Efficient scarless skin regeneration enabled by loading micronized amnion in a bioinspired adhesive wound dressing

Complete skin reconstruction is a hierarchical, physiological assembly process involving healing of the epidermis, dermis, vasculature, nerves, and cutaneous appendages. To date, few works have reported complete skin regeneration, particularly lacking vascular structures and hair follicles after full skin defects. In this study, a hydrogel derived from the skin secretion of Andrias davidianus (SSAD) that features adhesiveness was used as a bioactive scaffold to load micronized amnion (MA). The SSAD hydrogel was found to promote the migration and proliferation of amnion stem cells and human keratinocytes, as well as inhibit their apoptosis in vitro. In a rat full‐skin defect model, the regeneration of skin appendages was observed at the wound area, achieving scarless healing. Transcriptome analyses further validated that SSAD could positively regulate cell migration, proliferation, and differentiation. These functions might be attributed to the abundant growth factors present in the SSAD. Synergized by the delivery of MA, SSAD loaded with the MA could achieve a significantly better skin regeneration effect than SSAD or MA used alone, providing a simple yet highly effective means to obtain complete, scarless skin regeneration, suggesting favorable potential for clinical translation.

Unfortunately, large-area skin wounds or burns usually produce scars after treatments. [2]Disordered collagen fibers hinder the regeneration of hair follicles, sebaceous glands, and sweat glands. [3]This fact not only affects appearances and causes mental distresses but also leads to serious physiological dysfunctions, such as hinderance of the movement of joints, the secretion of sweat, and the skin temperature regulation. [4]or traumatic or burn wounds, the most commonly used clinical treatment method is autologous skin transplantation, which brings further secondary injury to the patients, making it especially unsuitable for treating large-area wounds. [5]On the other hand, xenografts always pose the risk of immune rejection and disease transmission.Many wound dressing materials have emerged in recent decades. [6]Nevertheless, the general successful outcomes that have been reported include accelerated wound closure, seldom involving skin appendage regeneration. [7]Complete wound regeneration preserves skin structure and physiological functions without causing fibrosis and scar formation. [8]Skin appendages play an important role in skin functions, the growth of which is also one of the signs of scarless skin regeneration. [9]Cell therapy seems to be another viable option. [10]For regenerative engineering applications, the cells are usually used in combination with designable biomaterials or as completely cultured cell sheets or aggregates to bypass the carrier scaffolds. [11]However, cell culture usually requires several weeks to separate and prepare cells tailored for each patient, [12,13] which imposes substantial costs of both time and capital. [14]he amnion is the innermost membrane of the placenta and constitutes a preformed cell sheet containing amniotic mesenchymal stromal cells (AMSCs).AMSCs have the potential to differentiate into numerous cell types, including keratinocytes and endothelial cells, [10] and contain various growth factors, such as basic fibroblast growth factor (bFGF), epidermal growth factor (EGF), platelet-derived growth factor (PDGF), vascular endothelial growth factor (VEGF), and hepatocyte growth factor (HGF), which promote stem cell recruitment, epithelialization, neovascularization, and fibroblast proliferation. [15]The low immunogenicity of the amnion makes it an ideal allograft for skin regeneration. [15,16]However, it is challenging to fix the amnion stably on the wound. [17]he hydrogels derived from skin secretions of Andrias davidianus (SSAD) have been shown to feature strong tissue adhesion and possess potential as surgical biological adhesives. [18]In this study, for the first time, we modified the SSAD hydrogels as a self-adhesive carrier for the micronized amnion (MA) (abbreviated as SA), where the structure of SA is shown in the Schematic 1. SSAD could fix MA on the wound, increase the bioactivity of MA by slowing down the apoptosis of cells, and protect the wound bed as a wound dressing.To evaluate the regenerative efficacies of SA, the rat critical full-skin defect model [19] was selected.Compared with other control groups, SA treatment exhibited almost scarless regeneration, which surprisingly recovered the key skin structures with exceptional homology to the native skin, suggesting its future clinical translational potential.

Preparation and characterization of SSAD materials
The SSAD materials were produced as described previously. [18,20]Briefly, fresh mucus from adult healthy Chinese giant salamanders was collected.After washing with pure water more than five times, the mucus was authenticated and lyophilized for 2 days and grounded into powders of different particle sizes (20-, 60-, or 300-mesh) (Figure 1A).A higher mesh value corresponded to a smaller pore size of the sieve, and hence a smaller particle size.When mixed with water, the SSAD powder could gel (Video S1), depending on the particle size; the smaller the particle size, the longer the gelation time.When the particle sizes surpassed 150 μm, they could swell into homogeneous SSAD hydrogels within 1 min.The SSAD-derived hydrogels further exhibited adhesive properties and porous structures.Larger particle sizes of SSAD-derived powders could form larger pore sizes in the resulting SSAD-derived hydrogels (Figure S1).The mechanism of the formation of the SSAD hydrogel is that when SSAD powders are mixed with water, hydrogen bonds and S-S bonds are formed between amino acid residues of SSAD proteins, inducing hydration and gelation. [18,21]he fresh amnion appeared whitish and translucent (Figure 1B) and is rich in proteins and other extracellular matrix molecules. [17]After micronization, it became fragmented into micro-sized pieces smaller than roughly 200 μm (Figure 1C).Hematoxylin-eosin (HE) staining showed that the amniotic epithelial cells attached to the basement membrane and that the amnion structure was still intact (Figure 1D).The MA loading into the SSAD hydrogel (derived from powder of 20-, 60-, or 300-mesh) was confirmed under scanning electron microscopy (SEM), where the MA pieces were wrapped tightly by the SSAD hydrogel matrix, forming tight bonding interfaces (Figure 1E and Figure S1). [22]

2.2
The effect of SSAD on MA To further evaluate the effect of SSAD on MA, a terminal deoxynucleotidyl transferase dUTP nick end labeling assay for cellular apoptosis was performed (Figure S2).Compared with pure MA (control group), cell apoptosis of MA loaded by the SSAD hydrogels formed from different meshes was much lower (Figure S2).Among them, the 60-mesh SSAD hydrogel showed the least cell apoptosis (Figure S2B).Therefore, the 60-mesh SSAD hydrogel was selected for the subsequent experiments.After 12 h, cell apoptosis in the control group started to become visible (Figure S3 and Figure 1I).However, even after 36 h, it was still almost impossible to detect cell apoptosis in the SA group (60mesh) (Figure 1H,I).Apoptosis is a common phenomenon of foreign cells. [23]When amniotic membrane was used alone to promote skin wound healing, the apoptosis of amniotic epithelial cells began on the 7th day and disappeared completely on the 28th day. [17]It is therefore believed that SSAD that could slow the apoptosis of cells might be able to make the amniotic epithelial cells retain even longer activities at the wound site than when they are used alone.Apoptotic bodies from stem cells promote cell division, tissue S C H E M AT I C 1 Schematic diagram showing the proposed process of SA to promote the healing of skin wounds.Skin secretions of Andrias davidianus (SSAD) loads micronized amnion (MA) to form SA, which is used towards promoting healing of a rat critical full-skin defect model.AMSCs, amniotic mesenchymal stromal cells; EGF, epidermal growth factor; IFESCs, interfollicular epidermis stem cells; PDGF, platelet-derived growth factor; SDF-1, stromal cell-derived factor-1, stromal cell-derived factor-1; VEGF, vascular endothelial growth factor.regeneration, and wound healing. [24,25]Accordingly, prolonging the survival time of stem cells have two advantages: preserving more active peptides secreted by MA [17] to promote cell proliferation, mesenchymal stem cell recruitment, epithelialization and neovascularization, [26][27][28][29] as well as decreasing infection by expressing antimicrobial molecules such as β3-defensin. [26]

The effect of SSAD on AMSCs and HaCaT cells
AMSCs are one of the key cellular components of the amnion, and keratinocytes are essential for reepithelialization and formation of the extracellular matrix. [30]he effects of SSAD-conditioned medium (SSAD group) and base culture medium (control group) on the migration and proliferation of AMSCs and HaCaT cells were examined.For HaCaT cells, both the scratch assay (Figure 2A) and the transwell assay (Figure 2B) validated that SSAD could promote cell migration.The wound healing ratios were 20.17% ± 3.21% in the control group and 61.82% ± 3.51% in the SSAD group (Figure 2C).The relative cell numbers were 100% ± 8.27% for the control group and 227.14% ± 24.57% for the SSAD group (Figure 2D).Moreover, proliferation was significantly enhanced by SSAD (p < 0.05) (Figure 2E).The results for HaCaT cells confirmed that the inclusion of SSAD could likely promote the migration and proliferation of keratinocytes.For AMSCs, scratch (Figure S4A), transwell (Figure S4B), and cell proliferation assessments (Figure S4E) also proved that SSAD could increase the duration and help AMSCs escape from the amnion.Both scratch assays and transwell assays were performed, which are complementary to each other in the horizontal and vertical directions.During the skin healing process, keratinocytes proliferate and migrate to cover the wound surface and then undergo differentiation and stratification to reconstruct the epidermal barrier. [30,31]To clarify the mechanism by which SSAD promotes skin healing, HaCaT cells were selected as the representative cell type for additional transcriptomic analyses and were cultured in SSAD-conditioned medium for 24 h.The Venn diagram showed that 11,014 genes were coexpressed in the two groups, while 353 genes were exclusively expressed in the SSAD group (Figure 2F).A volcano plot showed 393 notably differentially expressed genes (DEGs), of which 212 genes were upregulated, and 181 genes were downregulated (Figure 2G).We also studied the effect of SSAD on the expression of genes related to the promotion of proliferation and migration as well as the inhibition of apoptosis (Figure 2H).Some important genes that promote proliferation, such as LCK and IGF1, genes that promote migration, such as PRSS3 and CASS4, and genes that inhibit apoptosis, such as NR1H4 and FGG, were obviously upregulated after treatment with SSAD-conditioned medium.The top 20 Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis indicated that MAPK, VEGF, Ras, tight junction, Jak-STAT, and apoptosis signaling pathways were closely related to the mechanism by which SSAD promotes wound healing (Figure 2I).Among them, the MAPK signaling pathway is an intersection or pathway that transmits information from cell to cell and plays an important regulatory role in tissue fibrosis, cell proliferation, cell differentiation, cell apoptosis, and inflammation. [32]Remarkably, the MAPK signaling pathway was clearly activated after SSAD treatment (Figure S4F), indicating that SSAD could activate the MAPK signaling pathway by increasing Ras-related genes, thereby promoting cell proliferation and migration.

Effect of SA on skin wound healing in vivo
Rats are a common model animal for skin wound repair, and the use of the silicone ring splint model can reduce the effect of skin shrinkage in rats. [33,34]Full-thickness excisional cutaneous defects (diameter = 20 mm) were created on the backs of Sprague-Dawley rats treated with MA, SSAD, SA, hEGF (positive control), Pelnac (positive control), or nothing (blank control) (Figure 3A).At 7 days postsurgery, the wound closure ratio of the SA group (51.90% ± 12.26%) was significantly higher than those of the blank control group (15.23% ± 7.47%), the MA group (26.67% ± 10.56%), the SSAD group (33.96% ± 5.97%), the EGF group (46.13% ± 7.00%), and the Pelnac group (36.46% ± 5.28%).At 21 days after injury, the wounds in the SA group had almost completely healed; the average wound closure ratio at this time of the SA group (98.40% ± 1.01%) was also significantly higher than those of the blank control group (79.16% ± 7.00%), the MA group (80.30% ± 4.96%), the SSAD group (88.05% ± 2.61%), the EGF group (89.35% ± 1.03%) and the Pelnac group (83.99% ± 5.54%) (Figure 3B).As such, SA could greatly facilitate the effective healing of large skin defects within a short period of time.
At 35 days postsurgery, skin samples were collected, and paraffin sections were stained with HE (Figure 3C and Figure S5) and Masson's trichrome (Figure 3D).The results confirmed that the thicknesses of the new skin in the SSAD and SA groups were larger compared with the blank control group, the MA group, the EGF group, and the Pelnac group, with the former groups being similar to the normal skin thickness.In addition, the SA-and SSAD-treated groups had more mature new tissues, with characteristics of normal skin tissue structures, such as hair follicles, sebaceous glands, and blood vessels (Figure 3C).The morphology of collagen in the SA group appeared in the form of a woven basket (Figure 3D), indicating that the collagen deposition was sufficient.In contrast, the regenerated skins in the blank control group, the MA group, the EGF group, and the Pelnac group were incomplete, with immature and thinner collagen layers and insufficient collagen deposition (Figure 3D).The skin thickness ratio (expressed as thickness relative to normal skin) of the SA group was 95.84% ± 4.98%, that of the blank control group was 31.38% ± 10.04%, that of the MA group was 55.62% ± 13.58%, that of the SSAD group was 74.69% ± 3.85%, that of the EGF group was 57.95% ± 10.01%, and that of the Pelnac group was 59.95% ± 6.48%.(Figure 3E).Among them, the SA group had the highest numbers of new hair follicles (16.92 ± 2.34) (Figure 3F) and sebaceous glands (14.32 ± 1.33) (Figure 3G) in the regenerated newborn skin tissue compared with the other groups.The hair follicle number was 2.60 ± 0.47 per microscopic view, and the sebaceous gland number was 2.27 ± 0.33 for the blank control group in the regenerated newborn skin tissue; the numbers were 2.92 ± 0.54 and 3.12 ± 0.56 for the MA group in the regenerated newborn skin tissue; 7.67 ± 1.12 and 6.80 ± 1.01 for the SSAD group in the regenerated newborn skin tissue, the numbers were 4.90 ± 0.77 and 3.20 ± 0.48 for the EGF group in the regenerated newborn skin tissue, and the numbers were 4.28 ± 0.81 and 3.32 ± 0.37 for the Pelnac group in the regenerated newborn skin tissue.In general, the production of disordered extracellular matrix in the regenerated skin tissue could lead to fibrosis, followed by scar tissue. [35]Indeed, similar to the results above, the scarring ratio of the SA group was extremely low at merely 5.33 ± 1.62%, while those of the other groups were 63.87% ± 6.46% (the blank control), 54.68% ± 2.81% (MA), 18.08% ± 6.64% (SSAD), 50.93% ± 4.23% (EGF), and 48.68% ± 3.46% (Pelnac) (Figure 3H).
Vascularization [12] and remodeling of the extracellular matrix [36] have great significance in wound healing.New blood vessels can deliver nutrients and maintain oxygen homeostasis to ensure cell proliferation and tissue regeneration. [9]The amnion matrix contains numerous growth factors, such as EGF, nerve growth factor, HGF, keratinocyte growth factor, and transforming growth factors (TGF)-β1, β2, and β3, which can enhance the reepithelialization of chronic wounds.On the other hand, given that adhesive hydrogels could effectively close the incision and promote the healing process, [37] our SSAD hydrogel not only has good adhesion towards dressing purpose, but is also rich in a variety of growth factors related to wound regeneration, blood vessel formation, and re-epithelialization, such as VEGF, PDGF, EGF, HGF, and bFGF (Table S1, Supporting Information). [38,39]As such, it was hypothesized that the SA wound dressing could synergistically increase the efficacy of scarless skin healing.Indeed, compared with the blank control group and SSAD or MA applied alone, SA not only kept the wound surface moist but also avoided the disadvantages of amnion shedding, exhibiting enhanced generation of skin appendages as well as accelerated complete and scarless skin healing.
Moreover, studies have shown that fibroblasts can be differentiated into different lineages, in which reticulum fibroblasts are responsible for secreting collagens that form the scar tissue; after this, hair follicles cannot be regenerated. [40]This observation potentially explains why new hair follicles are rarely formed during uncontrolled wound healing.Reticulum fibroblasts can be identified based on the embryonic expression of engrailed-1 (EN-1). [41]Our results showed that the SA group had almost no EN-1 expression and that the SSAD group had a lower EN-1 content, while the control and MA groups, in contrast, both exhibited high expressions of EN-1 (Figure S6), consistent with our quantitative analyses of HE staining (Figure 3F).Accordingly, new hair follicles were missing in the healed skins in these latter two groups, whereas in the SA and SSAD groups, newly regenerated hair follicles could be readily observed, with the SA group showing more follicles.
Stromal cell-derived factor-1 (SDF-1) is the most abundant growth factor in SSAD, playing a key role in initial wound healing and significantly enhances the migration and recruitment of stem cells. [42,43]SDF-1 is downregulated in chronic wounds, which can lead to delayed wound healing. [44]The minimum effective concentration at which SDF-1 functions is 1 ng mL −1 , [45] while the concentration of SDF-1 in SSAD is as high as 35.79 ng g −1 .Growth factors can indirectly affect the behaviors of epithelial stem cells. [46]SDF-1 alone can promote cell migration, and SSAD was found to promote cell migration better than SDF-1 alone, suggesting that it might be the result of the combination of SSAD components in enhancing cellular behaviors. [47]nterfollicular epidermis stem cells (IFESCs) are crucial stem cells recruited to the wound site to promote skin healing. [48]It takes approximately 1 week for IFESCs to be recruited and migrate to the wound, and then they remain in place for about 35 days from the recruitment to help with skin repair. [49]As wounds heal, IFESCs disappear. [50]ccordingly, we stained the key biomarkers of IFESCs (integrin-α2 and β1) at Day 7 and Day 14 (Figure 4A).On Day 7, the positive cell ratio in the SA group was 8.20% ± 0.88% in each visual field, much higher than the values for the blank control group (1.93% ± 0.19%), the MA group (2.76% ± 0.98%), the SSAD group (4.91% ± 0.64%), the EGF group (2.47% ± 0.82%), and the Pelnac group (2.82% ± 0.57%).On Day 14, the positive cell ratio in the SA group (4.19% ± 0.80%) in each visual field was still much higher than the values for the blank control group (0.83% ± 0.17%), the MA group (1.59% ± 0.81%), the SSAD group (3.10% ± 1.02%), the EGF group (1.14 ± 0.27), and the Pelnac group (1.45% ± 0.28%) (Figure 4E).
The early stages of angiogenesis include the formation of excessive primitive networks, which need to be reorganized into secondary vascular networks with higher-level organizations.In the process of remodeling, new blood vessels are pruned, and abnormal vessels are removed through apoptosis to generate stable and well-perfused blood vessels, which can restore homeostasis and form resting endothelial cells.This might be the reason why the number of blood vessels observed on Day 35 was less than that observed on Day 14 (Figure 4B).
In addition, cytokeratin 14 (CK14) and cytokeratin 19 (CK19) were chosen as specific markers of hair follicles and sebaceous glands, respectively. [51]As shown in Figure 4C, noticeably more hair follicles and sebaceous glands were present in the SA group in the regenerated skin area, and a few regenerated hair follicles and sebaceous glands could be found in the SSAD group.However, almost no regenerated hair follicles or sebaceous glands were visible in the blank control group, the MA group, the EGF group, and the Pelnac group.With quantification, the number of new hair follicles in the SA group was 30.50 ± 3.01 in each visual field in the regenerated skin tissue, while that of the SSAD group was 9.83 ± 1.47 (Figure 4G).The number of new sebaceous glands in the SA group was 20.50 ± 3.08 in each visual field in the regenerated skin tissue, while that of the SSAD group was 6.67 ± 1.21 (Figure 4H).The improved regeneration of cutaneous appendages indicates the higher maturity of the regenerated skin tissue that can heal almost without scarring. [4]ecently, studies have suggested that myofibroblasts surrounding newly formed hair follicles can form dermal fat cells after injury, and this conversion of myofibroblasts to fat cells reduces the formation of scars. [52]Therefore, we examined the presence of adipocytes, which are known to guide hair follicle regeneration, [53] by immunostaining for perilipin-1 (PLIN1) (Figure 4D).On Day 35, few fat cells regenerated in the blank control group, the MA group, the EGF group or the Pelnac group; interestingly, the SSAD group revealed 5.47% ± 0.59% expression in each visual field, whereas the SA group demonstrated 8.36% ± 0.98% expression (Figure 4I).
HE staining images of the SA group at different time points illustrated that at 3 weeks after injury, small pockets of the epidermis grew downward to develop embryonal hair follicles in the wound center (Figure S7), which are the hair follicles at the early bud stage. [54]As the new hair follicles matured, their shapes became increasingly similar to those of normal hair follicles (Figure S7).Furthermore, HE examinations of the major organs (heart, liver, spleen, lungs, and kidneys) did not reveal any necrosis, hemorrhage, or congestion, indicating the good biocompatibility of SA (Figure S8).

Therapeutic mechanisms of SA for wound healing in rats
To further illustrate the underlying therapeutic mechanisms of SA, a rat skin defect model was chosen for further transcriptomic analyses.Compared with the blank control group, the SA group upregulated 2050 genes and downregulated 1229 genes; compared with the MA group, the SA group upregulated 1278 genes and downregulated 804 genes; and compared with the SSAD group, the SA group upregulated 1003 genes and downregulated 493 genes (Figure S9A).As shown in the Venn diagram (Figure S9B), we compared the gene set of the SA group with those of the other experimental groups and found that 414 genes were highly expressed in the SA group and coexpressed in the other experimental groups.We performed a clustering heatmap analysis of the expression levels of these 414 genes, through which we found that 304 of them were significantly upregulated in the SA group and relatively downregulated in the other groups (Figure S9C).This result illustrated that the expression profile of the SA group was significantly different from those of the other groups.We further performed gene function enrichment analyses on these DEGs.The KEGG pathway enrichment analysis revealed that the therapeutic mechanisms of SA were highly associated with extracellular matrix-receptor interactions, cell adhesion molecules (CAMs), the MAPK signaling pathway, the NK-κB signaling pathway, the Wnt signaling pathway, the Jak-STAT signaling pathway, the Ras signaling pathway, the VEGF signaling pathway, and the signaling pathways regulating pluripotency of stem cells (Figure S9D).The 20 significant gene ontology (GO) terms are shown in Figure S9E, including positive regulation of cell migration, positive regulation of keratinocyte differentiation, positive regulation of epithelial cell differentiation, stem cell development, epidermis development, stem cell division, and positive regulation of epithelial cell proliferation.
The extracellular matrix is a critical component of the skin microenvironment, and its synthesis and degradation are important processes for maintaining the ecological balance of the skin microenvironment. [55]The pathways related to regulating pluripotency, recruitment, proliferation, and migration of stem cells include the signaling pathway regulating pluripotency of stem cells, [56] the MAPK signaling pathway, [57] and the Ras signaling pathway. [58]The pathways related to angiogenesis include the Jak-STAT signaling pathway [59] and the VEGF signaling pathway. [60]We also observed that DEGs were enriched in the Wnt signaling pathway.Wnt signaling in epidermal keratinocytes is necessary for hair follicle regeneration, where excessive Wnt acting on the wound may promote the regeneration of hair follicles by changing the cell fate and increasing the number of cells capable of producing hair. [48]This fact explains to some extent why the SA group promoted skin healing without scarring.
The GO analyses further suggested that the upregulated genes coexpressed across the SA, SSAD, MA, and blank control groups were involved in epidermal development, stem cell division, stem cell development, cell proliferation, regulation of epithelial cell differentiation, regulation of vascular endothelial cell proliferation, and regulation of smooth muscle cell migration, among others.These biological processes promote the proliferation, migration, and differentiation of epidermal cells. [61]It has been reported that epidermal cells in the wound epidermis display the phenotype of hair follicle stem cells. [48]This observation also echoes the enrichment of DEGs in the SA group in the Wnt signaling pathway mentioned above.Through positive feedback regulating the proliferation and migration of vascular endothelial cells and the migration of smooth muscle cells, [62] SA treatment could significantly shrink the wound and promote the proliferation of blood vessels.The possible mechanism by which SA promotes skin regeneration is illustrated in Schematic 1.
IFESCs exist within their resident niche and respond to normal tissue homeostasis by being mobilized to repair a wound. [31]To more explicitly explore the mechanisms by which SSAD promotes skin healing, we also conducted single-cell RNA sequencing (scRNA-seq) analyses on sorted cells obtained from the wound dermis 14 days postwounding.Integrin-β1 is one of the marker genes of IFESCs.To enrich IFESCs, Integrin-β1-positive cells were enriched by flow sorting.Flow cytometry results revealed that the Integrin-β1-positive cell ratio in the control group was 17.2%, while the positive cell ratio in the SSAD group was 27.1% (Figure S10).These data suggested that there might be more IFESCs after SSAD treatment.
We then isolated cells from the samples of the control group and the SSAD group and applied them to the 10× scRNA-seq platform (Figure 5A).A total of 7130 cells in the control group and 11,085 cells in the SSAD group were captured.After cell filtering, 5874 single-cell transcriptomes were included in the final dataset (2491 for the control group and 3383 for the SSAD group) (Figure S11).Unsupervised clustering using Seurat categorized the cells into 20 clusters (Figure S12) based on global gene expression patterns (Figure S13), which were then assigned to eight main classes of cells (Figure 5B): fibroblasts, Langerhans cells, T cells, macrophages (MACs), myoepithelial cells, mast cells, interfollicular epidermis cells (IFECs), and endothelial cells (ECs).The composition of each cell cluster is listed so that the proportions of cells from the two groups could be identified across all cell clusters (Figure 5B).The percentage of IFECs was higher in the SSAD group (82.17%) than in the control group (17.83%) at 14 days postwounding (Figure 5B).Marker genes for the eight main classes of cells are shown in the heatmap in Figure 5C.Genes related to IFECs (Krt73, Igfbp3, Tchh, Krt75, Hspb1, Mt4, and Krt25) were substantially more highly expressed by the SSAD samples (Figure 5D).
Interfollicular epithelium (IFE) differentiation is gradual and consists of the following stages [63] : in the early differentiation stage, the basal layer of the epidermis forms, which is associated with mitosis on the basal plate; in the middle stage of differentiation, the basal cells withdraw from the cell cycle and form the spinous layer; in the terminal differentiation stage, the cells stop transcription activities and die off, flatten, and enucleate, and the impermeable cuticle forms.Therefore, we selected cells that were in the first-level clustering defined as IFECs (Figure 5B) and subjected them to a second round of unsupervised clustering as IFESCs, basal cells (BCs), spinous cells (SCs), keratinocytes 1 (KER 1), and keratinocytes 2 (KER 2) (Figure 6A) in both the control group and the SSAD group (Figure 6B).In general, the number of IFECs in the SSAD group was higher than that in the control group, especially for IFESCs, BCs, and SCs, while the number and proportion of KERs were almost the same (Figure 6C,D).Correlation analysis showed that the five cell subsets of IFECs were highly correlated (Figure 6E).Feacher plots (Figure 6F) and violin plots (Figure 6G) further illustrated the expression levels and distributions of key marker genes of IFECs (Figure S14).
To reconstruct the developmental trajectory during differentiation, pseudotime analysis of IFECs was conducted.In total, the pseudotime path had three branches (Figure 6H,I), and different cell clusters could be arranged relatively clearly at different branch sites of the pseudotime path (Figure 6I).The different developmental processes of IFE lineage cells could be observed from IFESCs to KERs (Figure 6H).The distributions of IFESCs and KERs in the development trajectory were relatively concentrated, but BCs and SCs could be found at several time points along the development trajectory (Figure 6I), suggesting that BCs and SCs might be at an intermediate stage of cellular development.For further understanding, we conducted heatmap analyses of the top 50 DEGs in the pseudotime analysis (Figure 6J).These DEGs likely activated IFESCs to differentiate separately into two directions.Indeed, by comparing DEGs in different clusters with the heatmaps, the IFESCs were divided into two branches after bifurcation: KERs and BCs/SCs (Figure 6H-J).Since IFESCs accounted for the highest proportion of IFECs, especially in the SSAD group, we subsequently focused on the enrichment analyses of DEGs in IFESCs.
GO analyses showed that the marker genes involved in biological processes associated with wound healing, epidermal development, hair follicle development, and stem cell migration were enriched in the IFESC population (Figure S15A).To further investigate the role of SSAD in wound healing, we performed GO analyses of DEGs of control versus SSAD.The regeneration and development of hair follicles, differentiation of epidermal cells, and migration and proliferation of stem cells related to scarless healing were all uniquely enriched in the SSAD group (Figure S16A), such as hair follicle development, positive regulation of epithelial cell differentiation, positive regulation of stem cell differentiation, positive regulation of stem cell proliferation, hair follicle morphogenesis, and MAPK cascade.KEGG pathway enrichment analyses also indicated similar results (Figure S16B); the Wnt pathway related to hair follicle regeneration, the extracellular matrix-receptor interaction pathway related to maintenance of dynamic balance of the skin microenvironment, the CAM pathway related to cell adhesion, and the MAPK signaling pathway related to stem cell proliferation and migration, were uniquely enriched in the SSAD group (Figure S16B).
The skin comprises tissue macrophages as the most abundant resident immune cell type. [64]Therefore, we selected cells that were in the first-level clustering defined as MACs (Figure 5B) and subjected them to a second round of unsupervised clustering as monocytes (Mo), M1 MACs (M1), M2 MACs (M2), and resident MACs (RM) (Figure 7A).In general, the number of M2 MACs in the SSAD group was higher than that in the control group (Figure 7B).Violin plots further showed the expression levels and distributions of key marker genes of MACs (Figure 7C-F).The violin plots indicated that the gene expression of each cell subset was significantly different from the others, and thus the different subsets could be distinguished from each other.GO and KEGG analyses suggested that MACs were enriched in inflammatory responses, immune regulation, as well as immune cell proliferation and survival, such as cytokine-cytokine receptor reaction, response to drug, response to bacteria, the nucleotide-binding oligomerization domain (NOD)-like receptor signaling pathway, the chemokine signaling pathway, and the Toll-like receptor signaling pathway, among others (Figure 7C-7F).
The pseudo-time analyses of MACs were also conducted.Interestingly, the pseudo-time path only had 1 line (Figure 7G,H), and different cell clusters could be arranged relatively clearly at different branch sites of the pseudotime path (Figure 7H).The different developmental processes of the various MACs cell lineages could be observed from monocytes to resident MACs (Figure 7H).After skin injury, a large influx of monocytes that are produced by adult myeloid progenitors infiltrate into the injured sites and differentiate into macrophage-like cells. [65]MACs receive constant input from circulating monocytes, especially when inflammatory response occurs. [64]After the early inflammatory phase subsides, the predominant MACs population assumes a wound-healing phenotype, that is, M2 MACs that is characterized by the production of numerous positive growth factors. [66]Many tissue-resident MACs are long-lived in mice and can proliferate within their tissue of residence, a mechanism involved in their maintenance at adulthood. [67]n some healthy tissues, MACs are constantly replenished by circulating monocytes, such as dermal MACs, and after inflammation occurs in local tissues, MACs differentiate to regulate immune response and promote wound healing, eventually turning into resident MACs. [68]Therefore, after skin injury, monocytes differentiate into M1-type MACs, and after early inflammation subsides, MAC population differentiates into M2-type MACs, which eventually turn into resident MACs (Figure 7G,H).
In general, the fraction of monocytes in the SSAD group was higher than that in the control group (Figure 7B), indicating that SSAD could mobilize more monocytes and initiate the immune responses more quickly.The fractions of M1 MACs were almost the same in the two groups.However, there were more M2 MACs in the SSAD group, suggesting that SSAD might promote M2-polarization and facilitate wound healing.In the end, only a few monocyte-derived MACs probably convert into resident MACs, while most MACs either egress from the inflammatory site or more likely, die. [69]The fraction of resident MACs in the control group was higher than that in the SSAD group, indicating that the SSAD group had already at the later stages of wound healing, while the control group was less involved in immunomodulatory responses.Correlation analysis further revealed that the four cell subsets of MACs were highly correlated (Figure 7I).
When tissues are injured, inflammatory response is induced by damage-associated molecular patterns and pathogen-associated molecular patterns, which are released by dead/dying cells and invading organisms, respectively. [70]hese molecular triggers induce a complex inflammatory response that is characterized by the recruitment, proliferation, and activation of a variety of hematopoietic and nonhematopoietic cells, including MACs, fibroblasts, epithelial cells, endothelial cells, and stem cells, which together make up the cellular responses that orchestrate tissue repairing. [71]lthough many cell types are involved in tissue repair, because of their highly flexible programming, [72] MACs have been shown to exhibit critical regulatory activities at all stages. [73]To explore potential interactions among MACs, IFECs, endothelial cells, and fibroblasts, we performed the CellChat analyses on these datasets. [74,75]According to pseudotime analyses, IFECs were divided into IFESCs, KERs, BCs, and SCs for CellChat analyses.
Based on the ligand-receptor heatmap (Figure S17), we subsequently focused on the analyses of cell interactions among M1 MACs, IFESCs, KERs, BCs, SCs, fibroblasts, and endothelial cells since they had the highest intercellular communications (Figure S17).For KERs (Figure 8A), M1-secreted SPP1 binds to the receptor CD44 of KERs and promotes stem cell enrichment, [76] inhibiting apoptosis and helping cells survive after injury. [77]SPP1 regulates cell adhesion and chemotaxis. [78]Two major functions of CD44 in the skin are the regulation of KER proliferation in response to extracellular stimuli and the maintenance of local hyaluronate homeostasis. [79]TGFB1 secreted by MACs binds to TGFBR3, which is a regulating growth factor dur-ing repairing, [80] promoting cell migration and epithelial cell proliferation together with FAM3C. [81]We then analyzed the CellChat between IFECs (Figure 8B) and found that the expressions of receptor ligands were very similar, suggesting that they may exhibit similar effects.In epidermis, NECTIN is expressed at cell-cell junctions mainly in the suprabasal layer. [82]Nectin-4 promotes cell adhesion, migration, proliferation, and increases stem cell stemness. [83]GFR2 and CD44 positively regulate each other's expression, maintaining cell stemness. [84]LAMP1 and FAM3C can promote epithelial cell proliferation. [81]For endothelial cells and M1 MACs (Figure 8C), CXCL9 secreted by endothelial cells is a pro-inflammatory chemokine that triggers immune cell migration and rapid relocation to boost antibody secretion at the infected site. [85]MACs AREG receptors bind to ICAM1 of endothelial cells and regulate stem cell pluripotency of endothelial cells. [86]For M1 MACs and fibroblasts (Figure 8D), TGFB1 secreted by MACs leads to the activation of EGFR receptors, which leads to the activation of the TGF-Smad pathway, resulting in fibrosis and cell proliferation. [87]GF-B1 is induced on tissue injury and regulates tissue remodeling and wound healing, but its dysregulated signaling results in excess extracellular matrix deposition and fibrosis. [80]For IFESCs and fibroblasts (Figure 8E), FGF2 secreted by IFESCs binds to CD44 of fibroblasts to induce genes that repair damaged epithelium and prevent fibrosis. [88]EGFA and SEMA3C secreted by fibroblasts bind to NPP2 receptor of IFESCs and promote regeneration in wound healing. [89]For fibroblasts and KERs (Figure 8F), TGFB1 secreted by IFESC and TGFB3 bind to the TGFBR3 receptor of fibroblasts to promote fibrosis. [80]IGF1 binds to IGF1R and regulates cell proliferation, differentiation, and migration. [90]n the 14th day after injury, wound healing would enter the stage of proliferation where M1 MACs transform into M2 MACs. [47]Of note, more M2-type MACs were observed in our SSAD group (Figure 7B), suggesting that the wound healing effect of the SSAD group might be better than the control group.The numbers of M1 MACs in the two groups were basically equal.However, the interactions of M1 MACs with IFECs, fibroblasts and endothelial cells would promote wound healing (Figure 8).IFESCs and fibroblasts are indispensable and important cells in wound healing. [41]IFESCs rapidly divide, proliferate, and migrate to close the wound, while fibroblasts migrate to the defect area and secrete collagens to repair the wound.Rapid repair by fibroblasts often results in scarring, especially full-thickness skin defects. [4]In healthy skin, collagen fibers are arranged in the form of a lattice or a woven basket, but during wound healing, fibroblasts arrange collagen fibers in parallel, resulting in hard and weak tissues, [4] which is consistent with the results of the control group (Figure 3D).Fibroblasts play a role in wound healing but are also thought to be responsible for the formation of scar tissues. [91]In our single-cell sequencing results, there were more IFESCs and fewer fibroblasts in the SSAD group, while the opposite was true in the control group.These data suggested that the wounds in the SSAD group healed differently from those in the control group.The GO and KEGG pathways uniquely enriched in the SSAD group in IFESC clusters also further confirmed that the application of SSAD was favorable in facilitating the proliferation and recruitment of more IEFSCs in skin wounds, activating the Wnt

CONCLUSION
This study shows that the SSAD hydrogel was feasible as a wound dressing to load MA.The thin thickness and poor mechanical strength of the amniotic membrane make MA itself difficult to be fixed at the wound site, which greatly limits its broad applications.The use of SSAD hydrogel as a scaffold to load the amniotic membrane solves this prob-lem.SSAD not only maintained the biological activities of the amnion and the AMSCs but also synergized with them in promoting healing by facilitating migration of IFESCs, recruitment of endogenous stem cells, formation of new blood vessels, and regeneration of skin appendages.This unique SA platform avoids the limitations of traditional biological materials such as amniotic membrane used to promote wound healing and brings new hope towards future clinical applications in promoting scarless healing of large-area skin traumas.According to the single-cell transcriptome analyses, the main cells involved in wound healing included MACs, fibroblasts, endothelial cells, and IFECs.IFESCs in IFECs play an important role in wound healing, and IFESCs can rapidly divide, proliferate, and migrate to close the wound.Ligands of MACs promote proliferation and migration of endothelial cells and IFECs and regulate stem cell pluripotency.Fibroblasts, on the other hand, are more susceptible to fibrotic signaling from other cells, which promotes wound healing to some extent but also causes fibrosis.Of note, in vivo experiments were conducted using splinting model of the rat full-skin defect.Since the skin of rats would contract during the healing process, which is different from human skin, although the splint model could be adopted to minimize the influence of the skin contraction of rats on the experimental results of skin healing, [92] it is necessary to consider the use of additional model animals such as pigs, which are more similar to the characteristics of human skin, for further experiments.Moreover, although in vivo experiments on rats have shown that SA is biocompatible, but more studies are needed to ensure its safety in humans before this wound dressing can be translated into the clinics.

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no competing interests.

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

F I G U R E 1
Characterizations of the materials and the bioactive effects of skin secretions of Andrias davidianus (SSAD) on micronized amnion (MA).(A) Photographs of SSAD powders and hydrogels.(B) Photograph of fresh amnion.(C and D) Hematoxylin-eosin (HE) staining images of (C) the amnion surface and (D) the longitudinal section of MA.The blue circle indicates the amniotic epithelial cell, and the purple triangle indicates the basement membrane.(E) Scanning electron microscopy (SEM) image showing the adhesion interface between the SSAD hydrogel (60-mesh) and the MA.The yellow pentagon indicates the SSAD hydrogel, and the red triangle indicates the MA.(F) Live/dead cell staining.The MA was cultured in normal medium (control) or SSADconditioned medium (SSAD).Live cells were stained green, and dead cells were stained red.(G) Quantified ratios of viable cells based on live/dead cell staining images.(H) Terminal deoxynucleotidyl transferase (TdT) dUTP nick-end labeling (TUNEL) assays of MA and SA (60-mesh).Nuclei were stained blue, and apoptotic cells were stained green.(I) Quantified numbers of apoptotic cells based on TUNEL assay images.(***p < 0.001).

F I G U R E 2
The effects of skin secretions of Andrias davidianus (SSAD) on HaCaT cell activities and transcriptome analyses of HaCaT cells cultured with SSAD-conditioned medium.(A) Scratch assay and (B) transwell migration assay of HaCaT cells.(C) Wound healing ratios from the scratch assay of HaCaT cells.(D) Relative cell number from the transwell assay of HaCaT cells.(E) CCK8 assays showing the proliferation profiles of HaCaT cells at 24 h.(F) Venn diagram of the transcriptomic profiles between the control group and the SSAD group.(G) Volcano plots showing the identified upregulated and downregulated genes induced by SSAD-conditioned medium.(H) Heatmaps of significantly upregulated genes involved in the proliferation, migration, and apoptosis of the cells in both groups.(I) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed on the identified differentially expressed genes.The figure shows the 20 most significantly enriched pathways.

F I G U R E 3
The effect of SA on skin wound healing in vivo.(A) Representative images showing wound healing in the different groups.Diameter of blue scales: 20 mm.(B) Quantification of wound closure ratios of the different groups (n = 6; ***p < 0.001).(C) Representative hematoxylin-eosin (HE) staining of the skin tissues at 35 days postwounding for observing the overall morphologies.The black triangles show the boundaries of the wounds.BV, blood vessel; HF, hair follicle; SG, sebaceous gland.(D) Masson's trichrome staining of the skin tissues at 35 days postwounding to observe the reorganization of collagen fibers.(E) Quantification of relative skin thicknesses ratios in the wound center of the different groups at 35 days postwounding.(F) Quantification of the numbers of hair follicles in the different groups at 35 days postwounding.(G) Quantification of the numbers of sebaceous glands in the different groups at 35 days postwounding.(H) Quantification of the scar ratios of the different groups at 35 days postwounding.(n = 6; ***p < 0.001).

F I G U R E 4
Immunofluorescence staining of the regenerated skin tissues and quantification results.(A) Interfollicular epidermis stem cell (IFESC) staining (brown color) at Days 7 and 14. (B) Staining for neovascularization (red circles) at Days 14 and 35.(C) Staining for sebaceous glands (SG) and hair follicles (HF) at Day 35.(D) Staining for adipocytes (green color) at Day 35.In all cases, nuclei were counterstained in blue.(E) Quantitation of the densities of IFESCs on Days 7 and 14. (F) Quantitation of the numbers of vascular rings on Day 14 and Day 35.(G) Quantitation of the numbers of hair follicles on Day 35.(H) Quantitation of the numbers of sebaceous glands on Day 35.(I) Quantitation of the percentage of adipose cells (PLNI1+) on Day 35.(n = 6; ***p < 0.001).

F I G U R E 5
Overview of the single-cell transcriptome analyses.(A) Workflow for single-cell experiments.PW, postwounding.(B) Cells from the wounds were categorized into eight main classes.The middle column represents the percentages of the cells in the 8 classes averaged for all the analyzed samples.Compositions of cells (%) in the control (Ctrl) and SSAD groups are listed on the right.(C) Heatmap showing marker gene expression for the cells in the 20 clusters and the eight classes.(D) Up-and downregulated genes in all clusters.BC, basal cell; IFESC, interfollicular epidermis stem cell; KER, keratinocyte; SC, spinous cell.

F I G U R E 6
Additional analyses of interfollicular epidermis cells (IFECs).(A) Subclustering of IFECs showing 5 subsets.(B) Cell distributions of IFECs in the control group and the skin secretions of Andrias davidianus (SSAD) group.(C) The fractions of the different subclusters of IFECs in the control and SSAD groups.(D) The numbers of the different subclusters of IFECs in the SSAD group.(E) Correlation analysis of IFEC subsets.(F) Expression levels of select genes that were enriched in IFESCs, basal cells (BCs), spinous cells (SCs), keratinocytes (KERs) 1, and KERs 2. (G) Violin plot displaying the expression of marker genes for the different IFEC subsets.(H) Distributions of cells on the pseudotime trajectory.The darker the color is, the earlier the stage of cell differentiation.(I) Pseudotemporal ordering of IFECs and the distributions of five subsets along the trajectory.(J) Gene expression heatmaps of the 50 top differentially expressed genes (DEGs) (cataloged in four clusters) in a pseudotemporal order.KERs are shown at the bottom, and SCs and BCs are both shown on top.

F I G U R E 7
Additional analyses of macrophages (MACs).(A) Subclustering of MACs showing four subsets.(B) The fractions of the different subclusters of MACs in the control and SSAD groups.(C) Violin plot displaying the expression of marker genes for monocyte (Mo) subset.GO and KEGG analyses of Mo in both control and SSAD groups.(D) Violin plot displaying the expression of marker genes for M1 MACs subset.GO and KEGG analyses of M1 MACs in both control and SSAD groups.(E) Violin plot displaying the expression of marker genes for M2 MACs subset.GO and KEGG analyses of M2 MACs in both control and SSAD groups.(F) Violin plot displaying the expression of marker genes for resident MACs subset.GO and KEGG analyses of resident MACs in both control and SSAD groups.(G) Distributions of cells on the pseudotime trajectory.The darker the color is, the earlier the stage of cell differentiation.(H) Pseudotemporal order of MACs and distributions of the four subsets along the trajectory.(I) Correlation analysis of MACs subsets.

F I G U R E 8
Analyses of cell interactions.(A) Cell interactions between keratinocytes 1 (KER 1) and M1.(B) Cell interactions between interfollicular epidermis cells (IFECs).(C) Cell interactions between M1 and endothelial cells.(D) Cell interactions between M1 and fibroblasts.(E) Cell interactions between IFESCs and fibroblasts.(F) Cell interactions between KERs and fibroblasts.pathway,and enhancing hair follicle regeneration and scarless healing.
This work was financed by the National Natural Science Foundation of China (grant number: 32070826 and 31871464), Science and Technology Research Project of Chongqing Education Commission (grant number: KJQN202200471), Chongqing Science and Health Joint Medical project (grant number: 2020GDRC017), CQMU Program for Youth Innovation in Future Medicine (grant number: W0075), Senior Medical Talents Program of Chongqing for Young and Middle-aged and the Key Research Cultivating Project of Stomatological Hospital of Chongqing Medical University (grant number: PYZD201603), and Program for Innovation Team Building at Institutions of Higher Education in Chongqing in 2016 (grant number: CXTDG201602006).Y.S.Z. was not supported by any of these funds; instead, support by the Brigham Research Institute is acknowledged.