MSC from the umbilical cord, stored in the Cell Bank of the Laboratory of Skin Physiology and Tissue Bioengineering of School of Arts, Sciences and Humanities of University of São Paulo (EACH-USP), were used in accordance with the Ethics Committee of the University of São Paulo (Number 4.351.554). After thawing, cells were plated in 75 cm² flasks at a density of 5 × 105 cells/mL and incubated in a humidified atmosphere, 5% CO₂, at 37°C in Dulbecco’s modified Eagle’s medium (DMEM) with low glucose concentration (2 mM) and 10% fetal bovine serum (FBS), named LD10, and called proliferation medium.

Human primary dermal fibroblasts from a 10 months-old male in first passage (code nh-skp-FB0050) and human basal keratinocytes from a 4 years-old female in second passage (code nh-skp-KT0085) were obtained from Rio de Janeiro cell bank (BCRJ). These cells were isolated from the epidermal and dermal layer, respectively, of human neonatal foreskin. Fibroblasts were cultured with LD10 using the same conditions as MSC. The keratinocytes were maintained in Keratinocyte Growth Medium (KGM) with its respective supplement (KBM™ Gold™ Basal Medium and KGM™ Gold™ SingleQuots™, according to manufacturer instructions, Lonza^®) under a humidified atmosphere, with 5% of CO₂ at 37°C.

To skin in vitro reconstruction, the air-liquid interface (ALI) technique was used, in which the dermal equivalent is in contact with the culture medium, and the epidermal equivalent is in contact with the air. The culture medium was changed every 48 h and the samples were analyzed in different periods of culture in ALI. Two different approaches were used to mimic the dermis: rat tail type I collagen (Corning Costar) with fibroblasts and the porcine decellularized skin matrix MatriXpec™ (TissueLabs™). For the dermal equivalent of collagen, human dermal fibroblasts (1.5 × 105/dermal equivalent) were resuspended in 3.0 mg/mL type I collagen solution. This solution was transferred to a 24-well plate and kept in an incubator at 37°C for 20 min. The plate was kept for 2 h at 37°C in a humidified atmosphere with 5% CO₂. For the decellularized skin matrix MatriXpec, dermal fibroblasts (1.5 × 105/dermis) were resuspended in 1 mL of the commercial product and incubated at 37°C in a humidified atmosphere with 5% CO₂ for gelatinization.

Epidermal equivalents were composed of 2.5 × 105 cells/epidermis: basal keratinocytes (positive control) were cultured in KGM and MSC were cultured in LD10 (proliferation medium) or KGM. For epidermis assembly, cells were first immersed, for 1 day, in KGM/DM (1:1) (DM, differentiation medium) and then exposed to ALI for 10 days in KGM/DM medium only facing the bottom of the dermal equivalent. DM was prepared by adding DMEM and Ham’s-F12 medium (GibcoBRL, Gaithersburg, MD) (3:1), 10% FBS and the supplements 10.1 nM cholerae toxin, 5 μg/mL insulin, 5 μg/mL apo-transferrin, 0.4 μg/mL hydrocortisone 21-hemisuccinate, 1 ng/mL epidermal Growth Factor (EGF), according to Catarino et al. (2018). MSC cultured with proliferation medium were used as negative control for in vitro skin reconstruction. The culture medium was changed every 48 h.

After the period of exposure to ALI of 10 days, the samples were fixed in 10% formalin buffered in PBS at 4°C for 4 h, dehydrated in increasing concentrations of alcohols and embedded in paraffin. The blocks were then cut on a Leica RM2165 automatic microtome (Leica, Wetzear, Germany) producing cross-sections of approximately 5 µm. The sections were placed on histological slides and dried in an oven at 60°C for 2 h to remove paraffin excess. Then, the samples were diaphanized in 100% xylene for 10 min, hydrated in increasing alcohol content (100%, 95%, 85%, and 70% ethanol), washed in distilled water for 5 min, and stained with hematoxylin-eosin (HE). Cover slides were mounted with Permount^® (Fisher Scientific, Pittsburgh, United States). The microscopic images were obtained using an Olympus BX 60 microscope with an attached digital camera (Axio CAM HRc—Zeiss^®).

After 4, 7, and 10 days of ALI, the organotypic cultures were collected and washed with PBS. For RNA extraction, the Mini Kit RNeasy (Qiagen, Frederick, MD, United States) was used, according to the manufacturer instructions. The samples were placed on the column containing a silica membrane, lysed with RLT buffer and homogenized. Ethanol was then added to the lysate, creating the conditions that promote selective binding of RNA to the membrane. Contaminants were washed in RW1 buffer, and RPE buffer and then the RNA was eluted in Ultra-pure water. After quantification, the complementary strand of DNA (cDNA) was synthesized from 2 ng of RNA using the RT2 First Strand kit (Qiagen, Frederick, MD, United States) in two steps: the first consisted of eliminating the genomic DNA using 8 μL of RNA (2 μg) and 2 μL of ×5 concentrated GE buffer, the mixture being incubated for 5 min at 42°C; the second step refers to reverse transcription, in which 10 µL of the product from the first step, 4 µL of 5x concentrated reaction buffer (BC3), 1 µL of P2 control (oligo dT) and dNTPs were used 10 mM, in a final volume of 20 µL. This solution was incubated for 15 min at 42°C and then for 5 min at 95°C in a thermocycler (Eppendorf, United States), according to the manufacturer recommendations.

To quantify the expression of CK14, CK10, involucrin, filaggrin, e-cadherin and vimentin, the quantitative PCR technique was used using the SYBR^® green PCR Master Mix (ThermoFisher Scientific, Massachusetts, United States) and the following program: 95°C for 10 min, 40 cycles of 95°C for 30 s, 60°C for 45 s and 72°C for 1 min. The final step was performed at 72°C for 10 min, according to the manufacturer instructions. The expression of the genes of interest was normalized by the expression of the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene. Dermal fibroblasts, MSC and keratinocytes two-dimensionally cultured, in addition to the control skin formed by stratified keratinocytes, were also analyzed for the expression of these proteins. The expression was analyzed in relation to the epidermis formed with MSCs grown in LD10 (proliferation medium), used as a negative control. The oligonucleotides were designed, according to Table 1, in order to avoid non-specific annealing and respecting the recommendations for use in quantitative PCR. Reactions were performed on the Eco™ Real-Time PCR System (Illumina, San Diego, CA, United States).

After deparaffinization and hydration, the slides were washed with PBS, permeabilized with PBS containing 0.1% Triton X-100 for 20 min and washed with PBS. Antigen retrieval was performed by incubating the slides in 0.01 M citrate buffer pH 6.0 for 30 min at 95°C. Non-specific sites were blocked by the incubation of samples with PBS containing 10% FBS for 20 min. Then, the primary antibodies anti-CK10, anti-CK5 and anti-involucrin were added (1:100), protected from light for 12 h at 4°C. For visualization of positively labeled cells, FITC-conjugated anti-IgG secondary antibodies were used. Slides were washed and analyzed under a fluorescence microscope (Zeiss, Germany). The pictures are representative of images analyzed in blinded tests, with n ≥ 3 in at least 10 random fields registered under the same exposure. The images were analyzed using image J software (https://imagej.nih.gov/ij/; Schneider et al., 2012). The mean fluorescence intensity per pixel was determined, and the threshold of 10 was defined as background.

To evaluate the enzyme activity of epidermal kallikreins, the dermal substitutes, consisting of collagen with fibroblasts or MatriXpec, were initially dissociated from the epidermis by mechanical separation, and the samples were lysed with 200 μL of 25 mM HEPES buffer, pH 7.5, 0.5% Triton X-100 (lysis buffer) producing extracts. The epidermis formed by basal keratinocytes was analyzed after 4, 7, and 10 days of ALI, while the epidermis potentially formed by MSC was analyzed only after 10 days of ALI. The activity of KLK in extracts was verified on the hydrolysis of fluorogenic substrates (FRET) (Angelo et al., 2006; Loura, 2011) suitable to detect KLK5 (Abz-KLRSSKQ-EDDnp), KLK6 (Abz-AFRFSQ-EDDnp) and KLK7 (Abz-KLYSSKQ-EDDnp) activities. 40 μL of samples were used in 50 μL of 2x concentrated enzyme reaction buffer (100 mM Tris, pH 8.0, 300 mM NaCl, and 0.01% Tween 20) and 10 μL of substrate (100 μM), in a final volume of 100 μL. Substrate hydrolysis (10 μM) was accompanied by an increase in fluorescence at λex = 320 nm and λem = 420 nm in the Synergy HT plate reader (Biotek, Winooski, VT, United States) according to Santos et al. (2019). Mean rate results for each reaction were obtained using Gen5™ software (BioTek, Winooski, VT, United States). Enzyme activity was measured in arbitrary fluorescence units (AFU) per min and expressed in terms of the mean reaction rate per 1 mg of protein present in the extracts.

Quantitative data were expressed as mean values ± standard error (SE) with n ≥ 3. For statistical comparison, the results were analyzed by the GraphPad Prism 5 software, using the two-way ANOVA plus Bonferroni post-test, with statistically significance of p < .05 (*); p < .01 (**), p < .001 (***) from the control. Under microscopy, the morphology of the cells, their three-dimensional arrangement and possible epidermal stratification were evaluated using the image J software (Schneider et al., 2012).

The cultivation of basal keratinocytes onto the dermal equivalent constituted by collagen with fibroblast resulted in a typical stratified epidermis I with all well-defined layers (Figures 1A, B), confirming that 10 days of culture in ALI, under the specified conditions, were sufficient for terminal differentiation of keratinocytes, being considered as a positive control.

MSCs cultured for 1 day with KGM/DM and exposed to ALI for 10 days proliferated and formed a stratified organotypic culture composed by cells distributed in a basophilic intercellular substance (Figures 1C, D). In these samples, the dermo epidermal separation was not clear and the presence of unstained vacuoles was verified. MSC grown with LD10 were arranged as a monolayer over the dermal equivalent (Figures 1E, F) and their cultivation resulted in the lowest epidermal thickness compared to the differentiation medium.

Macroscopically, MSC grown on MatriXpec™ with LD10, unlikely it was observed for the dermal equivalent of collagen with fibroblasts, promoted an intense contraction of the dermal support suggesting that MSC exhibited great interaction with this material. Also, the cells showed exacerbated migration/proliferation towards the dermal equivalent (Figures 2A, B), when compared to MSC cultured onto collagen (Figures 1C–F) resulting in some cases in the complete wrapping of MSCs by the matrix. This phenomenon was not observed when cells were cultured with KGM/DM (Figures 2C, D). In this substrate, MSCs are able to proliferate and form layers resulting in a stratified organotypic culture with no typical epithelial phenotype, similar to the observed when MSC were grown onto the dermal equivalent of collagen with fibroblasts. The cultivation of basal keratinocytes onto MatriXpec resulted in a typical stratified epidermis (Figures 2E, F) similar to the observed when these cells were cultured onto the collagen support.

The expression of CK14 and CK10, present in the basal and suprabasal strata of the epidermis, respectively, was evaluated after different periods of exposure to ALI. The expression of CK14 was verified on the 7th day of culture in the potential epidermis obtained with MSCs cultivated with KGM/DM, with no significant expression being observed in the other analyzed periods (Figure 3A), when compared to MSCs cultivated in LD10 at the same day of ALI. The presence of this protein was not observed in MSCs and fibroblasts cultured with LD10 in 2D systems whereas basal keratinocytes both in 2D and 3D systems (control skin) showed CK14 expression.

After 7 days of ALI, there was an increase in the expression of CK10 in skins formed by MSCs induced to differentiation. On the 10th day of culture in ALI, it is still possible to observe the significant expression of this protein, which is absent in MSCs cultivated with LD10 for the same period (Figure 3B). The expression verified on the 4th day of culture was similar to the expression in basal keratinocytes, suggesting the commitment of MSCs to the epidermal lineage at the beginning of the culture period, which did not occur in MSCs and fibroblasts cultured with LD10.

Involucrin expression was also investigated after 4, 7, and 10 days of exposure to ALI. Similar to CK10, involucrin had its expression increased on the 7th day in potential skins formed with MSC induced to differentiation compared to organotypic cultures formed MSC cultivated with LD10 (Figure 3C). 2D-cultured basal keratinocytes showed involucrin expression 16.3 times lower than the control skins while 2D-cultured MSC in LD10 did not show involucrin expression.

Filaggrin expression presented a different expression profile in comparison to the other analyzed proteins showing an increase at the end of the cultivation period (10 days) in MSCs induced to differentiation (Figure 3D). 2D-cultured basal keratinocytes showed lower expression (8.2-fold) of filaggrin than control skin, and MSC dimensionally cultured in LD10 did not show filaggrin expression, as expected.

The expression of those epidermal markers was accompanied by E-cadherin up regulation (Figure 4A) and vimentin down regulation (Figure 4B), suggesting the transition to epithelial phenotype and loss of the mesenchymal marker vimentin by MSC induced to epidermal differentiation.

After 10 days of exposure to ALI, the expression of different epidermal markers was analyzed by immunofluorescence allowing access to the localization of those proteins in the potential reconstructed skin. MSCs grown in LD10 onto collagen with fibroblasts or MatriXpec™ were not positively labeled for CK5 (Figures 5A–D), a protein found in the basal layer of the epidermis (Figure 5C). This protein was only identified in cells induced to differentiate onto collagen (Figure 5B), although it was not specifically related to an epidermis stratum. Specifically, the expression of CK5 was barely verified when the cells were cultured with KGM/DM onto the MatriXpec™ support (Figure 5E). The mean fluorescence intensity per group is shown in the Figure 5F, in agreement with the visual detection.

Regarding CK10, a protein expressed in the suprabasal layers of the epidermis as identified in the control epidermis (Figure 6C), it is possible to observe the expression of this protein randomly distributed in the organotypic culture formed by MSCs induced to differentiation with KGM/DM using both matrices as dermal support (Figures 6D, E). Nevertheless, CK10 was barely detected when MSCs were cultured with LD10 onto collagen or MatriXpec™ supports (Figures 6A–D, respectively), as demonstrated by the fluorescence quantification (Figure 6F).

Similarly observed for CK10, involucrin, a protein expressed in the terminal differentiated layers of the epidermis (Figure 7C), was not identified when MSC were cultured with LD10, either onto collagen (Figure 7B) or (Figure 7D) MatriXpec™ dermal equivalents. However, under differentiation onto the collagen matrix (Figure 7C), cells showed involucrin expression, which was verified mainly in the upper layers of the organotypic culture. The expression was also detected in cells differentiated on MatriXpec™, but involucrin positive cells were distributed along the potential epidermis without a characteristic arrangement (Figure 7E). The fluorescence intensity per group is shown in the Figure 7F, being possible to observe involucrin expression, at least, 3-fold higher in the organotypic cultures of MSC induced to differentiation onto collagen support in comparison to the ones cultivated with LD10 (control). This difference was less marked between MSC cultivated onto MatriXpec™ with KGM/DM or LD10.

For all KLK, at the different periods of time, except for KLK6 on day 7, it was observed a difference of at least 1.5 times between activity in the control epidermis and dermis, formed by keratinocytes and collagen with fibroblasts, respectively. The highest difference (up to a 3-fold) in KLK expression in these two layers occurred on day 4 (Figure 8). There was no difference between KLK activity in the control epidermis at different time periods.

Regarding the KLK5, KLK6 and KLK7 activities in the potential epidermis formed by MSC, there is no significant difference in comparison to the control epidermis composed by basal keratinocytes when grown for 1 day immersed in KGM/DM and then submitted to ALI in DM medium (Figure 9).

Specifically, KLK5 activity was 1.3-fold higher in cells induced to differentiate onto collagen and 1.8-fold higher over MatriXpec™ when compared to samples cultured in LD10 (Figure 9A). Additionally, KLK5 activity in MSC cultured onto collagen was 1.4 times higher than in cells cultured on MatriXpec™.

Even though KLK6 activity was not well detected in MSC induced to differentiate into keratinocytes onto MatriXpec™, the enzyme activity was higher in comparison to cells cultured with LD10 on the same dermal substrate (Figure 9B). Importantly, this activity was similar between MSC cultured with KGM/DM onto collagen and control epidermis, and it was 50% smaller in MSC cultured with LD10, indicating that epidermal differentiation was triggered.

KLK7 activity in differentiated MSC cultivated onto MatriXpec™ was 5.3 times higher than in cells not induced to differentiate on the same dermal support. In the collagen matrix, the activity was 1.9 times higher in cells induced to differentiate in comparison to cells cultured with LD10 during the same period. Furthermore, this activity was 2.8-fold higher in cells cultured on MatriXpec™ in relation to cells cultured on collagen considering the same period of cultivation (Figure 9C).