Advances in scarless foetal wound healing and prospects for scar reduction in adults

Abstract Healing after mammalian skin injury involves the interaction between numerous cellular constituents and regulatory factors, which together form three overlapping phases: an inflammatory response, a proliferation phase and a remodelling phase. Any slight variation in these three stages can substantially alter the healing process and resultant production of scars. Of particular significance are the mechanisms responsible for the scar‐free phenomenon observed in the foetus. Uncovering such mechanisms would offer great expectations in the treatment of scars and therefore represents an important area of investigation. In this review, we provide a comprehensive summary of studies on injury‐induced skin regeneration within the foetus. The information contained in these studies provides an opportunity for new insights into the treatment of clinical scars based on the cellular and molecular processes involved.

beautify or reduce their scars. According to available reports, it has been estimated that up to tens of billions of dollars are expended for treatments and consequences of scars. 1,4 Despite the application of various methods for the treatment of hypertrophic scars or keloids, the effectiveness of these protocols remains quite limited. Accordingly, a better understanding of the mechanisms involved with scar formation is sorely needed. One potential approach to investigate the root cause of scars involves an examination of this process in the foetus. In the human foetus, a scarless wound repair process is present, a phenomenon that was reported over 30 years ago. 5 Subsequent studies have revealed that this scarless wound repair depends on age and that there is a transition from a scar-free to scar formation process that occurs in the later period of foetal development. This transition occurs at around 24 weeks in human pregnancy and at day 18 of gestation in mice. 6,7 This extraordinary ability for the initial prevention of scar formation followed by a transition to a scar formation process in the foetus suggests that adult skin may also possess a similar mechanism, which could be activated to reduce adult scar formation like that observed in the foetus.
The molecular basis of scarless wound healing in mammalian foetuses has been partially reported before, 8 but research on the basis of small rodent models has not been comprehensively sorted out. In this review, we summarized research mainly in the recent years on the healing mechanisms as observed in the foetus and adult from the perspective of tissues, cells and molecules, and explored the possibility of applying these findings to scar treatment. We focused on the wound healing process within the foetal model due to its remarkable capacity for scarless healing and evaluated the advantages and disadvantages that can be garnered from this model. Compared with the previous review, 9 up-to-date findings on foetal skin and adult scars using gene sequencing technology, and mechanisms of various non-coding RNAs in the formation of scars and their potential application for clinical treatment are also included in this review.
Differences in tissue structure, cell types and molecules between foetal wound and adult wound are briefly illustrated in Figure 1.

| E VOLUTI ON OF THE FOE TAL WOUND MODEL
Since the initial discovery of foetal skin regeneration, a variety of animal wound healing models have been developed, including lamb, monkey, rabbit and opossum, with the goal that these models will provide an understanding of the mechanisms involved in this regeneration process. [10][11][12] As one example, Stelnicki and colleagues used foetal lambs to evaluate the neural dependence of scarless healing after denervation. 13 Although these large animal models are technically more advantageous, the expenses involved severely limit their use in these experiments. The substantially lower costs and shorter gestations of rats and mice, along with well-identified genomes and transgenic technology that can be used to examine loss or gain of function in mice account for the preferential use of these rodent models.
A considerable amount of valuable information has been obtained from in vitro experiments using major functional cell lines of F I G U R E 1 Differences and components involved in the repair process of foetal mice and adult mice after wound modelling keratinocytes and fibroblasts subjected to injury. 14-16 However, a major deficit in these studies was their failure to include all components of the skin and lack of assessing these injury responses in an in vivo environment. 17,18 Therefore, the development of an effective in vivo wound healing model is required for exploring the complex mechanisms involved with these responses.
In 1992, a graft of human foetal skin was observed to demonstrate scar-free wound healing when applied onto the skin of mice.
In this study, the amniotic fluid environment was not a necessary condition for this scar-free regeneration of foetal skin. 19 These findings provided the feasibility and foundation for the development of in vivo models to study foetal injury repair in mammals. One of the first attempts was that of Stelnicki and colleagues in 1997 and involved use of incisional wounds on the hind limbs of 14-day-old mice 20 ; however, the limited nature of their incision failed to affect the integrity of the skin in these mice, thus preventing an assessment of the healing process of this dermal deficiency. Over the next ten years, a number of in vivo models of excisional wounds were developed with a rat model in 2002 and the first in vivo excisional mouse model in 2006. 7,21,22 The recent scar-free wound healing model using foetal mice has been updated with the protocol of timed superovulation, while emphasizing the importance of amniotic fluid supplementation during surgery. 23 In Figure 2, we show the exploration of foetal wound healing models, especially rat and mouse models, in the above-mentioned time sequence since the 1990s.

| From macro to histological levels
Over the years, there have been a substantial number of studies directed at examining various aspects of tissue samples after foetal damage. In Table 1, we summarize the studies using the in vivo rat and mouse wound models that have been used to unearth the mechanisms of foetal skin regeneration. 22,[24][25][26][27][28][29][30][31][32][33][34][35] Needless to say there are an extensive number of genes differentially expressed during foetal development, regardless of whether the foetus is damaged or not. With the maturation of gene sequencing technology, the mechanisms by which non-coding RNAs participate in this process have also emerged. For example, dicer, the key enzyme for miRNA synthesis, is significantly increased in late embryonic tissue. 36 A particularly interesting mouse model of relevance to this review is the African spiny mouse (Acomys), which shows regenerative responses in their skin following damage. 37 These regenerative responses show a similar phenotype to that of the wound healing observed in the foetus with lower levels of inflammatory responses and no significant excessive amounts of collagen secretion. 38,39 When compared to normal mice, Acomys show differences in expression levels of phosphorylation-related proteins, macrophage markers and immunomodulators after injury. Ubiquitination and phosphorylation are major mechanisms significantly enhanced in protein synthesis and degradation. 40 Cellular and molecular differences after foetal injury will be discussed below.

| Key roles of fibroblasts
Fibroblasts represent one of the most critical cells that participate in the entire process of wound repair through their effects on proliferation, differentiation, synthesis of collagen, interactions with other cells 41 and their capacity to form different fibrotic environments in various organs. 42 Results obtained using sophisticated cell isolation methods 43 have revealed that enormous differences in gene expressions within fibroblasts are present between the mid-and lategestational periods in mice. These changes in fibroblasts at different developmental stages suggest that they may be closely related to the distinct responses observed after injury at these times, 44,45 as related to the accompanying activation and inhibition of multiple pathways. 46 Cyclooxygenase 1 was one of the 12 differentially expressed inflammatory genes, which was expressed at higher levels in E18 versus E15 fibroblasts. 45 When comparing human foetal versus adult dermal fibroblasts, as achieved with use of protein expression profiling, it was found that the heat-shock cognate 71 kDA protein, which is active in organism survival, and tubulin alpha-1A showing key studies and models reported since the 1990s chain, a vital component of cell cytoskeleton, both showed greater expressions in foetal fibroblasts that conceivably account for the rapid regeneration of foetus, whereas cofilin-1 and peroxiredoxin-1 were expressed at comparatively higher concentrations in adult fibroblasts. 47 An area of investigation that has attracted considerable attention of late is that of research on fibroblast heterogeneity and the tracking of different fibroblast lineages. It seems clear that distinct lineages can play diverse roles at different stages of skin development. For example, the reticular fibroblasts located in the lower portions of the dermis are responsible for synthesizing extracellular matrix and begin to repair immediately after injury. In contrast, papillary fibroblasts in the upper portion of the dermis regulate hair follicular formation and may be related to the loss of regional hair follicles after adult injury and play a role in the later stages of wound healing. 48,49 As based on the gene, engrailed 1 (En1), which is expressed in embryonic cells, two related fibroblast cell lines have been identified, En1-lineage-past fibroblasts TA B L E 1 Summary of studies using the in vivo rat and mouse wound models

| Differences in keratinocytes
As a critical component of the epidermis, keratinocytes secrete keratin filaments to participate in the keratinization and protection of skin from the invasion of external microorganisms. 56 Of particular significance to the present review, keratinocytes also undertake the major mission of re-epithelization and wound closure after skin damage 57 and are a key factor in the delay of wound healing in adults. 58 Pastar and colleagues reported some notable changes that occur in keratinocyte genes during embryonic development with 24 genes being down-regulated in keratinocytes at E16 as compared to E18 and no obvious evidence for any genes being up-regulated. In E18 keratinocytes, specifically up-regulated β-catenin-dependent Wnt pathways were demonstrated to play a key role in the scarring process of late-gestational age. 44 In contrast, focal adhesion kinase (FAK) was previously manifested a close connection with mechanotransduction and fibrosis.
Interestingly, keratinocyte FAK-deleted mice were found to obtain thinner skin and lessened collagen density, while FAK-deleted keratinocytes actively participated in mechanotransduction and extracellular matrix production, by overexpressing Igtbl, Mmpla and Col4a1. 59,60 Receptor-interacting protein kinase 4 (RFPK4) 61,62 and interferon regulatory factor 6 (IRF6) 63-65 both play a vital role in the differentiation of keratinocytes, with mutations in either of these two having the potential of resulting in epidermal developmental disorders. [66][67][68] As compared to that observed in RIPK4 full knockout mice, in neonatal mice with a specific removal of RIPK4 in keratinocytes, the epidermis shows a normal structure, but the lack of a skin barrier function leads to accelerated water loss and severely limits the lifespan of these newborns. 69  In addition to these components of foetal skin that can modulate this wound healing process, the amniotic fluid, cord blood and amniotic membrane present during pregnancy contain a rich source of stem cells, which can also contribute to the repair and scar-suppressing effects. 83 by reduced level of inflammatory cytokine expressed by UCB-MSCs and closely analogous collagen type III proportion to that of the foetal wound triggered by hAFs. Simultaneously, exosomes, as part of the paracrine activity of stem cells, also promote wound healing and limit scar formation. 87,88 Recent evidence indicates that peripheral cells play a role in scar formation and regulation of vascular regeneration. 89 As peripheral cell therapy has been shown to reduce the accumulation of inflammatory cells and organize collagen in remodelling phases, this suggests one possible mechanism for this effect of peripheral cells in this scar formation. 90

| Proteins displaying key functions
In foetal sheep, a diverse array of molecular responses has been observed in response to wounds and trauma. For example, IL-10 effectively reduces the number of inflammatory cells and significantly lessened scars at the location of a severe trauma. 91

| Non-coding RNAs implicated in dermal fibrosis
Along with this research focusing on proteins, in recent years, the emergence of non-coding RNAs, which actively participate in every stage of wound healing, has also attracted a considerable amount of attention. 9 Gene therapy involved with the application of noncoding RNAs will inevitably represent a major breakthrough in the clinical treatment of various diseases. With regard to skin fibrosis, the mechanisms responsible for the extreme hyperplastic scar and keloid in trauma sequelae are closely related to non-coding RNAs, in particular miRNA and lncRNA. Accordingly, the mechanisms of non-coding RNA involvement in scar development and their prospects for being targeted in clinical treatment warrant discussion.

| Effects of miRNAs on scar formation
miRNAs are small molecules of approximately 22-nucleotide lengths, which target the inhibition of mRNA translation. Interestingly, miR-NAs are characterized by dynamic changes during embryonic development and usually suppress expression of factors in foetal wounds. 99 In Table 2, we present a summary of research primarily focused on the impact of miRNAs in scar formation during the adult stage.  Differentially expressed miRNAs have been detected between scar lesions in adult versus embryonic skin. 36,121 Such discriminatory responses in core miRNAs as related to scar development after injury suggest promising new targets for clinical treatment.   [124][125][126][127][128] In contrast, the axes lncRNA COl1A2-AS1/miR-21/smad7, lncRNA 8975-1 and lncRNA AC067945.2 inhibit scar formation through exerting opposite effects on fibroblast pathological changes and collagen synthesis as that observed in scar formation. [128][129][130] Moreover, in cell co-culture experiments, lncRNA ASLNCS5088 exosomes as secreted from M2 cells endogenously bind miR-200c-3p, thus promoting fibroblast differentiation and pathological function, an effect which can be reversed by GW4869. 131 As these complicated functions of lncRNAs are gradually revealed, their application and effectiveness in the treatment of hypertrophic scars and keloids will be realized.

| CON CLUS IONS
Research on scarless wound healing and scar reduction has been ongoing for decades, with innovative techniques and in vitro and in vivo models used to identify numerous cells, factors and novel genes related to this process. Armed with this information, as well as findings from future studies, effective clinical treatment, such as targeted drugs for key molecules or the development of engineered cells, will gradually be revealed. However, the complex and overlapping pathways may require more investigations on the regulation mechanisms of genes with unidentified functions. Newly emerging technologies focusing on detection of non-coding RNA, single-cell transcriptome and RNA methylation can be regarded as a breakthrough in further exploration of unknown areas. Even if the impeccable regeneration of foetal skin cannot be fully exerted in clinical practice, at present, we feel this review will promote the evolution of this process to bring new perspectives for future treatment and cosmetology.

CO N FLI C T S O F I NTE R E S T S
The authors have no conflicts of interest to declare.

DATA 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 reasonable request.