Multifunctional electrospun asymmetric wettable membrane containing black phosphorus/Rg1 for enhancing infected wound healing

Abstract Bacterial infection is one of the most frequent complications in the burn and chronic wounds. Inspired by natural existing superhydrophobic surface structures, a novel asymmetric wettable membrane was prepared using the electrospinning technique for facilitating the bacteria‐infected wound healing. Herein, the prepared membrane consists of two layers: The hydrophobic outer layer was composed of poly (lactic‐co‐glycolic) acid (PLGA) and black phosphorus‐grafted chitosan (HACC‐BP), while the hydrophilic inner layer was composed by using a mixture of gelatin (Gel) with ginsenoside Rg1 (Rg1). Biological studies in vitro showed BP@PLGA/Gel (BP@BM) membrane with excellent antibacterial activity could significantly inhibit the adhesion of bacteria, and Rg1 could facilitate the migration and tube formation of human umbilical vein endothelial cells (HUVECs). Compared to Aquacel Ag dressing, the result in vivo revealed that the Rg1/BP@BM could facilitate better wound healing by triggering phosphoinositide 3‐kinase (P‐PI3K/PI3K) and phosphorylation of protein kinase B (P‐AKT/AKT) signaling pathways, upregulating Ki67, CD31, α‐SMA, and TGF‐β1, and downregulating TNF‐α, IL‐1β, and IL‐6, promoting M2 polarization (IL‐10, CD206, and Arg‐1) of macrophages, inhibiting M1 polarization (iNOS) of macrophages. These findings suggested that the asymmetric wettable membrane have the huge potential for wound healing.

wound healing, and high bacterial levels will hinder the progression of wound healing. 1 Unfortunately, prior study reported that infection caused by antimicrobial-resistant pathogens may lead to 10 million annual deaths by the year 2050. 2 Bacterial infection was usually accompanied by inflammation. 3 Persistent inflammation will eventually cause cell death and tissue necrosis, leading to inflammation or injures of surrounding soft tissues. 4 Furthermore, excessive inflammation response was one of the significant reasons for delayed wound healing, which may trigger severe complications. 5 Different wound dressings such as gauze, foam, and bandages have been presented to obtain the temporary barriers. 6,7 However, most of them were not compatible with the compatible environment to completely remodel cell layer and functions of skin. 8,9 Therefore, there is an urgent need for developing an ideal wound dressing in chronic infected wounds.
Asymmetric membrane was an emerging functional dressing since they can mimic construction of skin and furnish the compatible environment for wound healing. 10 Asymmetric membranes have a compact outer layer which similar to skin epidermis, exhibit hydrophobicity that could inhibit the adhesion of the bacteria and prevent the loss of water. 1 Moreover, the hydrophilic inner layer played the role of skin dermis, which could provide an ideal microenvironment of wound healing, and further promote delivery of nutrition and remove the excess exudates. 11 Owing to their extraordinary properties containing water retention, high mechanical properties, and biocompatibility, asymmetric membranes have been applied in the biomedical field such as wound healing 12 and sensor. 13 As a novel and grow rapidly technique, electrospinning has attracted tremendous attention of the biomaterial field owing to its simplification, multifunction, and high efficiency. 8 Besides, electrospun membranes have an ideal loading capacity of drug owing to the high surface area. 14 Many kinds of polymers such as polyurethane (PU), 15 polyacrylonitrile (PAN) 16 and poly(ethylenimine) (PEI), 17 have been applied to produce the asymmetric membranes by electrospinning. However, some polymers like PU, having completely dense structure which made the membranes with low water vapor transmission rate (WVTR), may cause the accumulation of exudate and complications of infection-related in clinic. 18 Poly (lactic-coglycolic) acid (PLGA), due to its biodegradability and biocompatibility properties, has been approved applications in therapeutic devices by the Food and Drug Administration. 19 Besides, it has adjustable water vapor permeability 20 and excellent mechanical properties, 21 making it has great potential to be applied in wound dressing.
Black phosphorus (BP), a rising two-dimensional (2D) nanomaterial, could kill the bacteria under the near-infrared ray (NIR) and cause irreversible damage through high photothermal conversion efficiency. 22 BP had huge potential in the field of biomaterials because of its remarkable biocompatibility. Phosphorus was inherent elements in the human body, 23 and the degradation products of BP were phosphates and phosphonates, which were harmless to human. 24 Furthermore, ginsenoside Rg1 (Rg1) was one of the protopanaxtriol-type saponin, which could be extracted from Traditional Chinese Medicine Panax ginseng. 25 Previous studies reported that the Rg1 had abundant pharmacological activities such as cardiovascular protection, anti-oxidative and anti-inflammation, 26 and also made progress in the treatment of diabetes, 25 nervous system diseases 27 and cardiovascular diseases. 26 Unfortunately, to the best of our knowledge, asymmetric membranes loading BP and Rg1 to facilitate the infected wound healing has not been reported. Therefore, we prepared the electrospinning asymmetric wettable membrane consisting of outer layer and inner layer (Scheme 1). The first layer was composed of PLGA and BP graft chitosan (HACC-BP), while the second layer was composed by using a mixture of gelatin (Gel) with Rg1. The prepared asymmetric wettable membrane exhibited excellent mechanical properties and suitable moisture, which was not only available for cell proliferation and water vapor permeation, but also had a good antibacterial effect. Antibacterial activity in vitro showed that BP@PLGA/Gel (BP@BM) hydrophobic outer layer could prevent bacterial colonization, and had good antibacterial activity. Besides, cytotoxicity in vitro revealed that Rg1 could facilitate the migration and tube forma-

| Preparation of asymmetric wettable membrane
The asymmetric wettable membrane was prepared by continuous electrospinning technique. Briefly, PLGA nanofiber membrane was obtained by electrospinning, and the Gel or Gel/Rg1 was spun continually and covered onto the surface of the PLGA nanofiber membrane, which could obtain an uncrosslinked asymmetric wettable membrane.
Then, the above membrane was crosslinked by the EDC/NHS method. Finally, the crosslinked asymmetric wettable membrane could be obtained after drying, named as BM.

| Water contact angle measurement
The hydrophobic layer and the hydrophilic layer were dried overnight.
Briefly, a volume of about 5 μl of distilled water was dispensed from the syringe onto the membrane's surface. After 5 s, the images of water droplets on the membrane's surface were captured by a water contact angle instrument (DSA25; KRUSS, Germany), and water contact angle of each membrane was calculated.

| Water vapor transmission rate
The WVTR of the asymmetric membrane was measured by the standard method described in the American Society for Testing and Materials (ASTM) E96-00. 28 In brief, the membranes were cut into a suitable size, and then mounted on the mouth (diameter 13 mm) of glass bottles containing 10 ml of deionized water. Then the membrane was transferred to a constant temperature and humidity incubator (temperature set 37 C, relative humidity was maintained at 79%).
A glass bottle without membranes was used as a blank control. The weight of the sample was recorded at predetermined time points. The WVTR is calculated according to the following formula: where Δm/Δt represents the weight of water loss in 24 h (g/day) and A represents the surface area of the bottle mouth (m 2 ).

| Mechanical properties
The mechanical properties of the materials are measured according to the test method of the pharmaceutical industry standard YY/T 0471.4-2004. 29 The tensile strength and Young's modulus were measured by an universal tester (ELF3200, U.S.) with a stretching rate of 6 mm/min. Rectangular specimens of each membrane (1 cm in width Â 7 cm length) were stretched.

| In vitro swelling behavior and degradation behavior
The electrospinning membrane was cut into a 20 mm Â 20 mm square. The Gel, PLGA, and PLGA/Gel were immersed in PBS solution (pH = 7.4) to assess the swelling behavior, and in PBS with 0.02 U/ml collagenase to evaluate the degradation behavior at 37 C, respectively. The swelling ratio (SR) was calculated according to the following formula: where M 0 represents the initial weight and M t is the weight at predetermined time points of the membranes, respectively.
The Gel, PLGA, and PLGA/Gel membranes (20 mm Â 20 mm) were placed in PBS containing 0.02 U/ml of collagenase at 37 C and 100 rpm. At predetermined time intervals, the membranes were collected, lyophilized, and weighed. The degradation ratio was calculated according to the following formula where W 0 represents the initial weight of the membranes and W t is the weight of the degraded membranes at predetermined time points, respectively.

| In vitro drug release of Rg1
The drug release of Rg1 was assessed as follows: The dried membrane with a mean weight of 0.1 g was immersed in 10 ml of PBS (pH = 7.4) at 37 C and irradiated under an 808 nm NIR laser for 5 min.
Subsequently, 1 ml of incubation liquor was taken out at predetermined time points, and an equal volume of fresh medium was replenished. The released amount of Rg1 was detected by UV spectrophotometry at 203 nm.

| Scratch wound assay
The effects of asymmetric wettable membranes on HUVECs migration were evaluated with a wound healing scratch assay. HUVECs cells were seeded into 24-well plates and grown until forming confluent monolayer. Subsequently, the cell monolayer was gently scratched in a straight line using a tip of sterile pipette, and then washed in PBS to remove residual cell debris. Afterward, to the artificial wound was added 500 μl of serum-free medium and then a sterilized membrane was added. At predetermined time points, the bright field images of cells were taken by a microscope, and the cell migration rate was calculated according to the following formula: where A 0 represents the scratch wound area at 0 h and A t is the wounded area measured at each time point after scratching.

| Cell attachment study
The membranes were put into a 48-well plate and soaked in 2 ml PBS to reach a saturated state. Afterward, the 3T3 and HUVECs cells were seeded on the membranes at a density of 2 Â 10 4 cel-

| Angiogenesis test
Matrigel and 96-well plates were thawed at 4 C overnight. Afterward, 50 μl of Matrigel was dropped into a precooled 96-well plates, and then placed statically at 4 C for 5 min. Subsequently, 100 μl HUVECs were seeded on each well at a density of 1 Â 10 4 cells/ml, and the FBS-free medium containing different membranes were added. After incubation for 4, 6, and 12 h, the angiogenic capacity of membranes on HUVECs was recorded by a microscope.

| In vivo wound healing experiments
All Mannitol salt agar selection medium was used for the selection of S. aureus (yellow gold colony), respectively.

| Statistical analysis
Quantitative data sets were performed as mean ± standard deviation (SD). Differences among the groups were assessed by the software SPSS (version on 19.0) using one-way analysis of variance (ANOVA).
3 | RESULTS AND DISCUSSION

| Characterization of the asymmetric membrane
The TEM image showed that the HACC-BP have a uniform 2D sheetlike morphology ( Figure 1A). The DLS showed that the diameter of HACC-BP was at the range about 238 ± 5.6 nm ( Figure S1). As shown in Figure 1B, the developed asymmetric membrane consisted of bilayer structure that could be easily distinguished. The optical images of the asymmetric membrane are exhibited in the Figure S2. As shown in Figure 1C, the cross-section of SEM showed the layer presented the remarkably uniform and smooth nanofiber structure, which exhibited the characteristics of porous three-dimensional, bead-free, and branch-free. Meanwhile, the nanofiber diameters of Gel and PLGA nanofibers were about 500-1000 nm. Besides, the water contact angle can be utilized to assess hydrophilic and hydrophobic behaviors of the gel layer and PLGA layer. The surface of solid represented hydrophilic when the contact angle assessed below 90 , while the solid surface exhibited hydrophobic if the contact angle exceeded 90 . 30 In addition, the preparation of hydrophobic layer is inspired by the "lotus leaf effect." As shown in Figure 1D, the contact angle of PLGA layer was 147.5 ± 2.96 while the Gel layer was 21.2 ± 1.36 , which demonstrated that the asymmetric wettable membrane was successfully prepared. Figure S3 showed the water droplet could be diffused on the surface of hydrophilic Gel layer, but the spherical shape can be maintained on the surface of hydrophobic PLGA layer, which further revealed the excellent asymmetric wettability of the membrane.

| Physical properties of the asymmetric membrane
The ideal wound dressings should have a suitable SR, which will help to absorb exudates of wound site and keep moisture environment on wound bed. 12 As shown in Figure 1E,  of Gel and porous structure of membrane. 31 However, the low SR for the Gel/PLGA membrane was related to strong hydrophobicity of the PLGA. 32 WVTR is a critical parameter of the dressing, which can maintain moisture microenvironment during wound healing. 33 As shown in Figure 1F, the measured WVTR values for Gel, PLGA, and Gel/PLGA were 3112, 1129, and 2252 g/m 2 Á24 h, respectively. In contrast to the other, the high WVTR value for Gel was due to the large pore size of Gel membrane. However, the low SR for PLGA membrane was related to its hydrophobic surface. Prior study reported that the wound dressings with WVTR range about 2000-2500 g/m 2 Á24 h could keep moisture environment, which could contribute to promoting proliferation of epidermal cells and fibroblasts. 30 All results indicated that prepared asymmetric membrane was close to this range and will help to promote wound healing and prevent the scars.
As shown in Figure 1G, the tensile strength of Gel/PLGA membrane was 2.29 ± 0.06 MPa, which is significantly higher than   34 Our results showed that the Gel/PLGA with appropriate mechanical properties will be suitable for wound healing applications.
The biodegradation of biomaterials was a critical factor in tissue engineering applications. 35 As shown in Figure 1I, the degradation rate of Gel was 71.6% at 10 days, which was because the collagenase could decompose the gelatin chain.  36 The standard curve of Rg1 is shown in Figure S4. As the results presented in Figure 1J, the Rg1/BP@BM membrane could promote the release Rg1 under the NIR excitation. Infrared thermal images exhibited the temperature at six time points (0, 1, 2, 3, 4, and 5 min) ( Figure 2A). As shown in Figure

| Antibacterial activity in vitro
Photo thermotherapy exhibited a broad spectrum of antibacterial effect, and was not prone to produce drug resistance. The antimicrobial efficacy of asymmetric wettable composite membrane was evaluated by E. coli, P. aeruginosa, and S. aureus with and without NIR irradiation. As presented in Figure 2C, Figure 2D, and Figure 2E, the BM group showed no antibacterial effect with or without NIR irradiation. Meanwhile, the BP (1.5) @BM group showed the limited antibacterial effect because the temperature was too low to restrain the bacteria. However, the BP (5.0) @BM and BP (3.0) @BM groups showed excellent antibacterial effect, which was owing to the fact that BP could convert light energy into high enough temperature under the NIR irradiation, resulting in the devastating destruction of bacteria. 23 Notably, the BP (5.0) @BM and BP (3.0) @BM groups exhibited moderate antibacterial activity without NIR irradiation, which was related to the sharp edges, resulting in the damage of the bacterial membrane, which is similar to graphene and MoS 2. 38 The images of agar plates with E. coli ( Figure 2F), P. aeruginosa ( Figure 2G), and S. aureus ( Figure 2H) showed direct antibacterial activity, which was consistent with the above results. As shown in Figure 2I, E. coli, S. aureus, and P. aeruginosa of the control groups showed smooth and intact membrane. However, the bacteria of the BP (3.0) @BM + NIR group showed an abundant presence of cracks as well as bacterial shrink and collapse. Figure S6 shows the hydrophobic surface of BM showing bacterial antiadhesion ability, which caused the weak cell affinity because of a lack of cell recognition sites, further inhibiting the adhesion of bacteria. 39

| Cytotoxicity, cell migration, and angiogenesis in vitro
The biocompatibility of membranes was evaluated by live-dead staining and CCK-8 assay. Figure 3A,B showed the live-dead staining of 3T3 and HUVECs cells, respectively. Most of the 3T3 and HUVECs cells were alive after co-culturing with different membranes. Moreover, the cell viability was consistent with the staining results ( Figure 3C and Figure 3D). The cell viability of 3T3 in Rg1/BP@BM group was 97.0%, 102.6%, and 98.3% after culturing day 1, day 2, and The Rg1/BP@BM showed slightly lower viability, which was due to that the degradation products of BP are phosphates and phosphonates, which are of low toxicity to cells and harmless to humans. 24 All results revealed that all membranes exhibited no distinct toxicity to 3T3 and HUVECs, which were similar to the previous study. 40 The Rg1/BP@BM with superior biocompatibility has potential prospect in the biomaterials field.
The influence of the different membranes on HUVECs migration was determined using scratch assay. The photographs of scratch and the percentage of wound closure are shown in The H&E staining of skin tissues was used to analyze the effects of membrane. As shown in Figure 6A was used to assess the collagen deposition. 42 As shown in Figure 6B, the collagen of all groups was gradually increased during the wound healing. On the 3rd day, the collagen deposition of all groups exhibited no significant difference in the dermis. However, the collagen fiber in Rg1/BP@BM showed basket-weave organization on the 7th day, and other groups were still low and irregular ( Figure S7). On the 14th day, the group treated with Rg1/BP@BM exhibited a large amount of collagen deposition, and the collagen fiber was regular and thick as normal tissue.

| The mechanism of wound healing
Angiogenesis was a crucial mechanism in transportation of oxygen and nutrients to the wound area and new tissue. 43 Besides, CD31 As shown in the Figure 7A-C, the Rg1/BP@BM presented more mature blood vessels than other groups at 7 days post-operation, which was due to that Rg1 could promote vascularization according to previous study. 25 WB of CD31 and α-SMA further supported immunohistochemistry results ( Figure 7D). Previous study reported that the wound healing is a complex and multi-faceted biological process, which can be schematically separated into four phases: hemostasis, inflammation, proliferation, and remodeling. 45 Classical M1 macrophages could generate pro-inflammatory mediators to resist infection, while M2 macrophages promoted disappearance of inflammation, and enhanced the tissue remodeling. 46 The immunofluorescent analyses of M1 maker (iNOS) and M2 maker (IL-10, CD206 and Arg-1) were used to estimate the polarization state. As shown in Figure 7E, the Rg1/BP@BM exhibited higher expression of IL-10, CD206, and Arg-1 and lower expression of iNOS. The ratio of fluorescence intensity of iNOS/Arg-1 to measure the dynamic regulation of M1/M2 polarization of macrophages by different materials at the 7th day. Figure S8 showed that the Rg1/BP@BM has the lowest value of M1/M2, which indicated that Rg1/BP@BM promoted M2 polarization of macrophages and inhibited M1 polarization of macrophages, resulting in promoting the wound healing.
The expression of several pro-inflammatory cytokines could be induced by inflammatory cells and played the important role in defense responses in early wound healing process. 47 TGF-β1 is the anti-inflammatory cytokine that could promote wound healing. As shown in Figure 8A, the BP@BM, Rg1/BP@BM, and Aquacel Ag groups exhibited high expression of TGF-β1 at the 7th day, while the gauze, BM, Rg1@BM group showed low expression at 14th day. TNFα and IL-1β were the typical pro-inflammatory factors that could assess the anti-inflammatory effects during wound healing process.
As shown in Figure 8B

| WB analysis and signal pathway
WB analysis was used to test the protein expression level of ki67, P-PI3K/PI3K, P-AKT/AKT, and PERK/ERK. The signal pathway of ki67, PI3K/PI3K, P-AKT/AKT, and PERK/ERK is a typical antiapoptosis pathway, which could regulate the promotion of cell growth, proliferation, and differentiation. 48 As shown in Figure 9A, the Rg1/BP@BM showed higher protein expression than other groups, which indicated that it could effectively enhance cell growth, proliferation, and differentiation during wound healing process. In conclusion, the asymmetric wettable Rg1/BP@BM membrane could enhance the wound healing through regulating the signal pathway by inhibiting inflammation and promoting cell/fibroblast proliferation, collagen deposition, and angiogenesis ( Figure 9B).

| CONCLUSIONS
In summary, we successfully developed a symmetric wettable mem-