An Autologous Dermal-Epidermal Composite Engineered Skin Repairs Wounds in a Pig Model

1. Introduction

Anatomic and physiological similarities of humans and pigs make porcine skin a commonly studied and valuable wound repair model. Epidermal thickness in pigs (30 to 140 µm) is similar to that in humans (50 to 120 µm) and both epidermises present with well-developed dermal rete-ridges, subdermal adipose tissue and sparse body hair. In addition, both species manifest a comparable skin blood vessel network [1,2]. Importantly, similar to humans, pig skin wounds are repaired with re-epidermisation, whereas rodents heal mostly via wound contraction.

In order to test engineered skin grafts in pigs, two approaches have been adapted: grafting human-derived engineered skin to an immunosuppressed pig [3], or grafting an autologous porcine engineered skin [4,5]. The latter was chosen in this study, providing the advantage of a physiologically relevant graft integration and remodeling in a fully immunocompetent animal.

Our group has previously shown that a dermal – epidermal composite skin can be prepared by seeding human fibroblasts and keratinocytes into a hydrogel made from human platelet material. The bioengineered skin, a Platelet hydroGel Human Skin Equivalent (PG-HSE), presented with similar architecture and features to human native skin [6]. We have demonstrated effective wound closure by grafting PG-HSE in immune-compromised nude mice [7]. The contribution of grafted keratinocytes to wound closure in this xenograft model was shown using human-specific antibodies.

One of the challenges in creating a holistic wound healing model is the ability to track the cells that close the wound as a result of the intervention. In this study, in vitro DNA labelling using thymidine analogues was used to trace the autologous engineered skin post grafting in pigs. Although this method has been effective in tracking injected autologous mesenchymal stem cells and fibroblasts post grafting in pigs [8,9], keratinocyte tracking in neo-epidermis has not been reported before. We investigated epidermisation and full-thickness wound closure using 5-Ethynyl-2′-deoxyuridine (EdU)-labelled autologous engineered skin in a two-stage surgical procedure.

The two-stage skin grafting in pigs reflects the clinical treatment of deep wounds [10]. To simulate real life clinical conditions, large full thickness wounds were initially grafted with a dermal template, which is vascularised over time, providing a robust wound bed. During this period, autologous, engineered skin was constructed through minor modifications of the methods used to produce the human version of the bioengineered skin, PG-HSE, to form a Platelet hydroGel Porcine Skin Equivalent (PG-PSE). Similar to the human version, PG-PSE incorporated autologous basal keratinocytes and dermal fibroblasts in a novel Platelet hydroGel (PG) as the scaffold. Grafts were labelled with EdU, and the labelled composite dermal-epidermal composite skin was grafted to close wounds. Epidermisation, vascularisation and collagen composition were then examined. The current study provided a pig model to test the autologous bioengineered composite skin against the gold standard, native meshed split skin graft (SSG), in a clinically relevant manner.

2. Materials and Methods

2.1. Isolation and Expansion of Porcine Keratinocytes and Dermal Fibroblasts

Fresh skin was processed to isolate pig fibroblasts and keratinocytes as described for human skin previously, with minor modifications (9). Isolated fibroblasts were expanded in fibroblast medium containing low glucose Dulbecco’s Modified Eagle Medium (DMEM) with 4% foetal bovine serum (FBS, Cytiva, Australia), hydrocortisone (0.5 µg/ml, Millipore, Burlington, MA, USA), insulin (50 IU/ml, Merck, Australia), EGF (10 ng/ml, R&D Systems, Minneapolis, MN, USA) and gentamicin (50 µg/mL, Thermo Fisher Scientific, Australia). Epidermal sheets were digested with trypsin (0.25%, Thermo Fisher Scientific, Australia) at 37 °C for 5-7 min to release the basal keratinocytes. Keratinocytes were expanded on irradiated feeder cells CCL92 (ATCC, Manassas, VA, USA) in cFAD medium. (DMEM and Ham’s F12 (3:1 ratio, Life Technologies, Australia), L-Glutamine (4 mM, Life Technologies, Australia), adenine (0.18 mM, Merck, Australia), hydrocortisone (0.4 µg/ml), insulin (5µg/ml), 3,3′,5-triiodo-L-thyronine sodium salt (2 x 10-9 M, Sigma-Aldrich, Australia), transferrin (5 µg/ml, Sigma-Aldrich, Australia), gentamicin (50 µg/mL), isoproterenol (0.2 µg/ml, Hospira, Australia), EGF (10 ng/ml), and 10% FBS.

2.2. EdU / BrdU Labelling Optimisation in Monolayer

Porcine keratinocytes were seeded on irradiated feeder cells in chambered slides (Sarstedt, Australia). Porcine fibroblasts were directly seeded on chamber slides. Once attached, cells were fed with 10 μM EdU (Invitrogen, Australia) or 10 μM BrdU (Sigma-Aldrich, Australia) in culture medium for 48 hours. Afterwards, cells were tracked for label retention and stained on days 3, 8 and 16 post-labelling.

2.3. Preparation of Platelet Hydrogel Pig Skin Equivalent (PG-PSE)

PG hydrogel scaffolds were prepared as previously reported with some modifications (6). Briefly, 2.4 IU/ml thrombin was added to human platelet precipitate (Australian Red Cross Lifeblood, Alexandria, NSW, Australia) to form a hydrogel. Expanded porcine dermal fibroblasts were seeded at 0.25 x 106/cm2, 24 h post gelation, and left for 72 hours to spread over PG scaffold in submerged culture. This was followed by seeding of basal keratinocytes (0.5 x 106/cm2). PG-PSEs were grown at the air-liquid interface in UCDM1 maturation medium consisting of low glucose DMEM (Thermo Fisher Scientific, Australia) 1:1 with Ham’s F12 medium (Thermo Fisher Scientific, Australia) plus L-glutamine (4.5 mM), L-serine (1.5 mM, Sigma-Aldrich, Australia), CaCl2, (1 mM Sigma-Aldrich, Australia), adenine (1.8 mM), O-phosphoryl ethanolamine (5 µM, Sigma-Aldrich, Australia), gentamicin (50 µg/ml), strontium chloride (1 mM, Sigma-Aldrich, Australia), 1 x ITS (Sigma-Aldrich, Australia), linoleic acid (10 µg/ml, Sigma-Aldrich, Australia), ascorbic acid (0.1 mM, Sigma-Aldrich, Australia), 3,3′,5-triiodo-L-thyronine sodium salt (2 x 10-12 M,), hydrocortisone (0.5 µg/ml,), KGF (5 ng/ml, R&D Systems, Minneapolis, MN, USA), EGF (10 ng/ml) and FGF-2 (1 ng/ml, R&D Systems, Minneapolis, MN, USA) with 0.3% FBS (serum-supplemented) or 0.2% platelet lysate (serum-free) for five days.

To label 3D PG-PSEs, fresh UCDM1 culture medium was supplemented with 10 μM EdU and added to the PG-PSE cultures 48 hours before surgery. PG-PSEs were washed in fresh low glucose DMEM, supplemented with amphotericin / gentamicin prior to surgery and transported to the animal operating theatre for grafting.

2.4. Animal Surgery

A total of five (2 animals in pilot stage / 3 animals in main stage), 2-3 months old, female large white pigs (S. domesticus) were sedated and transported to the surgical theatre. An endotracheal tube was inserted and isoflurane anaesthesia administered through a face mask (2-3.5% isoflurane per litre), with Buprenorphine (0.01 mg/kg i.v.) injected for perioperative pain management. Pigs were placed in sternal recumbency on the operating table. Anaesthetic delivery, and vital signs were continuously monitored. During the initial surgery, a 7 x 7 cm full thickness donor skin was recovered under aseptic conditions to prepare PG-PSE. The wound was grafted with Novosorb^® Biodegradable Temporising Matrix (BTM) as a dermal template (PolyNovo, Melbourne, Australia) and dressings were applied. Fentanyl ear patches (Apotex, Australia) were provided for postoperative pain management.

A week after taking donor skin, seven full thickness wounds were created on the dorsal side of each pig (Supplementary Data 1). Wounds were grafted with BTM as described above. The animals underwent routine dressing changes twice weekly, while PG-PSE autografts were prepared. It has been shown by others that BTM is fully integrated in three weeks in young pigs [11]. After three weeks, the temporary seal of the BTM was removed, and wounds were splinted to minimise infiltration of the host epidermis. The seven wounds were grafted with a randomised allocation of 3 serum-free PG-PSEs, 3 serum-containing PG-PSEs and 1 meshed SSG per pig. A wound chamber was piloted on one wound per graft type (i.e., 1 out of 3) in two pigs to create a barrier against epidermisation from the wound margins. Skin grafts were secured in place over the vascularised wound bed and held in place on the edges using sutures with Conformant 2 wound veil (Smith & Nephew, Australia) backing.

Dressings were applied over each grafted wound in the following sequence: Mepitel (Mölnlycke Health Care, Australia), wet gauze (soaked in PBS with 50ug/ml gentamicin and 1ug/ml amphotericin B), Combine (Mediflex, Australia), Conformant 2, Tegaderm (Solventum, Australia), Opsite (Smith & Nephew, Australia), Hypafix (BSN Medical, Hamburg, Germany), and Exu-dry (Smith & Nephew, Australia). VetRap (Solventum, Melbourne, Australia) and a body coat were fitted for graft protection. Tissue samples were collected using an 8 mm punch biopsy on Day 8 and Day 14 after skin grafting for processing and assessment. At each surgery and dressing change, standardized photographs were taken to assess healing, total wound size reduction, graft survival, and presence of infection by comparison with the baseline values. On Day 14, pigs were humanely euthanised. All wounds were excised and taken to the lab for processing and analysis.

2.5. Histological Staining

Skin graft biopsies were fixed in 4% paraformaldehyde prior to cryopreservation. Specimens were sectioned and stained with haematoxylin and eosin as previously described (9). For Masson’s trichrome staining, sections were processed according to standard protocols as previously described (9). Whole stained sections were imaged using a Nikon Eclipse Ts2 inverted microscope (10x/0.30 objective) and Nikon DS-Fi3 camera. An Image J macro (NIH) was used to quantify collagen by assigning thresholds for the following colours: black (0-8), blue (120-177, collagen) and purple (178-220). Collagen was quantified separately in the papillary and reticular dermis. Papillary dermis was determined as areas below the epidermis, while reticular dermis encompassed deeper areas directly above BTM. Five randomly selected regions within the papillary (below epidermis and within 200 µm of dermoepidermal junction) and reticular dermis (> 600 µm below dermoepidermal junction) were selected for collagen analysis.

For collagen composition, tissue sections were stained with Picrosirius Red (Abcam, UK) and imaged using an Olympus VS200 slide scanner under polarised light. A modified ImageJ macro (PR-BRF Analyser v2.20, CCM Lab, Darlinghurst, NSW, Australia) was used to measure areas occupied by each colour [12]. Images were converted to 8-bit with colour ranges assigned to collagen fibre types as follows: red (0-27, 240-255, long/mature), yellow (28-47, mixed/intermediate), and green (48-140, thin) to reflect controls including collagen I, native human skin, and native porcine skin. Five randomly selected regions of 360 x 240 µm within the papillary and reticular dermis were selected and analysed for collagen composition.

2.6. Immunofluorescence

Fixed tissue sections or cells were permeabilised in with 0.2% (v/v) Triton X-100 (Sigma, Australia) for 10 min and washed in PBS. Sections were blocked using 3% normal goat serum or normal donkey serum, 30 min and washed. Mouse anti-BrdU antibody (1:300, Invitrogen, Australia); rabbit anti-pancytokeratin (1:100, Novus Biologicals, Centennial, USA); mouse anti-pig CD45 (1:200, GeneTex, Irvine, USA); rabbit anti-vimentin (1: 600, Cell Signaling, Danvers, USA), rabbit anti-CD31 (1:100, Abcam, Cambridge, UK) or rabbit anti-cytokeratin 5 (1: 100, BioLegend, San Diego, USA) primary antibody was added to the sections and incubated overnight at 4 °C in a humidified chamber. The primary antibody was washed off and sections were incubated with goat anti-mouse Ig - AF 488 (1:400, Invitrogen, Australia), goat anti-rabbit Ig - AF 488 (1:400, Invitrogen, Australia), donkey anti-mouse Ig-AF568 (1:400, Invitrogen, Australia), donkey anti-rabbit-AF647 (1:400, Invitrogen, Australia) and Hoechst 33342 (1:500) stains for 45 min at room temperature before multiple wash with PBS. Coverslips were mounted coverslips using ProLong Gold solution (Invitrogen, Australia). For BrdU detection, slides were initially pretreated with 4M HCl for 3 min and washed in PBS prior to permeabilisation. For EdU detection, sections were processed as per Click-iT Plus kit instructions (Invitrogen, Australia). Widefield Nikon Ti-E and confocal Nikon A1r microscopes were used to visualise EdU presence in far red channel.

Images acquired using an Olympus VS200 slide scanner were analysed using QuPath version 0.7.0 and Fiji/ImageJ version 1.54 to quantify vessel diameter and vessel abundance within the papillary and reticular dermal layers.

For CD31 quantification, in QuPath, regions of interest (ROIs) were manually annotated to define the papillary dermis, corresponding to the superficial 150 µm of the dermis, and the reticular dermis, corresponding to the 250 µm region directly above the BTM (> 400 µm deep). Vessel segmentation was performed in QuPath using the pixel classifier machine-learning tool [13]. Segmented objects with a minimum diameter below 3 µm were excluded from downstream analysis. For each ROI, QuPath was used to measure the total annotated area, total CD31-positive vessel area, vessel area fraction, and number of detected vessel objects. Vessel coverage was calculated as:

Vessel area (%) = total CD31-positive segmented vessel area / total ROI area × 100.

Binary masks of the ROIs and vessel detections were exported from QuPath as TIFF images for vessel diameter analysis in Fiji/ImageJ. Masks were batch processed using a standardised macro workflow. Each mask was converted to a binary image, skeletonised to reduce each vessel object to its medial axis, and used to generate a Euclidean distance map from the corresponding binary vessel mask. Local vessel diameter was calculated by multiplying the distance map value along the vessel skeleton by two, thereby converting the radius from the medial axis to the vessel boundary into a diameter measurement. Vessel diameter measurements were then summarised for each ROI.

3. Results

3.1. An Autologous Bioengineered Skin Graft to Treat Full-Thickness Wounds in a Pig Model

In order to test the autologous dermal - epidermal composite bioengineered skin graft, PG-PSE, multiple full-thickness wounds were created on the dorsal side of each young female large white pig (Figure 1A-B). All wounds were grafted with polyurethane-based acellular BTM dermal template. BTM provided a template for vascularisation, cell infiltration and nerve regeneration over three weeks (Figure 1C). Wounds were then grafted with either PG-PSE or traditional meshed SSG as control. PG-PSE was prepared from autologous basal keratinocytes and fibroblasts, which were isolated from pig skin and expanded in culture. Autologous cells were seeded on a Platelet hydroGel (PG) to form the mature PG-PSE. Human Platelet Lysate (PL) was investigated as an alternative media supplement to foetal bovine serum (FBS) during skin maturation phase. PG-PSE architecture was monitored by Haematoxylin and Eosin staining in vitro, showing similar architecture between PG-PSE matured in serum-free (Figure 2A) and in serum-containing conditions (Figure 2B).

One out of seven wounds per pig was grafted with traditional meshed SSG, as the gold standard control group. Three wounds were grafted with autologous PG-PSE either matured in serum-free media or PG-PSE matured in serum-supplemented media (Figure 1D-F). Skin grafts were secured with sutures and several layers of wound dressings to minimise shearing.

Graft integration and wound repair were monitored macroscopically and histologically over the course of 14 days with regular dressing change. On Day 14, grafted areas with autologous bioengineered skin were epidermised (Figure 1H and 1I) similar to the gold standard meshed SSG (Figure 1G). No macroscopic differences were observed between the wounds closed with the PG-PSE matured in serum-free (Figure 1H) and serum-supplemented (Figure 1I) media.

3.2. Grafted Keratinocytes Formed a Traceable Neo-Epidermis in Pigs

Both BrdU and EdU thymidine analogues were tested for labelling autologous PG-PSE. Initially, keratinocytes and fibroblasts, were labelled with either a BrdU or EdU in monolayer (Supplementary data 2). Although, both thymidine analogues la-belled the cells, to retain and track cells for longer, EdU labelling was performed on the 3D PG-PSE construct prior to grafting. EdU detection is antibody independent and does not rely on DNA denaturation. EdU labelled cells were observed throughout the depth of PG-PSE. There appeared to be more labelled cells in the PG-PSE matured un-der serum-free conditions (Figure 2C) both in dermis and epidermis, likely due to slower proliferation and turnover of cells, compared to PG-PSE matured in se-rum-containing conditions on the grafting day (Figure 2D). Seven days post-grafting of PG-PSEs, labelled keratinocytes and fibroblasts were detected in the serum-free (Figure 3B) and serum-containing (Figure 3C) grafts. The numbers of EdU-labelled cells were reduced post-grafting, possibly due to: 1) rapid cell proliferation leading to label dilution, 2) epidermal turnover and cell death, as mature keratinocytes lost their nuclei and were shed off and 3) immune-mediated detection and removal of EdU-labelled cells in grafts. In fact, an influx of CD45+ scavengers were observed in PG-PSE grafts suggesting that the immune-mediated removal of the labelled cells and debris was a likely mechanism. EdU signals were largely undetected by two weeks post grafting.

3.3. Bioengineered Skin Graft Forms a Neo-Dermis and Neo-Epidermis Enriched in Progenitor Keratinocytes and Displays Rete Ridge Architecture

Haematoxylin and eosin-stained graft biopsies presented with moderate thickening of the epidermis in meshed SSG. Both PG-PSE graft groups exhibited thicker epidermis, indicative of epidermal hyperplasia, with rete ridge elongations growing deep into dermal space. Hyperplasia was confirmed by immunofluorescent staining using cy-tokeratin 5 (CK5) antibody. Porcine native skin exhibited CK5 signal confined to the basal keratinocyte layer (Figure 4E). Conversely, in SSG (Figure 4F) and in both serum-free and serum-containing PG-PSE (Figure 4G-H), CK5 signal was detected across the entire thickness of epidermis. A similar outcome was reported by Bruen et al. in mice wherein their model of EGFP, synthesized under a CK5 promoter, recorded maximum signal in the advancing wound margin re-epithelialisation between days 12 and 15 post-burn injury [14]. In the dermis, signals from vimentin-expressing cells were ob-served in SSG (Figure 4J) and both PG-PSE grafted regions (Figure 4K-L) when com-pared to native pig skin (Figure 4I).

3.4. Neo-Vascularisation Supports Bioengineered Skin Graft Survival

Immunofluorescence staining for the endothelial marker CD31 enabled visualisation and quantification of the vascular network within each grafted area. Vessel density and vessel thickness were assessed in the papillary and reticular dermis (Figure 5). Numerous microvessels were observed in the papillary neo-dermis at two weeks following skin grafting (Figure 5E). The increased abundance of microvessels relative to uninjured skin suggests that the grafts remained in an active phase of integration and remodelling at this time point. Wounds grafted with PG-PSE matured in platelet lysate-supplemented medium under serum-free conditions exhibited significantly greater microvessel density in the papillary dermis, immediately beneath the neo-epidermis, compared with uninjured native skin (p = 0.0077), indicating that platelet lysate may promote neovascularisation.

Most microvessels identified within both the papillary and reticular neo-dermis meas-ured between 5 and 20 µm in diameter, with vessels in the 10–20 µm range being the most prevalent. Compared with native skin, PG-PSE grafts matured under serum-free conditions contained significantly more vessels with diameters of 10–20 µm (Figure 5G, p = 0.0208). This shift towards larger calibre microvessels may indicate ongoing maturation of the neo-dermal vascular network. Notably, residual fragments of the biodegradable temporising matrix (BTM) were still evident within the deep reticular dermis of both split-thickness skin graft (SSG) and PG-PSE-treated wounds at five weeks following BTM implantation). This observation is consistent with previous re-ports demonstrating the relatively slow degradation kinetics of BTM in vivo. [15].

3.5. Collagen Deposition and Remodelling

Collagen deposition by dermal fibroblasts, was detected with Masson’s Trichrome staining two weeks post skin grafting (i.e., five weeks post dermal template grafting). In native pig skin, tightly packed, dark blue staining throughout the dermis indicates the presence of dense collagen (Figure 6A). Conversely, relatively loosely packed collagen was observed by form of lighter blue staining in SSG (Figure 6B) and both PG-PSE matured in serum-free (Figure 6C) and serum-supplemented (Figure 6D) media.

Quantification of collagen in the papillary neo-dermis showed as expected SSG contained significantly higher amounts of collagen present, when compared to freshly deposited collagen in both types of PG-PSEs. Collagen density in all native and bioengineered grafts still lagged behind the collagen density of uninjured healthy papillary neo-dermis two weeks post-grafting (Figure 6E). In the deeper reticular neo-dermis, there was no significant difference between the amount of collagen deposition in SSG and both types of PG-PSE. However, all three treatment groups had significantly lower collagen deposition when compared to uninjured native pig skin (Figure 6F), providing further evidence that grafts are still in active integration phase at this time point.

Collagen fibre structure and type were visualised with Picrosirius red staining, allowing to discriminate between collagen type I and type III deposition when imaged under polarised light. As expected, mature, basket-weave structured collagen bundles were observed in native reticular dermis, while some random collagen bundles were present in papillary dermis (Figure 7A).

Both types of collagens, type I collagen (Col I, red and yellow fibres) and type 3 collagen (Col III, green fibres) were detected in uninjured and skin grafts. Notably, significantly higher levels of Col I were observed in the papillary neo-dermis of SSG when compared to both types of PG-PSE grafts (Figure 7E, p-value = 0.0027 for se-rum-free PG-PSE, p-value = 0.0009 for serum-supplemented PG-PSE). Similar amount of Col III was detected in uninjured and all skin grafts. This suggested that unlike native skin grafts, Col III contributes proportionately more to the freshly deposited collagen in PG-PSE grafts than Col I. Collagen deposition and integration still outweighs remodelling in papillary dermis two weeks post-grafting, compared to SSG.

Conversely, in the deep reticular neo-dermis that was formed over five weeks since the BTM grafting, there was no significant difference observed in Col I nor Col III levels between three types of grafts and native skin (Figure 7F). Overall, similar composition and structure in the grafts reticular neo-dermis suggests that remodelling in deep neo-dermis in all grafts is occurring in a similar fashion, regardless of whether the wounds are closed with SSG or PG-PSE. Five weeks post grafting, all grafts have reached similar pre-grafting collagen composition in reticular dermis.

4. Discussion

In this study, we developed a large white pig model to evaluate the repair of large full-thickness wounds using bioengineered skin grafts. The bioengineered skin is prepared from autologous fibroblasts and keratinocytes and a bioactive hydrogel as the scaffold. In this model, the bioengineered skin graft was compared to the gold standard native meshed SSG in terms of neo-epidermisation, vascularisation, and graft integration / remodelling.

Epidermal stratum corneum, stratum granulosum, stratum spinosum and stratum basale layers were detected in bioengineered skin graft, similar to the gold standard meshed SSG, albeit thicker than uninjured skin. Thickened epidermis in all skin grafts is consistent with increased keratinocyte proliferation during graft integration. CK5 expression was expanded to supra basal layers in bioengineered skin grafts and meshed SSG, compared to uninjured skin, reflecting enrichment of proliferating keratinocytes during skin graft integration and remodelling [16]. This is in agreement with a study by Yamaguchi et al. showing increased number of proliferative Ki67+ keratinocyte in skin grafts [17]. The proliferating keratinocytes were supported by abundance of vimentin+ fibroblasts in regenerated dermis.

Two variants of PG-PSE were analysed in this study. Platelet lysate was tested to replace foetal calf serum during skin maturation in culture. Notably, both types of PG-PSE grafts, regardless of culture condition, formed rete ridges within two weeks post grafting, reflecting progressive maturation of the neo-dermis and dermal–epidermal junction [18]. Rete ridges contribute to the mechanical integrity of skin by enhancing resistance to shear forces at the dermal–epidermal junction in pigs. In con-trast, rodents and other loose-skinned animals in general lack rete ridges and exhibit a relatively flat dermal–epidermal interface [2]. Rete ridges were absent in the bioengineered skin that was grafted in Nude mouse model in an earlier study [7].

Timely vascularisation is fundamental to graft survival and graft integration, yet excessive vascularisation is associated with fibrosis. Interestingly, we observed higher number and thicker microvessels in PG-PSE papillary neo-dermis, when PG-PSE was matured in platelet lysate during culture. This supports effectiveness of platelet lysate as a serum substitute for skin bioengineering, although the functional scarring consequences of having increased vascularisation is to be determined in a longer study.

We had previously shown that this novel bioengineered skin closes wounds in an immune-compromised Nude mouse model [7]. This study provided an independent animal model describing relevant surgical procedures that are adaptable for clinical application. Particularly, a two-stage procedure, that mimics skin grafting in clinics, was adapted in this large animal study. Full thickness wounds were initially grafted with an acellular dermal template. Three weeks post-surgery, the wounds were definitively closed with a skin graft.

Collagen deposition, types and distribution in bioengineered skin grafts were fully analysed using Masson’s trichrome and Picrosirius Red staining. By two weeks post skin grafting, similar ratio of Col I and Col III is reached in deep dermis, compared to uninjured skin, albeit at lower levels. However, in more superficial papillary dermis, there is still a lower level of Col I in PG-PSE, compared to SSG and native skin. This data suggests fresh collagen deposition in papillary dermis in PG-PSE grafts is active and therefore collagen deposition and integration phase may be extended in PG-PSE grafts compared to native SSG. Longer-term studies are required to evaluate tissue remodelling and scarring outcome following PG-PSE grafting. Nevertheless, the absence of excessive collagen deposition during the early post-grafting period suggests a low propensity for pathological fibrosis or hypertrophic scar formation.

In accordance with good manufacturing practice (GMP) guidelines, foetal calf serum was replaced with platelet lysate during PG-PSE maturation, enabling the complete elimination of animal-derived components from the final stage of production. Further studies are underway to achieve full removal of foetal calf serum throughout the entire manufacturing process. Other bioengineered composite skin grafts under development such as denovoSkin™ (Cutiss AG, Switzerland) and SASS (LOEX, Cana-da) also demonstrate the potential of autologous bioengineered skin grafts. However, our novel bioengineered skin has several advantages: unlike bovine collagen-based hydrogel in denovoSkin, PG-HSE has a unique human platelet hydrogel scaffold with inherent pro-vascularisation and pro-epidermisation components [19,20]. Moreover, PG-HSE, similar to other platelet-derived materials, is beneficial in possessing antimicrobial activity [6,21]. PG-HSE is rapidly prepared, requiring 3 weeks for a scaled-up production, in comparison to SASS, which requires 6 weeks to prepare from a small skin biopsy [22]. This is a clinically significant advantage in a setting where time to wound healing has major survival, scarring and cost implications. Moreover, SASS requires foetal calf serum during production.

The bioengineered skin was generated by the timely expansion of autologous dermal fibroblasts and epidermal progenitor keratinocytes. Grafted autologous keratinocytes were subsequently traced within the regenerated epidermis, providing direct evidence of their engraftment leading to skin regeneration. Importantly, real-time pro-duction aligns well with the two-stage clinical management of deep wounds, in which a dermal template, such as BTM, is applied as the first step prior to definitive epidermal coverage [23].

5. Conclusions

Collectively, these data demonstrate the feasibility of scalable, clinically relevant production of autologous bioengineered skin for the treatment of large full-thickness wounds. It provides clinically meaningful evidence for the effective repair of deep wounds with the novel bioengineered skin and has potential applications in treating extensive skin defects, such as severe burns, reconstructive surgical defects, and giant congenital melanocytic nevi.

In addition, this pig model can be used to investigate the pathways activated during wound repair following skin grafting. Improved understanding of the mechanisms governing wound repair and regeneration can support future advances in skin bioengineering.