendrimers; erlotinib; non-small-cell lung cancer; cytotoxicity

## **1. Introduction**

Cancer is the second largest cause of death worldwide. Among different types of cancer, lung cancer is considered the most common source of cancer-related death [1]. Among all lung cancers, non-small cell lung cancer (NSCLC) is responsible for 90% of cases, where the majority are found to be in a late stage when diagnosed [2]. Epidermal growth factor receptor (EGFR) is strongly expressed in NSCLC. Therefore, extensive efforts have been focused on the involvement of tyrosine kinase inhibitors of EGFR in the treatment of NSCLC [3,4]. Erlotinib (ERL), the EGFR tyrosine kinase inhibitor (TKI), is a quinazoline derivative targeted antiplatelet drug. It was approved in 2004 as a second-line treatment of NSCLC irrespective of EGFR genotype and a first-line treatment of activated mutations in EGFR [5]. ERL exhibits its antitumor activity by selectively occupying the adenosine triphosphate (ATP)-binding sites of EGFR, thus reversibly hindering EGFR activation. This leads to it suppressing the downstream signaling pathways, mainly by the two pathways of phosphatidylinositol 3-kinase (PI3K) and mitogen activated protein kinase (MAPK), and thus inhibiting cancer cell proliferation, angiogenesis, and metastases [6,7].

**Citation:** Fatani, W.K.; Aleanizy, F.S.; Alqahtani, F.Y.; Alanazi, M.M.; Aldossari, A.A.; Shakeel, F.; Haq, N.; Abdelhady, H.; Alkahtani, H.M.; Alsarra, I.A. Erlotinib-Loaded Dendrimer Nanocomposites as a Targeted Lung Cancer Chemotherapy. *Molecules* **2023**, *28*, 3974. https:// doi.org/10.3390/molecules28093974

Academic Editor: Artur J. M. Valente

Received: 20 February 2023 Revised: 4 May 2023 Accepted: 4 May 2023 Published: 8 May 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The aqueous solubility of ERL hydrochloride depends on pH, with an increase in solubility at a pH that is lower than 5, where maximal solubility of approximately 0.4 mg/mL arises at a pH of 2 [8]. Therefore, any alteration in the pH of the upper gastrointestinal (GI) tract may affect ERL solubility [9]. Since ERL has low aqueous solubility and high permeability (logP of 2.7), ERL is classified as class II in the biopharmaceutical classification (BCS) system [8]. In fact, many of these limitations are associated with ERL especially as the marketed ERL is administered orally as a film-coated tablet (Tarceva®, Astellas Pharm Global Development, Inc., Tokyo, Japan). These limitations include side effects, pH-dependent solubility, low permeability, drug interaction, and drug resistance. ERL's poor solubility constrains its dissolution in the GI fluid, which restricts its absorption, thus resulting in a rise in its peak plasma concentration (Cmax) and area under the drug concentration–time curve (AUC), causing large inter-patient variability [10]. Moreover, food has been found to raise the bioavailability of the oral dosage form upon the administration of a 150 mg tablet for a level up to 100% from 60% in healthy volunteers, taken without food [11]. Additionally, studies showed that ERL was metabolized by the human liver primarily by CYP3A4, with a percentage of 70%, along with a secondary contribution from CYP1A2 of approximately 20%. For that reason, various factors that affect these hepatic enzymes may have an impact on ERL plasma concentration level [12,13].

Genomic variability between some ethnicities, drugs which either induce or reduce the enzymes, and smoking cigarettes are all factors that affect CYP3A4 and CYP1A2 levels and thus ERL plasma concentration levels [9]. Although ERL is similar to other chemotherapies in effectiveness, receivers of this medication were still experiencing ERL resistance after one year of treatment [14]. In addition, severe toxicities may appear, such as rashes, diarrhea, GI perforations, and Stevens–Johnson syndrome, which hamper its clinical application [15]. The pharmacokinetic parameters of the marketed tablet, Tarceva®, include a Tmax of four hours and an elimination half-life of 36.2 h. All these limitations necessitate the willingness to overcome these obstacles by using proper nano delivery systems to enhance ERL efficacy and decrease undesirable adverse events and resistance. The proposed formulation can be assessed for future administration as an inhaled drug delivery system. This allows targeted drug delivery with minimized systemic side effects.

Dendrimers are nanosized composite materials and known as dendrimer nanocomposites. To this end, polyamidoamine (PAMAM) dendrimers have emerged as promising, well-defined nanocarriers for targeted anticancer drug delivery; they have exceptional mono-dispersibility, nano-size, biocompatibility, the ability to enhance drug solubility, along with permeability, and retention. These dendrimers originate from an ethylene di-amine core; different dendritic branches are then inserted by exhaustive Michael addition, depending on the chosen number of generation (G0–G10) [16]. Dendrimers possess a unique molecular architecture consisting of a central core, dendrons, and peripheral functional groups, located on the outer surface of the macromolecule, that control DNA formation, drug encapsulation efficacy, and cellular targeting. Cellular ligands such as folic acid, hyaluronic acid, transferrin, peptides, and antibodies have been extensively applied for the development of tumor-selective drug delivery systems [16,17]. The dendrimer architecture has three main mechanisms for drug loading, namely, molecular entrapment within (i) the dendritic nanocapsules; (ii) branching points (making hydrogen bonds between dendrons and drug molecules); (iii) and external surface groups by electrostatic interactions [18,19]. The binding strength (H-8 is 100% and H-2 is 70%) of the encapsulated 6-mercaptopurine (6-MP) within the nanocavities of the amine-terminated dendrimers was quantitatively expressed using epitope maps [20]. Furthermore, the binding constants (logKa 3.85–4.74) of internal nanocavities of dendrimers and catecholamines were measured by UV–Vis, fluorescence, and 1D and 2D NMR spectroscopy [21]. A study was conducted by including the anticancer drug tetra methyl scutellarein entrapped inside a PAMAM fourth generation (G4) dendrimer [22]. The results showed that the blank PAMAM G4 dendrimer enhanced the water solubility of the medication. In addition, enhancement in encapsulation efficiency with a percentage 77.8 ± 0.69% and drug loading

rates with a percentage 6.2 ± 0.06% were observed [22]. Another study showed that a multifunctional drug delivery system was developed to co-deliver ERL in combination with gene-recombinant short hairpin RNA-expressing plasmids (Survivin-shRNA) in modified PAMAM with chloroquine [23]. Chloroquine was added to enhance the endosomal escape capability of the AP/ES gene for effective gene transfection to obstruct survivin, where downregulation of survivin may overcome epidermal growth factor receptor-tyrosine kinase inhibitor resistance (EGFR-TKI). Additionally, it exhibited high vessel-normalization, which increases tumor microcirculation, which then promotes drug delivery to cancerous cells and improves ERL efficacy, especially in ERL-resistant cancer cells [24]. Other supramolecular drug delivery systems have also been investigated for cancer targeting and other therapeutic applications [25–28]. The aim of this work is to design non-invasive erlotinib nanocomposites using polyamidoamine dendrimers for the treatment of non-small cell lung cancer.

#### **2. Results**

#### *2.1. HPLC Analysis of ERL*

The proposed HPLC method for ERL analysis was linear at the range of 1–100 µg/mL. The linear regression equation was predicted as y = 50,678x + 4472, in which x is the concentration of ERL and y represents the measured peak area for ERL. The determination coefficient (R<sup>2</sup> ) for ERL was found to be 0.991. The accuracy of the proposed HPLC method for ERL analysis was estimated as 98.21–101.43%. The intra-day and intermediate precision values for the anticipated HPLC technique were estimated as 0.68–1.23 and 0.74–1.42%, respectively. The relative standard deviation (% RSD) for robustness analysis was found to be 0.62–0.97%. The limit of detection (LOD) and limit of quantification (LOQ) values were determined to be 0.38 µg/mL and 1.14 µg/mL, respectively. The previous findings suggested that the proposed HPLC method for the determination of ERL analysis was linear, accurate, precise, robust, and sensitive.

#### *2.2. Formulation, Loading and Characterization of ERL in PAMAM Dendrimers*

ERL was inserted individually into G4-FITC and G5-FITC PAMAM dendrimers, after which it was lyophilized. Data from previous reports have been considered to choose the most fit molar ratio of ERL to G4-FITC and G5-FITC PAMAM dendrimers. The chosen molar ratio of ERL to dendrimer was 25:1 [29]. Transmission Electron Microscope (TEM) imaging was utilized for morphological characterization of the blank and conjugated dendrimers. TEM images for different dendrimers at higher magnifications are presented in Figure 1. The TEM images for different dendrimers were also recorded at low magnifications and results are shown in Figure S1. The calculated percentage amount of encapsulation efficiency (EE) was 99.96% and 99.92% for formulations G4-FITC and G5-FITC, respectively. However, the percentage amount of drug-loading capacity (DLC) was 14.05% and 7.52% for formulations G4-FITC and G5-FITC, respectively. Figure 2 illustrates the in vitro release of ERL from G4-FITC and G5-FITC ERL-loaded dendrimers in addition to 150 mg ERL tablets. The two generations of PAMAM dendrimer formulations revealed sustained release patterns up to seven days while the tablet was released after 10 min.

The physicochemical characteristics of blank and ERL-loaded PAMAM dendrimer G4-FITC and G5-FITC, such as particle size and polydispersity index (PDI), were measured (Table 1). The particle size distribution by intensity for blank and ERL-loaded G4-FITC and G5-FITC dendrimers are presented in Figure S2. Additionally, the exterior charge found on the surface of the dendrimers was estimated at various stages and environmental conditions. The zeta potential values for blank and ERL-loaded G4-FITC and G5-FITC dendrimers at pH 5.4 and 7.4 are presented in Figure S3 and Figure S4, respectively. The cationic G4-FITC plain dendrimer demonstrated a positive zeta potential in phosphate buffer solution (PBS) pH 5.4 and pH 7.4 (8.9 ± 1.1 and 10.6 ± 0.1 mV), respectively. These values were expected to be from the cationic surface groups on their surface due to the positive net charge. After loading with ERL, the zeta potential increased to 15.7 ± 1.34

and 18.2 ± 1.9 mV. Meanwhile, the neutral G5-FITC plain dendrimer demonstrated a negative zeta potential in PBS pH 5.4 and pH 7.4 (−5.66± 0.76 and −10.2 ± 0.21 mV). These values were expected to be from the anionic surface groups on their surface due to the negative net charge. After loading with ERL, the zeta potential value increased to −3.82 ± 0.66 and −5.78 ± 0.7 mV, respectively. In previous studies, the zeta potential of plain G5 PAMAM dendrimers was also recorded as a negative value [29]. After loading a weak base drug, such as ruboxistaurin, the zeta potential value of G5 dendrimers was increased [29]. Another study also indicated the increased zeta potential of G5 dendrimers after loading a weak base drug, namely, vardenafil [30]. In the present study, the selected drug ERL is a weak base. The zeta potential of G5-FITC dendrimers was also increased after drug loading in the present study. Therefore, the obtained results of plain and drug-loaded FITC-G5-dendrimers were in accordance with those reported in the literature [29,30]. The increase in the value of zeta potential supports effective loading of ERL within the two dendrimer generations. *Molecules* **2023**, *28*, x FOR PEER REVIEW 4 of 15

**Figure 1.** TEM images at higher magnifications to inspect the morphology of (**a**) blank Generation 4-FITC (G4-FITC) polyamidoamine (PAMAM) dendrimers; (**b**) erlotinib conjugated G4-FITC PA-MAM dendrimers proving erlotinib placement within the dendrimer; (**c**) blank generation (G5- FITC) PAMAM dendrimers; and (**d**) erlotinib conjugated G5-FITC PAMAM dendrimers proving encapsulation of erlotinib within the dendrimer cavity. **Figure 1.** TEM images at higher magnifications to inspect the morphology of (**a**) blank Generation 4-FITC (G4-FITC) polyamidoamine (PAMAM) dendrimers; (**b**) erlotinib conjugated G4-FITC PAMAM dendrimers proving erlotinib placement within the dendrimer; (**c**) blank generation (G5-FITC) PA-MAM dendrimers; and (**d**) erlotinib conjugated G5-FITC PAMAM dendrimers proving encapsulation of erlotinib within the dendrimer cavity.

**Figure 2.** In vitro release of erlotinib, assessed as a percentage of cumulative release from dendrimers G4-FITC and G5-FITC, showing comparable prolonged release patterns. (**a**) Comparative in vitro release study at pH 5.4 for PAMAM dendrimer G4-FITC and G5-FITC; (**b**) comparative in vitro release study at pH 7.4 for PAMAM dendrimer G4-FITC and G5-FITC; and (**c**) comparison in vitro release study for ERL tablet between two media: 0.01 N HCl and 0.02% Tween 80 in 0.01 N HCl. **Figure 2.** In vitro release of erlotinib, assessed as a percentage of cumulative release from dendrimers G4-FITC and G5-FITC, showing comparable prolonged release patterns. (**a**) Comparative in vitro release study at pH 5.4 for PAMAM dendrimer G4-FITC and G5-FITC; (**b**) comparative in vitro release study at pH 7.4 for PAMAM dendrimer G4-FITC and G5-FITC; and (**c**) comparison in vitro release study for ERL tablet between two media: 0.01 N HCl and 0.02% Tween 80 in 0.01 N HCl.

The physicochemical characteristics of blank and ERL-loaded PAMAM dendrimer G4-FITC and G5-FITC, such as particle size and polydispersity index (PDI), were measured (Table 1). The particle size distribution by intensity for blank and ERL-loaded G4- FITC and G5-FITC dendrimers are presented in Figure S2. Additionally, the exterior **Table 1.** Physicochemical characteristics of blank and ERL-loaded PAMAM dendrimer G4-FITC and G5-FITC. The particle size and polydispersity index (PDI) were measured. Data represent averages ± SD, (*n* = 3).


and 18.2 ± 1.9 mV. Meanwhile, the neutral G5-FITC plain dendrimer demonstrated a neg-

#### ative zeta potential in PBS pH 5.4 and pH 7.4 (−5.66± 0.76 and −10.2 ± 0.21 mV). These *2.3. In Vitro Cytotoxicity*

values were expected to be from the anionic surface groups on their surface due to the negative net charge. After loading with ERL, the zeta potential value increased to −3.82 ± 0.66 and −5.78 ± 0.7 mV, respectively. In previous studies, the zeta potential of plain G5 PAMAM dendrimers was also recorded as a negative value [29]. Afterloading a weak base drug, such as ruboxistaurin, the zeta potential value of G5 dendrimers was increased [29]. Another study also indicated the increased zeta potential of G5 dendrimers after loading a weak base drug, namely, vardenafil [30]. In the present study, the selected drug ERL is a weak base. The zeta potential of G5-FITC dendrimers was also increased after drug loading in the present study. Therefore, the obtained results of plain and drug-loaded FITC-MTT results have shown that ERL has decreased A549 cells viability significantly with all the doses (10, 20, 40, 80 µg/mL) (*p* < 0.0001) (Figure 3a). Interestingly, PAMAM G4-FITC dendrimers at all the mentioned doses have no significant effect on A549 cells' viability, while ERL G4-FITC complex (10, 20, 40, 80 µg/mL) exhibited a significant dosedependent decrease in the A549 cells' viability in comparison to the control group and the matched doses in blank PAMAM dendrimer G4-FITC groups (20 µg/mL; 49.06%; *p* < 0.0001, 40 µg/mL; 35.61%; *p* < 0.01, and 80 µg/mL; 42.26%; *p* < 0.0001) (Figure 3b). On the other hand, blank PAMAM dendrimer G5-FITC experiments have shown that the blank PAMAM dendrimer G5-FITC at doses of 20, 40, and 80 µg/mL have diminished the viability of A549 cells significantly after 72 h incubation (25.34%; *p* < 0.001, 23.98%; *p* < 0.001, and 28.92%; *p* < 0.0001, respectively). Additionally, ERL-G5-FITC complex at all the mentioned

doses have significantly decreased the viability of A549 cells in a dose-dependent manner compared to the control group (18.44%; *p* < 0.01, 16.79%; *p* < 0.05, 32.27%; *p* < 0.0001, and 39.37%; *p* < 0.000, respectively). However, surprisingly there was no significant difference between the PAMAM dendrimers G5-FITC groups and the matched doses in the ERL G5-FITC complex groups (Figure 3c). To facilitate the comparisons between the ERL, blank PAMAM dendrimers G4-FITC, ERL-G4-FITC complex, blank PAMAM dendrimers G5- FITC, and ER-G5-FITC complex, the highest dose of (80 µg/mL) was chosen for all the mentioned groups and the results are illustrated in Figure 3d, concluding that this dose will be considered for further experiments. *Molecules* **2023**, *28*, x FOR PEER REVIEW 7 of 15

**Figure 3.** (**a**) MTT results showing Erlotinib decreased the A549 cells' viability significantly with all the doses (10, 20, 40, 80 µg/mL) (*p* < 0.0001); (**b**) effect of PAMAM dendrimers G4-FITC and ERL-G4- FITC complex after 72 h. Exposure on the cell viability of A549 cells under controlled conditions; (**c**) effect of PAMAM G5-FITC and ERL-G5-FITC complex after 72 h. Exposure on the cell viability of A549 cells under controlled conditions; (**d**) a dose of 80 µg/mL was selected to compare the cell viability between PAMAM dendrimers G4-FITC, ERL-G4-FITC complex, PAMAM dendrimers G5- FITC, and ERL-G5-FITC complex. \*—Represents a comparison between each group of treatment to the control (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). #—Represents a comparison between each dose of G4 or G5 PAMAM dendrimer and the matched dose of ERL G4 or G5 complexes (## *p* < 0.01, ### *p* < 0.001, #### *p* < 0.0001). **Figure 3.** (**a**) MTT results showing Erlotinib decreased the A549 cells' viability significantly with all the doses (10, 20, 40, 80 µg/mL) (*p* < 0.0001); (**b**) effect of PAMAM dendrimers G4-FITC and ERL-G4-FITC complex after 72 h. Exposure on the cell viability of A549 cells under controlled conditions; (**c**) effect of PAMAM G5-FITC and ERL-G5-FITC complex after 72 h. Exposure on the cell viability of A549 cells under controlled conditions; (**d**) a dose of 80 µg/mL was selected to compare the cell viability between PAMAM dendrimers G4-FITC, ERL-G4-FITC complex, PAMAM dendrimers G5-FITC, and ERL-G5-FITC complex. \*—Represents a comparison between each group of treatment to the control (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). #—Represents a comparison between each dose of G4 or G5 PAMAM dendrimer and the matched dose of ERL G4 or G5 complexes (## *p* < 0.01, ### *p* < 0.001, #### *p* < 0.0001).

The surface bound and internalized particles was measured using flow cytometry analysis (*n* = 3/group) (Figure 4). Following treatment for 72 h, it was concluded that there was a significant uptake of ERL-G4-FITC complex by the cells in comparison to other

drimer ERL-G5-FITC groups, whereas the G5-FITC plain and PAMAM dendrimer ERL-

G5-FITC had minimal uptake by the cells.

*2.4. Cellular Uptake Analysis*

#### *2.4. Cellular Uptake Analysis*

The surface bound and internalized particles was measured using flow cytometry analysis (*n* = 3/group) (Figure 4). Following treatment for 72 h, it was concluded that there was a significant uptake of ERL-G4-FITC complex by the cells in comparison to other groups. In addition, we found that the cells showed significant uptake of blank G4- FITC PAMAM dendrimer compared to blank G5-FITC PAMAM dendrimer and PAMAM dendrimer ERL-G5-FITC groups, whereas the G5-FITC plain and PAMAM dendrimer ERL-G5-FITC had minimal uptake by the cells. *Molecules* **2023**, *28*, x FOR PEER REVIEW 8 of 15 *Molecules* **2023**, *28*, x FOR PEER REVIEW 8 of 15

**Figure 4.** Flow cytometry analysis after treatment for 72 h (*n* = 3/group); \*\*\*\* *p* < 0.0001, ### *p* < 0.001. **Figure 4.** Flow cytometry analysis after treatment for 72 h (*n* = 3/group); \*\*\*\* *p* < 0.0001, ### *p* < 0.001. **Figure 4.** Flow cytometry analysis after treatment for 72 h (*n* = 3/group); \*\*\*\* *p* < 0.0001, ### *p* < 0.001.

#### *2.5. Stability 2.5. Stability*

*2.5. Stability* ERL was found stable in the samples which were stored in the auto sampler at 4, 25, 37, and 50 °C for 6 months. The ERL contents were measured at 1, 3, and 6 months and results are included in Figure 5. The formulations at different temperatures were found to be stable. The freeze–thaw temperature cycles which were measured for three cycles were also stable. ERL was found stable in the samples which were stored in the auto sampler at 4, 25, 37, and 50 ◦C for 6 months. The ERL contents were measured at 1, 3, and 6 months and results are included in Figure 5. The formulations at different temperatures were found to be stable. The freeze–thaw temperature cycles which were measured for three cycles were also stable. ERL was found stable in the samples which were stored in the auto sampler at 4, 25, 37, and 50 °C for 6 months. The ERL contents were measured at 1, 3, and 6 months and results are included in Figure 5. The formulations at different temperatures were found to be stable. The freeze–thaw temperature cycles which were measured for three cycles were also stable.

**Figure 5.** Stability of erlotinib from formula obtained from four different temperatures (4 °C, 25 °C, 37 °C, 50 °C) for (**a**) PAMAM dendrimer G4-FITC complex; (**b**) PAMAM dendrimer G5-FITC complex for a period of 6 months. **Figure 5.** Stability of erlotinib from formula obtained from four different temperatures (4 °C, 25 °C, 37 °C, 50 °C) for (**a**) PAMAM dendrimer G4-FITC complex; (**b**) PAMAM dendrimer G5-FITC complex for a period of 6 months. **Figure 5.** Stability of erlotinib from formula obtained from four different temperatures (4 ◦C, 25 ◦C, 37 ◦C, 50 ◦C) for (**a**) PAMAM dendrimer G4-FITC complex; (**b**) PAMAM dendrimer G5-FITC complex for a period of 6 months.

#### **3. Discussion 3. Discussion 3. Discussion**

NSCLC is a type of cancer which does not often respond to therapy. This may be associated with different kinds of mutations in genes such as anaplastic lymphoma kinase (ALK), proto-oncogene B-Raf, discoidin domain receptor tyrosine k 2 (DDR2), and EGFR NSCLC is a type of cancer which does not often respond to therapy. This may be associated with different kinds of mutations in genes such as anaplastic lymphoma kinase (ALK), proto-oncogene B-Raf, discoidin domain receptor tyrosine k 2 (DDR2), and EGFR [31,32]. Upon these mutations, EGFR mutations have the highest rate of incidence, which NSCLC is a type of cancer which does not often respond to therapy. This may be associated with different kinds of mutations in genes such as anaplastic lymphoma kinase (ALK), proto-oncogene B-Raf, discoidin domain receptor tyrosine k 2 (DDR2), and

takes about 10–35% in the exons 18–21 [32]. The NSCLC patients with mutations in the EGFR gene are sensitive to the TKIs such as ERL. Inhibition of the tyrosine kinase do-

NSCLC [33]. In fact, many limitations are associated with ERL, especially as the marketed

[31,32]. Upon these mutations, EGFR mutations have the highest rate of incidence, which takes about 10–35% in the exons 18–21 [32]. The NSCLC patients with mutations in the

NSCLC [33]. In fact, many limitations are associated with ERL, especially as the marketed

EGFR [31,32]. Upon these mutations, EGFR mutations have the highest rate of incidence, which takes about 10–35% in the exons 18–21 [32]. The NSCLC patients with mutations in the EGFR gene are sensitive to the TKIs such as ERL. Inhibition of the tyrosine kinase domain of EGFR by ERL has been the mainstream treatment for advanced and/or metastatic NSCLC [33]. In fact, many limitations are associated with ERL, especially as the marketed ERL is administered orally as a film-coated tablet. These limitations include side effects, pH dependent solubility, low permeability, drug interaction, and drug resistance.

These side effects, however, can be reduced by developing a sustained and targeted delivery of ERL. Targeting ERL directly to cancerous cells through the lung may enhance its therapeutic effect [33].

The results showed successful formulations of ERL-loaded PAMAM dendrimer of G4 and G5. Drugs such as ERL could be found encapsulated or conjugated to dendrimers. The surface morphology and size of both generations of G4-FITC and G5-FITC ERL complexes were observed by TEM. It was found that both the plain and loaded dendrimers were found with a distinct spherical morphology. Investigations revealed that the drug–polymer binding showed that ERL was encapsulated inside PAMAM dendrimers and not binding to the PAMAM surface [20]. The mean particle size values of G4-FITC and G5-FITC loaded with ERL were 301.5 ± 8.42 nm and 302.0 ± 8.47 nm, respectively, indicating that the loaded dendrimers are suitable as an inhaler formulation [33,34]. Particle size of the polymer increased after loading of ERL. This may be attributed to the entrapment of ERL molecules within the cavities of PAMAM dendrimers. Sizes of complexes of each generation have enlarged in a range between 100 and 180 nm [20]. In this study, the PDI of empty and loaded PAMAM dendrimers was determined to estimate the average uniformity of the nanocomposite solutions [35]. The values of PDI for blank dendrimers G4-FITC and G5-FITC were 0.05 and 0.300, respectively. After ERL conjugation, the value of PDI changed to 0.02 for ERL-PAMAM dendrimers G4-FITC and G5-FITC. A study by Peng et al. demonstrated that PDI of PAMAM formulation was found to be in a range of 0.23–0.339 [36]. A study conducted by Bielski et al. established triphenylphosphonium (TPP) inside G4-FITC PAMAM dendrimer as an inhaled formulation. They reported a PDI of 0.36, which also supports our results [37]. In a study conducted by Conti et al., where siRNA was introduced into cationic PAMAM dendrimer G4-FITC and was delivered by hydrofluoroalkane (HFA)-based pMDI, the PDI values of these stable dispersions of the dendriplexes were between 0.4 and 0.6, showing satisfactory aerosol physical characteristics and suitability for deep lung deposition, with respirable fractions up to 77% [38]. PDI is related to the size uniformity in nanoformulations. Low PDI values indicated the uniformity in size distribution. However, higher PDI values indicated a large variation in the uniformity of particle size. These results illustrate that no significant changes in PDI were obtained of the dendrimer after loading the dendrimers with ERL. ζ-potential is defined as measuring the charge located on the surface of the particles. An imaginary shear plane outlines the boundary between the diffuse layer ions, which are unaffected by movement of the fluid, and those that are trimmed off by fluid motion. The net charge present at the shear plane of the diffuse layer is quantifiable as zeta potential; higher values of zeta potential indicate a higher rate of repulsion between particles [39]. This characteristic is essential in determining the stability of nanocomposites as well as knowing the intensity of electrostatic attraction between biomolecules and the nanocomposites [40]. The elevation of zeta potential value determines the effective entrapment of ERL within the dendrimers [21,41]. The findings of physicochemical investigations were in accordance with those reported for the PAMAM-dendrimers of neratinib and ruboxistaurin [29,42].

The in vitro release dissolution study revealed a prolonged release pattern of ERL from the formulations for more than seven days. According to other studies, it was shown that the slow liberation of drug molecules from dendrimer nanocomposites was due to two main reasons, namely, low aqueous solubility of drug and their confinement within the nanocomposites [42]. The percentage of ERL release from the dendrimers G4-FITC and G5-FITC at pH 5.4 were approximately 70% and 60%, respectively. The medium used pH 5.4, which is a mild acidic pH which mimics the environment around the cancerous cells, and the percentages of ERL release from the dendrimers G4-FITC and G5-FITC at pH 7.4 were approximately 34.97% and 41.9%, respectively. This medium represents extracellular physiological pH [24]. Furthermore, the stability of G4-FITC and G5-FITC ERL nanocomposites were inspected by the eye in order to assess any change in color, precipitation, or turbidity. This was conducted to determine the disintegration of the proposed nanocomposites when exposed to light and/or high temperatures. Results obtained from this study after a specified time show that there was no color change or precipitation detected. Our findings revealed that the nanocomposite was found to be physically stable at 4 ◦C. The results indicated high formulation stability. The findings of stability studies were in accordance with those reported for the PAMAM dendrimers of ruboxistaurin [29]. A significant increase in cellular uptake of the cationic ERL-G4-FITC PAMAM dendrimer was achieved after 72 h of treatment in comparison to the neutral G5- FITC-ERL dendrimer. Furthermore, in vitro cytotoxicity studies have shown that not only have loaded dendrimers demonstrated a significant decrease in cell viability; dendrimers of both generations G4-FITC and G5-FITC have also. This finding is supported by the hypothesis that dendrimers possess an anticancer effect [43]. The cell viability has decreased significantly by blank dendrimer G4-FITC more than blank PAMAM dendrimer G5-FITC. This is due to the presence of the negative charge on the modified surface of dendrimer G4-FITC in comparison to the neutral charge found on the surface of dendrimer G5-FITC. The main limitation of this work is the efficacy of prepared dendrimer nanocomposites in animal and human models. In the near future, pharmacodynamics and pharmacokinetics studies can be performed in animal and human models to explore the complete potential of developed dendrimer nanocomposites.

### **4. Materials and Methods**

#### *4.1. Materials*

ERL hydrochloride was obtained from Jazeera Pharmaceutical Industries Ltd. (Riyadh, Saudi Arabia). Dialysis cellulose membrane (MWCO 14,000 Da), reusable plastic sample cuvette, and folded capillary ζ-cells were obtained from Sigma Aldrich (St. Louis, MO, USA). Polyamidoamine dendrimer generation 4 (PAMAM dendrimers G4, 1 g vial) and polyamidoamine dendrimer generation 5 (PAMAM dendrimers G5, 1 g vial) were purchased from Nanosynthons (Mt. Pleasant, MI, USA). The HPLC-PDA purity of PAMAM dendrimers G4 and G5 is included in Figure S5. All chemicals and reagents were of analytical grade.

#### *4.2. Instrumentation and Chromatographic Conditions*

The analysis of ERL was performed at 25 ± 1 ◦C by means of Waters HPLC system 1515 isocratic (Waters, Milford, MA, USA) pump, 717 auto sampler, a programmable UV– Visible variable wavelength detector, a column oven, a SCL 10AVP system controller, and an inline vacuum degasser was used. The HPLC system utilized the use of the Millennium software, version 32, for data processing and analysis. The column utilized for this analysis was a Nucleodur (150 mm × 4.6 mm) RP C18 with particle size of 5 µm. The eluent/solvent system was composed of methanol: water (5:1 % *v/v*). The solvent system flowed with a flow rate of 1.0 mL/min. The detection of ERL was carried out at 254 nm. The samples (20 µL) were introduced into the system via a Waters auto sampler. The suggested method was validated for linearity, accuracy, precision, robustness, and sensitivity.

### *4.3. Synthesis of Dendrimers of ERL in G4-FITC PAMAM Dendrimer and G5-FITC PAMAM Dendrimer*

The lyophilized form of PAMAM dendrimers (G4 and G5) were dissolved in sterilized Milli-Q deionized water (10% PAMAM). Each generation was mixed separately with excess amount of ERL solution in a molar ratio of 25:1 of ERL to dendrimer. The 25:1 molar ratio of ERL:denfrimer was selected based on our previous studies with ruboxistaurin [29]. In our previous studies, various molar ratios of drug:dendrimer were investigated and

25:1 ratio was found to be the best. As a result, this ratio was selected in this work [29]. The mixtures were stirred overnight at room temperature for equilibration. Thereafter, the amount of ERL attached on the dendrimer was removed by placing it in cellulose membrane (MWCO14,000 Da) and dialyzing it against deionized water for 24 h to eliminate unconjugated or free ERL. The mixtures were lyophilized for 60 h [42]. Briefly, G4 and G5 PAMAM dendrimers were dissolved in deionized water and then mixed with FITC (E-Merk, Darmstadt, Germany) in 1:6 molar ratios. The components of dendrimers are expected to dissolve FITC in dendrimers. Then the formulations were stirred for 24 h [42].

### *4.4. Morphological Studies, TEM*

TEM imaging was performed to illustrate the morphological characteristics of the loaded dendrimer (size and shape of the particles) and compared with the plain dendrimer. The blank and loaded dendrimers were visualized by TEM (JEM1230EX; Tokyo, Japan). A drop of each blank and loaded dendrimer was placed onto a grid and set aside for air drying for 15 min then images of the particles was taken by TEM [42].

#### *4.5. Characterization of ERL Loaded Dendrimer*

The surface potential of different nanocomposites was evaluated by Malvern Zetasizer Nano-ZS (Malvern Instruments Ltd., Malvern, UK). Samples were suspended in water or PBS (pH 5.4 and pH 7.4). They were under analysis at 25 ◦C. The hydrodynamic diameter of the nano conjugates was also assessed using a Malvern Zetasizer Nano-ZS (Malvern Instruments Ltd.).

## *4.6. Particle Size Distribution, PDI, and ζ-Potential*

The two generations of nano conjugates were tested to characterize the physicochemical properties, which include particle size distribution, zeta potential, and PDI. The size of particles was assessed according to light scattering and the data were analyzed by an attached software, giving a measure of distribution of particle size. The particle size was measured at neutral pH (water). For the measurement, one mL of each formulation was diluted 100 times with water and subjected to particle size measurement. The ζ-potential was measured at PBS pH 5.4 and 7.4. For ζ-potential measurement, one mL of each formulation was diluted 100 times with each buffer and subjected for the ζ-potential measurement. ζ-potential determined the net charge presented on particles, which were moving in an electric field [42].

#### *4.7. Drug Loading, Entrapment Efficiency*

Approximately 0.002 g of G4-FITC-ERL and G5-FITC-ERL PAMAM dendrimer were dissolved in Milli-Q deionized water and placed in cellulose membrane (MWCO 14,000). Then, they were placed in 100 mL medium of deionized water. The mixture was vortexed for 1 h at 100 RPM. After overtaxing, 5 mL of solution was analyzed by HPLC to determine the amount of free drug which was not entrapped in the nano formulation. The following equations were used to measure the PAMAM dendrimer EE and ERL DLC:

$$\text{EE } (\%) = \frac{\text{Total } drug - \text{Free drug}}{\text{Total } drug} \times 100;\tag{1}$$

$$\text{DLC } (\%) = \frac{\text{Drug weight in nanopparticles}}{\text{Nanoparticles weight}} \times 100. \tag{2}$$

#### *4.8. In Vitro Drug Release Studies Using a Dialysis Method in PBS (pH 5.4 and pH 7.4)*

In vitro drug release studies were performed using the dialysis method [42]. An aliquot (5 mL) of each individual sample was placed in cellulose membrane dialysis tubing (MWCO 14,000). The mixtures were subjected to rotary machine stirring at 100 RPM in 50 mL of 40% methanol PBS pH (5.4) and pH (7.4) media at 37 ± 0.5 ◦C. At various time

intervals, an aliquot of (3 mL) was withdrawn from the released media and was replaced with an equivalent volume of fresh media. Triplicate samples (*n* = 3) of each PAMAM generation dendrimer were measured. Meanwhile, the in vitro dissolution study for the reference tablet was examined using Pharma test (DT 70, Germany) dissolution apparatus, utilizing type II apparatus (paddle) at stirring speed of 75 RPM in 1000 mL of 0.01 N HCl, and compared with 0.02% Tween 80 in 0.01 N HCl [44]. The amount of ERL released from each medium was then analyzed with HPLC-UV [24].

#### *4.9. Cell Culturing and Cytotoxicity Assay*

Non-small lung carcinoma (A549) cells were cultured at a density of ~ 3 <sup>×</sup> <sup>10</sup><sup>5</sup> cells/mL in basal medium containing 1:1 Dulbecco's modified Eagle's medium 1× (DMEM 1×): 10% fetal bovine serum (FBS); and 1% streptomycin/penicillin (100 µg/mL and 100 Units/mL, respectively). Cells were maintained in a humidified incubator containing 5% CO<sup>2</sup> at 37 ◦C. Once cells became confluent, they were passaged by trypsinization (300 µL) into T-75 flasks. Cells from passages 3 to 10 were utilized to conduct all proposed experiments. Ninety six well plates were utilized to perform MTT assay. All treatments were accomplished in full culture media containing FBS and streptomycin/penicillin. After every 48 h, the medium was replaced, regardless of the presence or absence of various treatments. MTT assay is considered a colorimetric process conducted to assess both cell proliferation and survival. This assay was used to evaluate the cytotoxicity of either ERL, ERL G4- FITC complex, ERL G5-FITC complex, or vehicles alone (PAMAM dendrimers G4-FITC or PAMAM dendrimers G5-FITC) on A549 cells. Cells were cultured at the conditions described above. Then, cells were harvested and trypsinized to be counted. Additionally, <sup>1</sup> <sup>×</sup> <sup>10</sup><sup>4</sup> cells per well in 96 wells plates were seeded. The plates were incubated for 24 h. In the survival experiments, viability was assessed in cells incubated with either ERL, ERL-G4-FITC complex, ERL-G5-FITC complex, or vehicles alone (10, 20, 40, and 80 µg/mL) for 72 h. Once cells were incubated at the proposed conditions, 10 µL of MTT (5 mg/mL PBS) reagent was added to each individual well for 30 min until purple precipitate was viable. Then, we added 100 µL of DMSO at room temperature while shaking for 5 min. At the end, absorbance at 570 nm was documented by using a micro-plate reader.

## *4.10. Cellular Uptake Analysis*

A549 cells were seeded in 12-well plates at 1 <sup>×</sup> <sup>10</sup><sup>5</sup> cells per well. After cells had attached and proliferated for 24 h, the culture medium was replaced with fresh media and treated with FITC-labeled dendrimers at a concentration of 80 µg/mL for 72 h. Treated cells were washed three times with PBS then harvested by trypsinization. Following that, the cells were collected, centrifuged, and suspended in 500 µL of PBS. Then, the cells were analyzed by flow cytometry (Cytomics FC 500; Beckman Coulter, CA, USA). Mean fluorescence intensities (MFI) of different cell exposure groups were compared and analyzed.

#### *4.11. Stability Studies*

Accelerated stability studies were performed in this work. Hence, the stability studies were conducted for 1, 3, and 6 months at accelerated temperatures of 4, 25, 37, and 50 ◦C. The stability of the tested samples was expressed as a percentage recovery relative to the freshly prepared solution. All stability studies were conducted in triplicate. Lastly, freeze/thaw stability tests were assessed on both formulations, after they had been frozen at (−30 ◦C) then thawed at room temperature for three cycles. The drug content was measured at different temperatures of storage [42].

#### *4.12. Statistical Analysis*

Diverse physicochemical variables, drug delivery and biological data were analyzed using Graph Pad InStat® software (San Diego, CA, USA) by the student *t*-test for two groups and aone-way analysis of variance (ANOVA) for multiple groups. Differences between each of the two related parameters were statistically significant for a *p*-value of equal or less than 0.05.

#### **5. Conclusions**

Erlotinib is an antineoplastic agent that has been marketed since 2004 as a film-coated tablet for the treatment of non-small cell lung cancer and pancreatic cancer. There are a vast number of obstacles to the current marketed erlotinib being delivered effectively to the cancerous cells; these include pH dependent solubility, inter-patient variability, and drugdrug interaction. To overcome these limitations, a non-invasive therapy of erlotinib was designed. This was performed utilizing nanocomposites of PAMAM dendrimers. Additional assessments of the formulations, including in vitro characterization, were conducted to ensure the feasibility of the formulation to be administered as a targeted inhaled drug delivery system. The physicochemical investigations suggested the proper formation of PAMAM dendrimers. The drug release studies showed prolonged release of erlotinib from both generation dendrimers. However, G4-FITC dendrimers showed higher drug release compared to the G5-FITC dendrimers. The dendrimer nanocomposites were found to be stable at 4 ◦C. Cytotoxicity studies revealed the suitability of dendrimer nanocomposites in the treatment of NSCLC. The proposed nanocomposites will overcome the unwanted adverse events, low bioavailability, and resistance associated with the currently marketed orally administered erlotinib tablet, and they will increase the efficacy by ensuring the sustained exposure of the cancerous cells to erlotinib. As a result, the frequency of administration of erlotinib will be remarkably reduced.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules28093974/s1. Figure S1: TEM images at low magnification to inspect the morphology of (a) blank G4-FITC PAMAM dendrimers; (b) erlotinib conjugated G4-FITC PAMAM dendrimers; (c) blank G5-FITC PAMAM dendrimers; and (d) erlotinib conjugated G5-FITC PAMAM dendrimers; Figure S2: Particle size distribution by intensity for (a) blank G4-FITC PAMAM dendrimers; (b) erlotinib conjugated G4-FITC PAMAM dendrimers; (c) blank G5-FITC PAMAM dendrimers; and (d) erlotinib conjugated G5-FITC PAMAM dendrimers; Figure S3: Zeta potential measurements for (a) blank G4-FITC PAMAM dendrimers; (b) erlotinib conjugated G4-FITC PAMAM dendrimers; (c) blank G5-FITC PAMAM dendrimers; and (d) erlotinib conjugated G5-FITC PAMAM dendrimers at pH 5.4; Figure S4: Zeta potential measurements for (a) blank G4-FITC PAMAM dendrimers; (b) erlotinib conjugated G4-FITC PAMAM dendrimers; (c) blank G5-FITC PAMAM dendrimers; and (d) erlotinib conjugated G5-FITC PAMAM dendrimers at pH 7.4; Figure S5: HPLC-PDA purity of polyamidoamine (PAMAM) dendrimers G4 and PAMAM dendrimers G5.

**Author Contributions:** Conceptualization, I.A.A. and F.S.A.; methodology, W.K.F., F.Y.A., M.M.A., A.A.A. and F.S.; software, H.A. and H.M.A.; validation, N.H., F.S. and F.Y.A.; formal analysis, H.M.A. and M.M.A.; investigation, W.K.F., N.H., F.S. and H.A.; resources, I.A.A.; data curation, F.S.; writing original draft preparation, W.K.F.; writing—review and editing, F.S., F.S.A. and I.A.A.; visualization, F.Y.A.; supervision, I.A.A. and F.S.A.; project administration, I.A.A. and F.S.A.; funding acquisition, I.A.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research project was funded by the Researchers Supporting Project (number RSP2023R340) at King Saud University, Riyadh, Saudi Arabia and The APC was funded by RSP.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** This study did not report any data.

**Acknowledgments:** The authors extend their sincere appreciation to the Researchers Supporting Project (number RSP2023R340) at King Saud University, Riyadh, Saudi Arabia for supporting this research.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples of the compounds erlotinib are available from the authors.

## **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Review* **Smart Nanocarriers as an Emerging Platform for Cancer Therapy: A Review**

**Madhuchandra Kenchegowda <sup>1</sup> , Mohamed Rahamathulla <sup>2</sup> , Umme Hani <sup>2</sup> , Mohammed Y. Begum <sup>2</sup> , Sagar Guruswamy <sup>1</sup> , Riyaz Ali M. Osmani <sup>1</sup> , Mysore P. Gowrav <sup>1</sup> , Sultan Alshehri <sup>3</sup> , Mohammed M. Ghoneim <sup>4</sup> , Areej Alshlowi <sup>4</sup> and Devegowda V. Gowda 1,\***


**Abstract:** Cancer is a group of disorders characterized by uncontrolled cell growth that affects around 11 million people each year globally. Nanocarrier-based systems are extensively used in cancer imaging, diagnostics as well as therapeutics; owing to their promising features and potential to augment therapeutic efficacy. The focal point of research remains to develop new-fangled smart nanocarriers that can selectively respond to cancer-specific conditions and deliver medications to target cells efficiently. Nanocarriers deliver loaded therapeutic cargos to the tumour site either in a passive or active mode, with the least drug elimination from the drug delivery systems. This review chiefly focuses on current advances allied to smart nanocarriers such as dendrimers, liposomes, mesoporous silica nanoparticles, quantum dots, micelles, superparamagnetic iron-oxide nanoparticles, gold nanoparticles and carbon nanotubes, to list a few. Exhaustive discussion on crucial topics like drug targeting, surface decorated smart-nanocarriers and stimuli-responsive cancer nanotherapeutics responding to temperature, enzyme, pH and redox stimuli have been covered.

**Keywords:** cancer; smart nanocarriers; drug targeting; nanoparticles; stimulus for drug release

## **1. Introduction**

Cancer is defined as uncontrolled cell growth and the lack of cell mortality, resulting in an abnormal cell mass, i.e., tumour, apart from haematological malignancy, where tumour cells multiply and proliferate throughout the lymph, blood and bone marrow systems [1]. Chemotherapy is the use of chemicals to kill or inhibit tumour progression, because tumour cells develop considerably faster than normal cells, and chemotherapy medications target those rapidly developing cells. However, some of the normal cells are also growing rapidly, so chemotherapy drugs target those rapidly multiplying normal cells [2,3]. The destruction or alteration of proto-oncogenes, which encode proteins involved in cell growth and division and tumour suppressor genes, which encode proteins that give inhibitory signals to cell development and trigger cell death, are the most common causes of cancer. Mutations in tumour susceptibility genes, which code for proteins involved in DNA damage regulation, are required for tumour formation and are encouraged by mutations in oncogenes and tumour suppressor genes. The mutations which cause tumours are clonally chosen to favour abnormal and unregulated cell growth, the lack of abnormal cell growth inhibition, the minimization of the immune system, the obliteration of cell mortality and transmission and the build-up of genetic information defects [1,4].

**Citation:** Kenchegowda, M.; Rahamathulla, M.; Hani, U.; Begum, M.Y.; Guruswamy, S.; Osmani, R.A.M.; Gowrav, M.P.; Alshehri, S.; Ghoneim, M.M.; Alshlowi, A.; et al. Smart Nanocarriers as an Emerging Platform for Cancer Therapy: A Review. *Molecules* **2022**, *27*, 146. https://doi.org/10.3390/ molecules27010146

Academic Editor: Ildiko Badea

Received: 3 December 2021 Accepted: 22 December 2021 Published: 27 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Although radiotherapy and surgery are the most appropriate and beneficial therapies for non-metastatic and local malignancies, they are ineffective when the tumour cells are spread to other parts of the body. Cancer medications (such as biological, chemotherapy and hormonal treatments) may reach each organ in the body through the circulation, they are the present treatment of options for metastatic cancers [1,5]. Conventional medicines have minimal aqueous solubility, bioavailability and therapeutic benefits. In greater doses, this substance is required to cause toxicity. The advancement of nanotechnology has a significant impact on cancer treatment [6].

To recognize tumour regions nanocarriers use physiochemical differences between tumour and normal cells. There are two methods for determining the location of tumour cells. The Enhanced Permeability (EPR) effect is used in passive targeting to determine the tumour location indirectly. Cancer cells are killed by employing an overexpressed cell surface receptor as a guided missile in active targeting. The next stage is to deliver medications in a specific place and at a specific concentration. Based on the nature and intelligence of the nanocarriers, drugs can be delivered by internal or external stimuli [3,7]. Smartnanocarriers are colloidal nano-scale particles capable of delivering anticancer drugs, such as medicine, that contain low molecular weight components such as genetics or enzymes [1]. Nanocarriers (10–400 nm) were chosen as drug carriers because of their ability to carry large amounts of medication, provide prolonged flow times and preferentially target tumour location due to enhanced permeability and retention effect (EPR). The P-glycoprotein is a drug efflux transporter that is commonly expressed on the surfaces of tumour cells produces MDR (multidrug resistance), smart-nanocarriers are used to combat MDR [8,9]. This review overviews the current advances in smart nanocarriers such as dendrimers, liposomes, mesoporous silica nanoparticles, quantum dots, micelles, superparamagnetic iron-oxide nanoparticles, gold nanoparticles and carbon nanotubes. This review discusses various topics like drug targeting, smart-nanocarriers and targeting moieties that respond to several stimuli including temperature, enzyme, pH and redox stimulus.

#### **2. Drug Targeting**

Smart-nanocarriers which are used for tumour targeting result in improved drug release, increased intracellular and internalization delivery, pharmacodynamic and pharmacokinetic profiles, controlled and higher specificity and most importantly lowers toxic effects [8]. Some of the common features of tumours are leaky blood vessels and poor lymphatic drainage. Two important types of drug targeting include active targeting and passive targeting.

#### *2.1. Passive Targeting*

Passive targeting to tumour cells can be done by EPR effect, which is exhibited by tumour cell. Due to the leaky endothelium of the tumour vasculature, the rate of drugloaded nanocarriers accumulating in a tumour is substantially greater than the healthy tissue. This is referred to as enhanced permeability effect. A defect in the lymphatic system causes nanoparticle retention in the tumour. This is referred to as the enhanced retention effect. The EPR effect refers to both phenomena [10]. Passive targeting is mostly determined by carrier characteristics such as tumour leakiness and vascularity, as well as size and circulation time. When compared to healthy organs, passive targeting significantly improves in specificity by 20–30%. Furthermore, EPR-based passive targeting to tumours is influenced by nanocarrier features like charge, size and surface chemistry, as well as the limitations imposed by improbable cell targeting within malignant tumours [11,12]. The EPR effect will be excellent if smart-nanocarriers can avoid immune surveillance and circulate for a long time. At the tumour location, very relatively high concentrations of drug-loaded smart-nanocarriers can be achieved in 1–2 days, 10–50 times higher than in normal cells [13]. Figure 1 shows the schematic representation of drug targeting via passive targeting mode and active targeting mode.

[11,12]. The EPR effect will be excellent if smart-nanocarriers can avoid immune surveillance and circulate for a long time. At the tumour location, very relatively high concentrations of drug-loaded smart-nanocarriers can be achieved in 1–2 days, 10–50 times higher than in normal cells [13]. Figure 1 shows the schematic representation of drug targeting

via passive targeting mode and active targeting mode.

Figure 1. Schematic representation of drug targeting via passive targeting (EPR effect) mode and active targeting mode. **Figure 1.** Schematic representation of drug targeting via passive targeting (EPR effect) mode and active targeting mode.

Hansen et al. created copper-64-loaded liposomes (PEGylated) and used in imaging to evaluate their EPR effects. Despite the fact that EPR had a dominant effect in only a few tumours, the outcome of high liposome deposition in 11 dogs with various solid tumours could not be extrapolated to any tumour [14]. Interstitial fluid pressure (IFP) is one obstacle to the efficient deposition of drug-incorporated nanocarriers in cancer cells [15]. Successful nanocarrier advancements can overcome various biological obstacles, such as IFP (interfacial fluid pressure) and RES (reticuloendothelial tissues) [16]. Hansen et al. created copper-64-loaded liposomes (PEGylated) and used in imaging to evaluate their EPR effects. Despite the fact that EPR had a dominant effect in only a few tumours, the outcome of high liposome deposition in 11 dogs with various solid tumours could not be extrapolated to any tumour [14]. Interstitial fluid pressure (IFP) is one obstacle to the efficient deposition of drug-incorporated nanocarriers in cancer cells [15]. Successful nanocarrier advancements can overcome various biological obstacles, such as IFP (interfacial fluid pressure) and RES (reticuloendothelial tissues) [16].

#### 2.2. Active Targeting Tumour cells and surface-modified targeted nanoparticles are used in active target-*2.2. Active Targeting*

ing [12,17]. Certain on-surface cells, such as cell surface antigens and folic acid, have been shown to be increased and overexpressed by tumour cells. Active ligands are coupled with drug-induced nanocarriers, where these ligands will recognise their overexpressed target on the surface of tumour cells. Aptamers, transferrin, peptides, folate and antibodies are the most commonly studied ligands [3]. Immunoliposomes, or antibody conjugated liposomes, are another technique in the active targeted delivery of anti-tumour medicines. Immunoliposomes, like liposomes, encapsulate anti-tumour medicines, but due to the associated tumour-specific antibody, they provide high concentration cancer cell targeting. Doxorubicin-loaded anti-human epidermal growth factor receptor-2 immunoliposomes have been proven to have a higher therapeutic effects against numerous breast cancer, when compared to naked PEGylated liposomes [18]. 3. Nanocarriers Used in Cancer Therapy Tumour cells and surface-modified targeted nanoparticles are used in active targeting [12,17]. Certain on-surface cells, such as cell surface antigens and folic acid, have been shown to be increased and overexpressed by tumour cells. Active ligands are coupled with drug-induced nanocarriers, where these ligands will recognise their overexpressed target on the surface of tumour cells. Aptamers, transferrin, peptides, folate and antibodies are the most commonly studied ligands [3]. Immunoliposomes, or antibody conjugated liposomes, are another technique in the active targeted delivery of anti-tumour medicines. Immunoliposomes, like liposomes, encapsulate anti-tumour medicines, but due to the associated tumour-specific antibody, they provide high concentration cancer cell targeting. Doxorubicin-loaded anti-human epidermal growth factor receptor-2 immunoliposomes have been proven to have a higher therapeutic effects against numerous breast cancer, when compared to naked PEGylated liposomes [18].

#### The nanoparticles (NPs) have recently received a lot of interest because of their drug carrier systems, bio-medicine potential as targeting systems, bio-imaging and controlled **3. Nanocarriers Used in Cancer Therapy**

drug releases. Functional organic solutes are typically encapsulated into NPs to overcome their limited water solubility. The hydrophilic coatings on Nanoparticle surfaces can also be coupled with amphiphilic surfactants, allowing insoluble organic solutes to be readily The nanoparticles (NPs) have recently received a lot of interest because of their drug carrier systems, bio-medicine potential as targeting systems, bio-imaging and controlled drug releases. Functional organic solutes are typically encapsulated into NPs to overcome their limited water solubility. The hydrophilic coatings on Nanoparticle surfaces can also be coupled with amphiphilic surfactants, allowing insoluble organic solutes to be readily supplied and distributed in an aqueous phase [19,20]. Nanocarriers protect medications from degradation, reduce their half-life in the bloodstream and renal clearance, increase the utility of cytotoxic medications, regulate the release kinetics of antitumour medications and increase the solubility of chemicals [1]. In terms of structure and intelligence, several fascinating smart-nanocarriers are described in detail below.

#### *3.1. Liposomes*

Liposomes are phospholipid-enclosed concentric bilayer vesicles with a hydrophilic centre [21]. Since they may entrap a wide range of medicines, both hydrophilic and lipophilic in nature, liposomes have been intensively researched as a preferred carrier for the delivery of therapeutic drugs in recent decades [22–24]. Liposomes have various advantages, including active group protection, cell-like membrane structure, minimal immunogenicity, biocompatibility, safety, efficacy and increased half-life [25]. Although, typical liposomes are also having drawbacks, including low entrapment and a higher likelihood for hydrophilic and amphiphilic medications to escape from liposomal vesicles, as well as accelerated blood clearance (ABC) through the reticuloendothelial system (RES). When liposomes are recognized as foreign to the body resulting in taken up by the mononuclear phagocyte system (MPS) and RES macrophages. The physicochemical characteristics of liposomes such as size, charge, hydrophilicity and hydrophobicity all influence their removal from the body's systemic circulation. PEGylated liposomes solve the RES uptake problem. Then, targeted liposomes were created to allow for the selective delivery of drugs to the appropriate region. Peptide, transferrin, folate, mannose, antibody and asialoglycoprotein are some of the ligands used in liposome targeting [26,27]. Using stimuli-triggered drug delivery systems, components of the tumour microenvironment (such as hypoxia, slightly increased temperature and acidic pH) have recently been utilized to deliver payloads in tumour tissues [28–32]. Drug localization, bio-distribution and therapeutic efficacy can all be tracked using theragnostic systems, which incorporate both a diagnostic and a therapeutic moiety in a liposomal system [33].

The primary issue with using liposomal delivery systems are ABC and RES uptake. There are many efforts are made to prevent liposomes uptake by RES and to increase the systemic circulation duration through a size adjustment or liposome surface modification. Liposomes of the second generation are a sort of customized liposome containing oligosaccharides, glycoproteins, synthetic polymers and polysaccharides are added to the surface to enhance circulation time. To achieve extended blood circulation of liposomes, many approaches have been used such as PEG coating on the liposomal surface. These PEGylatedliposomes demonstrated improved blood circulation time, greater biodistribution, good stability and better antitumour effectiveness (Long circulatory liposome) [34,35]. Molecules 2022, 27, x FOR PEER REVIEW 5 of 27 Table 1. Liposomal formulation approved by FDA for cancer therapy. Sr. No. Product Name Type Drug Uses/Treatment Ref. Acute myeloid

The pH of the tumour microenvironment has been found to be different from the pH of healthy cells as a result, using a pH-sensitive liposomal formulation to increase medication accumulation at the tumour site (extracellular or intracellular) could be a potential strategy [36]. The pH-stimuli liposomes can stay constant at physiological-pH (pH 7.5), but once inside the tumour (PH 5.7) causes pH-triggered drug release due to lipid layer break down [37]. The hyaluronic acid targeted PH-stimuli liposomes were produced and shows improved effectiveness against CD44 receptor overexpressing cells and lower toxic effects towards healthy cells than free doxorubicin (stimuli sensitive liposome) [38]. The schematic representation of different types of liposomes are shown in Figure 2. The FDA approved liposomal formulation for cancer therapy as show in Table 1. 1 Vyxeos® Liposome Daunorubicin and Cytarabine leukaemia [38] 2 Doxil® PEGylated liposome Doxorubicin Ovarian and breast cancer [39] 3 Lipo-Dox® PEGylated liposome Doxorubicin Multiple myeloma, Ovarian and breast cancer [40] 4 Onivyde® PEGylated liposome Irinotecan Metastatic pancreatic cancer [41] 5 Marqibo® Liposome Vincristine sulfate Acute lymphoblastic leukemia [41]

Figure 2. Schematic representation of different types of liposomes. (A) Conventional liposome, (B) ligand targeted liposome, (C) PEGylated liposome and (D) theranostic liposome. **Figure 2.** Schematic representation of different types of liposomes. (**A**) Conventional liposome, (**B**) ligand targeted liposome, (**C**) PEGylated liposome and (**D**) theranostic liposome.

Dendrimers are also known as dendron, which is derived from a Greek word, which

As nanocarriers for cancer therapy, poly-amidoamine (PAMAM), poly-l-lactide, poly-lysine, peptide dendrimer, poly-propylene-imine, poly-caprolactone and poly-ethylene glycol are currently employed [46]. For cancer treatment, paclitaxel, doxorubicin (DOX), methotrexate and cisplatin are loaded into dendrimer nanosystems, as are iron oxide nanoparticles and gold nanoparticles incorporated in dendrimer for imaging and diagnostic [47]. By connecting specific molecules to the dendrimer, which has the potential to attack cancer cells, the efficiency of cancer therapy can be increased, resulting in a decrease in toxicity and aiding in the control of cancer therapy side effects. Typically, ligands (galactose, Dextran and folate) and antigens are used as targeting molecules [48].

these layers are called generations because they are radially homocentric. [2,3] A central core of pharmacokinetic profiles and biocompatibility are determined by a functional group at the dendrimer's periphery [43]. Since the cationic dendrimers cause cell lysis, which damages the cell membrane due to interaction between the negatively charged cell membrane and the positively charged dendrimer surface, PEGylation and glycosylation enhance dendrimer biocompatibility [44,45]. Figure 3 shows the schematic representation

3.2. Dendrimers

of dendrimers.


These ligands are used to deal with the cationic toxicity of dendrimers, as well as to target

Investigators are currently investigating stimuli-responsive dendrimers, in which drug release happens when a specific stimulus is delivered by the external environment. Temperature, magnetic, light, pH and other sorts of stimuli are available, with dendrimer being responsive to pH-sensitive [50]. The dendrimers which are PH-sensitive used for cancer cell-specific delivery have hydrolysable connections that remain intact during circulation, but disintegrate quickly once inside the cancer cell, releasing medication for an-

**Table 1.** Liposomal formulation approved by FDA for cancer therapy. Poly-amidoamine dendrimer loaded with doxorubicin conjugation for tumour therapies was reported by Lai et al.; at 4.5 pH, nanocarriers release drugs faster (47 percent in

Molecules 2022, 27, x FOR PEER REVIEW 6 of 27

#### *3.2. Dendrimers*

tumour cells [49].

ticancer action [51].

Dendrimers are also known as dendron, which is derived from a Greek word, which means "tree", since it has a similar branching structure as a tree [42]. Dendrimers are made up of three main components [1]. A unit that repeats itself and is linked to the central core; these layers are called generations because they are radially homocentric [2,3]. A central core of pharmacokinetic profiles and biocompatibility are determined by a functional group at the dendrimer's periphery [43]. Since the cationic dendrimers cause cell lysis, which damages the cell membrane due to interaction between the negatively charged cell membrane and the positively charged dendrimer surface, PEGylation and glycosylation enhance dendrimer biocompatibility [44,45]. Figure 3 shows the schematic representation of dendrimers. <sup>3</sup>Folic acid-PAMAM dendrimer PPI #-dendrimer Methotrexate Epithelial cancer 4 PAMAM-PEG dendrimer PEGylated dendrimer 5-Flouro uracil Pancreatic cancer # PAMAM dendrimer—Poly(amidoamine) dendrimer, PAMAM-PEG dendrimer—Poly(amidoamine)-poly(ethylene glycol), PPI-dendrimer—Poly(propylene imine) dendrimers, RNA—Ribonucleic acid.

thyroid cancer

Figure 3. Schematic representation of dendrimer. **Figure 3.** Schematic representation of dendrimer.

As nanocarriers for cancer therapy, poly-amidoamine (PAMAM), poly-l-lactide, polylysine, peptide dendrimer, poly-propylene-imine, poly-caprolactone and poly-ethylene glycol are currently employed [46]. For cancer treatment, paclitaxel, doxorubicin (DOX), methotrexate and cisplatin are loaded into dendrimer nanosystems, as are iron oxide nanoparticles and gold nanoparticles incorporated in dendrimer for imaging and diagnostic [47]. By connecting specific molecules to the dendrimer, which has the potential to attack cancer cells, the efficiency of cancer therapy can be increased, resulting in a decrease in toxicity and aiding in the control of cancer therapy side effects. Typically, ligands (galactose, Dextran and folate) and antigens are used as targeting molecules [48]. These ligands are used to deal with the cationic toxicity of dendrimers, as well as to target tumour cells [49].

Investigators are currently investigating stimuli-responsive dendrimers, in which drug release happens when a specific stimulus is delivered by the external environment. Temperature, magnetic, light, pH and other sorts of stimuli are available, with dendrimer being responsive to pH-sensitive [50]. The dendrimers which are PH-sensitive used for cancer

cell-specific delivery have hydrolysable connections that remain intact during circulation, but disintegrate quickly once inside the cancer cell, releasing medication for anticancer action [51].

Poly-amidoamine dendrimer loaded with doxorubicin conjugation for tumour therapies was reported by Lai et al.; at 4.5 pH, nanocarriers release drugs faster (47 percent in 24 h) than at 7.4 pH (8 percent in 24 h), although PAMAM-amide-DOX releases drugs slower than PAMAM-hyd-DOX at 4.5 pH. PAMAM-hyd-DOX dendrimer nanocarriers are more harmful to malignant cells than PAMAM-amide-DOX nanocarriers [52] as shown in Table 2.


**Table 2.** Dendrimer for cancer treatment in clinical trials [53].
