**In Situ Hydrogel-Forming**/**Nitric Oxide-Releasing Wound Dressing for Enhanced Antibacterial Activity and Healing in Mice with Infected Wounds**

**Juho Lee 1, Shwe Phyu Hlaing 1, Jiafu Cao 1, Nurhasni Hasan 1, Hye-Jin Ahn 2, Ki-Won Song <sup>2</sup> and Jin-Wook Yoo 1,\***


Received: 6 September 2019; Accepted: 25 September 2019; Published: 27 September 2019

**Abstract:** The eradication of bacteria from wound sites and promotion of healing are essential for treating infected wounds. Nitric oxide (NO) is desirable for these purposes due to its ability to accelerate wound healing and its broad-spectrum antibacterial effects. We developed an in situ hydrogel-forming/NO-releasing powder dressing (NO/GP), which is a powder during storage and forms a hydrogel when applied to wounds, as a novel NO-releasing formulation to treat infected wounds. An NO/GP fine powder (51.5 μm) was fabricated by blending and micronizing S-nitrosoglutathione (GSNO), alginate, pectin, and polyethylene glycol (PEG). NO/GP remained stable for more than four months when stored at 4 or 37 ◦C. When applied to wounds, NO/GP absorbed wound fluid and immediately converted to a hydrogel. Additionally, wound fluid triggered a NO release from NO/GP for more than 18 h. The rheological properties of hydrogel-transformed NO/GP indicated that NO/GP possesses similar adhesive properties to marketed products (Vaseline). NO/GP resulted in a 6-log reduction in colony forming units (CFUs) of methicillin resistant *Staphylococcus aureus* (MRSA) and *Pseudomonas aeruginosa,* which are representative drug-resistant gram-positive and -negative bacteria, respectively. The promotion of wound healing by NO/GP was demonstrated in mice with full-thickness wounds challenged with MRSA and *P. aeruginosa*. Thus, NO/GP is a promising formulation for the treatment of infected wounds.

**Keywords:** in situ hydrogel-forming powder; nitric oxide-releasing formulation; S-nitrosoglutathione (GSNO); antibacterial; wound dressing; wound healing

#### **1. Introduction**

Cutaneous wound infections are a global problem whose cost of treatment runs into millions of dollars per year, and these infections can lead to severe complications including sepsis, for which mortality remains around 30% in the United States [1,2]. Wound healing proceeds spontaneously through three sequential phases; inflammation, proliferation, and remodeling [3–6]. However, when wounds are infected, the healing process is delayed during the inflammation phase, since bacteria induce continuous inflammation at the infected site [7,8]. Thus, eradicating bacteria from the injured site is essential for the treatment of infected wounds.

In recent years, nitric oxide (NO) has gained attention as a novel agent for the treatment of infected wounds because it facilitates wound healing processes such as skin cell proliferation and tissue remodeling, and it also exerts broad-spectrum antibacterial effects [9–12]. NO enhances re-epithelialization and collagen synthesis during wound healing [5,13]. In addition, NO possesses broad-spectrum antibacterial properties through the formation of reactive nitrogen species such as peroxynitrile, nitrogen dioxide, and dinitrogen trioxide, which can interact with various bacterial proteins, DNA, and enzymes to result in bacterial cell death [14]. Since NO exerts bactericidal effects via multiple biochemical pathways, it exhibits antibacterial effects against drug-resistant bacteria, including methicillin resistant *Staphylococcus aureus* (MRSA) [15]. Moreover, NO does not develop drug resistance due to its diverse antibacterial mechanisms, which require multiple mutations for resistance to develop [16]. However, despite these beneficial effects of NO, its clinical application remains challenging because of its short half-life and gaseous nature. Therefore, the development of an NO-releasing formulation with a sustained NO-release and good storage stability is required.

We hypothesized that a powder dressing that forms a hydrogel in situ maintains a powder state during storage and converts to hydrogel immediately when applied to wounds could be an ideal NO-releasing formulation for the treatment of infected wounds. An in situ hydrogel-forming/NO-releasing powder dressing possesses the benefits of both powders and hydrogels. Since a water-free powder ensures high storage stability, the stability of easily hydrolyzable pharmaceutical ingredients can be markedly improved. When applied to the wound site, an in situ hydrogel-forming powder dressing can fit the irregular surface of the wound without wrinkles or fluting [17–19]. In addition, in situ hydrogel-forming powders are easy to apply to bendable areas of the human body, such as elbows, finger joints, ankles, or wide wound beds resulting from burn and pressure ulcers. Once applied to the wound, they can absorb the wound fluid and form an adhesive hydrogel, which can protect the wound site from the external environment and maintain a humid environment to aid wound healing [20,21]. At the same time, a NO release triggered by the absorbed wound fluid may assist wound healing via the elimination of bacteria and stimulation of wound-healing processes [22].

In this study, a novel in situ hydrogel-forming/NO-releasing powder dressing (NO/GP) possessing the benefits of both powders and hydrogels was developed using S-nitrosoglutathione (GSNO), alginate, pectin, and polyethylene glycol (PEG) for the treatment of infected wounds. GSNO is a widely used endogenous NO donor that can generate NO over several hours under physiological conditions, which may be beneficial properties for NO releasing wound dressings [23]. Sodium alginate was selected to form a bioadhesive hydrogel. In addition, it can absorb wound exudate up to approximately 20-times its weight [24]. Pectin was used to form a hydrogel structure and accelerate the powder-to-hydrogel transition [19]. PEG was added to modulate the ability of the formulation to uptake fluid to prevent potential wound drying caused by the excessive absorption of exudate. Following the characterization of NO/GP, the storage stability, in situ hydrogel-forming ability, rheological properties, and NO-release profiles were evaluated in a series of in vitro and in vivo experiments. The antibacterial effects of NO/GP were examined against MRSA and *Pseudomonas aeruginosa*, which are drug-resistant gram-positive and -negative bacteria, respectively. The in vivo therapeutic effects of NO/GP were evaluated using a bacteria-challenged full-thickness wound mouse model.

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

#### *2.1. Materials*

Glutathione (reduced form) was purchased from Wako Pure Chemical (Osaka, Japan). Pectin and agar were purchased from Yakuri Co., Ltd. (Osaka, Japan). Sodium alginate, PEG (average molecular weight: 8000), sodium nitrite, crystal violet, Lugol's solution, 2,2,2 tribromoethanol, *tert*-amyl alcohol, Mayer's hematoxylin, and eosin-Y disodium were purchased from Sigma-Aldrich (St. Louis, MO, USA). Bacto™ tryptic soy broth (TSB) and Difco™ cetrimide-agar media were purchased from BD Biosciences (San Jose, CA, USA). A Twort's Gram stain set was purchased from Newcomer Supply (Middleton, WI, USA). Masson's trichrome stain kit was purchased from Abcam (Cambridge, MA, USA). The LIVE/DEAD® *Bac*Light™ bacterial viability kit was purchased from Thermo Fisher

Scientific (Waltham, MA, USA). All other reagents and solvents were of the highest analytical grade commercially available.

#### *2.2. GSNO Synthesis*

GSNO was synthesized following a previously reported method with some modifications [25]. Briefly, sodium nitrite and reduced glutathione were added to a cold HCl solution with stirring for 40 min in an ice bath (the final concentration of NaNO2, glutathione, and HCl was 0.625 M). To precipitate GSNO, acetone was added and stirred for 20 min. The precipitate was collected by filtration and washed once with 80% acetone, twice with 100% acetone, and three times with diethyl ether. After drying, GSNO was stored in a −20 ◦C refrigerator for subsequent experiments.

#### *2.3. Preparation and Characterization of the NO-Releasing In Situ Hydrogel-Forming Powder*

To prepare the in situ hydrogel-forming agent, sodium alginate, pectin, and PEG were micronized, sieved (90 μm), and blended. The ratio of alginate, pectin, and PEG used for the in situ hydrogel-forming powder (GP) was optimized with sodium alginate:pectin:PEG at a 2:1:6 ratio in several pilot studies. To prepare the NO-releasing in situ hydrogel-forming powder (NO/GP), GSNO was added to the GP (final content of GSNO was 4 wt %). GSNO was sieved before blending with GP to eliminate the GSNO aggregates, which could induce undesirable effects in the content uniformity. NO/GP was stored in a −20 ◦C refrigerator.

NO/GP was characterized by determining particle size, GSNO content, content homogeneity, and powder flowability. Briefly, the particle sizes of NO/GP and the non-micronized mixture were analyzed by ImageJ software (National Institutes of Health, Bethesda, MA, USA) using microscopic images. More than 100 particles were used to assess particle size. GSNO contents were determined using 10 samples of each NO/GP and non-micronized mixture, which were dissolved in distilled water (DW), and then the absorbance was detected using a UV/Vis spectrophotometer (U-5100, Hitachi, Tokyo, Japan) at 335 nm. From the absorbance values, GSNO content and relative standard deviation (RSD) were calculated to investigate GSNO content and content homogeneity, respectively. Powder flowability was evaluated as previously reported, with some modifications [26]. Briefly, 40 mg of NP/GP was loaded in a 1 mL syringe and tapped until no change in volume was detected. The bulk and tapped density were calculated from the tapped and untapped volume of NO/GP, and the Hausner ratio was calculated based on the tap density/bulk density ratio.

#### *2.4. Rheological Properties of NO*/*GP in Hydrogel Form*

The rheological properties of the hydrogel-form of NO/GP were evaluated as previously reported, with some modification [27–29]. The steady shear and dynamic viscoelastic properties of the hydrogel-form of NO/GP were measured using a strain-controlled rheometer (Advanced Rheometric Expansion System [ARES], Rheometric Scientific, Piscataway, NJ, USA) equipped with a parallel-plate fixture with a radius of 12.5 mm and a gap size of 1.0 mm. All rheological measurements were performed at a fixed temperature of 37 ◦C over a wide range of shear rates and strain amplitudes. In this study, simulated wound fluid (SWF) [30,31] was used to induce NO/GP swelling. The SWF consisted of 0.64% NaCl, 0.22% KCl, 2.5% NaHCO3, and 0.35% NaH2PO4 in double distilled water with pH 7.4. Three different conditions of NO/GP in hydrogel form (NO/GP that absorbed 200%, 350%, and 500% of SWF absorbed per weight) were examined with SWF. Before initiating the experiments, 2, 3.5, or 5 mL of SWF were added to 1 g of NO/GP and mixed well to obtain the homogeneous hydrogel. In all experiments, a fresh sample was used and rested for 15 min after loading to allow for material relaxation and temperature equilibration. To evaluate the steady shear flow behaviors of the hydrogel-form of NO/GP, steady rate-sweep tests were performed over a range of shear rates from 1 to 1000 s−<sup>1</sup> with a logarithmically increasing scale. Next, strain-sweep tests were conducted to investigate both the linear viscoelastic region and nonlinear viscoelastic behavior over a strain amplitude range of 0.0625%–500% at a fixed angular frequency of 10 rad/s.

#### *2.5. Storage Stability*

The storage stability of NO/GP was evaluated by determining GSNO degradation under two temperature conditions: 4 and 37 ◦C. Two milligrams of NO/GP were placed in each tube, which were then stored in either a 4 ◦C refrigerator or a 37 ◦C incubator. At previously set time points, three tubes from each group were sampled, and the absorbance at 335 nm was measured to determine the GSNO content following dilution with DW.

#### *2.6. Behavior of NO*/*GP After Exposure to Wound Fluid*

#### 2.6.1. Morphological Changes of NO/GP at the Wound Site

All animal experiments in this study have been reviewed and approved by the Pusan National University Institutional Animal Care and Use Committee (PNU-IACUC) on 01 February 2018 (Approval Number: PNU-2018-1800). To investigate the ability of NO/GP to form hydrogel in situ, macroscopic images of wounds treated with NO/GP were taken at each time point. Briefly, imprinting control region (ICR) mice (7 weeks old, male, Samtako Bio Korea) were purchased and acclimated for 7 days. To induce anesthesia, 0.5–0.6 mg/g of avertin (tribromoethanol) were administered intraperitoneally. Then, hair was removed from the dorsal side of the mouse by electric trimmers and hair removal cream (Veet for sensitive skin, Reckitt Benckiser, France). After hair removal, a full-thickness wound was created on the dorsal area of the mouse via an 8 mm diameter disposable biopsy punch (Kai medical, Japan). Macroscopic images were taken at each time point following treatment of the wound with 28.5 mg NO/GP.

#### 2.6.2. Fluid Uptake Ability

The ability of NO/GP to uptake fluid was measured as previously described, with some modifications [26,32]. A three-station Franz diffusion cell apparatus (PermeGear, Inc., Hellertown, PA, USA) was used to measure the water uptake ability. A regenerated cellulose membrane (pore size = 0.45 μm) was placed between the donor and receiver compartments. In the receiver compartment, SWF was filled and thermostated at 37 ◦C. After SWF was loaded into the receiver compartment, 40 mg of NO/GP were placed on the regenerated cellulose membrane (donor compartment). The amount of SWF was maintained at 8 mL. At each time point, the weight of the donor compartment was measured to calculate the amount of absorbed fluid.

#### 2.6.3. NO Release from NO/GP

The NO release from NO/GP was calculated by measuring the GSNO decomposition. The amount of GSNO remaining was determined using a UV/Vis spectrophotometer at a wavelength of 335 nm. Fifty milligrams of NO/GP powder were placed in a 2 mL microtube, and different amounts of SWF were added to mimic swelling (NO/GP that absorbed 200%, 350%, and 500% of SWF per weight). All microtubes were placed in a 37 ◦C incubator. At the set time points, the remaining GSNO was measured by determining the absorbance of the supernatant at 335 nm after dilution and centrifugation. The NO released from NO/GP at each time point ([NO]t) was calculated using an Equation (1) ([GSNO]0: Initial GSNO concentration; [GSNO]t: GSNO concentration at each time point) [25,33].

$$\text{[NO]}\_{\text{t}} = \text{[GSNO]}\_{\text{0}} - \text{[GSNO]}\_{\text{t}} \tag{1}$$

#### *2.7. Antibacterial Assay*

The bactericidal effect of NO/GP was evaluated against *P. aeruginosa* PAO1 (wild-type prototroph) [34] and MRSA (USA 300) [35]. Each pathogen was incubated overnight in TSB at 37 ◦C, and the bacterial suspension was adjusted with TSB media to approximately 108 colony forming unit (CFU)/mL until the optical density at 600 nm reached 0.15–0.2 (0.5 of the McFarland scale) [36]. The adjusted bacterial suspension (100 μL) was inoculated into each tube. Then, 28.5 mg of NO/GP and GP were added to each well and incubated for 24 h at 37 ◦C. To calculate the number of living bacterial cells, tubes were diluted with an additional 1.9 mL of the TSB medium. After serial dilution, 100 μL of each aliquot were plated on the TSB agar and incubated for 24 h at 37 ◦C. CFUs were determined by counting the colonies on the agar plates after incubation. To visualize the antibacterial activity of NO/GP, bacteria treated with or without GP and NO/GP were stained with SYTO9 (Thermo Fisher Scientific, Waltham, MA, USA) (final concentration was 66.8 μM) for 15 min. After incubation, bacteria were collected by centrifugation at 3000 *g* for 10 min. Each sample was washed three times and resuspended in 5 mL of normal saline. Green fluorescence from stained bacteria was imaged by an in vivo imaging system (FOBI, Neoscience, Suwon, Korea). For confocal laser scanning microscopy, bacteria were washed three times with normal saline after 24 h incubation with or without GP and NO/GP. Then, the bacteria were stained with SYTO 9 dye and propidium iodide (LIVE/DEAD® *Bac*Light™ bacterial viability kit) according to the manufacturer's protocol. Images were obtained at 20× magnification using the LSM 800 (Carl Zeiss, Oberkochen, Germany).

#### *2.8. In Vivo Wound Healing Study in a Bacteria-Challenged Full Thickness Wound Model*

#### 2.8.1. Evaluation of Wound Size Reduction

In this study, mouse models of *P. aeruginosa-* and MRSA-challenged full thickness wounds were used to evaluate the ability of NO/GP to heal infected wounds. For *P. aeruginosa*-challenged wound healing study, 8 mm-sized wounds were created using above-mentioned method. Then, 109 CFU of *P. aeruginosa* suspension was inoculated at the wound site. Each wound was covered with Tegaderm film and fastened by surgical tape (Durapore™, 3M) for protection. Mice were then incubated for 2 days with no treatment for wound infection. Two days after wounding, mice were treated with 28.5 mg of NO/GP and GSNO-free NO/GP every 2 days. Untreated mice were used as a control group (changing Tegaderm film and surgical tape only). Photographs were obtained every 2 days, and the size of the wound was analyzed by ImageJ software. For the MRSA-challenged wound healing study, a 2 <sup>×</sup> 10<sup>6</sup> CFU of MRSA suspension was used instead of *P. aeruginosa,* and the procedure described above was followed.

#### 2.8.2. Quantification of *P. aeruginosa* at the Wound Site

In the *P. aeruginosa*-challenged wound study, bacteria were quantified at the wound as previously reported, with some modifications [37]. Briefly, wound samples were harvested from representative mice on predetermined days with an 8 mm diameter biopsy punch. Each sample was placed in 1 mL of PBS, chopped, and sonicated to detach bacteria from the tissue samples. Then, 100 μL of each tissue-bacteria suspension was plated on an agar plate after serial dilution (1:10). To quantify the amount of *P. aeruginosa* at the wound site, a cetrimide-agar plate was used as a pseudomonas-selective media [38]. CFUs were determined by counting the colonies on the agar plates after 24 h of incubation.

#### 2.8.3. Histological Examination

In the *P. aeruginosa*-challenged wound healing study, mice were euthanized 14 days after the initiation of drug treatment, and each wound site was sampled with an 8 mm diameter biopsy punch. Each sample was immediately immersed in 10% buffered formalin for fixation. Fixed wound samples were placed in paraffin blocks, sectioned to obtain 5 μm wound samples, and prepared for hematoxylin and eosin (H&E) staining, Twort's Gram staining, and Masson's trichrome staining. Each

staining procedure was performed according to the manufacturer's protocol with some modifications. After staining, each slide was photographed using a light microscope at 20× magnification for H&E and Masson's trichrome staining, as well as 100× for Twort's gram staining.

#### *2.9. Statistical Analysis*

The statistical analysis was performed using a one-way analysis of variance (ANOVA) with a Bonferroni posttest in GraphPad Prism 5.0 (GraphPad Software, Inc., La Jolla, CA, USA). *p*-values less than 0.05 were considered statistically significant.

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

#### *3.1. Powder Characterization*

NO/GP was successfully prepared by micronizing, sieving, and blending GSNO, sodium alginate, pectin, and PEG. To examine the homogeneous fabrication of NO/GP, particle size distribution was analyzed using more than 100 particles, and GSNO contents were measured from 10 different samples. As shown in Figure 1A, pink GSNO particles were homogeneously dispersed in NO/GP. After micronizing, the average particle size and standard deviation decreased (from 85.2 ± 47.3 to 51.5 ± 22 μm, respectively) (Table 1, Figure 1A,B). In addition, the RSD of GSNO contents was reduced (16.06–4.77) due to the removal of large GSNO particles which interfere with content homogeneity (Table 1, Figure 1C). Flowability was measured to investigate the movement of NO/GP at the wound site before conversion to hydrogel. The flowability of in situ hydrogel-forming powders should be low because highly flowable powders can be easily cleared from the wound site prior to hydrogel formation. In this experiment, the flowability of NO/GP was expressed by the Hausner ratio. After the blending and micronizing process, the Hausner ratio of NO/GP increased from 1.34 to 1.97 (Table 1). Due to the decreased particle size, flowability decreased after micronization. In general, powders with a Hausner ratio exceeding 1.6 possess extremely poor flowability [39]. Low flowability can prevent the removal of the powder and is essential for the in situ hydrogel-forming powder system; the powder should remain on the wound site until it is converted to a hydrogel. For this reason, the Hausner ratios of previously developed powder dressings are around 1.7 [18,19,26], as this ratio allows the powders to resist the flow. Therefore, NO/GPs with a Hausner ratio of 1.97 could also resist flow away from the wound site before hydrogel formation.

**Table 1.** Content homogeneity and flowability of in situ hydrogel-forming/nitric oxide (NO)-releasing powder dressing (NO/GP).


Particle size was analyzed by ImageJ software from microscopic images (*n* > 100). S-nitrosoglutathione (GSNO) contents and relative standard deviation (RSD) were calculated from 10 samples of NO/GP. The Hausner ratio was calculated from the tap/bulk density ratio (*n* = 3).

**Figure 1.** Characterization of NO/GP in powder form. Microscopic images (**A**), particle size distribution (*n* > 100) (**B**), and S-nitrosoglutathione (GSNO) contents of non-micronized mixture and NO/GP (*n* = 10) (**C**). Red arrows indicate GSNO particles (pink particles).

#### *3.2. Rheological Properties of NO*/*GP in Hydrogel Form*

Rheological properties, including viscosity, storage modulus, and loss modulus, were evaluated to investigate the adhesiveness of NO/GP in hydrogel form against various steady shear strains or oscillatory shear strains. Adhesiveness is an important factor of hydrogel dressings because hydrogels should adhere to the damaged site to protect the wound and maintain humid conditions, which is an essential role of hydrogel dressings. Figure 2A,B show the dependence of shear stress and steady shear viscosity on shear rate for NO/GP following the absorption of 200%, 350%, and 500% SWF per weight. In all steady rate-sweep tests, the shear stress tended to level off and approach a limiting constant value (usually referred to as "yield stress") as the shear rate approached zero. Yield stress plays an important role in predicting the adhesiveness of semi-solid formulations because stress is related to the level of internal structures that can exhibit resistance to flow. Steady shear viscosity decreased sharply as the shear rate increased, indicating that NO/GP exhibited a marked non-Newtonian shear-thinning flow behavior. In a previous study, petroleum jelly, which is a widely used and marketed product, demonstrated similar behavior in the 350% swollen condition, and the values of shear stress and steady shear viscosity were also similar to those of the swollen condition [40]. These results indicate that the NO/GP hydrogel could maintain a gel-like structure and resist small shear stress such as gravitational force or brushing against clothes. Figure 2C,D,E indicate the storage modulus, *G'*, and loss modulus, *G"*, as a function of strain amplitude under three different swelling conditions of NO/GP hydrogels with a fixed angular frequency of 10 rad/s. The storage modulus was found to be larger than the loss modulus within a relatively smaller strain amplitude, indicating that the rheological behavior in this region is dominated by an elastic (solid-like) rather than a viscous (liquid-like) property. However, as the strain amplitude gradually increased, viscous behavior became superior to elastic behavior because the storage modulus demonstrated a sharper decrease with increasing strain amplitude compared with the loss modulus. These results indicate that adhesiveness to a relatively large imposed deformation (such as scrubbing motion) is weakened; therefore, the NO/GP hydrogels could easily flow and be removed from the wound site.

**Figure 2.** Rheological properties of NO/GP in hydrogel form. Shear stress and viscosity (**A** and **B**, respectively) of NO/GP following absorption of 200%, 350%, and 500% of simulated wound fluid (SWF) per weight as a function of shear rate. Storage modulus (*G'*) and loss modulus (*G"*) of each swollen condition as a function of strain amplitude (**C** and **D**, respectively). *G'* and *G"* of 350% swollen NO/GP hydrogel as a function of strain amplitude (**E**). Strain-sweep tests were performed at a fixed angular frequency of 10 rad/s.

#### *3.3. Storage Stability*

The storage stability of NO/GP was evaluated by detecting GSNO decomposition under two temperature conditions (4 and 37 ◦C). Since GSNO can be easily degraded by hydrolysis, storage stability is an important factor in the development of GSNO-containing formulations. Though several GSNO-containing hydrogel dressings have been developed to accelerate wound healing [41–44], those formulations did not present long-term stability because GSNO hydrolysis is inevitable in water-containing formulations. Conversely, no significant GSNO decomposition was noted in NO/GP up to 140 days under both the 4 and 37 ◦C conditions (Figure 3). Since GSNO in the NO/GP remained in a powder state, which was a water-free condition during storage, the hydrolysis of GSNO was prevented. Thus, NO/GP could remain stable whilst in storage without any concerns relating to GSNO degradation. In addition, there were no significant changes in particle size, Hausner ration and rheological properties during the storage period (data not shown).

**Figure 3.** Storage stability of NO/GP at 4 (blue) and 37 ◦C (red). Remaining GSNO percentage compared with that of initial NO/GP was calculated to evaluate storage stability. Each point represents the mean ± standard deviation (*n* = 3).

#### *3.4. Behavior of NO*/*GP After Exposure to Wound Fluid*

#### 3.4.1. Morphological Changes in NO/GP at the Wound Site

To investigate the ability of NO/GP to form hydrogel in situ, morphological changes in NO/GP were observed following application of 28.5 mg of NO/GP to the full-thickness wounds in mice. The amount of NO/GP was sufficient to cover 1 cm<sup>3</sup> of a full-thickness wound. As shown in Figure 4A, following its application to the wound, the NO/GP powder was immediately converted to a glittering hydrogel, and more than 50% of the NO/GP powders converted to a hydrogel within 1 min. All of the NO/GP applied was converted to hydrogel within 10 min. After 10 min, no morphological changes in NO/GP were observed due to the completion of the hydrogel structure.

**Figure 4.** Behavior of NO/GP following exposure to wound fluid. Morphological changes in NO/GP applied to the mouse full-thickness wound model (**A**). Fluid uptake profile of NO/GP (amount of absorbed SWF is presented as the percent per initial NO/GP powder weight) (**B**). NO release profiles of NO/GP under different swelling conditions and the GSNO solution at the equivalent concentration to 350% swollen NO/GP (**C**). Each point represents the mean ± standard deviation (*n* = 3).

#### 3.4.2. Fluid Uptake Ability

Since hydrogel formation is initiated by the absorption of wound fluid, an investigation of fluid uptake ability is essential for evaluation of powder dressings that form hydrogels in situ. To evaluate the fluid uptake ability of NO/GP, the dressings were exposed to SWF at 37 ◦C, and the amount of absorbed fluid was calculated by measuring the change in weight of NO/GP. The amount of absorbed fluid was presented as a percentage of weight gained by fluid uptake per initial NO/GP. As shown in Figure 4B, NO/GP absorbed SWF rapidly, and around 200% of SWF was absorbed within 20 min. After the initial rapid absorption of the SWF, the rate of fluid uptake by NO/GP was decreased, and NO/GP absorbed up to 375% of SWF in 270 min. After that, no further significant fluid absorption was observed during the experiment. In the initial state, the swellable polymers in NO/GP (pectin and alginate) absorbed SWF and rapidly formed a hydrogel structure. Following hydrogel formation, SWF was slowly captured in the intermolecular space of the hydrogel structure, because hydrophilic polymers in NO/GP are able to trap SWF by hydrogen bonding between polymers and water molecules. Finally, since the intermolecular space was filled with SWF, no more fluid could be absorbed. Since NO/GP could efficiently absorb fluid, hydrogel formation and subsequent NO release was initiated rapidly.

#### 3.4.3. NO Release from NO/GP

Since NO released from GSNO in NO/GP exerts therapeutic effects, the NO release profiles of NO/GP were investigated by exposing NO/GP to SWF at 37 ◦C. The NO release from NO/GP was investigated under three swollen conditions (200%, 350%, and 500% of SWF per initial NO/GP weight) that represented an amount of low, medium and high wound exudate, respectively. As GSNO generates NO via hydrolytic cleavage of the S–N bond, the NO release from NO/GP is initiated by wound fluid. Thus, the amount of NO released from NO/GP was calculated by measuring GSNO degradation (released NO = initial GSNO − remaining GSNO). Regardless of swelling, NO/GP exhibited a linear NO release without a burst release (Figure 4C). In addition, lower levels of NO/GP swelling resulted in a slightly faster NO release rate compared with higher levels of swelling, and NO was released up to 18, 22, and 26 h in the 200%, 350%, and 500% conditions, respectively. Due to the presence of polymeric compounds (alginate, pectin, and PEG) in the NO/GP hydrogel structure, GSNO molecules or NO radicals were surrounded by polymeric molecules which restrict the diffusion of radicals in what is termed the "cage effect" [45,46]. Compared with the GSNO solution, NO/GP presented a prolonged NO-release profile, which was due to the restricted diffusion of radicals over the cage (NO was released 100% within 12 h at the GSNO concentration equivalent to 350% swollen NO/GP). The release profiles different under different levels of swelling because the rate of NO release from GSNO was affected by its initial concentration; thus, the higher initial concentration resulted in faster NO release due to the increased amount of radicals contributing to the degradation of GSNO molecules [47]. Therefore, 200% SWF added to the NO/GP group (high GSNO concentration) resulted in a faster NO release compared with the 350% and 500% groups. In addition, since NO/GP exhibited a linear NO release under all levels of swelling, this indicates that dressings can be changed any time without concerns of toxicity caused by burst-released NO.

#### *3.5. In Vitro Antibacterial Assay*

Antibacterial efficacy is the most important characteristic of a dressing for the treatment of infected wounds. To investigate the antibacterial activity of NO/GP, an in vitro antibacterial assay was performed using the CFU method against MRSA and *P. aeruginosa*, which are representative drug-resistant gram-positive and -negative bacteria. Following incubation for 24 h with or without NO/GP in TSB media, a 6-log reduction in bacterial CFUs was observed in the NO/GP-treated group compared to the GP-treated group against both MRSA and *P. aeruginosa* (Figure 5A). After CFU examination, the antibacterial activity of NO/GP was visualized by staining with SYTO 9, which is a green fluorescence dye that can stain bacterial DNA. Since only living bacteria were collected by centrifugation, green fluorescence indicated the presence of live bacteria. As shown in Figure 5B, distinct green fluorescence was detected in the untreated and GP-treated groups for both MRSA and *P. aeruginosa*. Furthermore, signals from the NO/GP-treated group were significantly reduced, owing to the high number of bacteria killed by NO in both the MRSA and *P. aeruginosa* groups. The antibacterial effect of NO/GP was also examined via confocal microscopy using the LIVE/DEAD® *Bac*Light™ bacterial viability kit. Since propidium iodide can only penetrate damaged bacterial membranes, living bacteria were stained with SYTO 9 (green fluorescence) and damaged bacteria were stained with propidium iodide (red fluorescence). As shown in Figure 5C, confocal images of the GP-treated and untreated groups exhibited distinct green fluorescence, while those of the NO/GP-treated group exhibited strong red fluorescence. This indicates that most of the bacteria survived in the GP-treated and untreated groups, while few bacteria survived in the NO/GP-treated group. Since NO possesses broad-spectrum antibacterial effects and NO/GP can release NO in a sustained manner, these results indicate that NO/GP exhibited significant bactericidal activity against both gram-negative *P. aeruginosa* and gram-positive MRSA without bacterial re-growth for 24 h. Broad-spectrum antibacterial and potent bactericidal effects against drug-resistant bacteria are essential for the treatment of infected wounds, since the infection of cutaneous wounds by drug-resistant bacteria is increasing and it is hard to immediately distinguish bacterial species. Moreover, the multiple antibacterial mechanisms of NO may prevent the emergence of NO-resistant bacteria [16]. Thus, these findings indicate that NO/GP possesses desirable antibacterial properties and may be beneficial for the treatment of infected wounds.

#### *3.6. In Vivo Wound Healing Study*

#### 3.6.1. Evaluation of Wound Size Reduction Effect

The therapeutic effects of NO/GP were evaluated in mice using the bacteria-challenged full-thickness wound model. The acceleration of infected wound recovery with NO/GP was evaluated by observing morphological changes in the wound and measuring wound size change every 2 days. In both *P. aeruginosa*- and MRSA-challenged full-thickness wound models, the NO/GP treatment resulted in a significant reduction in wound size compared with GP treatment and no treatment after 4 days (Figure 6). In NO/GP-treated groups, wound size was reduced to less than 20% of the initial size 14 and 8 days after treatment initiation in the *P. aeruginosa*- and MRSA-challenged models, respectively. Conversely, in the GP treated groups, no significant acceleration of wound healing was observed in either the *P. aeruginosa*- or MRSA-challenged models compared to the untreated groups. Accelerated wound healing in the NO/GP groups can be attributed to the action of NO released from GSNO in NO/GP [22]. In particular, broad and potent antibacterial effects could effectively eradicate infection with gram-positive or -negative bacteria [48,49]. Moreover, NO facilitates wound healing by promoting fibroblast proliferation, collagen formation, and tissue remodeling [13,50]. Therefore, NO/GP may facilitate wound healing in *P. aeruginosa*- and MRSA-challenged full-thickness wounds in mice.

**Figure 5.** Antibacterial effects of NO/GP. Antibacterial effects of NO/GP against *Pseudomonas aeruginosa* and methicillin-resistant *Staphylococcus aureus* (MRSA) by the colony forming unit (CFU) method. The results are presented as the mean ± standard deviation (*n* = 3) (**A**). Visualization of bacteria by SYTO9 staining. Green fluorescence indicates the presence of live bacteria (**B**). Confocal laser scanning microscopic images of cells treated and untreated with GP and NO/GP. Green fluorescence indicates live bacteria and red fluorescence indicates dead bacteria. Images were taken at 20× magnification. Scale bar represents 50 μm (**C**).

**Figure 6.** Representative macroscopic images of *P. aeruginosa*-challenged full-thickness wounds in mice 0, 6, and 14 days after treatment initiation (**A**). Changes in wound size in mice with *P. aeruginosa*-challenged full-thickness wounds (**B**). Representative macroscopic images of MRSA-challenged full-thickness wounds in mice 0, 6, and 14 days after treatment initiation (**C**). Changes in wound size in mice with MRSA-challenged full-thickness wounds (**D**). Values are presented as the mean ± standard deviation (*n* = 4).

#### 3.6.2. Quantification of P. aeruginosa at the Wound Site

To investigate the in vivo antibacterial effects of NO/GP, wound samples were harvested 2, 8, and 14 days after treatment initiation, and CFUs were assessed with *Pseudomonas*-selective agar plates, which exclude other bacterial species. As shown in Figure 7A, there was no decrease in the number of *P. aeruginosa* 2 days after treatment initiation; however, on day 8, significant bactericidal effects were observed in the NO/GP-treated group (around 3-log CFU reduction). Only 4.1 and 4.8 CFU/cm2 *P. aeruginosa* were observed in the NO/GP-treated group, while 6.7 and 7.6 CFU/cm2 and 6.4 and 6.5 CFU/cm<sup>2</sup> *P. aeruginosa* were observed in untreated and GP-treated groups 8 and 14 days after treatment initiation, respectively. To visualize the antibacterial effects in vivo, Twort's Gram-staining was performed. Because *P. aeruginosa* is a rod-shaped gram-negative bacteria 1–2 μm in size, it can be detected by Twort's gram staining in tissue samples containing more than 105 CFU (low levels of bacteria are hard to detect by Gram staining) [51,52]. Fourteen days after treatment initiation, rod-shaped, brown-colored bacteria (*P. aeruginosa*) were observed in GP-treated and -untreated groups in the damaged epidermal region, indicating that at least 105 CFU *P. aeruginosa* was present in the samples (Figure 7B). Conversely, no bacteria were observed in the NO/GP-treated group or in the healthy control. Since NO released from NO/GP was able to efficiently eradicate *P. aeruginosa* from infected wound sites, wound healing may have occurred subsequent to inflammation. Furthermore, high numbers of *P. aeruginosa* in GP-treated and untreated groups resulted in consistent inflammation and, consequently, in impaired re-epithelization.

**Figure 7.** Bacterial quantification at the wound site (*n* = 3) (**A**). Twort's gram staining of wound samples from mice treated with or without GP or NO/GP. The epidermal region of the tissue samples was imaged with a microscope at a magnification of 100×. Arrows indicate *P. aeruginosa* (1–2 μm sized, rod-shaped, and brown). Scale bar represents 10 μm (**B**).

#### 3.6.3. Histological Examination

Tissue regeneration and collagen synthesis in full thickness wounds challenged with *P. aeruginosa* were evaluated by H&E and Masson's trichrome staining. Fourteen days after the initiation of drug treatment, more organized skin morphology and higher collagen abundance were observed in the NO/GP-treated group compared with the GP-treated and untreated groups (Figure 8). Well-differentiated epidermis was observed in the NO/GP-treated group, whilst damaged epidermis was observed in the GP-treated and untreated groups following H&E staining. In addition, skin cells, such as keratinocytes and fibroblasts, were abundant in the NO/GP-treated group. Conversely, granulation and large numbers of immune cells were observed in GP-treated and untreated groups. The amount of collagen in wound samples was visualized by Masson's trichrome staining (blue color indicates collagen). As shown in Figure 8, samples from the NO/GP-treated group exhibited a prominent blue color similar to that of healthy skin tissue. However, GP-treated and untreated groups exhibited less collagen in the dermis region. Since inflammation was ongoing in these groups, the collagen synthesis and tissue remodeling processes were inhibited, resulting in delayed wound healing.

**Figure 8.** H&E (hematoxylin and eosin) and Masson's trichrome straining of wound tissues from *P. aeruginosa*-challenged full-thickness wound mice threated with or without GP or NO/GP for 14 days. A: Adipose tissue; C: Cell debris; E: Epidermis; F: Fibrous tissue; G: Granulation tissue; H: Hair follicle; I: Immune cells; M: Muscle. Scale bar represents 200 and 100 μm for the H&E and Masson's trichrome images, respectively.

#### **4. Conclusions**

In this study, we successfully developed an in situ hydrogel-forming/NO-releasing wound dressing (NO/GP) composed of alginate, pectin, PEG, and GSNO, with a controlled NO release property and good storage stability for the effective treatment of infected wounds (Figure 9). Since NO/GP maintained a water-free powder form until use on the wound, the degradation of GSNO in NO/GP was prevented for more than 3 months when stored at 4 and 37 ◦C. When applied to wounds, NO/GP absorbed up to 350% of wound fluid and was quickly transformed from a dry powder to an adhesive hydrogel. Simultaneously, a NO release was triggered by absorbed wound exudates, followed by a sustained NO release over 24 h without an initial burst release. Rheological studies indicated that the hydrogel structure of NO/GP exhibited sufficient adhesiveness to remain stable on the wound surface. The results of an in vitro antibacterial study demonstrated that NO/GP leads to a 6-log reduction in MRSA and *P. aeruginosa* over 24 h. Finally, in vivo antibacterial effects and accelerated wound healing were observed in mice with infected wounds treated with NO/GP. These results suggest that the in situ hydrogel-forming/NO releasing formulation presented in this study can be fabricated by a simple and cost-effective manufacturing process and thus would be a promising alternative to dressings for the treatment of infected wounds.

**Figure 9.** Schematic illustration of NO/GP for the treatment of infected wounds.

**Author Contributions:** Conceptualization, J.L. and J.-W.Y.; methodology, J.L., N.H., S.P.H., H.-J.A. and J.C.; formal analysis, J.L. and H.-J.A.; investigation, J.L.; writing—original draft preparation, J.L.; writing—review and editing, J.L., K.-W.S. and J.-W.Y.; supervision, J.-W.Y.; project administration, J.-W.Y.

**Funding:** This research was supported by a grant from the Korean Healthcare Technology R&D Project, Ministry for Health and Welfare Affairs, Republic of Korea (HI15C2558) and by a Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2019R1I1A3A01057849).

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

#### **References**


© 2019 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Dermal Delivery of the High-Molecular-Weight Drug Tacrolimus by Means of Polyglycerol-Based Nanogels**

**Fiorenza Rancan 1,\*, Hildburg Volkmann 1, Michael Giulbudagian 2, Fabian Schumacher 3,4, Jessica Isolde Stanko 1, Burkhard Kleuser 3, Ulrike Blume-Peytavi 1, Marcelo Calderón 2,5,6 and Annika Vogt <sup>1</sup>**


Received: 30 June 2019; Accepted: 31 July 2019; Published: 5 August 2019

**Abstract:** Polyglycerol-based thermoresponsive nanogels (tNGs) have been shown to have excellent skin hydration properties and to be valuable delivery systems for sustained release of drugs into skin. In this study, we compared the skin penetration of tacrolimus formulated in tNGs with a commercial 0.1% tacrolimus ointment. The penetration of the drug was investigated in ex vivo abdominal and breast skin, while different methods for skin barrier disruption were investigated to improve skin permeability or simulate inflammatory conditions with compromised skin barrier. The amount of penetrated tacrolimus was measured in skin extracts by liquid chromatography tandem-mass spectrometry (LC-MS/MS), whereas the inflammatory markers IL-6 and IL-8 were detected by enzyme-linked immunosorbent assay (ELISA). Higher amounts of tacrolimus penetrated in breast as compared to abdominal skin or in barrier-disrupted as compared to intact skin, confirming that the stratum corneum is the main barrier for tacrolimus skin penetration. The anti-proliferative effect of the penetrated drug was measured in skin tissue/Jurkat cells co-cultures. Interestingly, tNGs exhibited similar anti-proliferative effects as the 0.1% tacrolimus ointment. We conclude that polyglycerol-based nanogels represent an interesting alternative to paraffin-based formulations for the treatment of inflammatory skin conditions.

**Keywords:** tacrolimus formulation; nanogels; skin penetration; drug delivery; human excised skin; Jurkat cells

#### **1. Introduction**

Tacrolimus, also known as FK 506, is an immunosuppressive drug acting predominantly via inhibition of T cell proliferation. This macrolide molecule builds a complex with the FK binding protein, which in turn binds to calcineurin and inhibits the dephosphorylation of the nuclear factor of activated T cells (NFATs) and its translocation to the nucleus [1–3]. As a consequence, the expression of IL-2, which is required for T cell proliferation, is inhibited. Tacrolimus also binds to isoforms of the FK binding protein, which have a cell type-specific expression pattern. This results in inhibitory or toxic effects towards other types of cells, e.g., mast and Langerhans cells [4–6]. In addition, induction of T cell

apoptosis was found by Hashimoto and co-workers [7] and confirmed by Chung et al., who showed that tacrolimus induced apoptosis in Jurkat cells via activation of caspases 3 and 12 [8]. Furthermore, interference with the activation of the mitogen-activated protein kinase (MAPK) signaling pathways in primary human T lymphocytes was described by Matsuda et al. [9].

While the compound is widely used as topical agent to treat certain inflammatory skin diseases, its limited penetration across the skin barrier is a major limitation for a wider clinical use. Because of the poor solubility, the marketed formulation (Protopic®) consists of a paraffin-based ointment containing 0.03% or 0.1% tacrolimus in hard, liquid and white soft paraffin. In addition to insufficient drug penetration rates, mis-sensations such as skin stinging, burning and pruritus at the application site frequently lead to treatment discontinuation [10]. These side effects probably result from drug-induced chemical irritation and release of pro-inflammatory cytokines such as IL-6, as demonstrated for UVB-activated keratinocytes [11] and fibroblasts [12].

The hydrophobic paraffins in the ointment can reduce skin water loss and have therefore moisturizing properties. However, besides the fact that greasy excipients may be unpleasant for daily skin care, increasing concerns have been expressed because of symptom exacerbation in eczema sufferers. Furthermore, paraffin is a flammable excipient and consumers are often not aware of this fact [13]. Thus, several aspects indicate that new alternative formulations could result in significant improvements of current therapeutic strategies. Consequently, new formulations for the delivery of tacrolimus into the skin have been investigated in the last years, including lipid-based nanocarriers [14], liposomes [15], micelles [16–18], or transferosomes [19]. In all these works, the investigated nanocarrier formulation improved skin penetration of tacrolimus with respect to the marketed formulation.

Almost all studies used lipid-based carriers or delivery systems formulated with surfactants or permeabilizing components like ethanol that may cause side effects when applied on inflamed skin. In contrast, in this study, we investigated an alternative formulations based on thermoresponsive nanogels (tNGs) made of dendritic polyglycerol, which is cross-linked with thermoresponsive polymers, namely poly(glycidyl methyl ether-co-ethyl glycidyl ether) (tPG) and poly(*N*-isopropylacrylamide) (pNIPAM). During inflammatory skin diseases, skin temperature is slightly enhanced compared to unaffected skin. The concept of thermoresponsive nanogels implies sensing of such temperature differences with preferential release of drug in diseased skin. In fact, the progressive temperature increment after topical application has been shown to influence nanogel softness and the release of the drug [20]. More specifically, nanogels are soft when applied on skin surface and can penetrate deep in the stratum corneum (SC), where local temperature increase induces a change of conformation that results in the release of the incorporated active compound. The investigated nanogels were shown to enhance SC hydration [21–23] and exhibited promising drug delivery properties [24] along with low cytotoxicity [25,26].

In most previous studies, tacrolimus penetration was investigated using animal skin [18,27], which is anatomically and immunologically different from human skin, with little consideration of skin barrier impairment or skin inflammatory status. In this study, we used excised human skin from two distinct anatomical regions, breast as well as abdomen, and different methods to induce barrier disruption and inflammatory reactions. In this way, we could monitor the effects of barrier-disruption on tacrolimus skin penetration and release of inflammatory markers like IL-6 and IL-8. The additional use of a trans-well set up to co-culture ex vivo human skin and Jurkat cells enabled the assessment of tacrolimus anti-proliferative effects.

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

#### *2.1. Nanogel Preparation and Characterization*

Commercially available chemicals from standardized sources were used as delivered. Solvents were purchased as reagent grade and distilled if necessary. Anhydrous solvents were either purchased as ultra-dry solvent (Acros Organics®, Geel, Belgium) or received from solvent purification system. For the tPG polymerization reactions, dry toluene was obtained from MBRAUN SPS 800 solvent purification system (Garching, Germany). Water was purified by Millipore water purification system. Dendritic polyglycerol with average molecular weight of 10 kDa (PDI = 1.27) was purchased from Nanopartica GmbH (Berlin, Germany). Glycidyl methyl ether (85%) and ethyl glycidyl ether (98%; both TCI Europe, Eschborn, Germany) were dried over CaH2, distilled, and stored over molecular sieves (5 Å). The crosslinking reagent (1R,8S,9s)-bicyclo[6.1.0]non-4-yn-9-ylmethyl (4-nitrophenyl) carbonate was purchased from Synaffix (Oss, the Netherlands). The cyanine dye indodicarbocyanine (IDCC) was purchased from Lumiprobe GmbH (Hannover, Germany).

The tNGs were synthesized according to previously reported methods. For detailed synthesis and characterization description, refer to the following publications [22,26,28]. Briefly, for the synthesis of poly(*N*-isopropylacrylamide) (pNIPAM)-based nanogels (NG-pNIPAM), NIPAM (66 mg), acrylated dendritic polyglycerol (dPG-Ac10%) (33 mg), sodium dodecyl sulfate (SDS) (1.8 mg), and ammonium persulfate (APS) (2.8 mg) were dissolved in 5 mL of distilled water. Argon was bubbled into the reaction mixture for 15 min, which was followed by stirring under argon atmosphere for another 15 min. The reaction mixture was transferred into a hot bath at 68 ◦C and polymerization was activated after 5 min with the addition of a catalytic amount of *N*,*N*,*N* ,*N* -tetramethylethane-1,2-diamine (TEMED; 120 μL). The mixture was stirred at 500 rpm for at least 4 h, prior to purification by dialysis. For the synthesis of fluorescently labeled nanogels, a mixture of unlabeled dPG-Ac10% (30 mg) and indodicarbocyanine (IDCC)-labelled dPG-Ac10% (3 mg) was used.

For the synthesis of tPG-based nanogels (NG-tPG), dPG functionalized with (1R,8S,9s) bicyclo[6.1.0]non-4-yn-9-ylmethyl carbonate (dPG-BCN8%) (10 mg) and di-azide functionalized tPG (tPG-(N3)2) (20 mg) were mixed in 1 mL of dimethylformamide (DMF), cooled in an ice bath and injected with a syringe into 20 mL of water at 45 ◦C. The mixture was stirred for 3 h and the unreacted alkynes were quenched with azidopropanol or alternatively with IDCC-N3.

For loading tacrolimus, highly concentrated tNGs (10 mg) were added to a suspension of tacrolimus (5 mg) in 2 mL of MilliQ water. The nanogels were let to swell, followed by sonication in an ice water bath for 30 min. The suspension was stirred overnight at 25 ◦C and the encapsulated fraction was separated from the free drug by filtration with a 0.45 μm regenerated cellulose syringe filter. Following, 100 μL of the encapsulated fraction were lyophilized for determination of the total concentration, while the encapsulation capacity was determined by LC-MS/MS measurements. The nanogels with the encapsulated drugs were stored at 4 ◦C. The characterization of the investigated nanogels is reported in Table 1.



performed in triplicates; intensity average mean value presented; <sup>b</sup> Cloud point temperature determined as the temperature at 50% transmittance by UV-Vis (λ = 500 nm); <sup>c</sup> Refer to Gerecke, et al., *Nanotoxicology* **2017**, *11*, 267–277 [26] for detailed information; <sup>d</sup> ζ potential determined by Zeta-sizer in phosphate buffer. Measurements were performed in triplicates; intensity average mean value presented. NG-pNIPAM: poly(*N*-isopropylacrylamide-based nanogels; NG-tPG: poly(glycidyl methyl ether-co-ethyl glycidyl ether)-based nanogels.

#### *2.2. Skin Samples*

Breast and abdominal skin were obtained after informed consent from healthy donors undergoing plastic surgery. The study was conducted after approval by the Ethics Committee of the Charité–Universitätsmedizin Berlin (approval EA1/135/06, renewed on January 2018) and in accordance with the Declaration of Helsinki guidelines. The excised skin was used between 3 and 6 h after surgery and was examined to exclude injured parts, including macroscopic skin surface damages but also deeper structural defects like stretch marks. Subcutaneous fat tissue was partially removed (approximately 0.5 cm was kept) and skin was stretched and fixed on a Styrofoam block

using needles. Skin areas of 2 cm<sup>2</sup> were marked. In order to induce skin barrier disruption, three methods were used: tape stripping 50 times (50 × TS), laser poration (LP), and sodium lauryl sulfate (SLS). TS was performed with adhesive, polyethylene tape (19-mm diameter, TESA film no. 5529, Beiersdorf, Hamburg, Germany). For LP, a P.L.E.A.S.E® Professional system (Pantec Biosolutions, Ruggell, Lichtenstein) was used with the following setting: delivered energy 73.1 J/cm2, pulse length 125 μs, repetition rate 300 Hz, 10 pulses per pore, treated surface 10 × 10 mm, pore density 8%. For SLS treatment, 20 μL/cm2 of 5 % SLS (*w*/*v*) in deionized distilled water were applied on a filter paper disc (12 mm2 diameter for Finn Chambers®, SmartPractice, Hillerød, Denmark) and incubated for 4 h at 37 ◦C. Thereafter, skin was carefully cleaned with a paper towel.

A total amount of 5 μg tacrolimus per square centimeter was applied either incorporated in an aqueous suspension of nanogels or as 0.1% marketed formulation (Protopic®, manufactured by LEO Laboratories Ltd., Dublin, Ireland). Safety margins of at least 0.5 cm were left. Controls consisted of skin treated with 20 μL/cm<sup>2</sup> of sterile 0.9% NaCl solution. Samples and controls were placed in humid chambers (plastic boxes with lid filled with humidified towels) and incubated for 24 h at 37 ◦C, 5% CO2 and 100% humidity. After incubation, non-penetrated material was removed with a paper towel, the untreated safety margins were cut and the treated tissue blocks were plunge frozen in liquid nitrogen and stored at −80◦C for further processing.

#### *2.3. Preparation of Skin Extracts*

In order to separate epidermis from dermis and prepare extracts for tacrolimus and cytokine analyses, skin samples were cut horizontally (thickness 50 μm) with a microtome (Frigocut 2800 N, Leica, Bensheim, Germany): the first 100 μm corresponding roughly to epidermis and the remaining 900 μm to dermis. The sections were put in 500 μL of extraction buffer (100 mM Tris-HCl; 150 mM NaCl; 1 mM EDTA; 1 g Triton-X-100; 10% EtOH), homogenized (GLH OMNI homogenizer, Kennesaw, GA, USA) for 10 s, and incubated on ice for 45 min. Then, samples were sonicated at 4 ◦C for 10 min, vortexed and centrifuged for 5 min at 450× *g*. The pellets were then added of 500 μL of extraction buffer and the above described procedure was repeated once more. The supernatants were pooled and stored at −80◦C.

#### *2.4. Determination of Tacrolimus in Skin Extracts by Isotope-Dilution Liquid Chromatography Tandem-Mass Spectrometry (LC-MS*/*MS)*

Skin extracts were thawed at room temperature, spiked with stable-isotope labeled internal standard [13C1,D4]tacrolimus (Alsachim, Illkirch-Graffenstaden, France), vortexed, and centrifuged at 9300× *g* for 3 min (4 ◦C). Analyses were conducted with an Agilent 1260 Infinity LC system coupled to an Agilent 6490 triple quadrupole-mass spectrometer (both from Waldbronn, Germany) interfaced with an electrospray ion source operating in the positive ion mode (ESI+). Chromatographic separation was carried out using an Agilent Zorbax SB-C18 column (1.8 μm, 2.1 × 50 mm). Aqueous ammonium formate (20 mM, pH 3.5) and methanol (VWR, Darmstadt, Germany) were used as eluents A and B, respectively. Samples (5 μL) were injected into a mobile phase consisting of 90% eluent A. Tacrolimus and its internal standard [13C1,D4]tacrolimus were eluted from the column, which was tempered at 30 ◦C, with a 4-min linear gradient to and a subsequent isocratic stage for 5 min at 2:98 (*v*:*v*) eluent A/B at a flow rate of 0.35 mL/min. Tacrolimus and [13C1,D4]tacrolimus co-eluted from the separation column at 5.3 min. The total run time for one analysis was 13 min, including re-equilibration of the LC system. The following ion source parameters were taken from Koster et al. [29], who used a comparable mass spectrometric configuration for detection of four immunosuppressants including tacrolimus: drying gas temperature = 200 ◦C, drying gas flow = 13 L/min of nitrogen, sheath gas temperature = 200 ◦C, sheath gas flow = 12 L/min of nitrogen, nebulizer pressure = 18 psi, capillary voltage = 4500 V, and nozzle voltage = 0 V. The ion funnel parameters were: high pressure RF voltage = 150 V and low pressure RF voltage = 60 V. Quantification of tacrolimus in relation to the internal standard [13C1,D4]tacrolimus was carried out using the multiple reaction monitoring (MRM) approach. Ammonium adducts [M+NH4] <sup>+</sup> were selected as precursor ions by the first quadrupole. The following mass transitions were recorded (optimized collision energies in parentheses): tacrolimus: *m*/*z* 821.5 → 786.5 (16 eV), *m*/*z* 821.5 <sup>→</sup> 768.5 (20 eV), *m*/*z* 821.5 <sup>→</sup> 576.2 (24 eV); [13C1,D4]tacrolimus: *m*/*z* 826.5 <sup>→</sup> 791.5 (16 eV), *m*/*z* 826.5 → 773.6 (20 eV), *m*/*z* 826.5 → 581.4 (24 eV). Thereby, the loss of two hydroxyl groups and ammonium from the precursor ion, represented by *m*/*z* 821.5 → 768.5 for tacrolimus and *<sup>m</sup>*/*<sup>z</sup>* 826.5 <sup>→</sup> 773.6 for [13C1,D4]tacrolimus, was used for quantification. The dwell time for each of the six mass transitions recorded was 150 ms.

#### *2.5. Enzyme-linked Immunosorbent Assay (ELISA)*

Human IL-6 and IL-8 were investigated using ELISA kits (Human IL-6 and IL-8 CytoSetTM (CHC1263, CHC1303) Invitrogen Corporation, Carlsbad, CA, USA) following the manufacturer instructions. The amounts of cytokines were normalized to total protein content measured with Pierce 660 nm Protein Assay (Thermo Fisher Scientific Inc., Rockford, IL, USA). Absorbance was measured with EnSpire® Multimode plate reader (Perkin Elmer, Akron, OH, USA).

#### *2.6. Preparation of Cryosections and Fluorescence Microscopy*

Skin samples treated with fluorescent nanogels were placed in tissue freezing medium (Leica Microsystems, Wetzlar, Germany) and plunge-frozen in liquid nitrogen. Cryosections of 5 μm thickness were prepared and observed by means of a fluorescence microscope (Olympus BX60F3, Olympus, Hamburg, Germany). The following filter combinations were used: bright pass = 545–580 nm, long pass > 610 nm for IDCC (red) and bright pass = 470–490 nm, long pass > 550 nm for FL. Pictures (magnification of 200×) of at least 20 randomly chosen skin sections per donor and skin sample were taken, and the mean fluorescence intensity of areas in the SC, viable epidermis, and dermis was calculated using the ImageJ software (Version 1.47).

#### *2.7. Isolation of Cells, Flow Cytometry, and Confocal Fluorescence Microscopy*

After incubation with nanogels, skin samples were cut in small pieces (0.2 × 0.2 cm) and incubated overnight at 4 ◦C in 2.4 U/mL dispase (Roche Applied Science, Penzberg, Germany) in order to detach epidermis from dermis. The epidermis sheets were incubated for 10 min at 37 ◦C in 5 mL of trypsin solution (0.025% trypsin and 1.5 mM CaCl2 in PBS). Dermis tissue was digested by incubation for 2 h at 37 ◦C with an enzyme cocktail made of 0.6 g/mL collagenase II (Biochrom, Berlin, Germany), 0.3 g/mL hyaluronidase (Sigma-Aldrich, Hamburg, Germany), and 1 μg/mL DNase (Roche Diagnostics, Berlin, Germany). Enzymes were stopped with RPMI-1640 cell culture medium (PAA, Heidelberg, Germany) containing 10% fetal calf serum (PAA, Heidelberg, Germany) and cells were harvested by repeated pipetting, filtering through a 70 μm cell strainer (FalconTM, Becton Dickinson, Heidelberg, Germany) and washing twice with PBS (PAA, Heidelberg, Germany). After centrifugation at 300× *g* for 10 min, cells were fixed with 4% paraformaldehyde (Sigma-Aldrich, Taufkirchen, Germany) and stored at 4 ◦C until analysis by flow cytometry (FACS Calibur, BD, Heidelberg, Germany). At least 20,000 events were collected in the selected gate. The software FCS Express (De Novo Software, Version 3.1, Glendale, CA, USA) was used for data analysis. Pictures of isolated cells were taken by means of a confocal laser microscope (LSM Exciter, Zeiss, Jena, Germany).

#### *2.8. Isolation and Culture of T cells*

T cells were isolated from dermis after digestion as described above. Dermis cell suspension was cultured overnight in RPMI 1640 medium (Gibco, Darmstadt, Germany) supplemented with 10% fetal calf serum, 100 μg/mL streptomycin, and 100 I.E./mL penicillin. Non-adherent cells were collected and red blood cells were lysed with 1% Triton X-100 in PBS for 10 min on ice. Cells were then washed with PBS and incubated with carboxyfluorescein succinimide ester (CFSE, CellTraceTM, Life Technologies, Darmstadt, Germany) for 20 min at room temperature. Cells were washed with 1% bovine serum albumin and re-suspended in supplemented RPMI 1640 medium. Cells were left untreated, incubated

with IL-2 (2.5 μg/mL) only, or incubated with IL-2 and tacrolimus in 10% ethanol or tacrolimus-loaded NG-tPG nanogels (final tacrolimus concentration of 5 μg/mL). Cell proliferation was measured after 5 days by flow cytometry. T cells were gated in forward vs. side scatter dot plots and at least 10,000 events were collected.

#### *2.9. Co-culture of Full-Thickness Human Skin and Jurkat Cells*

After topical application of tacrolimus formulations, skin samples were transferred on 8 μm-pore inserts (Cell Culture Inserts, BD Falcon™, Corning, New York, NY, USA). These were placed in a 6-well plate with 2 mL supplemented RPMI 1640 medium and 106 Jurkat cells (clone E6.1) per well. Jurkat cells had previously been stained with CFSE as described above. Skin was maintained at the air-liquid interface and co-cultured with Jurkat cells for 2 or 5 days (37 ◦C, 5% CO2). 500 μL of fresh medium were replaced every 2 days. Jurkat cells were collected and analyzed by flow cytometry.

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

#### *3.1. Tacrolimus Penetration and Inflammatory Reaction in Ex Vivo Human Skin with Intact or Disrupted Barrier*

In order to compare the new nanogel-based tacrolimus formulations to the commercially available 0.1% tacrolimus ointment, we measured the amounts of penetrated drug and released inflammatory markers (IL-6 and IL-8) in ex vivo human skin. Different methods (TS, LP, and SLS) were applied to simulate a barrier dysfunction and improve drug penetration. Laser poration (at the used energy) induces a local removal of stratum corneum (micropores), whereas tape stripping reduces the thickness of the stratum corneum. Both methods result in a partial removal of both lipids and corneocytes [30,31]. On the other hand, the surfactant SLS acts predominantly on the lipid layers disrupting their structure [32]. Skin from eight donors and two different areas was used: abdomen (see Appendix A, Figure A1(D1–D3)) and breast (Figure 1(D4–D6), and Figure 2(D7,D8). After incubation, skin samples were processed as described in Material and Methods and the amounts of drug as well as cytokines were quantified by means of LC-MS/MS and ELISA, respectively. In general, the ointment formulation resulted in higher concentrations of penetrated tacrolimus. This better performance might be due to the permeabilizing effects of some of the excipients in the ointment formulation. However, it has to be considered that the detection of tacrolimus in epidermis extracts does not provide any spatially resolved information, i.e., the distribution of the drug between the SC and the viable epidermis remains unknown.

Low penetration of tacrolimus was measured in intact abdominal skin. Pretreatment with TS had no significant effects on drug penetration, independently on the tested formulation (Figure A1). Little effect was measured also with regard of the inflammatory markers IL-6 and IL-8. On the contrary, higher tacrolimus concentrations were detected in intact breast skin as compared to intact abdominal skin (Figure 1). In addition, slightly increased drug penetration and cytokine production were found after barrier disruption by LP.

These results suggest that breast skin is more permeable than abdominal skin, probably due to a thinner SC and a higher density of hair follicles, which have been shown to contribute to the skin permeation of topically applied substances [33–35]. The fact that skin from breast areas expressed also much more IL-6 and IL-8 than skin from abdominal region might be due to a higher number of immune active cells.

With regard to tacrolimus penetration in breast skin, TS pretreatment resulted in a marked increase of tacrolimus penetration for both ointment and nanogel formulations (Figure 2(D7,D8)), while application of 5% SLS, a standard procedure for chemical barrier disruption, had no effects. The results obtained after different time points (Figure 2(D8)) show that the penetration of tacrolimus in barrier disrupted skin treated with the ointment increased exponentially with time. On the contrary, the amount of drug penetrated in nanogel-treated samples after 100 min of incubation were similar

to that measured after 24 h of incubation (Figure 2(D7)), suggesting a slower but constant delivery of drug by the nanogel formulations. Thus, the lower amounts of penetrated drug observed for the nanogel formulations after 24 h of incubation may also be a result of the slower drug delivery rate of the nanogel formulations.

**Figure 1.** Tacrolimus skin penetration and expression of the inflammatory cytokines IL-6 and IL-8 after topical application of tacrolimus ointment or nanogel formulation (tacrolimus final dosage 5 μg/cm2) on breast skin from three different donors (D4–6). Skin barrier was left intact or treated with LP prior to the application of the test formulations. Control skin was treated with 0.9% NaCl solution. TAC: tacrolimus; LP: laser poration; NG-tPG: nanogels based on thermoresponsive polyglycerol.

When considering cytokine release, a clear increase of both inflammatory markers was observed in control skin after LP or TS especially in the experiments performed on breast skin (Figures 1 and 2). An additional increase of IL-8 and IL-6 was registered in tacrolimus-treated barrier disrupted skin. This reaction might be attributed to activating or irritating effects of the nanogels themselves or of tacrolimus. While in previous experiments no toxicity was detected for the investigated nanogels [25,26], it is known that tacrolimus may have irritating effects [12]. Our findings point towards the fact that enhanced penetration, while beneficial for therapeutic effects, may pose new challenges with regard to tolerability, especially when the applied substances are capable of reaching cells of the viable epidermis that might be in an activated state as observed in inflammatory skin [36,37].

Overall, these results indicate that breast skin is more permeable to tacrolimus than abdominal skin, regardless of whether skin barrier was disrupted or not. After 24 h of incubation, higher amounts of tacrolimus were measured in ointment-treated samples. Nevertheless, biological effects, like the release of IL-6 and IL-8, were detected also in response to tacrolimus-loaded nanogel formulations.

**Figure 2.** Effects of tape stripping (TS) on tacrolimus penetration and release of IL-6 and IL-8 in breast skin. The amounts of tacrolimus, IL-6, and IL-8 were measured in the epidermis and dermis of breast skin pre-treated to disrupt the skin barrier and incubated with the investigated tacrolimus formulations. (D7) Breast skin was treated with 50 × TS or with 5% SLS previous topical application of 0.1% tacrolimus ointment and tacrolimus-loaded NG-tPG (tacrolimus end concentration 5 μg/cm2) after 24 h. (D8) Tacrolimus penetration and cytokine release in breast skin pre-treated with 50 × TS after 10, 100, and 1000 min of incubation with tacrolimus ointment and comparison with skin pre-treated with 50 × TS or LP and incubation for 100 min with NG-tPG or ointment formulations. Control skin was treated with 0.9% NaCl solution. 50 × TS: tape stripping 50 times; TAC: tacrolimus; LP: laser poration; 5% SLS: 5% sodium lauryl sulfate; NG-tPG: nanogels based on thermoresponsive polyglycerol.

#### *3.2. Nanogel Skin Penetration and Cellular Uptake after Di*ff*erent Degrees of Barrier Disruption by TS*

The SC is a key barrier which hinders the penetration of topically applied compounds. Especially size is one of the major determinants of penetration. With regard to nanoparticulate drug delivery systems, deformability and elasticity further contribute to their penetration properties. Large amounts of carrier material typically remain on the skin surface or in superficial SC compartments. However, the question as to how deep single nanogel particles are capable to penetrate is important with regard to the likelihood of exposure of viable cells to nanogels. Thus, possible translocation of small amounts of nanogel particles to the viable skin layers after skin barrier disruption by TS was investigated using IDCC-tagged nanogels loaded with fluorescein (FL) (Figure 3).

The degree of skin barrier disruption was assessed by measuring the SC thickness in images of skin sections from different regions of the sample. In Figure 3a,d, representative images and the average of at least 15 measurements from three different donors (D9–11) are shown. After 50 TS, a mild skin barrier disruption was achieved for one donor (Figure 3a(D9)), whereas a more severe skin barrier disruption with a stronger reduction of SC thickness was obtained in skin from two other donors (Figure 3d, D10 and D11). Such a different extent of barrier disruption reflects the individual variability of SC thickness and strength. The spatially resolved skin penetration of nanogels (IDCC, red fluorescence) and delivered model dye (fluorescein, FL, green fluorescence) served to clarify the drug delivery mechanism of nanogels. In the representative fluorescence images (Figure 3b,e), it is

to recognize that the two fluorophores co-localized in the SC, which resulted in yellow fluorescence, whereas the released FL penetrated deeper in the viable epidermis, especially in the skin with severe barrier disruption. To quantify the extent of FL penetration, the mean fluorescence intensity (MFI) in the green channel was calculated for different skin areas of at least 20 skin sections per sample. For D9 (Figure 3b), a higher fluorescence signal was detected in the SC for the tNG sample as compared to free FL, confirming the ability of NG-tPG to create a depot in the outermost layer of the epidermis [22]. In skin from donors 10 and 11, where TS procedure had induced a severe skin barrier disruption, a better skin penetration of the released dye was observed not only in the SC but also in the viable epidermis and dermis (Figure 3e). The analysis of NG-tPG penetration (IDCC signal) by fluorescence microscopy showed penetration only in the SC for donor 9 and also in the viable epidermis for donors 10 and 11 (Figure 3c,f). To detect possible cellular association with penetrated nanogels, part of the treated skin was processed to isolate cells and analyze them by flow cytometry and confocal fluorescence microscopy. While after mild skin disruption (D9) only a small percent of cells had high fluorescent signal, over 50% of epidermis cells and 10% of dermis cells were associated with nanogels after severe barrier perturbation (D10 and D11). Especially in the two donors with severe skin barrier disruption, NG-tPG were found to be associated also with Langerhans cells.

**Figure 3.** Skin penetration of topically applied nanogels and released dye depends on the degree of barrier disruption. Mild (**a**–**c**) and severe (**d**–**f**) skin barrier disruption was induced in breast skin by TS previous application of free fluorescein (green) or NG-tPG tagged with IDCC (red) and loaded with FL. After 16 h of incubation, skin was processed to prepare cryosections and isolate cells. (**a**,**d**) Representative transmission light microscopy images of skin sections showing the different degrees of barrier disruption and diagrams showing the average SC thickness; (**b**,**e**) analysis of FL penetration in SC, viable epidermis and dermis of skin sections by measurement of mean fluorescence intensity; (**c**,**f**) analysis of IDCC fluorescence on skin sections (diagrams) and in cells (flow cytometry, dot plots, and images of single cells) isolated from NG-tPG-treated skin and stained with anti-CD1a antibody (Langerhans cells). FL: fluorescein; SC: stratum corneum; VE: viable epidermis; D: dermis; MFI: mean fluorescence intensity; NG-tPG: nanogels based on thermoresponsive polyglycerol; IDCC: indodicarbocyanine.

These results clearly show that, when skin barrier is compromised, nanogels can penetrate to viable skin layers, be taken-up by epidermal and dermal cells and that uptake in dendritic cells is favored in pro-inflammatory environment. This observation may be of special advantage for the treatment of inflammatory skin diseases where, besides T cells, dendritic cells are key targets of therapeutics like tacrolimus. [5,6].

#### *3.3. E*ff*ects of Tacrolimus Nanogels and Ointment on T Cells and Skin*/*Jurkat Cell Co-Cultures.*

As outlined in the previous sections, assessment of a potential benefit of enhanced tacrolimus penetration on the cellular level is limited by its irritative effects on keratinocytes and fibroblasts. Given that immune cells are main therapeutic targets, we created an experimental set-up of skin tissue/T cell co-cultures to measure the effects of penetrated tacrolimus on T cell proliferation [3,10]. First, tacrolimus anti-proliferative effects was measured in vitro on T cells isolated from human excised skin and stimulated with recombinant human IL-2 (Figure 4a–c).

**Figure 4.** Effects of tacrolimus in solution or formulated in nanogels on Jurkat and T-cell proliferation in vitro. (**a**–**c**) Preliminary experiments on isolated dermal T-cells stimulated with IL-2 showed the inhibitory effects of tacrolimus (5 μg/mL) both in solution and in nanogels. Different cell populations were detected by flow cytometry (**a**) and the percentage of cells in each gate at day 5 of culture are plotted (**b**). Values in gates D and E are reported in (**c**) using a different axis scale. (**d**–**f**) Effects of tacrolimus were also detected in Jurkat cells after incubation with 5 and 10 μg/mL tacrolimus solution for 4 days. Cells were gated according to fluorescence intensity (**d**). The percentages of cells in each gate (e) as well as the normalized mean fluorescence intensity of all cells (**f**) showed a decrease of proliferation after treatment with tacrolimus. MFI: mean fluorescence intensity.

Approximately 20% of cells were stimulated to proliferate after addition of IL-2. Both tacrolimus in solution and loaded on NG-tPG reproducibly reduced the percentage of proliferating cells (Figure 4b,c). Tacrolimus was also reported to exert anti-proliferative effects on T cells and Jurkat cells via induction of apoptosis [7,8]. Accordingly, when Jurkat cells were incubated with tacrolimus, a concentration dependent anti-proliferative effect could be measured (Figure 4d–f). These results show the ability of nanogels to deliver tacrolimus to T cells in vitro and induce anti-proliferative effects in a degree similar to that of the free drug. In addition, it was shown that the pro-apoptotic effects of tacrolimus on Jurkat cells were measurable by CFSE assay (Figure 4e,f).

As next step, we cultured ex vivo skin at the air/liquid interface with Jurkat cells using a trans-well set-up (Figure 5a).

We hypothesized that, when tacrolimus is applied topically on skin, the penetrated drug would reach the medium compartment and affect the proliferation of Jurkat cells. In fact, all treated samples had higher MFI than the controls, as shown in the representative histogram in Figure 5b, confirming the anti-proliferative effects of the penetrated drug. A concentration-dependent effect was visible when Jurkat cells were co-cultured for 4 days with skin topically treated with 10 and 20 mg/cm2 of 0.1% tacrolimus ointment, corresponding to 10 and 20 μg/cm2 of tacrolimus, respectively (Figure 5c). When the ointment was compared to the NG-tPG tacrolimus formulation using a final tacrolimus concentration of 5 μg/cm2 (Figure 5d,e), a clear reduction of proliferation was visible at day 5 (Figure 5e). Interestingly, at day 5, ointment and nanogels had similar effects for donors D14 and D15, while for donor D13 nanogels had even a higher anti-proliferative activity than the ointment. Thus, even if in the 24 h penetration experiments, low tacrolimus amounts were measured in the skin treated with nanogels, after longer incubation time in the skin/cell co-culture set up, nanogel formulations could exert anti-proliferative effects comparable to that of tacrolimus ointment.

**Figure 5.** Penetration of tacrolimus across full thickness skin and inhibition of Jurkat cell proliferation. (**a**) Typical experimental procedure. Tacrolimus ointment or nanogels were applied topically on ex vivo skin with disrupted barrier (tape stripping 50 times) that was co-cultured with CFSE-labelled Jurkat cells in a trans-well set up. (**b**) Flow cytometry histogram of cells after treatment with tacrolimus ointment. (**c**) Normalized mean fluorescence intensity of cells cocultured with skin treated with 10 and 20 mg/cm<sup>2</sup> of ointment. (**d**,**e**) Normalized mean fluorescence intensity of Jurkat cells co-cultured for 2 (**d**) or 5 days (**e**) with skin treated with 5 μg/cm<sup>2</sup> of tacrolimus formulated as ointment or nanogel. Both ointment and tacrolimus-loaded NG showed inhibitory effects on Jurkat cell proliferation after topical application on co-cultured ex vivo skin. TAC: tacrolimus; CFSE: carboxyfluorescein succinimide ester; MFI: mean fluorescence intensity; tNGs: thermoresponsive nanogels.

#### **4. Conclusions**

This In this work, we show that tacrolimus formulated as ointment or nanogel suspension penetrates skin with different efficiency in dependence on SC thickness and integrity. Irritation effects of tacrolimus ointment and nanogel formulations, reflected by the released inflammatory cytokines IL-6 and IL-8, were more pronounced in barrier-disrupted and immuno-activated skin. Extensive barrier disruption by mechanical removal of SC resulted also in enhanced penetration of topically applied materials, including deformable macromolecules like nanogels, with consequent interaction and uptake by skin immune active cells. These results support the key role of the SC as barrier for drug and nanocarrier penetration. Our results further demonstrate the critical balance of penetration enhancement and potential increase of side effects. Slow release systems as observed herein already addresses this aspect. The slow drug release and delivery combined with SC hydration and special interaction with skin cells in inflamed skin may explain the good performance of nanogels with respect to tacrolimus ointment. Thus, we conclude that tNGs represent an attractive alternative

water-based formulation for the topical delivery of high molecular weight, poorly water-soluble drugs like tacrolimus.

**Author Contributions:** Conceptualization, F.R., M.C., and A.V.; methodology, F.R., H.V.; formal analysis, F.R., M.G., and H.V.; investigation, H.V., M.G., F.S., and J.I.S.; resources, B.K., M.C., U.B.-P., and A.V.; writing—original draft preparation, F.R.; writing—review and editing M.G., F.S., M.C., and A.V.; visualization, F.R.; project administration, A.V., M.C., and B.K., funding acquisition, B.K., U.B.-P., M.C., and A.V.

**Funding:** This research was funded by the German Research Foundation (DFG), grant SFB1112, projects C04, A04, and Z0. We also acknowledge support from the German Research Foundation (DFG) and the Open Access Publication Fund of Charité—Universitätsmedizin Berlin for publication costs.

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

#### **Appendix A**

Tacrolimus penetration and inflammatory reaction in abdominal skin with or without barrier disruption. Tape stripping had no significant effects on drug penetration in abdominal skin. Only in skin from one donor and only for the ointment a high amount of tacrolimus, approximately 6 ng per μg of total extracted proteins, was detected after tape stripping (Figure A1, Donor 3). Little effect was measured also with regard of the inflammatory markers IL-6 and IL-8, with values around 0.2 pg per μg of total proteins and maximal of 4 pg/μg for NG-tPG in donor 3. While a moderate increase of IL-6 could be observed in tape-stripped skin controls, almost no IL-8 increase was measured in controls after barrier disruption, indicating that IL-6 might be a more sensitive marker for measuring the effect of mechanical damage to the SC. A moderate increase of IL-6 with respect to controls was observed only in the samples treated with the nanogel formulations, whereas a slight decrease of IL-6 with respect to control was measured in skin treated with 0.1% tacrolimus ointment.

**Figure A1.** Tacrolimus penetration after topical application on abdominal skin with or without barrier disruption. Abdominal skin from three different donors (D1-3) was left intact or treated with tape stripping times (50 × TS). NG-pNIPAM (0.9 wt.% tacrolimus), NG-tPG (2.5 wt.% tacrolimus), and 0.1% tacrolimus ointment were applied for 24 h at a final tacrolimus concentration of 5 μg/cm2. Tacrolimus as well as the inflammatory markers IL-6 and IL-8 were detected in skin extracts. TAC: tacrolimus.

#### **References**


© 2019 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 (http://creativecommons.org/licenses/by/4.0/).

*Article*
