Novel nanofibrous membrane‐supporting stem cell sheets for plasmid delivery and cell activation to accelerate wound healing

Abstract The integration of biomaterials with cells for high overall performances is vitally important in tissue engineering, as scaffold‐free cell sheet lacks enough mechanical performance and cell viability while cell‐free scaffold possesses limited biological functions. In this study, we propose a new strategy to strengthen cell sheets and enhance cell activity for accelerating wound healing based on a novel sandwich structure of cell sheet‐plasmid@membrane‐cell sheet (CpMC). Specifically, the CpMC contains two adipose‐derived stem cell (ADSC) sheets on outer surfaces and an electrospun gelatin/chitosan nanofibrous membrane (NFM) encapsulating vascular endothelial growth factor (VEGF) plasmids in between. The physicochemical properties of NFM including swelling, stiffness, strength, elasticity, and biodegradation can be tailored by simply adjusting the ratio between gelatin and chitosan to be 7:3 which is optimal for most effectively supporting ADSCs adhesion and proliferation. The swelling/biodegradation of NFM mediates the sustained release of encapsulated VEGF plasmids into adjacent ADSCs, and NFM assists VEGF plasmids to promote the differentiation of ADSCs into endothelial, epidermal, and fibroblast cells, in support of the neoangiogenesis and regeneration of cutaneous tissues within 2 weeks. The proposed membrane‐supporting cell sheet strategy provides a new route to tissue engineering, and the developed CpMC demonstrates a high potential for clinical translation.


| INTRODUCTION
Regenerative medicine is promoted as a promising treatment for the difficult-to-treat diseases with physically impaired function in patients. 1 Replacement of injured or lost tissues with appropriate cells or tissues is one of most desired treatments in regenerative medicine. Tissue engineering is a promising strategy to overcome inadequate cell number and cell loss after transplantation. 1 The integration of scaffold with functional cells is an important route to enhance regeneration efficacy, but their biocompatibility and cellular activity maintenance are still challenging. 2 On the other hand, Okano et al. proposed the concept of scaffold-free cell sheet engineering to bypass the effect of scaffold. 3 But it is challenging to facilely harvest intact cell sheet because of low strength and poor stability of extracellular matrix (ECM). Moreover, the layering of multiple cell sheets severely confines cell growth and nutrient supply, impairing cell viability and differentiation potential. Therefore, we here proposed a new strategy to strengthen cell sheets and activate cells based on porous membrane supporting cell sheets for the first time.
Since collagen is the primary constituent of ECM, 4 gelatin is a denatured form of collagen and therefore is being popularly used in tissue engineering owing to its high biodegradation, bioresolvability, biocompatibility, commercial availability, and nonantigenicity. 5 Chitosan is also a kind of promising biopolymers for tissue engineering because of its biocompatibility, biodegradability, and antibacterial and antifungal activity. 6 The electrospinning technique is a simple and versatile method to fabricate nanofibrous membrane (NFM), whose morphology can closely mimic the ECM structure and thus facilitate cell adhesion, proliferation, and differentiation. [7][8][9][10][11] The use of electrospinning technique to make NFM with cross-linked gelatin/chitosan mixture can well overcome the shortcomings of overquick biodegradation of gelatin and low spinnability of chitosan, 12 but the effect of the gelatin/chitosan proportion on the physicochemical and biochemical properties of gelatin/chitosan NFM is unknown so far.
In this work, we synthesized a biodegradable gelatin/chitosan NFM by an electrospinning method, and used it to construct a novel sandwich structure by supporting two adipose-derived stem cell (ADSC) sheets on S C H E M E 1 Schematic illustration of the fabrication of the sandwich structure of cell sheet-plasmid@membrane-cell sheet (CpMC) with electrospun gelatin/chitosan nanofibrous membrane (NFM) and two-layer adipose-derived stem cell (ADSC) sheets for wound healing (b), and the mechanisms for skin regeneration (a) both sides (cell sheet-membrane-cell sheet, abbreviated as CMC) for efficiently repairing wound (Scheme 1). High porosity, hydrophilicity, and mechanical properties of the NFM with an optimal proportion of gelatin/ chitosan (7:3) effectively strengthened cell sheets and activated supported ADSCs, including the enhancement of cellular activity and the facilitation of ADSCs differentiation into various skin cells. Moreover, the electrospinning technique enabled local encapsulation of vascular endothelial growth factor (VEGF) plasmids into the nanofibers (cell sheet-plasmid@membrane-cell sheet, abbreviated as CpMC), whose excellent swelling/biodegradation features ensured sustained release of VEGF plasmids, which promoted the proliferation of various skin cells and their migration from normal tissue to wound site, and also enhanced the differentiation of endothelial cells to promote neoangiogenesis, in great support of wound healing and skin regeneration. The water swelling and biodegradation capability of NFM is highly desired to keep the humidity in the wound site and to enable degradation-induced sustained drug release in favor of wound healing.
Though the gelatin/chitosan NFMs can be fabricated by the electrospinning method, the non-cross-linked NFMs will quickly degrade/ resolve in the phosphate-buffered solution (PBS) within half an hour and therefore cannot be applied for tissue engineering. By comparison, the covalent cross-linking can remarkably improve the stability of the gelatin/  From Figure 3f, it can be found that the sandwich structure of CpMC had been successfully constructed.

| NFM assisted VEGF plasmid to promote the differentiation of ADSCs into endothelial, keratinocyte, and fibroblast cells
To check whether VEGF plasmid released from CpMC can promote the endothelial differentiation of ADSCs, we cultured ADSCs on the surface of plasmid@NFM for 3 weeks. From Figure S10, it can be found that differentiated ADSCs treated with NFM and plasmid@NFM showed obvious morphological changes (Figure S12), and both NFM and plasmid@NFM caused a significant increase in the transcription levels of endothelial cell-specific markers including vWF, CD31, VE-cadherin, eNOS, and Flk-1 ( Figure 4c). In parallel, we also observed an upregulation in the expression of specific proteins (vWF, CD31, and VE-cadherin), whose levels in the plasmid@NFM-treated ADSCs were almost the same with that in human microvascular endothelial cells (HMECs) (Figure 4a Figure S16). Similarly, NFM assisted released VEGF plasmids to promote the expressions of collagen I, vimentin, and S100A4 significantly (Figures S17 and S18).
In a word, NFM assisted VEGF plasmids to promote the differentiation of ADSCs toward endothelial, keratinocyte, and fibroblast cells in the simulated skin microenvironment. Such a differentiationenhancing effect of NFM might be attributed to that gelatin as a main component of NFM is one of the ECM capable of promoting the differentiation of stem cells. In addition, we also found that VEGF can significantly promote the migration of HaCaT cells ( Figure S19) and HSF cells ( Figure S20), which would favor wound healing with CpMC.

| DISCUSSION
In this study, we constructed a novel sandwich structure of CpMC by supporting two ADSC sheets on both sides of VEGF plasmid@NFM for efficiently repairing wound. The electrospinning technique we used can produce porous NFM with large specific growth surface, which allowed for cell proliferation and functionalization. The architecture of our electrospinning NFM made from natural biodegradable and biocompatible materials (gelatin and chitosan) was similar to the ECM of skin tissue, which can satisfy physiological requirements.
However, chitosan is a hard electrospinning biomaterial, and the strength of electrospinning chitosan fiber is too weak to form nanofibers, 14 while gelatin is too unstable and subject to biodegradation. 15,16 In this work, the blending of gelatin with chitosan can increase the spinnability of chitosan and improve the mechanical properties of chitosan fibers. Other researchers also used them to develop nanofibers, 17 Our results showed that the physicochemical properties and biocompatibility of NFMs changed with the variation of gelatin/chitosan proportion. The microstructure of gelatin/chitosan NFMs clearly demonstrated that gelatin concentration was very important in obtaining fine gelatin/chitosan nanofibers, and more gelatins achieved bigger pore size, higher water absorption/swelling, and higher mechanical strength. The mechanical property of skin is important to its structure, appearance, and functionality, 19 so the mechanical strength of NFM is also important for the skin substitute construction and wound healing.
The cross-linking treatment of gelatin/chitosan NFMs with glutaraldehyde can avoid their fast degradation. 12 The biodegradation property of NFM was beneficial for skin regeneration as it can support and regulate skin regeneration. 15 Gelatin absorbed an amount of water and made the NFM more hydrophilic and sensitive to degradation, while more chitosans exhibited slower degradation. 20 Additionally, the degradation products of gelatin and chitosan are relatively nontoxic small molecules, which can easily be excreted directly or after entry and exit from various metabolic pathways. 15,21 So, all the investigated NFMs had good biocompatibility with ADSCs. ADSC is a promising adult stem cell for clinical therapy; our previous study has indicated that ADSC can be easily isolated from adipose tissue and has strong proliferation and differentiation ability. 22 It was reported that both fiber diameter and pore size of NFM can affect cell proliferation and differentiation. 23 The NFM with the gelatin/chitosan proportion of 7:3 exhibited appropriate fiber diameter and pore size as well as good mechanical and swelling/biodegradation properties, and thus caused most ideal ADSC adhesion and proliferation.
The gelatin/chitosan NFM was positively charged owing to amino groups, and thus negatively charged plasmids can be bound to the NFM by electrostatic attraction, which can protect plasmid from the destruction caused by enzymes and other harmful factors. 24 So the NFM could be a promising nonviral vector for gene transfection, and enhance the formation of engineered tissue. 25 Because of good swelling/biodegradation property of the 7:3 gelatin/chitosan NFM, we chose it for the VEGF plasmid loading. The release of VEGF plasmid DNA was unlikely to be simple diffusion but mediated by biodegradation. The sustained biodegradation/release behaviors ensured a long-term skin repairing.
The use of biodegradable plasmid@NFM as supporting membrane was quite easy to construct and transplant CpMC for wound healing, because our previous study suggested that ADSC sheet was too thin and fragile so that it was hard to transplant intact cell sheet. 26 A largearea (2 cm in diameter) full-thickness cutaneous defect in mouse was used to investigate the skin regeneration potential of CpMC. Complete wound closure and re-epithelialization was firstly observed after implantation of CpMC for 13 days, much faster than the other transplantation. Although NFM only just had a little beneficial effect on skin regeneration, highly porous network structure and good swelling property of NFM can provide a moisture environment and enhance ADSC activity. So, both NFM and ADSCs were key elements for wound healing and skin regeneration in this work. The wound skin transplantation of CpMC had thicker and complete epithelial layer than that of CMC, which indicated that VEGF also played an important role in wound healing and skin regeneration.
VEGF is a multitasking cytokine which stimulates cell survival, proliferation, migration, and differentiation, 27 which is also confirmed in this work. VEGF released from CpMC locally promoted the differentiation of ADSCs into endothelial, keratinocyte, and fibroblast cells.
Interestingly, NFM also had a little effect on the differentiation of ADSCs. One of possible reasons may be that gelatin, as a denatured form of collagen and one component of the ECM, can promote the differentiation of stem cells in specific microenvironment. 28,29 Besides, VEGF also can enhance the migration of keratinocyte and fibroblast cells, which may be beneficial for the wound repairing, because the whole wound-healing process includes cell adhesion, proliferation and migration, and eventually leads to skin regeneration. 30 Endogenous angiogenic factor, VEGF, is naturally produced in response to tissue hypoxia during wound healing. 31 Because of inadequate secretion of VEGF, this restorative process of wound is insufficient to prevent tissue ischemia and necrosis. During wound healing, rebuilding damaged blood vessels (angiogenesis) is extremely important since repairing activities are highly dependent on newly formed blood vessels to supply oxygen and nutrients. 32 Wound treated with CpMC showed larger diameter of capillary, indicating more oxygen and nutrients supply for skin tissue regeneration. VEGF plays an important role in endothelial cell differentiation and vascular repair. 33 The enhanced expressions of CD31 and α-SMA, two widely used markers for blood vessels, 34 implied that the CpMC-mediated VEGF expression promoted endothelial cell differentiation and accelerated the maturation of blood vessels to accommodate the urgent demand of oxygen and nutrients for wound regeneration. It is well known that angiogenesis is modulated by VEGF. 35 Our results suggested that CpMC facilitated angiogenesis and skin cell differentiation as well as migration.

| Mechanical property
The NFM were collected using cellophane, and then the thickness of each sample was measured by SEM. These samples were measured, cut carefully and glued on the frame, and then the cellophane was removed. The paper frame was cut before initiating the tensile strength measurements. The tensile properties of these samples were characterized by a tensile testing machine (UTM150, Advanced Nanomeasurement Solutions, Agilent Technologies). The velocity was 100 μm/s. The data were analyzed by using the Linksys32 software.
The reported Young's modulus represented average results of the five samples.

| Furrier transform-infrared spectroscopy
The functional changes in different NFM were evaluated using Thermo Scientific TM Nicolet-6700 FTIR (Shanghai, China). For this, each non-cross-linked and cross-linked nanofiber were scanned from 4000 to 400/cm with step resolution of 0.4 and 32 number of scans.

| Differential scanning colorimetry (DSC)
DSC was carries out on the DSC Q200 differential scanning colorimeter (TA Instrument, Delaware, USA). Each sample was analyzed between 25 C and 300 C with 5 C/min of heating rate. Nitrogen gas was used as inert gas and slandered aluminum crucibles were used for experiment.    Table S1. HMEC was used as positive control. To differentiate ADSC into KLC, we used the cocultured system as previously described. 36 ADSCs and HaCaT cells were inoculated in the two chambers sepa- To investigate the effects of VEGF on the migration of HaCaT and HSF, transwell cell culture insert was applied and cultured for 24 h. Wound-healing assay was also applied to test the migration ratio of HaCaT cells.

| Preparation and transplantation of CpMC
According to the physical characteristics and biocompatibility with ADSC, we chose gelatin/chitosan (7:3) NFM or plasmid@NFM to support ADSC sheets. After culturing ADSCs on the surface of temperature-responsive culture dish to a confluent layer, all media were aspirated and fresh media were added to prevent the cells from

| Immunofluorescence staining and western blotting
For the tissue section, after deparaffinization, hydration, and blocking of the nonspecific binding sites with 5% bovine serum albumin (BSA) at 37 C for 30 min, tissue sections were incubated with a rabbit polyclonal antibody to CD31 (1:200, ab28364, Abcam), a mouse polyclonal antibody to α-SMA (1:100, ab7817, Abcam), and a rabbit polyclonal antibody to VEGF (1:100, ab28364, Abcam) at 4 C overnight. Then, sections were washed with PBS, followed by incubation with a donkey anti-rabbit IgG Alexa Fluor ® 488-conjugated secondary antibody (1:1000, ab150073, Abcam) and a donkey anti-mouse IgG Alexa Fluor ® 647-conjugated secondary antibody (1:1000, ab150111, Abcam) at 37 C for 60 min in the dark. After being stained with DAPI and mounted with anti-fluorescence quenching reagent, the tissue sections were observed and imaged with a Nikon confocal laser microscope (Nikon, A1 PLUS, Tokyo, Japan). For the immunocytochemistry, the cells were fixed with 4% formaldehyde, and incubated the primary and secondary antibody according to the above procedure.
Total protein was obtained by lysing tissues or cells in radioimmunoprecipitation assay buffer containing protease and phosphatase inhibitors. Following mixing with loading buffer and boiling, the prepared samples were separated on 10% or 8% SDS-PAGE gels and

| Quantitative real-time polymerase chain reaction
Total RNA from the differentiated cells was obtained using Trizol (Invitrogen). The RNA was reverse transcribed to complementary DNA (cDNA) using the First Strand cDNA kit (Takara) following the manufacturer's protocol. Quantitative polymerase chain reaction analysis was then performed using the Quantitect SYBR Green PCR Master Mix (Takara). For amplification process, the sense and anti-sense primers (Table S1) were used. Standard curves were generated, and quantities of each transcript were normalized to GAPDH as an internal control. Each assay was performed in triplicate with an n of 3 independent experiments.

| Statistical analysis
All data were shown as the mean ± SD. All experiments were independently repeated for at least three times unless otherwise stated.
Differences between groups were assessed by independent-samples t test or by one-way analysis of variance followed by Tukey's multiple comparison test. All analyses were performed in SPSS 22.0 software and p < 0.05 was considered to be significant unless otherwise specified.

| CONCLUSIONS
In this study, we fabricated a biodegradable NFM to locally encapsulate VEGF plasmid and support multilayer stem cell sheets transplantation for wound healing. The gelatin/chitosan (7:3) NFM had appropriate biodegradation and mechanical strength, as well as good biocompatibility with ADSCs, and could release VEGF plasmid in a sustained way to promote the migration of skin cells, proliferation and differentiation of stem cells. The gelatin/chitosan NFM supported the multilayer stem cell sheets and made it easy to transplant into the wound site. This NFM can be a good plasmid delivery platform, and the proposed membrane-supporting cell sheet strategy provides a new route to tissue engineering, and the developed CpMC demonstrates a high potential for clinical translation.