Mesenchymal Stem Cells Regulate IL8 and TGFA Expression in a Novel Leucocytes Depleted Platelet-Rich Plasma-Skin Equivalent in a Preliminary in vitro Study of Chronic Wound Healing

Background: Chronic leg ulcerations are associated with Haemoglobin disorders, Type 2 Diabetes Mellitus, and long-term venous insuciency. Mesenchymal stem cells (MSCs) ability to modulate the inammatory response represents the fundamental requisite for their applicability as a treatment of chronic wounds. Methods: This study aimed to develop a novel bioactive platelet-rich plasma (PRP)-leukocytes-depleted scaffold to reproduce typical clinical wound of patients with poor chronic skin perfusion and low leucocytes inltration. After scratching the wound model to mimic injury three conditions were compared; an untreated condition, a condition treated with recombinant TNF to mimic an inammatory state and a condition treated with TNF and also with MSCs to evaluate how the latter’s immunomodulatory properties affect the therapeutic outcomes in an inammatory state. Gene expression of IL8 and TGFA were analysed in biological triplicates of the three conditions. Statistical analysis was done through a paired student t-test and a p <0.05 was considered signicant. Results: We set up a skin model that consisted of a leukocyte-depleted, platelet-rich plasma scaffold, with embedded broblasts as dermal equivalent and seeded keratinocytes on it as multi-layered epidermidis. IL8 expression increased upon scratching (p=0.014) and continued to increase up to day 1 (p=0.048). IL8 expression decreased upon administration of TNF (p=0.005) but then increased again. IL8 expression decreased in the untreated condition after day 1 as the natural healing process took place and was lower than in treated conditions in day 8 (p=0.048). TGFA expression decreased upon scratching (p=0.006) and increased again in day 1, more so in the untreated than in the treated conditions (p=0.02). TGFA expression decreased again in day 4 in the study group before increasing sharply (p=0.027) in day 8 to reach pre-scratch levels. Conclusion: This study found that a leukocyte-depleted PRP-based skin equivalent can be useful in the study of treatments of chronic wounds. This study also indicates that MSCs appear to modulate the expression of IL8 by switching from an immunosuppressive phenotype to a pro-inammatory phenotype. These results indicate that the administration of MSCs could offer a potential therapeutic approach for the treatment of leg ulcers in patients with poor skin perfusion.

and in long-term venous insu ciency where poor perfusion and altered metabolism set up a chronic in ammation that impairs repair and wound closure [6,7]. Although many therapies have been explored for leg ulcers, none has been completely satisfactory. In our clinic, we obtained complete rapid closure of a longstanding indolent ulcer in the ankle of a βthalassaemia patient with a homozygous β globin mutation using a mixture of Platelet-Derived Wound Healing Factor (PDWHF) prepared from a pooled platelet concentrate of a matching blood group. The application of the PDWHF resulted in a remarkable change in ulcer appearance which was completely cured after 141 days of treatment. The locally applied PDWHF appears to have stimulated the granulation tissue formation and accelerated the reepithelization.
Due to their ability to differentiate into various cell types and their immunomodulatory properties [8][9][10] mesenchymal stem cells (MSCs) are thought to have therapeutic potential in the healing of various types of skin defects including leg ulcers, trauma, burn wounds, and scar excision [11] due to their involvement in every stage of the wound healing process [12]. The resolution of in ammation is essential to successful wound healing, and chronic in ammation can lead to poor healing outcomes. Studies have shown that MSCs possess both anti-in ammatory and pro-in ammatory properties, promoting in ammation when the immune system is under-activated and exerting immunosuppressive effects when the immune system is over-activated to avoid damage secondary to in ammation [12,13]. MSCs stimulated by pro-in ammatory cytokines, including tumor necrosis factor (TNF) and interferon-gamma (IFN-γ), modulate immunosuppression by secreting high levels of soluble factors, including indoleamine 2,3-dioxygenase (IDO), prostaglandin E 2 (PGE2), nitric oxide (NO), hepatocyte growth factor (HGF), haem oxygenase (HO), cyclooxygenase-2 (COX-2), interleukin-4 (IL-4) and interleukin-1 (IL-1) receptor antagonist preserving T-cell activation. In addition, the MSCs exhibit a pro-in ammatory phenotype by secreting proin ammatory chemokines, including metalloproteinase-1 (MIP-1), interleukin 8 (IL-8) and interleukin 8 (IL-10) which enhance the T-cell response [14][15][16].
To study wound healing a skin substitute scaffold can be used. This is a three-dimensional organotypic culture [17] typically used for skin biology studies and testing of topically applied products. The epithelial cells and the broblasts of the skin equivalent secrete cytokines, chemokines, and growth factors that favour skin regeneration and provide a protective layer over the wound [18,19].
Our objectives were a) to generate a novel bioactive platelet-rich plasma (PRP)-leukocytes-depleted scaffold to develop an in vitro model of the typical clinical wound of patients with poor chronic skin perfusion and low leucocytes in ltration and b) to determine whether MSCs affect cytokine expression in this wound model. An air-liquid interface model was used to reproduce a full-thickness skin consisting of a co-culture of epithelial cells and broblasts, seeded into a PRP-leukocytes-depleted scaffold.
The PRP leukocytes-depleted scaffold was stimulated with calcium chloride (CaCl 2 ) which primes and degranulates platelets, hence providing the support structure [20] and being a source of growth factors involved in haemostasis, wound healing and tissue regeneration in various injured tissues [21]. The PRP-leukocytes-depleted scaffold aimed to avoid the excessive cytokine expression from white blood cells that may delay or impede the wound healing process. Although the in ammatory response is fundamental for wound healing, a subset of in ammatory cells can result in delayed healing [22]. Furthermore, leukocyte-depleted PRP is more predictable, as growth factor release and brin scaffold integrity are conserved when exposed to in ammatory conditions [23].
TNF is a cytokine with pleiotropic effects upon cell growth, in ammation, and immune responsiveness.
Whereas the local effects of TNF are usually bene cial to the host, when generated at higher concentrations within chronic in ammatory lesions, the proin ammatory effects of TNF often become deleterious and systemic [24]. Recombinant TNF (rTNF) was therefore administered to mimic an in ammatory condition.
Expression of the genes encoding IL-8 and transforming growth factor-α (TGFA) involved in the in ammatory phase [25] and the proliferative phase [26] of wound healing was analysed to evaluate how the leucocyte depleted PRP-based skin equivalent, the rTNF administration, and the MSC infusion affect the nal therapeutic outcomes in an in ammatory state.

Materials And Methods
Skin tissue biopsies discarded after medical interventions were obtained to create a skin equivalent model through a co-culture of broblasts and epithelial cells. Pooled Buffy coats were used to isolate peripheral blood-derived MSCs.

Isolation and cell culture of primary epithelial cells and broblasts
The skin biopsies were washed in Dulbecco's phosphate-buffered saline (PBS) and then suspended in Dulbecco's Modi ed Eagle Medium (DMEM) supplemented with antibiotics (50 µg/mL gentamycin, 0.25 µg/mL amphotericin B, 100 units/mL penicillin, and 100 µg/mL streptomycin) as well as the antimycotic amphotericin B (Sigma-Aldrich, Milan, Italy). This was followed by overnight digestion at 4 °C with 1 mg/mL Collagenase/Dispase (COLLDISP-RO Roche, UK). The following day, they were incubated at 37 °C for 1 h with 0.25% trypsin-EDTA solution (Sigma-Aldrich) to separate the epidermis from the dermis.
The dermis and epidermis were cut into small pieces of approximately 1 mm 2 each and incubated separately at 37 °C for 2 h with 0.25% Trypsin-EDTA solution and mixing every 10 min. The enzymatic action was stopped by adding complete medium supplemented with 10% fetal bovine serum (FBS). The digested tissues were centrifuged at 100 × g for 10 minutes at room temperature to obtain cell pellets. The epithelial cells derived from the epidermis were plated in a 6-well plate and cultured with Stemline Keratinocyte medium II supplemented with 5 µg/mL hydrocortisone and 10 ng/mL human recombinant keratinocyte growth factor (KGF) (Sigma-Aldrich). The dermal cells were plated in a 6-well plate and cultured with DMEM/F12 (1:1 mix) supplemented with 50 µg/mL L-ascorbic acid and 5 ng/mL human recombinant broblast growth factor (FGF-basic) (Sigma-Aldrich

Cell count and viability
Cell count and viability of the three cellular components involved in the study were performed before the assembly of the experimental model and administering the MSCs to the wound model. The assay was performed with Countess TM II Automated Cell Counter using a 0.4% Trypan Blue solution (Thermo Fisher Scienti c) as per the manufacturer's protocol. The analyses were performed in triplicate.

Preparation of the leukocyte-depleted PRP-based skin equivalent
Apheresis donor platelet concentrates were transferred into 50 mL centrifuge tubes and centrifuged for 20 min at 300 x g at 4 C. The platelet pellet was resuspended in 5 mL residual plasma to form the PRP. The epithelial cells were seeded on the PRP-based derma equivalent with a density of 6 × 10 4 cells/cm 2 to establish the prototype skin equivalent. This skin model was cultured with an air-liquid interface system allowing the development of a multi-layered skin equivalent. Once fully developed, the skin equivalents were collected and xed for 24 h at room temperature in 10% neutral buffered formaldehyde for sectioning and processing using a short histology protocol, with an overnight incubation step in 60% ethanol. The tissue-equivalent sections were embedded in wax blocks, cut at 5 µm, and transferred to slides and stained with Mayer's Haematoxylin-Eosin solutions (Sigma-Aldrich).

The in vitro experimental model of wound healing
In order to create an in vitro experimental model of wound healing, a scratch injury was in icted. Administration of human rTNF (Bachem AG, Switzerland) was used to mimic an in ammatory condition. The infusion of peripheral blood-derived MSCs was done 24 h after the administration of the rTNF to evaluate how their immunomodulatory properties affect the nal therapeutic outcomes in an in ammatory state. The three different conditions of wound healing created in the study were: Control: a scratch injury was in icted but no rTNF or MSC was subsequently administered.
Treated: A wound was in icted and 30 ng/ml of rTNF was administered 3 h after wound in iction. The purpose of this condition was to reproduce an in ammatory condition.
Study: A wound was in icted, 30 ng/ml of rTNF was administered 3 h after wound in iction, and 6 × 10 4 cells/cm 2 MSCs were infused 24 h after the administration of rTNF. The purpose of this condition was to study whether the presence of MSCs alters the in ammatory response.
Skin equivalent models from all conditions were collected at day 0 (3 h after wound in iction) and then at 1, 2, 4-and 8-day intervals after the wound was in icted. The in vitro model was repeated in biological triplicate.
RNA extraction and qPCR RNA was extracted from 5 × 10 adherent MSCs, 5 × 10 adherent broblasts, and the leukocyte-depleted PRP-based skin equivalents using the Pure Link® RNA Mini Kit (Thermo Fisher Scienti c). In brief, leukocyte-depleted PRP-based skin equivalents were harvested quickly and were immediately digested and homogenised with 45 µl of Digestion Buffer and 5 µl of Proteinase K using a homogenizer (Speed Mill PLUS Analytic Jena AG, Germany and normalized against the endogenous reference gene ACTB (encoding β-Actin), and the data are reported as the mRNA fold change. The expression levels of the skin equivalent models before scratching were used as calibrators.

Statistical analysis
The data are presented as mean ± standard deviation (SD). The statistical signi cance was determined using the Wilcoxon signed-rank test, a nonparametric alternative to paired t-test, or a paired student t-test. P < 0.05 was considered statistically signi cant.

Results
Successful isolation and culture of primary epithelial cells, broblasts, and skin equivalent.
The epithelial cells displayed a short and wide spindle-shape and epithelial morphology like a "pavement stone" (Fig. 1A). The cultures reached con uence and were ready to passage within two weeks. The growth of broblasts from the dermal tissue was observed by day 10 after tissue plating. The dermal tissue was removed from the plate 15 days after plating, where the proliferating broblasts initiated the primary culture reaching con uence at day 20 after plating (Fig. 1B) The skin equivalent model was assembled through a co-culture of epithelial cells and broblasts-seeded into the PRP-leucocyte depleted scaffold and cultured in an air-liquid interface system allowing the development of a multi-layered skin equivalent (Fig. 1C).
Microscopical observation of the scaffold con rmed the presence of con uent and multi-layered epithelial cells (Fig. 1D), and the presence of proliferating spindle-shaped broblasts (Fig. 1E) on the PRP scaffold. Histological examination of the skin equivalent showed the presence of different epithelial cells organized in a typical skin tissue. The nuclei of the epithelial cells and the broblasts were stained in violet (Fig. 1F).
Characterization of primary culture of epithelial cells and broblasts.
The presence of epithelial stem cells and proliferating epithelial cells in the skin equivalent allows for the development of a multi-layered epithelium. In order to identify the presence of epithelial stem cells and committed epithelial progenitors on the primary epithelial cell culture we evaluated the expression of wellestablished markers such as the epithelial stem cell markers CD34 (CD34 molecule) [27] and the CD133 (Prominin-1) [28], CD90 [29] and the committed progenitor epithelial cells marker CD326 (epithelial cell adhesion molecule) [30].
Epithelial cells were also identi ed through the presence of specif lineage markes such as CD29 and CD44.

Characterization of peripheral blood-derived MSCs
FITC-IgG1, PE-Cy5-IgG1, PE-Cy7-IgG1, PE-IgG1 and PerCP-IgG1 conjugated isotypes were used as negative controls to identify and quantify the percentage of positivity of the stained population for each marker. Flow cytometric analysis performed on peripheral blood-derived MSCs at the rst passage showed the presence of two separate populations, p1 (Fig. 3A) and p2 (Fig. 4A) with different immunophenotypic patterns.
Isolation, culture and gene expression of MSCs Peripheral blood isolated mesenchymal stem cells showed their properties to form colonies (Fig. 5A), to adhere ( Fig. 5B and Fig. 5C) and to have a spindle-shape broblastic picture ( Fig. 5B and Fig. 5C) The ability of the MSCs to differentiate into the three lineages was analysed by histological staining and microscopic observation. Adipocytic differentiation was reached at day 14 after stimulation, as determined by the presence of characteristic lipid vacuoles in the cytoplasm. The Oil Red O stained the fat in the lipid vacuoles while the nuclei were counterstained with Mayer's Haematoxylin solution.
Osteogenic differentiation was achieved at day 21 after stimulation and was con rmed by the presence of dark-coloured calcium deposits stained with Von Kossa solution. Chondrogenic differentiation was noted at day 21 after stimulation as determined by the presence of proteoglycans stained with Toluidine Blue solution.

Viability of the various cell types used for the wound model
Cell viability tests were performed in triplicate on the three cellular cell types: broblasts, epithelial cells, and MSCs. Viability (mean ± SD) was 85.89 ± 0.51% for broblasts, 85.67%±0.58% for epithelial cells and 85 ± 3.2% for MSCs.

Effects of scratch injury, rTNF, and MSC on gene expression in skin equivalent model
We investigated the response of the skin equivalent cells in our experimental, leukocyte-depleted PRPbased skin equivalent model of a chronic wound associated with low skin perfusion. For this purpose, we performed real-time qPCR analysis to evaluate the level of expression of genes involved in the in ammatory (IL8) (Fig. 6A) and proliferative (TGFA) (Fig. 6B) phases at various time points in response to the scratch injury, rTNF administration, and MSC infusion.
In these experiments, we monitored a control condition which received only the scratch and was used to verify the in ammation and proliferation of an untreated wound, while the treated condition and the study condition were administered a scratch injury and rTNF, or a scratch injury, rTNF, and MSC respectively (Fig. 6A,B).
As shown in Fig.xx, the expression of IL8 increased markedly within 3 h of the scratch injury, whereas the expression of TGFA decreased.
A decrease in the expression of IL8 was noted in day 1 in the condition treated with rTNF (30 ng/ml) when compared with the untreated control condition (Fig. 6A).
The expression of IL8 decreased steadily in the control condition from day 2 onwards but increased slightly in the treated condition. The expression of IL8 decreased signi cantly in the study condition in day 4 and almost reached the levels which were present prior to the injury but then increased again in day 8 (Fig. 6A).
The expression of TGFA increased in the rst two days and decreased in day 4 in all groups. A signi cant decrease was noted in the study group between day 4 and day 8 (Fig. 6B).

Discussion
Skin equivalents represent the rst examples of three-dimensional organotypic cultures which are often used for in vitro research study of normal and abnormal skin biology [17,31].
In this study, we developed a new wound model on which to observe the immune response during the wound healing process through the expression of IL8 and TGFA. The control condition was not treated, the treated condition was given an additional in ammatory stimulus through the administration of 30 ng/mL rTNF which regulates genes that code for in ammatory mediators [32], while the study condition was administered with 30 ng/mL rTNF and 6 × 10 4 cells/cm 2 MSCs. This model is shown in Fig. 7.
Studies on monolayer cultures of keratinocytes submerged into culture media do not resemble the true nature of the physiological process of wound healing [33]. Multi-layered differentiated models are comparable to native skin and produce excellent results when analysing epithelial attachment, proliferation, differentiation, and dermal remodelling [34]. Such models support further expertise of skin biology and skin diseases, without the complexity of the intrinsic interactions found in native skin [34,35].
One example of a skin model cultured in an air-liquid interface system is the In Vitro Reconstructed Human Epidermis [36], which does not contain broblasts. On the other hand, de-epidermidised dermis and collagen framework, which includes only the broblasts, basement membrane, and extracellular matrix (ECM) components and which are the most commonly used dermal equivalents [34] is not su cient to reproduce a skin model. It was shown that the inclusion of broblasts enhances the production of extracellular matrix proteins, generating a more normal epidermal architecture [34]. Therefore, the interaction between epidermal and dermal components is needed for adequate wound healing. In another study, a skin equivalent was developed by implanting keratinocytes onto the upper surface of a collagen scaffold, occupied with broblasts and culture at the air-liquid interface [37].
It is known that platelet-rich plasma contain more than 300 biologically active molecules containing growth factors and pro-in ammatory and immune-modulating cytokines that can activate the platelets themselves, perpetuating the in ammatory cycle [27].
Moreover, the platelets trapped into the brin matrix release the growth factors slowly over a duration of seven days, in contrast with the use of exogenous thrombin, where almost all growth factors are discharged during the rst hour [38].
Several studies reported positive results on the application of PRP in stimulating the wound healing process [39] and increasing keratinocyte migration [35]. Previous attempts to use platelet-derived products were tested in skin tissue engineering. One study described the use of platelet lysate, combined with chitosan and 107 hyaluronic acid dressing [40]. At the same time, others used platelet lysate in conjunction with a collagen/gelatin scaffold [41] or a collagen type I gel which was mixed with PRP [42]. These studies all showed promising outcomes, supporting the approach of the use of platelet-derived products in wound healing.
Our novel skin model consisted of a leukocyte-depleted, platelet-rich plasma scaffold, with embedded broblasts, as dermal equivalent and seeded keratinocytes on it as multi-layered epidermidis (Fig. 7). Calcium chloride was used as an activator to initiate the formation of autologous thrombin from prothrombin, forming a brin clot that provided a surface for keratinocyte seeding and enabled the skin cells to mature into stratum corner and basal, spinous and granular layers. The lack of leukocytes allowed for the mimicking of typical chronic wounds of patients with poor skin perfusion and low leukocyte in ltration. The leukocyte-depletion allowed for the evaluation of the immunomodulatory properties of the infused MSCs, which modulate the IL8 and TGFA secretion.
We subsequently used our new wound model to analyse cytokine gene expression under three conditions: control, treated, and study conditions.
In physiological wound process, the ECM components, such as bronectin, glycosaminoglycans and collagens, regulate the dynamic and interactive process of wound healing. [43] The platelets are early modulators of the healing process [44] and the blood clot formed upon platelets activation provides a provisional "scaffolding" containing brin molecule and plasma bronectin. This occurs during the rst 24 hours after the injury and enables formation of a temporary matrix in the wound bed. [45] Therefore, our PRP-based scaffold as dermal equivalent reassembles the physiological scaffolding formed during the hemostatic phase and required for the normal wound process.
The initial wave of in ammatory phase is characterized by IL8 production by platelet α-granules and skin resident cells to reduce blood loss and ll the tissue gap with a blood clot rich in platelets, macrophages, leukocytes and mast cells producing/secreting cytokines and growth factors [46][47][48][49].
The in ammatory response occurs within hours of the occurrence of the damage as a localized or systemic protective response. It is activated by molecules expressed by pathogens or associated with tissue injury and are recognized by Toll-like receptors (TLRs) present on skin resident cells [50]. TLRs activation in response to injury and in ammation is responsible for the upregulation of IL8 [51][52][53].
A signi cant upregulation of IL8 expression was noted three hours after the scratch injury when compared to the levels exhibited just before, thereby con rming the success of our scaffold in mimicking the wound. On the other and, the scratch injury exhibited a down regulatory effect on the expression of TGFA.
In addition to the induction of in ammation by chemokines, other molecules such as TNF promote the in ammatory response following wounding.
It has been shown that the prolonged stimulation of their TLR receptors causes downregulation of TLR2 and TLR4, most likely as a self-regulatory mechanism to prevent overactive skewing of the immune response [54]. In our model we noted a signi cant down regulation of IL8 following the administration of rTNF which appears to indicate the delayed the activation of the in ammatory response. The progression of IL8 in the treated group occurred in delay (i.e., at a later time point) when compared with the control group.
TLR ligation triggers the release of in ammatory mediators initiating innate immune responses mainly through the activation of macrophages, neutrophils, leucocytes, and stromal cells including MSCs, thus creating an in ammatory environment [55][56].
Neutrophils and monocytes/macrophages represent the key cells of the in ammatory phase [57] as their simultaneously release of large number of cytokines and growth factors are crucial to initiate the next phase of the healing process [58]. Neutrophils appear in the wound area a few minutes after the injury [59] and are replaced after two or three days by monocytes that undergo a transformation into macrophages [60]. Macrophages are cells of great importance for the healing process [61] as they participate in phagocytosis and are also the main source of cytokines and growth factors stimulating the proliferation of broblasts and collagen biosynthesis [62][63].
It was noted that a decreased in ux of neutrophils in the rst 4 days after the in iction of a wound has a negative impact on healing outcomes. [64] It is well known that macrophages switch phenotypes from an M1 pro-in ammatory phenotype to an M2 pro-repair phenotype leading to the reduction of in ammatory markers and the promotion of the proliferation phase. [65]. Moreover, macrophages secrete PDGF, TGF-, and bFGF, which modulate the epithelialization, collagen accumulation, and angiogenesis. [66]. During the proliferative phase there is an increase in migration and proliferation of broblasts and endothelial cells as well as keratinocytes, which secrete bFGF, EGF, VEGF, bFGF, and PDGF, TGF-and KGF. TGF-a mRNAs were isolated in both wound macrophages [67] epidermal keratinocytes at the wound edge. [68]. Based on its expression level, TGFA can be considered as a biomarker of the early phase of re-epithelialization. [69] The results obtained with our model indicate that all three conditions studied were in an in ammatory state throughout the study as shown by the lower expression of TGFA when compared with IL8.
The absence of leucocytes, which promote the resolution of the in ammation by releasing numerous potent cytokines, probably led to a delay of the proliferative phase.
In experimental models, the stimulation of MSCs with the pro-in ammatory cytokine TNF upregulates expression of a subset of TLRs, thus increasing the sensitivity of MSCs to the in ammatory milieu [70].
We postulate that the MSC infusion could modulate the expression of IL8 and that the decrease of IL8 expression in the study condition at day 2 and especially at day 4 could indicate that the presence of MSCs inhibited the in ammatory response in contrast with an increase in the treated condition.
TLR4 receptor activation triggers the MSC1 population which exhibits a pro-in ammatory pro le while activation of the TLR3 receptor activates the MSC anti-in ammatory phenotype MSC2 [14,15,71]. MSCs are known to display an anti-in ammatory phenotype in an in ammatory environment as characterized by increased mRNA expression of IL6 [72].
Our data suggest the possibility that the MSCs modulate the in ammatory response, switching from an immunosuppressive phenotype to a pro-in ammatory phenotype and regulating the IL8 expression. Presumably, this switch in our model occurred between day 4 and day 8, showing a substantial increase of mRNA expression in the study condition in day 8 when compared with the treated condition.
TGFA was down regulated throughout all time points in all the three conditions in our study. Interestingly, the changes in expression of TGFA had a similar pattern of to the changes in the expression of IL8 between days 2 and 8 in the study condition. We hypothesize that the modulation of IL8 could affect the expression level of TGFA.
Keratinocytes and broblasts could have both contributed equally to synthesize IL8.
We also suggest that the presence of a pro-in ammatory cytokine (TNF) stimulates MSCs to exert their immunomodulatory properties secreting directly IL8 or/and having a paracrine effect on IL8 and TGFA production by acting on the resident skin cells.
Further investigation is necessary to address the speci c cellular source of IL8 and TGFA production and the cellular target of the MSCs paracrine action and hence to evaluate the clinical relevance of the infusion of MSCs in a scenario of a low blood supply in the site of the wound.

Conclusion
A leukocyte-depleted, platelet-rich plasma scaffold was developed which allowed for the mimicking of typical chronic wounds of patients with poor skin perfusion and low leukocyte in ltration. The wound model indicated accurately how the resident skin cells produce IL8 and deregulate TGFA, in response to damage in a physiological and in ammatory state. The change in the expression level of IL8 con rms that an in ammatory condition was created. Notwithstanding the absence of leukocytes, MSCs modulated the IL8 expression in response to damage in an in ammatory environment by acting directly on skin resident cells. Therefore, MSCs could have therapeutic potential for the treatment of ulcers in patients with poor skin perfusion.   for CD90 (C), CD34 (D), CD326 (E), and CD133 (F) stemness markers; G) mRNA expression of genes involved in broblast characterization was determined by qPCR. Transcript levels were normalized to the ACTB reference gene using the 2-ΔCt method. The data are presented as mean ± standard deviation (SD).

List Of Abbreviation
The graph bar shows expression level of the genes CD90, CD73, CD105, CD45, and CD34 of cultured broblasts. Gene expression was con rmed by 1.5% agarose gels.    CD73, CD90 and low for CD105 and CD29. Gene expression was con rmed by analysis of the products on 1.5% agarose gels. IL8 and TGFA Gene Expression Bar graphs showing the log fold change expression levels of the genes IL8 (A) and TGFA (B) of the control, treated, and study conditions for the ve time points at which they were measured. The mRNA expression was determined by qPCR. Relative transcript levels were normalized to the ACTB reference gene. The expression levels of the skin models before scratching were used as a calibrators. Data presented as log fold change using the 2−ΔΔCt method. P values were worked out through a paired student t-test. p <0.05 was considered as statistically signi cant and is shown in the gure where applicable (* p<0.05, ** p <0.01, *** p<0.001).