Skin samples were obtained from adult patients who underwent abdominoplasty at the Red Cross Hospital in Beverwijk, Medical Clinic in Velsen or Spaarne Gasthuis in Haarlem. Samples from 17 different donors were used (donor age: 48 ± 13 years; sex: 93% female). Consent for the use of these anonymized, post-operative residual tissue samples was received through an informed opt-out protocol, in accordance with the national guidelines (https://www.coreon.org/) and approved by the institutional privacy officers. Subjects were actively informed of this procedure and were able to easily withdraw at any point. Split-thickness samples of 0.3 mm were harvested using a dermatome (Aesculap AG & Co. KG, Tuttlingen, Germany).

See Supplementary Table 1 for the contents of culture media. Harvested skin was incubated in 0.25% dispase (Gibco, ThermoFisher Scientific, Paisley, UK) at 37°C for 45 min. The epidermis was separated from the dermis using forceps. For fibroblast isolation, the dermal part of the split skin was cut into small pieces and submerged into a 0.25% collagenase A (Roche, Basel, Switzerland) solution at 37°C for 2 h. After addition of 1 mM EDTA (Life Technologies, Paisley, UK) + PBS (Gibco) to inhibit enzyme activity, the cell suspension was poured through a 500 µm cell strainer (PluriSelect, Leipzich, Germany) and centrifuged for 10 min at 360 × g. The cell pellet was resuspended in culture medium and poured through a 70 µm cell strainer (Starstedt AG & Co. KG, Nümbrecht, Germany) and cultured at 37°C with 5% CO₂. For keratinocyte isolation, the epidermis was transferred into 0.05% trypsin (Gibco) and incubated for 20 min at 37°C. The cell suspension was poured through a 70 µm cell strainer and centrifuged for 10 min at 110 × g. Next, the cell pellet was washed in culture medium and centrifuged for 10 min at 160 × g. The cell pellet was then resuspended in CnT-07 medium (CELLnTEC Advanced Cell Systems AG, Bern, Switzerland) and keratinocytes were transferred onto a 1 µg/cm² collagen type IV (Sigma-Aldrich, Saint Louis, MO, USA)-coated culturing flasks (Starstedt) at 37°C with 5% CO₂.

Our FSE development protocol was based on previous experiments (30). MatriDerm^® (MedSkin Solutions Dr. Suwelack AG, Billerbeck, Germany) with a thickness of 3 mm was cut into circular pieces of 1.13 cm². At day one, 2 × 10⁵ fibroblasts were seeded onto the matrix and the matrix was submerged in culture medium containing 65 µg/mL ascorbic acid for 4 days at 37°C with 5% CO₂ ( Figure 1A ). Subsequently, 1 × 10⁵ keratinocytes were seeded on the opposite side and the model was cultured submerged in FSE I medium containing 2 ng/ml KGF (ImmunoTools GmbH, Friesoythe, Germany) and 0.5 ng/ml EGF (R&D Systems, Inc., Minneapolis, MN, USA) for 4 days at 37°C with 5% CO₂. Next, the FSE was transferred to a transwell (Starstedt) and cultured air-exposed in deep well plates (Greiner Bio-One BV, Alphen aan den Rijn, the Netherlands) with FSE II medium containing 4 ng/ml KGF and 1 ng/ml EGF. From day 11 onward the FSE was cultured in FSE III medium containing 4 ng/ml KGF and 1 ng/ml EGF and from day 15 onward in FSE III medium that was refreshed twice weekly. At day 22, the FSE was ready to use for immune cell culture. Cell numbers and culture conditions are based on preceding experiments (30).

Our burn injury procedure was based on previous experiments (30). A copper plate (2 × 10 mm) attached to a PACE intelliHeat ST50 soldering iron (Vass, USA) was heated to 80-90°C and stably applied to the epidermal side of the FSE for 20 sec to make contact without exerting pressure or indenting the FSE samples ( Figure 1A ). The temperature of the copper device was measured by an external digital thermometer (Farnell InOne, Utrecht, the Netherlands). Using this procedure, we created a burn injury that covered about 19% of the surface area of the model.

PBMCs were isolated from buffy coats obtained from healthy donors (Sanquin, Amsterdam, the Netherlands) by density gradient centrifugation using Lymphoprep (Stemcell Technologies, Vancouver, Canada). The buffy coat was diluted in 0.5% Bovine serum albumin in PBS and layered over the density gradient medium. After centrifugation at 1000 × g for 15 min (without brakes), the PBMCs were collected in FSE I medium. Cells were resuspended in 50% fetal bovine serum (Gibco) + 40% FSE I medium + 10% dimethyl sulfoxide. After 24 h storage in Mr. Frosty (ThermoFisher scientific) with isopropanol at -80°C, cells were stored in liquid nitrogen until use.

PBMCs were incubated with anti-CD14 beads (Invitrogen, Waltham, MA, USA) at a bead/cell ratio of 2.5:1 at 2-8°C for 20 min on a tube roller. Monocytes were isolated from the PBMCs using a magnet (Invitrogen Dynal AS, Oslo, Norway). Monocytes were resuspended in FSE I medium and 2.5 × 10⁵ cells were added to the dermal side of the FSE. For burn-injured FSEs, the cells were added directly after burn injury was inflicted. Inverted FSE with monocytes was incubated at 37°C for 2 h and subsequently placed back into the transwell ( Figure 1B ). The FSE with monocytes was cultured for 7 more days with a medium change at day 3.

Lymphocytes were isolated by culturing PBMCs in a culture flask. After 24 h, adherent cells were removed. T cells were activated by adding anti-CD3/CD28 Dynabeads (Gibco) at a bead/cell ratio of 5:1 at 37°C for 4 h. After the activation, cells were resuspended in FSE I medium and 2.5 × 10⁵ cells were placed between the transwell membrane and the dermal side of the FSE ( Figure 1C ), based on previous findings (35). Of these cells, 71 ± 14% was CD3⁺. For burn-injured FSEs, the cells were added directly after burn injury was inflicted. The FSE with T cells was cultured for 3 more days.

The FSE dissociation procedure was based on a protocol from He et al. (36). Macrophage FSEs were incubated with 0.25 U/ml collagenase A (Roche) at 37°C in a shaking water bath for 20 min. Because enzymes affect the expression chemokine receptors (37), T cell models were not dissociated using collagenase A. FSEs were then put in C-tubes (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) with 5 mL of PBS containing 1 mM EDTA and (further) dissociated by running program “B” twice on a tissue dissociator (gentleMACS, Miltenyi Biotec GmbH). Samples were passed through a 500 µm cell-strainer (PluriSelect) and then a 40 µm cell strainer (Sarstedt) to obtain a single cell suspension.

Single cell suspensions were stained using the macrophage or T cell panel ( Supplementary Table 2 ). Zombie Aqua (BioLegend, San Diego, CA, USA) was used in the macrophage panel and propidium iodide (Miltenyi Biotec GmbH) was used in the T cell panel to determine viability of cells. Stained cell samples were acquired on the flow cytometer (MACS Quant Analyzer 10, Miltenyi Biotec GmbH) and gating ( Supplementary Figure 1 ) was performed in FlowLogic (Inivai Technologies, Victoria, Australia).

See Supplementary Table 3 for antigen retrieval and primary antibodies. Kryofix (50% ethanol + 7% PEG300 in demineralized water)-fixed paraffin-embedded samples were cut into sections with a thickness of 5 µm and rehydrated followed by hematoxylin and eosin staining or blocking of endogenous peroxidase using 1% hydrogen peroxide at room temperature for 15 min. After antigen retrieval was performed, sections were pre-incubated with 5% normal goat serum (Sigma-Aldrich) diluted in PBS + 1% bovine serum albumin (ThermoFisher). Sections were then incubated with primary antibodies at room temperature for 1 h followed by incubation with a poly-HRP-goat-anti-mouse or rabbit secondary antibody (BrightVision, VWR, Amsterdam, the Netherlands) at room temperature for 30 min. After washing, detection was established using 3,3′-diaminobenzidine (DAB). After DAB staining was completed, sections were counterstained with hematoxylin, dehydrated and mounted with Eukit Mounting Medium (Sigma-Aldrich).

Snap-frozen FSEs from -80°C were thawed and fixated in 1% paraformaldehyde (Sigma) for 2 h at 4°C. The FSEs were then put in 20% sucrose (Sigma) in PBS solution overnight at 4°C. FSEs were embedded in Tissue Tek OCT (Sakura Finetek Europe B.V., Alphen aan de Rijn, Netherlands) and sections of 10 µm were cut using a cryotome (Slee MNT, Adamas Instruments B.V., Rhenen, Netherlands). Dried sections were washed in PBS and incubated with LDH solution (2 mM Gly-Gly (Sigma); 0.75% NaCl (Sigma); 5% polypep (Sigma); 1.75 mg/ml β-nicotinamide adenine dinucleotide (Sigma); 3 mg/ml nitroblue tetrazolium (Sigma) in demineralized water of pH 8) for 3 h at 37°C. Sections were washed in tap water at 50°C and in PBS and then stained with Eosin Y for 4 min. Sections were then put in PBS for 1 sec, acetone for 30 sec, acetone/xylene (1:1) for 1 min and xylene for 1 min, before embedding with Eukit Mounting Medium.

Microscopic visualization was performed with a Zeiss Axioskop40FL microscope (Zeiss, Breda, The Netherlands). Images were acquired using a Nikon Eclipse TS2 camera and the NIS-Elements software version 4.4 (Nikon Instruments, Amsterdam, The Netherlands).

Length of re-epithelization in the FSEs was measured in microscopic images of H&E-stained sections by two assessors using NIS-Elements software. The mean of the two wound sides was used for analysis.

Cytokines, chemokines and growth factors were analyzed in samples of medium. Neat samples were measured using the Human Essential Immune Response LegendPlex Multi-analyte Flow Assay kit (cat. 740929, BioLegend), according to the manufacturer’s instructions and were acquired on the flow cytometer. This 13-plex immunoassay included: IFN-γ, IL-1β, IL-2, IL-4, IL-6, IL-8 (CXCL8), IL-10, IL-12p70, IL-17A, IP-10 (CXCL10), MCP-1 (CCL2), TNF-α and TGF-β1. Concentrations were determined using FlowLogic software. When cytokine levels were lower than the standard range, the lowest level of quantification was used. When cytokine levels were higher than the standard range, the levels were estimated based on the fluorescent signal in the assay.

We used the Shapiro-Wilk test in R (ggpubr and ggplot2 packages, open source) to determine distribution of data and found that the majority of data were not normally distributed. Therefore, differences in cell number/percentages and cytokines levels between different modeling conditions were explored using Mann-Whitney U test in R (ggpubr and ggplot2 packages, open source). Data was visualized using R (ggplot2 package, open source) and significant (p value of < 0.05) differences were indicated by asterisks.

FSEs were generated by seeding human keratinocytes and fibroblasts into a collagen-elastin containing matrix from MatriDerm (see Figure 1A for procedure), as we described previously (30). After 3 weeks of culture, the FSEs presented a well-established epidermis and dermis ( Figures 2A, B ). Burn injury inflicted on the FSE was visualized by microscopy. Three days post injury the burn wound was visible, characterized by detachment of the epidermis from the affected region of the dermis ( Figure 2C ). Staining for lactate dehydrogenase (LDH) (38) showed viable cells in the dermis (fibroblasts) and epidermis (keratinocytes) up to the wound edge but not in the wound, confirming that the injury resulted in cell damage ( Figure 2D ).

Unstimulated monocytes were introduced to full-established (burn-injured) FSEs to simulate an innate immune response. To prevent the cells from adhering to the transwell membrane, monocytes (about 2.5 × 10⁵) were administered directly to the dermal side of the FSEs (see Figure 1B for procedure). Monocytes cultured in suspension or in matrix without skin cells served as controls. Through microscopy analysis, we confirmed the presence of monocytes within the FSE in both uninjured and burn-injured models ( Figure 3A ). Monocytes seemed to downregulate or lose monocyte marker CD14 (data not shown) and upregulate the expression of macrophage marker CD68 in the cultured FSEs, regardless of burn injury ( Figure 3A ). This showed that the monocytes differentiated into macrophages within a span of 7 days. Rate of re-epithelization in the FSEs after 7 days was 347 ± 168 µm. The re-epithelization rate was slightly higher in FSEs with monocytes (439 ± 126 µm), but did not reach significance.

To study the effect of burn injury on these monocyte-derived macrophages in more detail, FSEs were dissociated after 7 days of culturing the full-established FSEs. Using flow cytometry, we identified the macrophages based on their expression of CD68 (macrophage marker), CD14 (monocyte marker), CD11b (activation marker), HLA-DR (M1 differentiation marker) and CD163 (M2 differentiation marker). In uninjured FSEs, an average of 8.0 ×10⁴ CD68⁺ macrophages were present ( Figure 3B ). There was high variability in the fraction of CD68⁺ macrophages that expressed CD14 or CD11b ( Figures 3C, D ), irrespective of burn injury. This variation in macrophage differentiation and activation within the FSEs, was presumably dependent on the donor (buffy coat, fibroblast or keratinocyte donor). Comparing the FSEs to macrophages cultured in the matrix without skin cells, we observed a smaller proportion of HLA-DR⁺ or CD163⁺ ( Figures 3E, G ) macrophages in the FSEs. Burn injury appeared to increase the average number of CD68⁺ macrophages in the FSE (1.6 × 10⁵; Figure 3B ), although not significantly. Interestingly, the percentage of CD14⁺ macrophages was significantly decreased after burn injury ( Figure 3C ). Furthermore, burn injury significantly increased the expression of HLA-DR on macrophages ( Figure 3F ) and appeared to decrease the percentage of CD163⁺ macrophages within the FSEs ( Figure 3G ). Thus, we generated a human FSE model incorporating monocytes capable of actively differentiating into macrophages during culture and observed that burn injury appeared to enhance M1 differentiation of macrophages.

At day 7 (when FSEs were terminated), the levels of 13 inflammatory cytokines in the culture media were analyzed ( Figure 4 ). In the absence of monocytes, FSEs secreted high levels of IL-6, IL-8 and MCP-1 ( Figure 4D, E ; Supplementary Figure 2 ). Burn injury significantly increased the level of IL-8 and IL-12p70 ( Figures 4E, G ). Because of the high MCP-1 levels in the FSEs, the cytokine assay reached maximum signals, making it impossible to detect differences in MCP-1 levels between burn-injured and uninjured models. When monocytes were incorporated into the FSEs, there was a slight increase in the levels of IL-4, IL-6, IL-8, IP-10 and TGF-β1 ( Figures 4C-E, H, I ). Burn injury on the monocyte incorporated FSEs led to a further increase of IL-8. Cytokines IL-2, IL-17A and TNF-α were not detected in any of the experimental conditions ( Supplementary Figure 2 ).

To simulate an adaptive immune response, we introduced CD3/CD28 bead pre-activated T cells into fully-established (burn-injured) FSEs. Approximately 2.5 × 10⁵ T cells were placed between the transwell membrane and the dermal side of the FSEs, and they were cultured for a duration of 3 days (see Figure 1C for procedure), following a previously established protocol (35). Pre-activated T cells cultured in suspension or in matrix without skin cells served as controls. Using microscopy, we could detect CD3⁺ T cells that had actively migrated into the FSEs ( Figure 5A ). As 3 days was too soon after burn injury, the re-epithelization rate in these FSEs could not be measured.

Following a 3-day culture period, FSEs were dissociated to perform flow cytometric analysis of T cells. T cell differentiation was examined based on their expression of CD3 (T cell marker), CD4 (effector T cell marker), CD25/CD127 (activation marker and regulatory T cell marker), CXCR3 (Th1 differentiation marker) and CCR4/CCR6 (Th17 differentiation marker). Only a small portion (2.8 × 10³) of T cells had migrated into the FSEs ( Figure 5B ). Among these migrated T cells, the majority (approximately 86.7%) were CD4⁺ T cells ( Figure 5C ). Most of these CD4⁺ T cells expressed CD25, indicating their activation and suggesting a correlation between T cell activation and migration ( Figure 5D ). The percentage of CD25⁺CD127¯ T cells, potentially indicating Treg differentiation, was higher in the FSEs compared to T cells cultured in the matrix alone ( Figure 5E ). Furthermore, the FSEs contained a higher percentage of CXCR3⁺ T cells, indicating enhanced Th1 activity ( Figure 5F ). Similarly, an increase in the percentage of CCR4⁺CCR6⁺ T cells was observed in the FSEs, suggesting augmented Th17 activity ( Figure 5G ). The average number of T cells in burn-injured FSEs was comparable to that in the uninjured FSEs ( Figure 5B ) and burn injury did significantly not affect the investigated T cell markers ( Figures 5D-G ). Together, our findings demonstrate that particularly activated T cells migrated into the FSEs, and there is a potential enhancement of Treg and Th1/Th17 activation, regardless of burn injury.

To investigate cytokine secretion in the T cell-incorporated FSEs, we analyzed the culture medium at day 3. FSEs without T cells produced high levels of IL-6, IL-8 and MCP-1 ( Figures 6C, D ; Supplementary Figure 3 ), consistent with the expression observed after 7 days of culture ( Figures 4C, D ; Supplementary Figure 2 ). In FSEs cultured without T cells, burn injury significantly increased the levels of IL-4, IL-6, IL-8, IL-12p70 and TGF-β1 ( Figures 6B-D, F, I ).

Introducing T cells into uninjured FSEs resulted in elevated levels of IFN-γ, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-17A, IP-10 and TGF-β1 ( Figures 6A-I ; Supplementary Figure 3 ). While burn injury did not further increase the levels of these cytokines, it slightly decreased the levels of IL-10 and IP-10 in the presence of T cells. IL-2 was only detected in the presence of T cells and no significant differences were observed for the levels of IL-1β and TNF-α ( Supplementary Figure 3 ). Overall, the inclusion of T cells in the FSEs appeared to further increase both pro- and anti-inflammatory cytokines, while burn injury specifically reduced the T cell induced levels of IL-10 and IP-10.