**1. Introduction**

The number of antibiotic resistant bacteria, as well as the number of immune deficient patients, is increasing. For this reason, infections of post-operative and chronic wounds are becoming a concern for many patients and health care providers [1]. Age-related immune deficiency, diabetes mellitus type 2, venous insufficiency, or immobility are the major conditions leading to chronic wounds [2–4]. Chronic wounds are often associated with microbial biofilms, i.e., organized communities of one or more microorganism species encased and shielded by extracellular polymeric substances [5,6]. It was shown that significantly higher concentrations of antibiotics are required to treat biofilm-associated

infections [7,8]. One of the reasons for biofilm drug-resistance is the extracellular polymeric matrix, which acts as a shield protecting bacteria from the external environment. In addition, the low concentrations of antibiotics that reach the bacteria in the biofilm favor the formation of persisters, i.e., physiologically inactive dormant cells that are less responsive to antibiotics [9]. Thus, for an e fficient treatment of a biofilm-associated infection, a high and sustained concentration of antimicrobial drugs should be achieved at the site of infection and within the biofilm. This is often not fulfilled by most of the available antimicrobial formulations because of unfavorable physicochemical properties of drugs like instability in di fferent biological environments, low solubility, or high molecular weight. New solubilization strategies and innovative pharmaceutical formulations, like polymer conjugates, nanocarriers, and membranes, have the potential to improve drug delivery and thus increase their concentration at the site of infection [10,11].

In this study, we investigated the drug delivery properties of polyvinylpyrrolidone (PVP)-based foils and nanofiber mats. A ciprofloxacin base was chosen as model drug because of its low water-solubility, detectability using fluorescence spectroscopy, and quorum sensing inhibiting properties in *Pseudomonas aeruginosa* (*P. aeruginosa*) at concentrations below the minimal inhibitory concentration (MIC) [12]. It is a broad-spectrum antibiotic that has been approved by the U.S. Food and Drug Administration (FDA) for the treatment of urinary tract and other infections, including skin and skin-structure infections. Nevertheless, it has severe side e ffects, and in 2016, the FDA recognized the existence of a rare side e ffect, a potentially permanent syndrome called fluoroquinolone-associated disability [13]. Because of the low water solubility, ciprofloxacin is formulated into tablets for oral use as monohydrochloride monohydrate salt. In this form, its absolute bioavailability is of 60% [14]. Alternatively, a ciprofloxacin base has been formulated in oil-based syrups [15] with a better bioavailability of up to 70% [16]. Formulations for topical use (e.g., ciprofloxacin ophthalmic ointment) are made by dissolving the drug in mineral oil. However, it was shown that drug crystals can still form and even cause the blockage of the bottle nozzle [17]. In general, antibiotics are not used for the topical treatment of chronic wounds because of insu fficient drug bioavailability and possible local side effects. However, topical drug delivery would be an advantage in those cases, like an infected diabetic foot, where blood circulation is reduced and systemic therapies are often ine fficient. In addition, it has been shown that locally applied antibiotics can reduce the risk of surgical site infections [18]. Finally, topical therapies reduce systemic toxicity and spare the gu<sup>t</sup> microbiome. Thus, a more e fficient topical drug delivery would improve both the managemen<sup>t</sup> and prevention of wound infections.

In a previous study, we used a pH modification strategy to solubilize a ciprofloxacin base and increase the amount loaded on PVP foils and nanofiber mats. Interestingly, the residual acetic acid bound to PVP conferred a peculiar transparency and elasticity to the foils. These delivery systems were shown to be safe and to have a good anti-microbial activity in vitro [19]. In this study, we tested further properties of these materials using a wound infection model based on full-thickness human ex vivo skin. Di fferent models are available to test the antimicrobial activity of a drug. In vitro grown biofilms give the possibility to study drug e fficacy toward bacteria grown in communities. Nevertheless, in vitro biofilms are often very di fferent from those found in in vivo infections. In vivo animal models have diverse advantages, including a tissue sca ffold, host immune response, and wound healing processes. However, besides ethical reasons, animal studies are expensive and results are not always reproducible in humans. To reduce the number of animal studies, ex vivo models have been proposed for preliminary studies prior to animal studies. Ex vivo human skin has been used for several years to test the skin penetration of chemicals and drugs [20]. Yet, in the last few years, ex vivo porcine or human skin has been used to develop wound infection models [21–27]. Human full-thickness skin represents not only a three-dimensional sca ffold where bacteria can grow, but also a complex environment with extracellular enzymatic activity, antimicrobial peptides, and several di fferent cell populations, including immune-active cells. Thus, human skin infection models are realistic animal-free systems; even if they cannot completely replace in vivo studies, they are useful tools to screen antimicrobial formulations [26].

Using this model, we could measure the drug delivery kinetics, monitor eventual toxic effects, and measure the antimicrobial activity of the tested materials.

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

#### *2.1. Preparation of Ciprofloxacin-Loaded PVP Foils and Nanofiber Mats*

Transparent films and nanofiber mats were prepared as described by Contardi et al. [15]. Briefly, films were produced using a solvent casting method starting from aqueous solutions of PVP (3% *w*/*v*) with a molecular weight (MW) of 360,000 g/mol (Sigma-Aldrich, Milan, Italy), monohydrochloride monohydrate free ciprofloxacin (≥ 98.0% measured by high performance liquid chromatography (HPLC), Sigma-Aldrich, Milan, Italy) and acetic acid (≥ 99.7%, Sigma-Aldrich, Milan, Italy). Three different initial quantities of ciprofloxacin (1.2, 30, and 60 mg) were combined with the polymer and dissolved in acetic acid 30% (*v*/*v*) to a final volume of 30 mL to reach different drug concentrations (2.2, 44, and 88 mmol). The solutions were cast on Petri dishes (diameter 8.75 cm) for 3 days under an aspirated hood under ambient conditions (16–20 ◦C and 40–50% r.h.). Then, the films were placed in a vacuum desiccator for 3 more days to complete the removal of excess acetic acid. The nanofiber mats were fabricated by using a vertical electrospinning set-up. Starting solutions of PVP, acetic acid (30% *v*/*v*), and ciprofloxacin were prepared (final volume 6.2 mL). A higher concentration of the polymer (25% *w*/*v*) was used with respect to the films to allow for the electrospinning process. Three different concentrations of ciprofloxacin (2.2, 44, and 88 mM) were also prepared for the nanofiber mats. Syringes with a stainless-steel needle (18 gauge) were filled with the three different solutions and connected to a syringe pump (NE-1000, New Era Pump Systems, Inc., New York, NY, USA) working at a constant flow rate (500 μL/h). The needles were clamped to the positive electrode of a high-voltage power supply generating 26 kV at a distance of 24 cm from an aluminum disk used as a collector (diameter of 8.75 cm). Only 2 mL of each solution were electrospun in order to obtain the same amount of ciprofloxacin on both the films and nanofibers.

The morphology of foils and nanofiber mats was analyzed using SEM with a variable pressure JSM-649 microscope (JEOL, Milan, Italy) equipped with a tungsten thermionic electron source working in high vacuum mode and an acceleration voltage of 10 kV. The cross-section of the films was obtained by cutting slices with a UCS ultramicrotome (Leica Microsystems, Wetzlar, Germany) equipped with a glass knife. The specimens were coated with a 10-nm thick film of gold using the sputter coater 208 HR (Cressington Scientific Instruments, Watford, U.K.).

#### *2.2. Skin Samples and the Creation of Superficial Wounds*

Abdominal skin was obtained after getting informed consent from healthy donors undergoing plastic surgery. The study was conducted according to the Declaration of Helsinki guidelines and after approval by the Ethics Committee of the Charité—Universitätsmedizin Berlin (approval EA1/135/06, renewed on January 2018). Skin explants were used within 2–4 h after surgery. Subcutaneous fat tissue was partially removed, keeping a layer of approximately 5 mm and skin pieces (1.5 × 1.5 cm) were stretched and fixed on a Styrofoam block covered with Parafilm (Bemis Company, Neenah, WI, USA) using needles. The surface of the ex vivo skin (free of injuries or redness) was cleaned with saline solution (0.9% NaCl). The epidermis was then removed with a ball-shaped milling cutter 6 mm in size (No. 28725, Proxxon, Föhren, Germany) mounted on a micro motor handpiece (Marathon N7, TPC Advanced Technology, Inc. Diamond Bar, CA, USA) and rotating at 16,000 rpm. In this way, superficial wounds of approximately 5 mm in diameter were produced [21].

### *2.3. Drug Penetration Kinetics*

Using a punch biopsy cutter, discs of 8 mm in diameter were cut out from foils, as well as nanofiber mats containing 44 mmol ciprofloxacin, so that each disc contained approximately 250 μg of ciprofloxacin. Wounds of 5 mm in diameter were produced on 1.5 × 1.5 cm pieces of skin. The disks

were applied on the top of the wounds on skin blocs of 1.5 × 1.5 cm that were stretched on a Styrofoam block, placed in a humid chamber, and incubated at 37 ◦C, 5% CO2, and 100% humidity for di fferent time points. Thereafter, non-penetrated material was removed with a cotton swab and the treated wound was removed from the rest of the tissue by means of an 8 mm punch biopsy tool. The tissue was chopped into small pieces and placed in 2-mL tubes filled with HCl (0.1 N, 1.5 mL) to extract the ciprofloxacin. The samples were gently mixed on a shaker for 24 h at room temperature. After centrifugation for 5 min at 300× *g*, the supernatant was collected and placed in triplicate in a 96-well plate (100 μL/well). Ciprofloxacin fluorescence (excitation wavelength: 275 nm, emission wavelength: 480 nm) was measured with an EnSpire ® Multimode plate reader (Perkin Elmer, Akron, OH, USA). A standard curve was prepared by dissolving ciprofloxacin in HCl (0.1 N) and preparing dilutions (0.5–10 μg/mL). The amount of penetrated drug was calculated on the basis of the standard curve. Results are presented as the means and standard deviations of three independent experiments.

#### *2.4. Metabolic Activity of Skin Cells after the Topical Application of Ciprofloxacin on Ex Vivo Skin Wounds*

Wounds were treated with 8 mm discs from PVP-foils loaded with 44 mmol of drug. Wounds treated with NaCl solution (0.9%, 20 μL) served as negative controls, and wounds treated with 50 μg/cm<sup>2</sup> PVP-coated silver nanoparticles (50 nm size, nanoComposix, San Diego, CA, USA) served as positive controls. Samples and controls were incubated in six-well plates for 20 h in 2 mL RPMI-1640 medium (Gibco, Darmstadt, Germany) without phenol red supplemented with fetal calf serum (FCS) (10%, Gibco, Darmstadt, Germany), glutamine (2mM, Gibco, Darmstadt, Germany), streptomycin (100 μg/mL, Gibco, Darmstadt, Germany), and penicillin (100 I.E./mL, Sigma-Aldrich, Hamburg, Germany). Thereafter, the old medium was replaced with fresh medium added with 500 μL of 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide (XTT) (Roche Diagnostic, Berlin, Germany). After 4 h, the medium from each sample and control was collected (3 × 100 μL) and placed in a 96-well microplate. The optical density at 450 nm was read with the EnSpire ® Multimode plate reader using 650 nm as the reference wavelength. Results are presented as a percentage with respect to the values from the control wounds. The averages and standard deviations of values from three independent experiments are reported.

#### *2.5. Histological Analysis of Wound Tissue after the Topical Application of Ciprofloxacin-Loaded Foils*

Wounds created on ex vivo skin were treated with ciprofloxacin in solution or loaded on foils and nanofiber mats (250 μg/wound). Wounds left untreated served as controls. Skin explant cultures were grown in supplemented RPMI-1640 medium as described in Section 2.4. Every 2 days, the medium (500 mL) was removed and fresh medium (500 mL) was added. After eight days of incubation, the skin was plunge frozen in Tissue Freezing Medium (Leica Microsystems, Wetzlar, Germany) and cryosections were prepared. Hematoxylin and eosin (H&E) staining was performed following the manufacturer's manual (Roth, Karlsruhe, Germany) and images were taken using optical microscopy with an Olympus IX 50 (OLYMPUS, Hamburg, Germany).

#### *2.6. Bacteria Inoculation and Characterization of the PAO1 Wound Infection*

An overnight culture of the *P. aeruginosa* strain PAO1 (ATCC 15692) was diluted with tryptic soy broth. The suspension (5 μL, 1 × 10<sup>7</sup> bacteria) was injected with a 10 μL syringe (26 gauge) with a tapered tip (SGE Analytical Science, Ringwood, VIC, Australia) from the edge of the wound into the dermis. As a control, sterile saline (5 μL, 0.9% NaCl) was injected in an uninfected control wound. Wound samples were incubated in a humidified chamber at 37 ◦C for 20 h. Biopsies (8 mm) were taken and fixed in a solution of formaldehyde (4%) and glutaraldehyde (0.5% in 50 mmol (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (HEPES) for 48 h at room temperature. Skin samples were then washed in HEPES (50 mmol), and dehydrated in 30, 50, 70, 90, 95, and 100% ethanol. Samples were infiltrated first with a LR White/ethanol solution (1:1, 10 min), and then with pure LR White (2 × 15 min). Successively, samples were transferred to polyallomer centrifuge tubes

(5 × 20 mm, Beckman Coulter, Inc., Brea, CA, USA) containing LR White with an accelerator (5 μL/mL monomer). The centrifugation tubes were capped with a gelatin capsule and the samples were left to polymerize for 1 h on ice, and then at 60 ◦C overnight. Five hundred nanometer sections were cut using a ultramicrotome (EM UC7, Leica, Wetzlar, Germany), mounted on poly-l-lysine slides, and incubated for 10 min on an 80 ◦C thermo-plate. Sections were stained for 4 min with Richardson's stain (1% Azure II, 1% Methyleneblue, 1% Borax, Sigma-Aldrich, Hamburg, Germany, then washed with double distilled water (ddH2O), and imaged using a microscope (Axiophot, Carl Zeiss Microscopy GmbH, Jena, Germany). For analysis with a scanning electron microscope (SEM), after fixation, one of the sample duplicates was cut with a scalpel in order to reveal the skin profile. The samples were then dehydrated as described above, critical-point dried, mounted on aluminum stubs, sputter-coated with a 12-nm layer of gold-palladium, and finally examined with an SEM (ZEISS 1530 Gemini, Carl Zeiss Microscopy GmbH, Jena Germany) operating at 3 kV using the in-lens electron detector. Images have been cropped, adjusted for optimal brightness, and contrasted using Photoshop Lightroom (version 6.0, Adobe Systems, San Jose, CA, USA).

#### *2.7. Antimicrobial Activity of Ciprofloxacin-Loaded Foils and Nanofiber Mats*

Each tested setting was done in triplicates with a total of at least three runs. Three di fferent dosages of ciprofloxacin were tested (11, 250, and 500 μg/wound). The treatments with PVP foils and nanofiber mats started either 1 h or 20 h after bacteria inoculation and lasted for 20 h. Thereafter, an 8 mm punch biopsy was used to collect the wound tissue, including some of the surrounding intact skin. The tissue was placed in a 1.5 mL microcentrifuge tube containing saline (0.2 mL) and homogenized for 3 min with a sterile steel pistil mounted on a digital overhead stirrer at 150 rpm (DSL, VELP Scientifica Srl, Usmate, MB, Italy). Thereafter, samples were sonicated for 10 min in an ultrasonic bath (BactoSonic1, Bandelin, Berlin, Germany) at 40 kHz using 200 We ff to detach the bacteria. Volumes of each sample were transferred to the wells of 96-well microplates and diluted in 1:10 steps (20 μL sample + 180 μL saline) by using a multichannel pipette. A volume (5 μL) of each well was plated on square tryptic soy agar plates. After incubation overnight at 37 ◦C, spotting areas with 5 to 50 colony forming units (CFU) were counted. Mean values of the triplicates were calculated and bacteria number per wound was calculated considering the used dilutions. In the diagrams, bacteria counts/wound are presented as the mean and standard deviation of three independent experiments.
