Mesenchymal stromal cell‐derived factors promote the colonization of collagen 3D scaffolds with human skin cells

Abstract The development of stem cell technology in combination with advances in biomaterials has opened new ways of producing engineered tissue substitutes. In this study, we investigated whether the therapeutic potential of an acellular porous scaffold made of type I collagen can be improved by the addition of a powerful trophic agent in the form of mesenchymal stromal cells conditioned medium (MSC‐CM) in order to be used as an acellular scaffold for skin wound healing treatment. Our experiments showed that MSC‐CM sustained the adherence of keratinocytes and fibroblasts as well as the proliferation of keratinocytes. Moreover, MSC‐CM had chemoattractant properties for keratinocytes and endothelial cells, attributable to the content of trophic and pro‐angiogenic factors. Also, for the dermal fibroblasts cultured on collagen scaffold in the presence of MSC‐CM versus serum control, the ratio between collagen III and I mRNAs increased by 2‐fold. Furthermore, the gene expression for α‐smooth muscle actin, tissue inhibitor of metalloproteinase‐1 and 2 and matrix metalloproteinase‐14 was significantly increased by approximately 2‐fold. In conclusion, factors existing in MSC‐CM improve the colonization of collagen 3D scaffolds, by sustaining the adherence and proliferation of keratinocytes and by inducing a pro‐healing phenotype in fibroblasts.

There are two main research directions for the treatment of chronic wounds: innovative biomaterials and the use of stem cells, each having their own advantages. The skin comprises cells and extracellular matrix, with the latter providing mechanical strength and modulating cell proliferation and differentiation. 5 When the tissue damage is extensive, the sole delivery of cells without a matrix results in poor survival and engraftment. 6 Therefore, in order to improve tissue regeneration, the use of biomaterials is particularly useful, with collagen being the most explored, due to its high biocompatibility, biodegradability, weak antigenicity, 7 availability and possibility to be processed into porous structures, which provide a suitable environment for cell growth and proliferation and facilitates the supply of nutrients and oxygen. 8,9 On the other hand, cellular therapy for wound healing using mesenchymal stromal/stem cells (MSC) is based on the differentiation potential, bioavailability, low immunogenicity and, most importantly, on the secretion of numerous soluble factors promoting cell migration, angiogenesis and granulation tissue formation by this type of cells. 10 However, major drawbacks in the use of cellular therapy are the time-consuming process of in vitro expansion and also the donor's variability associated with an uncertain therapeutic outcome. 11 To avoid these inconveniences, MSC secretome in the form of conditioned medium (CM) could be standardized for large scale use as an 'off the shelf product'. [12][13][14] In this context, we aimed to combine these two approaches by supplementing a type I collagen scaffold with MSC-CM, making use of the new concept of 'cell-free therapy', in order to obtain an improved biomaterial, with boosted therapeutic properties. Thus, in this paper, we provide evidence that factors within the MSC-CM support the colonization of a three-dimensional (3D) type I collagen scaffold with the main types of skin cells (keratinocytes and fibroblasts) and they have the capacity to attract epithelial and endothelial cells, which is important for enhancing the re-epithelialization and angiogenesis processes. Moreover, the presence of MSC-released factors induces a pro-healing phenotype in fibroblasts, enabling their contribution to wound contraction, granulation tissue formation and scaffold biodegradation. October 2008). Human bone marrow-derived MSC were isolated using a modified protocol established by our group. 15 Briefly, MSC seeded at 10 5 /cm 2 were grown in DMEM 4.5 g/L glucose supplemented with 15% foetal bovine serum (FBS), 1% non-essential amino acids, 300 UI/mL penicillin, 300 mg/mL streptomycin and 150 mg/ mL neomycin, at 37°C and 5% CO 2 . The cells were characterized following the indications of the International Society for Cellular Transplantation. 16 The multilineage potential of bone marrow-isolated MSCs was tested by inducing their differentiation into adipogenic, osteogenic and chondrogenic lineages. Adipogenic differentiation: pre-confluent cells were incubated for 3 weeks in adipogenic differentiation medium (DMEM with 10% FBS, 10 −6 M dexamethasone, 100 μM indomethacin and 1% insulin transferrin selenite supplement-ITS, Sigma-Aldrich).

| Cell culture techniques
The lipid accumulation was determined by Oil Red O (Sigma-Aldrich) staining. Osteogenic differentiation: confluent cells were incubated in osteogenic differentiation medium (DMEM with 10% FBS, 10 −7 M dexamethasone, 10 mM β-glycerophosphate and 0.3 mM ascorbic acid). After 3 weeks of incubation, the cells were exposed to von Kossa stain to highlight calcium deposits. Chondrogenic differentiation: cells were detached with trypsin and 2.5 × 10 5 cells were added to 15-ml Falcon tubes containing 500 μl chondrogenic differentiation medium (high glucose-DMEM with 10 ng/ml TGF-β3, 10 −7 M dexamethasone, 50 μg/ml ascorbate-phosphate, 40 μg/ml proline, 100 μg/ml pyruvate, 50 mg/ml ITS-Sigma-Aldrich). After centrifugation, the pellets were maintained at 37°C for three weeks. Next, the pellets were embedded in paraffin and 5 μm thick sections were obtained using a Leica microtome. The sections were dehydrated and stained with Alcian blue (Sigma-Aldrich) to highlight the acid mucopolysaccharides.
For CM collection, pre-confluent MSC from passage 4-7 were washed with PBS and incubated with DMEM supplemented with 1% non-essential amino acids for 24 h. After that, CM was collected and centrifuged at 2000 g for 25 minutes, and the supernatant stored at −80°C.

| Flow cytometry
MSCs were detached with accutase (Sigma-Aldrich) and exposed to the antibodies that fulfil the minimal criteria for the definition of human MSC (FITC coupled antibodies for negative markers-CD45, CD34-and PE coupled antibodies for the positive markers-CD44, CD73, CD90, CD105, all from BD Biosciences) according to the manufacturer's protocol. The samples containing 10 5 cells were analysed using a Beckman Coulter 3 laser Gallios cytometer and the data were analysed with Summit software v4.3 (Cytomation, Inc).

| Characterization of human dermal fibroblasts by immunocytochemistry
Human dermal fibroblasts cultured on glass coverslips were fixed and permeabilized in 4% PFA with 0.1% Triton X and blocked with 1% BSA, before staining with mouse anti-vimentin (Sigma-Aldrich), mouse anti-fibronectin (Sigma-Aldrich) and rabbit anti-collagen type I (Abcam) for 1 h at 37°C. After washing the primary antibody, the cells were incubated with secondary antibodies conjugated with: rabbit Alexa 488 or mouse Alexa 568 (Thermo Scientific). The coverslips were mounted with Fluoroshield with DAPI (Thermo Scientific) and visualized using a Zeiss Observer D1 microscope.

| Collagen scaffold preparation
The type I fibrillar collagen gel was extracted from calf hide by acid and alkaline treatments, Briefly, the gels containing 1.2% collagen and having a pH of 7.4 were cross linked with 0.5% glutaraldehyde (Merck, Germany), followed by freeze-drying at −40°C for 4 hours, using a Delta 2-24 LSC Christ (Germany) as previously described. 17,18 The spongious forms were characterized by scanning electron microscopy using a QUANTA INSPECT F SEM device equipped with a field emission gun (FEG) with a resolution of 1.2 nm, as previously described 17. For in vitro testing, the scaffolds were sterilized in 70% ethanol overnight, followed by several rinses in sterile PBS and maintained in DMEM without serum for at least 12 hours.

| Assessment of collagen scaffold capacity to support colonization of human keratinocytes and dermal fibroblasts
Keratinocytes and dermal fibroblasts were seeded on collagen scaffolds (3 × 10 5 and 2 × 10 5 , respectively) and incubated in complete medium at 37°C and 5% CO 2 atmosphere. The scaffolds were washed with PBS, fixed in 4% PFA, embedded in Shandon Cryomatrix (Thermo Scientific) and cryosectioned using a Leica cryotome in order to obtain 4 μm thick slices. The slices were subsequently subjected to haematoxylin-eosin, DAPI, eosin-Hoechst or Ayoub-Shklar staining. For the first staining, the slices were incubated for 7 minutes with haematoxylin and 5 minutes with eosin Y, mounted in glycerol and visualized using a Zeiss Observer D1 microscope. For eosin-Hoechst staining, after incubation (2 min) in eosin Y, the slices were differentiated with 70% ethanol, washed with distilled water, stained (10 min) with Hoechst (1 mg/ml), washed, mounted and visualized. For Ayoub-Shklar staining, the slices were incubated with acid fuchsin 5% for 5 min, followed by 45 min in aniline blue-orange G solution, washed and mounted. Additionally, keratinocytes and fibroblasts were stained with CellTracker ™ Red CMTPX Dye (Thermo Fisher Scientific), seeded on scaffolds and kept at 37°C and 5% CO 2 atmosphere. After 3 days, the scaffolds were washed with PBS and visualized by IVIS Spectrum CT System (Perkin Elmer, Caliper, LifeSciences).

| Qualitative assessment of the viability of keratinocytes and fibroblasts cultured on 3D collagen scaffolds
The collagen scaffolds cultured with keratinocytes and fibroblasts for 5 days were stained with LIVE/DEAD ™ Viability/Cytotoxicity Kit (Thermo Fisher Scientific). Briefly, the scaffolds were washed with warm PBS and incubated for 30 minutes in a mixture of 2 μM calcein-AM and 4 μM ethidium homodimer-1. Next, the scaffolds were fixed in 4% PFA, embedded in Shandon Cryomatrix (Thermo Scientific) and cryosectioned. The 4 μm thick slices obtained were subsequently visualized using a Zeiss Observer D1 fluorescence microscope after being mounted with Fluoroshield with DAPI (Thermo Scientific).

| Evaluation of the effect of MSC-CM on the adherence and proliferation of keratinocytes and fibroblasts in 2D culture system and 3D collagen scaffolds
The assessment of adherence was performed by seeding keratino-

| In vitro wound healing assay (scratch test)
Keratinocytes, fibroblasts and endothelial cells were grown to confluence in 96-well. Briefly, the monolayer was scratched transversely using a 200-μl pipette tip, washed with PBS and incubated in complete, MSC-CM and serum-free media. The cells were photographed immediately after the addition of the media-0 h and at 8 h for endothelial cells and 14 h for keratinocytes and fibroblasts. The migration of the cells was quantified by measuring the area covered by cells using ImageJ software (NIH).

| Chemotaxis assay to evaluate the chemotactic effect of MSC-CM
The chemotactic properties of MSC-CM were evaluated using the

| Real-time PCR
Total RNA was extracted from fibroblasts after 5 days in culture using the TRIzol reagent (Thermo Fisher Scientific) and cDNA was synthesized starting from 1 μg of total RNA employing SENSIFAST cDNA Synthesis Kit (Bioline). Real-time PCR was performed using The SensiFAST ™ SYBR Hi-ROX Kit (Bioline) optimized amplification conditions. The experiments were performed three times in triplicate for each gene. The primer sequences are given in the Supporting Information (Table S1). The analysis was done using the comparative CT method and β-actin was employed for internal normalization.

| Assessment of the cytokine profile of MSC-CM
The cytokine profile was analysed using Human Angiogenesis Array (R&D Systems). Briefly, MSC-CM was mixed with the biotinylated detection antibody cocktail and incubated with nitrocellulose membranes. The washed membranes were incubated with Streptavidin-HRP. Cytokine detection was performed using a FUJIFILM Luminescent Image analyser LAS-3000. The pixel density was quantified by TotalLab Quant software.

| Statistical analysis
Data were analysed with GraphPad Prism 5.0 (GraphPad Software, Inc) and presented as mean ± SD of three independent experiments, unless otherwise stated. Comparison of multiple groups was done by ANOVA. Two-group analysis was carried out by Student t test.

| Characterization of human bone marrowderived MSCs and dermal fibroblasts
MSCs isolated from human bone marrow aspirate were characterized following the guidelines of the International Society for In order to characterize the dermal fibroblasts isolated by the explant method from human skin samples, the expression of characteristic fibroblast markers: vimentin, collagen I and fibronectin, was assessed by immunocytochemistry. As shown in Figure 1C, all these proteins were present in the isolated cells.

| The 3D collagen scaffold allows the colonization with human skin cells
Scanning electron microscopy revealed a three-dimensional structure with interconnected pores, having diameters between 75 and 150 μm (Figure 2A). The macroscopic aspect of 96-wells plate tailored scaffold, pre-conditioned by incubation in DMEM for 24 hours prior to cell culture, with dimensions of ~4 mm in height and ~6 mm in diameter, is shown in Figure 2B. In order to establish the capacity of keratinocytes and dermal fibroblast to grow and survive on this scaffold, the CMTPX labelled cells were seeded on the pre-conditioned sponges. The collagen scaffold itself is auto-fluorescent, making difficult to reveal the presence of keratinocytes; the overall fluorescence was not statistically significant in comparison to the collagen scaffold alone ( Figure 2C).
However, the presence of fibroblasts was easily detectable, as the fluorescence intensity of the collagen seeded with labelled fibroblasts compared to bare collagen scaffold controls was twice as high and statistically relevant.
After 5 days in culture, both cell types were found inside the 3D collagen structure as shown by the presence of stained nuclei with DAPI on cryosections ( Figure 2D), supporting the capacity of the scaffold to be colonized.
Next, the cell viability was assessed using a live/dead cytotoxicity kit and the presence of viable cells is evident after 5 days of culture on collagen scaffolds with some rare dead cells indicated by the red-stained nuclei, positive for ethidium homodimer 1-EthD-1 ( Figure 2E).

| MSC-CM impacts the properties of keratinocytes and dermal fibroblasts in 2D classical culture system
Before testing the influence of MSC-CM on the colonization of the collagen scaffold with skin cells, we assessed the capacity of CM showed that MSC-CM was able to induce keratinocyte attachment faster than the serum supplement ( Figure S1). At 2 h post-seeding, the cellular index was significantly higher for keratinocytes incu-

Also, the proliferation of keratinocytes incubated in MSC-CM
was similar to the positive control (90 ± 14% versus 100% in serum control) and significantly higher than the negative control (45 ± 14.2%), as illustrated in Figure S1. area, compared to 54 ± 4% for positive serum control and 37 ± 5.7 for negative control, as revealed by the scratch tests ( Figure 3D).

| MSC-CM promotes the colonization of collagen scaffolds with keratinocytes and fibroblasts
Given the encouraging effect of MSC-CM on skin cells in classical 2D culture system, we subsequently assessed the effect of MSC-CM on the colonization of the collagen scaffold. As shown in Figure 4A Figure 4C) and Ayoub-Shklar histological staining ( Figure 4D), which revealed the presence of these cells attached onto the collagen fibres, both on the surface as well as inside the structure.
As for the adherence of dermal fibroblasts, as shown in Figure 4E During histological processing, we noticed that the fibroblasts colonized scaffolds seemed more damaged, especially for MSC-CM samples ( Figure S2). Therefore, we investigated the effect of MSC  Figure S3, the opposite mechanism was observed on collagen scaffolds ( Figure 4G).
However, the ratio between collagen III and I mRNAs was 2 to 3 times higher for the fibroblasts incubated in MSC-CM, both in 2D and 3D systems.
Subsequently, we evaluated the effect of MSC-CM on the expression of α-SMA, an early marker of fibroblast differentiation towards myofibroblast. It was previously shown that, when cultured on plastic, fibroblasts gain a proto-myofibroblast phenotype, including the expression of α-SMA. 19 As revealed by immunocytochemistry ( Figure S4

| MSC-CM has pro-angiogenic properties
The importance of the angiogenic process in the wound healing mechanism is well established, 21

| Evaluation of MSC-secreted factors
The secretory phenotype of MSCs was evaluated using a cytokine array ( Figure 6A), which provided insight on the growth factors implicated in the effect of MSC-CM on the proliferation of keratinocytes and fibroblasts. Thus, the growth factors traditionally implicated in the proliferation of keratinocytes, EGF and KGF, had low levels in MSC secretome ( Figure 6B). Therefore, the proliferative effect on keratinocytes could be explained by the presence of HGF, which has been shown to have mitogenic properties and to stimulate keratinocytes motility. [24][25][26] Furthermore, MSCs secreted only low levels of

| D ISCUSS I ON
The ideal therapeutic approach for chronic wounds is a cost-effective, 'off the shelf' and topically applied biomaterial, which provides Therefore, we investigated the therapeutic potential of a porous scaffold made of type I collagen improved by the addition of a powerful trophic agent in the form of MSC-CM in order to be used as an acellular scaffold for skin wound healing treatment. We chose to develop a type I collagen scaffold, as this is the most abundant and most readily available form of collagen, influencing various cellular properties and functions of fibroblasts and keratinocytes, including cell shape, adhesion, differentiation and migration 31,. 30 As they were first described by Friedenstein and collaborators, 31 MSC have been proposed as main candidates for regenerative therapy. Although they possess a certain differentiation capacity, recent insights in the mechanism by which these cells contribute to tissue regeneration indicate that their most important feature is the abundant secretion of bioactive trophic factors, 32 which can modulate the local immune system, promote angiogenesis, prevent cell apoptosis, and stimulate survival, proliferation and differentiation of resident cells. [33][34][35] The approach of combining the properties of a biomaterial and MSC-CM in order to obtain a boosted therapeutic response has been successfully employed in several experimental settings, mainly for hard tissue regeneration. Thus, Osugi and collaborators showed the application of an agarose gel with MSC-CM for calvarial defects had a superior regenerative effect compared to agarose gel mixed with MSC. 36 Similar outcome has also been reported for collagen scaffolds supplemented with CM for the same type of tissue application. 37 Recently, it was shown that a 3D structure containing carbon nanofibres functionalized with MSC-CM improved cell adhesion, proliferation and viability and provided a biomimetic stem cell niche for cartilage and bone regeneration. 38 Diomede and collaborators showed that a synthetic poly-(lactide) scaffold enriched with human gingival MSC-CM had an increased osteogenic action in a rat model of cranial defect. 39 To our knowledge, the present paper is the first to explore this combinatorial approach involving a collagen scaffold and MSC-CM for skin wound healing therapy. Here, we demonstrate that a natural-derived biomaterial, such as the type I collagen scaffold combined with MSC-CM has beneficial effects valuable in skin lesions treatment. Our data show that the secretome of bone marrow-derived properties. An interesting observation was the modification of the gene expression of dermal fibroblasts grown on the collagen scaffold: more exactly, the increase of the collagen III and I mRNAs ratio, of α-SMA and TIMP-1, TIMP-2 and MMP-14, but not MMP-2 and MMP-9 expression for the cells incubated in MSC-CM compared to serum control. This effect could be particularly important in diabetic chronic wounds, where the fibroblasts are dysfunctional: the collagen production is decreased, MMP-9 synthesis is increased and the overall balance between MMPs and TIMPs is disturbed, leading to the destruction of the extracellular matrix. 40,41 These data indicate that MSC-CM induces a pro-healing phenotype in fibroblasts, contributing to the contraction of the wound which requires α-SMA activation, to the formation of the granulation tissue by increased collagen III/I ratio and to the biodegradation of the scaffold under the action of MMP-14.
In conclusion, our data show that supplementing the collagen scaffold with MSC-derived secretome results in an improved biomaterial, with enhanced therapeutic properties, such as boosted chemoattraction of epithelial and endothelial cells, while also contributing to the re-epithelialization, matrix remodelling and angiogenesis processes. We stipulate that our approach, having the advantage of being an 'off the shelf product', as both the scaffold and the CM can be obtained in advance and stored until needed, could become an efficient therapeutic alternative for difficult to treat skin wounds.

ACK N OWLED G EM ENTS
Work supported by The Executive Agency for Higher Education,

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.