*2.4. Further Characterizations of HP Starch Films* 2.4.1. FESEM and FTIR

Having established the HP starch films as the optimal buccal film formulation at high CUR payload, further characterizations of the HP starch films were carried out. The macroscopic image of the HP starch film was presented in Figure A2 of Appendix A. The CUR–CHI nanoplex embedded in the HP starch films was visible in the FESEM image shown in Figure 5a using the HP starch film prepared at theoretical CUR payload of 5 mg/cm<sup>2</sup> as the representative sample. The nanoplexes in the film were shown to be well dispersed as individual nanoparticles in the size range of 150–300 nm with minimal agglomeration among them. The FESEM image showed that the CUR–CHI nanoplex was well preserved upon its incorporation into the HP starch films.

**Figure 5.** (**a**) FESEM; (**b**) FTIR spectra of the CUR–CHI nanoplex-loaded HP starch film.

The presence of CUR in the HP starch film was verified by FTIR analysis via the appearance of the characteristic peaks of CUR at 1590, 1570, and 1410 cm−<sup>1</sup> in the FTIR spectrum of the nanoplex-loaded HP starch film (Figure 5b). These three peaks were attributed to the (C=C) and (C=O) vibrations, C=C aromatic ring stretching vibration, and

OH bending of the phenol group of CUR, respectively [19]. The three peaks in the HP starch films were shifted from higher wavenumbers of 1620, 1600, and 1410 cm−<sup>1</sup> in the FTIR spectra of the native CUR and CUR–CHI nanoplex. These peaks were not visible in the FTIR spectrum of the blank HP starch film. The peaks at 900–1100, 2900, and 3250 cm−<sup>1</sup> were attributed to the C-O-H bending, C-H stretching, and OH vibration of the amylose groups of the HP starch, respectively [53].

#### 2.4.2. CUR Payload Uniformity and Folding Endurance

The CUR payload uniformity among the independent samples of HP starch films (*n* = 10) prepared at theoretical CUR payloads of 1 and 5 mg/cm<sup>2</sup> was examined (Table 4). The results showed that both HP starch films exhibited the CUR payload's acceptance values (AV) equal to lower or slightly higher than the 15% maximum threshold value set by USP for AV. Hence, the HP starch films met the USP's requirement for uniformity of a dosage unit [54]. Nevertheless, we recognized that the AV value at high CUR payload barely met the acceptance limit; thus, improvements in the HP starch film formulation will be needed in the future. In terms of their physical robustness, both HP starch films exhibited good folding endurance with values around of 2.5 to 2.7 denoting no film breakage was observed after ≥300 double folds (Table 4).

**Table 4.** CUR payload uniformity (*n* = 10) and folding endurance of HP starch films (*n* = 3).


2.4.3. Amorphous form Stability and Thermal Stability

PXRD analysis of the nanoplex-loaded HP starch film performed after the accelerated storage did not show the appearance of strong intensity peaks, which were present in the PXRD pattern of the native CUR crystals (Figure 6a). The HP starch film prepared at theoretical CUR payload of 5 mg/cm<sup>2</sup> was used as the representative sample for PXRD. Thus, the CUR–CHI nanoplex in the HP starch film maintained its amorphous form after the accelerated storage equivalent to twelve-month storage at ambient condition. Nevertheless, the amorphous halo at 2*θ* ≈ 15–25◦ visible in the PXRD pattern of the HP starch film before storage became less pronounced after storage. The amorphous halo was replaced by low-intensity peaks, indicating decreased amorphous contents as crystallization of some of the nanoplex took place during storage.

The TGA results showed that the native CUR and HP starch film started to decompose at temperatures above 280 ◦C (Figure A3 in Appendix A); thus, the DSC thermograph for thermal stability was analyzed at temperatures below the decomposition temperature. DSC thermograph of the HP starch film prepared at theoretical CUR payload of 5 mg/cm<sup>2</sup> showed the appearance of a sharp endothermic peak at around 176 ◦C, which was typical of the melting point of crystalline CUR (Figure 6b). Not unexpectedly, the same peak at 176 ◦C appeared in the DSC thermograph of the native CUR crystals. The DSC results indicated that the CUR–CHI nanoplex in the HP starch film experienced amorphous to crystalline transition upon heating above 170 ◦C. In contrast, the melting point peak was not evident in the DSC thermograph of the free CUR–CHI nanoplex, where only solid transition events at around 160–170 ◦C were recorded. This signified the higher thermal stability of the free CUR–CHI nanoplex compared to the nanoplex embedded in the film. Nevertheless, the CUR–CHI nanoplex embedded in the HP starch film remained stable upon heating up to 150 ◦C, hence, the drying step in the film preparation should not adversely affect the amorphous form stability of the CUR–CHI nanoplex.

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

#### *3.1. Materials*

*Materials for CUR–CHI nanoplex's preparation and characterization*: curcumin (CUR) from turmeric rhizome (>95% curcuminoid content) was purchased from Alfa Aesar (Singapore). Chitosan (CHI) (190–310 kDa, 75–85% deacetylation), potassium hydroxide (KOH), disodium phosphate (Na2HPO4.7H2O), ethanol, potassium dihydrogen phosphate (KH2PO4), sodium chloride (NaCl), hydrogen chloride (HCl), phosphoric acid (H3PO4), and glacial acetic acid (AA) were purchased from Sigma Aldrich (Singapore). *Materials for buccal film's preparation*: hydroxypropyl methylcellulose (HPMC) (MW = 26 kDa), polyvinyl alcohol (PVA) (90 kDa, 99% hydrolyzed), sodium alginate (AGN), glycerol (Gly), and propylene glycol (PG) were purchased from Sigma Aldrich (Singapore). Hydroxypropyl (HP) starch (LYCOAT® NG720) and pre-gelatinized starch (LYCATAB®) were generously provided by Roquette (Singapore).

#### *3.2. Methods*

#### 3.2.1. Preparation and Characterization of CUR–CHI Nanoplex

The amorphous CUR–CHI nanoplex was prepared following the protocols presented in Lim et al. [19]. Briefly, CUR was dissolved at 5 mg/mL in 0.1M KOH (pH 13) and separately, CHI was dissolved at 5.9 mg/mL in 1.2% (*w*/*v*) AA (pH 2.7), both at room temperature. Equal volumes of the CUR and CHI solutions (10 mL each) were then mixed immediately after their preparation under gentle stirring to minimize alkaline degradation of CUR. The resultant CUR–CHI suspension was ultrasonicated for 15 s at 20 kHz (VC 505, Sonics, New Town, CT, USA) to break up large agglomerates (if any). The nanoplex suspension then underwent two cycles of ultracentrifugation (14,000× *g*, 10 min) and washing with deionized water to remove free CUR and free CHI that did not form the nanoplex. The CUR encapsulation efficiency into the nanoplex was characterized by measuring the free CUR concentration in the supernatant after the first centrifugation step using high performance liquid chromatography (HPLC) as described below. Afterwards, the washed CUR–CHI nanoplex suspension was lyophilized for 24 h at −52 ◦C and 0.05 mbar in Alpha 1–2 LD Plus freeze dryer (Martin Christ, Osterode am Harz, Germany) for characterization purposes.

The size and zeta potential of the CUR–CHI nanoplex suspension were characterized in triplicates by dynamic light scattering (DLS) after 100× dilution, using Brookhaven 90 Plus Nanoparticle Size Analyzer (Brookhaven Instruments Corporation, Holtsville, NY, USA). The CUR content in the nanoplex, which was defined as the amount of CUR per unit mass of the CUR–CHI nanoplex, was determined in triplicates by dissolving 1 mg of the lyophilized nanoplex powder in 10 mL 80% (*v*/*v*) ethanol. The amount of CUR in the ethanol was subsequently determined by HPLC (Agilent 1100, Agilent Technologies,

Santa Clara, CA, USA) at CUR detection wavelength of 423 nm. The HPLC was performed using ZORBAX Eclipse Plus C18 column (250 × 4.6 mm, 5 µm particle size) and 75% (*v*/*v*) acetonitrile solution as the mobile phase at flow rate of 1 mL/min, resulting in CUR retention time of approximately 2.8 min. The HPLC chromatogram of the CUR detection was presented in Figure A4 of Appendix A. The physical appearances of the CUR–CHI nanoplex before and after incorporation into the buccal film were examined by field emission scanning electron microscope (FESEM) (JSM 6700F, JEOL, Peabody, MA, USA).

#### 3.2.2. Preparation of CUR–CHI Nanoplex-Loaded Buccal Film

The precursor solution for the HPMC film was prepared by overnight dissolution of HPMC in deionized water at 5% (*w*/*v*) under gentle stirring. The plasticizer, Gly or PG, was added to the HPMC solution at 5% (*w*/*w*). In the study on the effects of adjuvants inclusion, aqueous AGN or PVA solution was added to the HPMC + plasticizer solution at 0.5% (*w*/*v*). Freshly prepared CUR–CHI nanoplex was added last at theoretical CUR payloads of 1 or 5 mg CUR per cm<sup>2</sup> of the film. The precursor solution was then vortexed for 1 min to ensure its homogeneity. Afterwards, the precursor solution was casted onto a 9 cm diameter glass petri dish at a liquid height of 4 mm. Next, the petri dish was transferred to a convective laminar flow oven for drying at 60 ◦C for 3 h. The resultant dried HPMC film was peeled off the petri dish and stored in a sealed plastic bag for characterizations. The starch films were prepared by the same procedures at starch concentrations of 15% (*w*/*v*) for both unmodified and HP starches using 5% (*w*/*w*) Gly as the plasticizer for both. The HP starch was pre-gelatinized in deionized water at 80 ◦C prior to the film preparation. The CUR–CHI nanoplex was also added at theoretical CUR payloads of 1 and 5 mg CUR per cm<sup>2</sup> of the starch films.

#### 3.2.3. Experimental CUR Payload, CUR Payload Uniformity, CUR Entrapment Efficiency

The experimental CUR payload in the buccal film (in mg of CUR per cm<sup>2</sup> of film) was determined by dissolving a 2 <sup>×</sup> 2 cm<sup>2</sup> square film samples in 25 mL 80% (*w*/*v*) ethanol solution for 1 h. Afterwards, the amount of CUR dissolved in the ethanol was determined by HPLC as previously described. The CUR payloads of ten independently prepared films were determined from which the CUR payload uniformity was characterized. According to the United States Pharmacopeia's (USP) criteria, the payload uniformity of a drug dosage was deemed acceptable when the acceptance value (AV) was less than 15%. The definition for AV is elaborately explained in USP monograph "<905> Uniformity of Dosage Units" [54] and not repeated here for brevity. After the experimental CUR payload was determined, the CUR entrapment efficiency (*n* = 10) was calculated from the ratio of the experimental CUR payload to the theoretical CUR payload.

#### 3.2.4. Weight, Thickness, Folding Endurance of the Buccal Film

The weight and thickness of the buccal film were characterized from 9 cm diameter circular films using analytical balance and digital caliper, respectively. The variations in the buccal film's weight and thickness were characterized using ten independently prepared films. According to the USP's criteria, the variations in the weight and thickness among the independent samples were determined to be acceptable if each sample exhibited weight (or thickness) within ±10% of the arithmetic average of the weight (or thickness) [54]. The folding endurance was defined as the logarithmic (log10) of the number of double folds that resulted in the breakage of the buccal film. Briefly, triplicates of 2 <sup>×</sup> 2 cm<sup>2</sup> square film samples were manually double folded and the number of folds at which the film started showing signs of breakage was noted.

#### 3.2.5. CUR Dissolution from Buccal Film and Film Disintegration

CUR dissolution from the buccal film was characterized in six replicates in simulated saliva fluid (SSF) under a sink condition at <sup>1</sup> 4 of the thermodynamic saturation solubility

of CUR in the SSF (CSat). The SSF was prepared following the formulation of Peh and Wong [43] in which 4.49 g of Na2HPO4.7H2O, 0.19 g KH2PO4, and 8.00 g NaCl were dissolved in one 1 L of deionized water. The pH of the SSF was adjusted to pH 6.75 by the addition of H3PO4. CSat of CUR in the SSF was experimentally determined by incubating excess CUR in 100 mL SSF maintained at 37 ◦C in a shaking incubator. After 24-h incubation, the CUR concentration in the SSF was determined by HPLC as previously described, resulting in CSat equal to 6.2 µg/mL.

Briefly, 2 <sup>×</sup> 2 cm<sup>2</sup> square film samples were completely immersed in 100 mL SSF maintained at 37 ◦C in a shaking incubator. At specific timepoints over a 4-h period, 1 mL of aliquot was withdrawn from the dissolution flask and the same volume of fresh SSF was added back to the flask as replenishment. The aliquot was syringe filtered (0.22 pore size), followed by 3-min ultracentrifugation at 14,000× *g*. Afterwards, the CUR concentration in the supernatant was determined by HPLC as previously described.

In a separate experiment, the disintegration time of the buccal film was characterized in triplicates by immersing 2 <sup>×</sup> 2 cm<sup>2</sup> square film samples in 10 mL SSF in a petri dish maintained at 37 ◦C in a shaking incubator. The smaller volume of SSF used in the disintegration test was to simulate the aqueous environment of oral cavity more closely. Herein, the disintegration time was defined as the time at which the buccal film had been completely disintegrated to aqueous suspension of the nanoplex and polymers, where no visible film fragments were present.

#### 3.2.6. PXRD, DSC, and FTIR

The long-term stability of the amorphous form of the CUR–CHI nanoplex in the buccal film was examined by storing the nanoplex-loaded film in an open container placed inside a desiccator for three months under accelerated storage condition of 40 ◦C and 75% relative humidity. The accelerated storage condition was approximately equivalent to twelve-month storage under ambient condition (i.e., 25 ◦C and 60% relative humidity). The 75% relative humidity was generated inside the desiccator by placing an open container of saturated NaCl solution at 40 ◦C. At the end of the storage, the amorphous form of the nanoplex in the buccal film was examined by powder X-ray diffraction (PXRD) using D8 Advance X-ray Diffractometer (Bruker, Berlin, Germany) performed between 10◦ and 70◦ (2θ) with a step size of 0.02◦/s. For comparison, the PXRD analysis was also carried out for the native CUR, the free CUR–CHI nanoplex, and the nanoplex-loaded film before storage.

The thermal stability of the amorphous form of the CUR–CHI nanoplex in the buccal film was characterized by thermal gravimetric analysis (TGA) (Pyris Diamond TGA, PerkinElmer, Waltham, MA, USA) and differential scanning calorimetry (DSC) (DSC 822E, Mettler Toledo, Columbus, OH, USA). The TGA analysis was performed at heating rate of 10 ◦C/min between 30 ◦C and 400 ◦C. The DSC analysis was performed at heating rate of 2 ◦C/min between 30 ◦C to 300 ◦C. Lastly, the presence of CUR in the buccal film was verified by Fourier transform infrared spectroscopy (FTIR) from 450 and 4000 cm−<sup>1</sup> at 4 cm−<sup>1</sup> spectral resolution in Spectrum One (Perkin–Elmer, Waltham, MA, USA). The DSC/TGA and FTIR analyses were also performed for the native CUR and the free CUR– CHI nanoplex.

#### 3.2.7. Statistical Analysis

All experiments were performed with a minimum of three replicates and the results are presented as mean ± standard deviation. The statistical significance was analyzed using Student's *t*-test in GraphPad Prism software (GraphPad Software, San Diego, CA, USA). All *p*-values were two-sided and considered significant at *p* ≤ 0.05, unless stated otherwise.

#### **4. Conclusions**

The feasibilities of HPMC and pre-gelatinized starch-based films as buccal sustained release delivery systems of amorphous CUR–CHI nanoplex at high CUR payload were investigated. HPMC and starch films were found to exhibit similar CUR entrapment efficiencies (≈ 80% *w*/*w*), resulting in their similar CUR payloads in the range of 0.6 to 4 mg/cm<sup>2</sup> . Both HPMC and starch films were able to accommodate higher CUR payloads without any adverse effect on the CUR entrapment efficiency. The starch films were denser and thicker than HPMC films, as higher film precursor concentrations were needed to produce starch films with good physical integrity. Despite being denser and thicker, starch films disintegrated faster than HPMC films. The faster disintegration time of the starch films resulted in its significantly superior CUR release profiles at high CUR payload (4 mg/cm<sup>2</sup> ) compared to the HPMC films. This was because the faster disintegration time enabled faster outward CUR molecular diffusion across the swollen polymer matrix, which in turn reduced the precipitation propensity of the highly supersaturated CUR concentration generated in the film by the amorphous nanoplex dissolution. The superior CUR release exhibited by the starch films, nevertheless, were not as evident at lower CUR payloads (≤1 mg/cm<sup>2</sup> ) due to the lower CUR supersaturation level generated in the film. Between the unmodified starch and HP starch films, the HP starch films exhibited superior CUR release profiles in which at least 85% (*w*/*w*) of the CUR payload was released within 4 h, whereas the CUR release from the unmodified starch plateaued at around 65% (*w*/*w*) after the same period. At the high CUR payload, HP starch films exhibited the ideal sustained CUR release profile following the zero-order kinetics. The films prepared from different batches (*n* = 10) exhibited good CUR payload uniformity and minimal weight/thickness variations. The HP starch films were physically robust with high folding endurance and the embedded nanoplex was thermally stable with good long-term storage stability.

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

**Funding:** This work was funded by the Nanyang Research Programme (NRP SCBE01jr 2017) of Nanyang Technological University, Singapore.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors would like to thank Roquette Singapore Pte Ltd. for gratuitously providing the starch samples.

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

#### **Abbreviations**



#### **Appendix A**

**Figure A1.** Dissolution kinetics modelling of the CUR release from HP starch film prepared at 5 mg/cm<sup>2</sup> (**a**) zero order kinetics, (**b**) first order kinetics, and (**c**) Higuchi model.

**Figure A2.** Macroscopic image of the CUR–CHI nanoplex-loaded HP starch film.

**Figure A3.** TGA of the HP starch film prepared at theoretical CUR payload of 5 mg/cm<sup>2</sup> .

**Figure A4.** HPLC chromatogram of CUR detection.

#### **References**

