Amphiregulin promotes hair regeneration of skin‐derived precursors via the PI3K and MAPK pathways

Abstract Objectives There are significant clinical challenges associated with alopecia treatment, including poor efficiency of related drugs and insufficient hair follicles (HFs) for transplantation. Skin‐derived precursors (SKPs) exhibit great potential as stem cell‐based therapies for hair regeneration; however, the proliferation and hair‐inducing capacity of SKPs gradually decrease during culturing. Materials and Methods We describe a 3D co‐culture system accompanied by kyoto encyclopaedia of genes and genomes and gene ontology enrichment analyses to determine the key factors and pathways that enhance SKP stemness and verified using alkaline phosphatase assays, Ki‐67 staining, HF reconstitution, Western blot and immunofluorescence staining. The upregulated genes were confirmed utilizing corresponding recombinant protein or small‐interfering RNA silencing in vitro, as well as the evaluation of telogen‐to‐anagen transition and HF reconstitution in vivo. Results The 3D co‐culture system revealed that epidermal stem cells and adipose‐derived stem cells enhanced SKP proliferation and HF regeneration capacity by amphiregulin (AREG), with the promoted stemness allowing SKPs to gain an earlier telogen‐to‐anagen transition and high‐efficiency HF reconstitution. By contrast, inhibitors of the phosphoinositide 3‐kinase (PI3K) and mitogen‐activated protein kinase (MAPK) pathways downstream of AREG signalling resulted in diametrically opposite activities. Conclusions By exploiting a 3D co‐culture model, we determined that AREG promoted SKP stemness by enhancing both proliferation and hair‐inducing capacity through the PI3K and MAPK pathways. These findings suggest AREG therapy as a potentially promising approach for treating alopecia.


| INTRODUC TI ON
Alopecia resulting from various factors that decrease hair follicle (HF) regeneration 1 affects a significant fraction of the world population. Although not life-threatening, hair loss predisposes people to considerable psychological distress 2,3 and severe quality-of-life impairment. 4 Its pathogenesis varies and leads to symptoms, such as androgenetic alopecia, alopecia areata, diffuse alopecia and therapyinduced hair loss. 5 Several oral, surgical, topical and intralesional treatments have been progressively developed in recent decades to achieve delayed hair loss or hair restoration in alopecic areas. 6 Although hair transplantation is regarded as the gold standard for creating natural-appearing hair in androgenic alopecia (AGA), 7 there is an unavoidable controversy surrounding autologous transplantation owing to the limited source of donor hair, the decreased viability of cells extracted in this time-consuming procedure, and, most importantly, the impermanent outcomes as the disease progresses. 1,3 Regarding oral medicines, drugs approved by the Food and Drug Administration (FDA), such as finasteride, dutasteride and minoxidil for AGA, vary greatly in efficacy from person to person and provide partial and temporary benefits, as well as side effects. [8][9][10][11] Such surgical procedures and medications cannot always meet patient satisfaction given the finite sources, unfavourable side effects and partial benefits.
Stem cell-based therapies have gained tremendous attention owing to their advantages of infinite self-renewal capacity and multi-lineage differential potential, 6 thereby providing approaches to coping with the challenges posed by traditional alopecia therapies. Mesenchymal stem cells (MSCs) show considerable promise for intravenous injection of human MSCs into non-scarring alopecia 12 and co-culture with human dermal papilla cells (hDPCs) 13,14 in preclinical studies. Moreover, using conditioned medium from cells (eg, Wnt1a, 15,16 Wnt7a, 17 or Nanog-overexpressing MSCs 18 ) can accelerate the telogen-to-anagen transition of HFs. 19 Adipose-derived stem cells (ASCs) are another latent cell population used in regenerative medicine. With their advantage of being easier to obtain than MSCs and low immunogenicity, they play a vital role in the activation of epidermal stem cells or hDPCs by secreting growth factors, such as vascular endothelial growth factor, hepatocyte growth factor, platelet-derived growth factor, insulin-like growth factor 1, thymosin beta 4, stromal cell-derived factor 1α, proinflammatory chemokine C-C motif chemokine ligand 2 and endothelial growth factor, thereby providing significant promotion of HF morphogenesis both in vitro and in vivo. [20][21][22][23][24][25][26][27][28] Additionally, administering the stromal vesicular fraction (SVF) or lipoaspirate obtained from abdominal fat to the scalp graft yields favourable outcomes. 29,30 However, MSC and ASC therapies utilize the paracrine function, and the implanted cells cannot regenerate into HFs. Accordingly, considering that the treatment effect is not guaranteed for different types and severity of hair loss and their inability to naturally grow and differentiate into HFs specifically, identification alternate and more promising strategies for hair regeneration is required.
Skin-derived precursors (SKPs) are a multipotent precursor cell population from the adult mammalian dermis capable of differentiating into several lineages, such as dermal, neural and mesodermal progeny and show promise for therapeutic and regenerative medicine associated with HFs. [31][32][33][34] The DP of HFs appears to comprise SKPs based on identical patterns of gene expression (Nexin, Wnt5a and versican), and cells from adult whisker follicle papillae cultivated under SKP conditions can generate properties similar to SKPs. 35 SKPs from the HF niche can not only differentiate into dermal cell types, such as fibroblasts and myofibroblasts, but also induce HF morphogenesis. 36,37 These findings indicate SKPs as ideal "seed cells" to yield complete HF structures, 38 thereby providing a possible avenue to address the graft origin of hair loss.
However, following SKP extraction from their physiological environment, their ability to generate large amounts or functional HFs decreases over time. Furthermore, the critical signalling pathways necessary to maintain SKP properties, self-renewal and proliferation remain unclear. Methods to produce highly proliferative SKPs capable of modulating cell fate in HF generation are required. Therefore, we used a 3D co-culture model, where distinct cell types were cultured within the same confined environment to simulate the microenvironment in vivo, in order to maintain their proliferation and HF-formation potency. [39][40][41] In the present study, SKPs were cocultured with ASCs and epidermal stem cells (Epi-SCs) so as to imitate the surrounding niche of SKPs in the dermal layer of the skin and identify the crucial factors and pathways in the co-culture system. Furthermore, this model allowed the exploration of critical targets for preserving dermal stem cell-proliferative capacity.

| Isolation and culture of SKPs and epidermal keratinocytes from skin
All experiments involving live rodents conformed to appropriate governmental and institutional regulations and were performed according to the guidelines of the Animal Ethics Committee of Sun Conclusions: By exploiting a 3D co-culture model, we determined that AREG promoted SKP stemness by enhancing both proliferation and hair-inducing capacity through the PI3K and MAPK pathways. These findings suggest AREG therapy as a potentially promising approach for treating alopecia.
Yat-sen University (approval no. SYSU-YXYSZ-20210332). C57BL/6 and BALB/c nu/nu mice were obtained from the Laboratory Animal Center of Sun Yat-sen University (Guangzhou, China). Dorsal skin from CO 2 -asphyxiated postnatal day 2 (P2) juvenile C57BL/6 mice were dissected, washed with phosphate-buffered saline (PBS; Biofil Chemicals and Pharmaceuticals Ltd.) and cut into pieces (2-3 mm 2 ) that were digested with 0.35% dispase II (Sigma-Aldrich) at 4°C overnight on a rotator. After washing twice with PBS, one pair of forceps was used to hold down a corner of the dermis, and the other was used to gently lift the epidermis away from the dermal sheet.
Sheared as finely as possible, Epi-SCs were released into a 0.035% collagenase I (Sigma-Aldrich) solution, and dermal cells from the dermal layer were digested in a 0.35% collagenase I solution at 37°C for

| Isolation of ASCs
The CO 2 -asphyxiated 9-week-old C57BL/6 mice were positioned in dorsal recumbency. Areas to be incised for harvesting adipose tissue need to be depilated entirely to avoid hair contamination. Adipose tissue from the subcutaneous and inguinal tissues was meticulously dissected and harvested. The linea alba was then longitudinally incised down to the peritoneal cavity in order to expose numerous locations for harvesting adipose tissue. The harvesting procedure was completed within 20 minutes of animal sacrifice.
The harvested adipose tissue was placed in a TC-treated petri dish and finely minced into small pieces with a pair of sterile surgical scissors. The tissue was then rinsed with PBS containing 2% penicillin/streptomycin (Gibco) solution, and the sample was covered with 0.1% collagenase I solution and incubated at 37°C at 250 rpm for 1 hour. DMEM containing 10% FBS was then added to inactivate collagenase I, followed by centrifugation at 2000 rpm for 5 minutes to obtain a high-density pellet (constituting the SVF, separated from the remaining fat lobules and oil). The pelleted SVF was then resuspended in red blood cell lysis buffer for 5 minutes at RT to lyse contaminating red blood cells, collected by centrifugation and filtered through a 100μm nylon mesh to remove cell debris. After washing with PBS, the SVF was again passed through a 70μm cell strainer (Biofil) and centrifuged at 2000 rpm for 5 minutes. ASCs were then cultured in low-glucose DMEM (Gibco) containing 15% FBS, 1% Lglutamine (Gibco) and 1% penicillin/streptomycin solution. When the primary cells reached 80% to 90% confluence following changes in medium every 2 days, they were trypsinized and passaged in typical 6-well plates at a 1:3 dilution. P2 to P5 cells were used for in vivo and in vitro assays. 43,44 2.3 | 3D co-culture system P3 SKPs were cultured alone at a density of 5 × 10 4 cells/well in the 0.001%PF127-precoated middle compartment of the 6-well transwell co-culture dishes in the control group. For the dual-cell 3D co-culture system, SKPs (P3) were added to the 0.001%PF127precoated middle compartment of the culture dish, whereas ASCs Regarding the ternary-cell 3D co-culture system, Epi-SCs, SKPs (P3) and ASCs were separately cultured in the upper, middle and lower compartments of the transwell culture dish, respectively. Two inserts separated the well into three compartments with permeable membranes (the upper insert: 1μm pore size, 12-well format, polyethylene terephthalate [PET], cat no. 353103; and the lower insert: 1μm pore size, 6-well format, PET, cat no. 353102; Corning Inc) on the bottom and uniquely stacked into each other. This plate was inserted in the incubator on a shaker, and the 1μm-diameter pores in the porous membrane allowed molecules to diffuse between the cell layers without permitting cell-cell contact. The cells were cocultured for 6 days, with the medium changed on day 3.

| Alkaline phosphatase (AP) colourdevelopment experiment
Skin-derived precursors were allowed to attach the bottom by and induced pluripotent stem cells. 45,46,48 Therefore, AP staining is usually used to detect the reprogramming efficiency, the subset of undifferentiated pluripotent stem cells with extensive and likely unrestricted self-renewal potential in detailed. 49,50 With the differentiation of cells, AP activity rapidly decreased. 51 SKP stemness can also be detected with this approach, since they are multipotent stem cells as well. 52,53

| Skin embedding and haematoxylin and eosin (H&E) staining
The skin was collected, fixed with 4% PFA for 24 hours and transferred to 70% ethanol. After fixation, samples were embedded in paraffin and sectioned (10μm thickness), with the sections stained using an H&E staining kit (Abcam) according to standard procedures.

| Immunofluorescence (IF) staining
The skin or SKPs were fixed and processed for IF staining, as described previously, 54 to determine levels of AREG, the proliferation marker protein Ki-67, keratin-14 (K14) and phosphorylated EGF receptor (p-EGFR). The samples were photographed under an inverted laser scanning confocal microscope (LSM880 with Airyscan; Carl Zeiss).

| Western blot analysis
Briefly, cells were lysed with radioimmunoprecipitation assay lysis buffer (Beyotime), and cell lysates were subjected to 8% sodium and AREG (Table S3) were detected using horseradish peroxidaseconjugated anti-rabbit or anti-mouse secondary antibodies (Table   S4) with an enhanced chemiluminescence reagent (Advansta).

| Transcriptome RNA sequencing (RNAseq) analysis
Purified total RNA from SKPs was used for RNA-seq library preparation. For mRNA library construction, purified mRNA was fragmented into small pieces with fragment buffer at an appropriate temperature, followed by generation of first-strand cDNA using random hexamer-primed reverse transcription and second-strand cDNA synthesis. To end the repair, A-tailing mix and RNA index adapters were added, and the acquired cDNA fragments were amplified by polymerase chain reaction (PCR), purified using Ampure XP beads (Beckman Coulter), dissolved in elution buffer and validated using a 2100 bioanalyzer (Agilent Technologies) for quality control. The double-stranded PCR products were heated and circularized using the splint oligo sequence to obtain the final library. Single-strand circular DNA was formatted as the final library and amplified with phi29 DNA polymerase to create a DNA nanoball, with ≥300 copies of one molecule. The qualified products were then loaded into the patterned nanoarrays and processed for 50-bp paired-end sequencing on a BGISEQ-500 sequencer (BGI-Shenzhen). After filtering with SOAPnuke (v1.5.2), 55 clean reads were obtained and stored in FASTQ format and then aligned to the mouse genome using HISAT2

| Telogen-to-anagen transition assay
Mice were randomly divided into eight groups: control, AREG

| Statistical analysis
All statistical analyses were performed using GraphPad Prism software (v.7.0; GraphPad Software) and Origin Pro-2018 software (OriginLab). Statistical significance between two groups was measured using an unpaired t test. One-way analysis of variance (ANOVA) was used to compare three or more groups. All data are expressed as the mean ±standard error of the mean (SEM; n ≥ 3), with a P < .05 considered significant.

| Epi-SCs and ASCs promote SKP stemness in vitro and in vivo
To determine whether the co-culture system promotes SKP stemness, we divided the SKPs (P3) into four groups: the SKP group, the SKP and ASC group (the SA group); the SKP and epidermal stem cells (Epi-SCs) group (the SE group); and the group with SKPs, ASCs and Epi-SCs (the SAE group). We then determined cell numbers, levels of the proliferation marker Ki-67 and AP activity. As expected, we observed increased cell numbers ( Figure 1A), proliferation ratio ( Figure 1B) and AP activity ( Figure 1C) in the SAE group relative to the SA and SE groups, with SKPs cultured alone showing the lowest indices. In the meantime, we exploited the AP Assay Kit for the absolute quantification of their enzyme activity, with results depicted on the right-hand-side picture ( Figure 1C). Additionally, the results of the HF-reconstitution experiment in vivo confirmed that co-cultured SKPs increased the number of regenerated HFs in vivo ( Figure 1D), and H&E staining ( Figure 1E) showed intact structures without abnormalities and consistent hair-morphogenic trends with previous results. These results indicated that co-culturing with Epi-SCs and ASCs promoted SKP stemness both in vitro and in vivo.

| The co-culture system activates the ErbB pathway of SKPs
To understand the contribution of each cell type to the SKPs in the co-culture system, we performed transcriptome profiling by RNAseq on SKPs from different groups. Figure 2A  We then investigated AREG-expression patterns using Western blot and IF. Co-culturing with Epi-SCs or ASCs increased AREG levels in SKPs (P3) relative to levels in control cells, with SKPs in the SAE group showing the highest AREG expression ( Figure 2F). To further demonstrate the role of AREG in HF development, we assessed AREG-expression patterns, revealing that levels were higher during the anagen stage than during the telogen stage ( Figure 2G,H, Figure S1). Among the ErbB family of receptors, AREG has a high binding affinity for EGFR. To determine whether receptor activation by AREG occurred in the anagen phase, we investigated levels of p-EGFR, with the result exhibiting consistent changes in levels along with those of AREG, as expected ( Figure 2I, Figure S2). These results implied that the co-culture system activated the ErbB pathway and upregulated AREG expression in SKPs, leading to hair morphogenesis in the anagen phase by activating EGFR.

| AREG promotes SKP stemness in vitro and in vivo
To validate the role of AREG in enhancing SKP stemness, we divided

| AREG activates the PI3K and MAPK pathways in vitro
To determine which signalling pathway downstream of EGFR is activated, we performed global transcriptional profiling, followed by next-generation RNA-seq analysis of SKPs treated with AREG (80 ng/ mL for 3 days) or cultured alone. As shown in Figure 4A, AREG-  Table S2. SKPs undergoing AREG treatment or knockdown, co-culturing with both Epi-SCs and ASCs or either one of them, and culturing alone were then lysed to determine changes in phosphorylation levels of EGFR, PI3K, AKT, MEK, MAPK and ELK between the two pathways, revealing that phosphorylation levels matched the observed enhancement of SKP stemness from AREG or co-culture treatment ( Figure 4D-G).
These results indicated that AREG played an essential role in the coculture system through the PI3K and MAPK pathways in vitro.

| AREG induces the telogen-to-anagen transition in vivo through the PI3K and MAPK pathways
To evaluate the hair-promoting effect of AREG in vivo, multipoint subcutaneous injection of AREG (0.5 mg/kg body weight, every However, after applying one or both of the inhibitors, HF entry to the anagen stage was delayed, which could not be reversed by AREG. All in vivo outcomes were consistent with the in vitro results ( Figure 5). These results suggested that AREG induced stemness in vitro and the telogen-to-anagen transition in vivo through the PI3K and MAPK pathways.

| D ISCUSS I ON
This study used 3D co-culture models to mimic the in vivo microenvironment and concluded that AREG plays an essential role in enhancing the stemness of SKPs through the PI3K and MAPK pathways in vivo and in vitro. These findings suggest AREG as a promising therapeutic strategy for alopecia.
Based on the application of the co-culture system to remedy the shortcomings of monoculture fermentation and increase the biosynthetic efficiency of natural products, 62 we successfully generated highly proliferative spherical SKPs by establishing a ternary-cell 3D co-culture system. In this system, Epi-SCs, SKPs and ASCs are separately plated in three-layer chambers mimicking skin epidermis, dermis and hypodermis in order to facilitate the diffusion of small molecules between inserts and divide cells into their designated tiers without physically contacting with one another. The 2D model is thus not able to simulate such a normal anatomical position. Upon addition, the 2D culture settings hinder us from isolating the SKPs from other cells, posing the subsequent difficulties for the separate analysis for SKPs.
Moreover, the SKPs are isolated via an adherence-separation method, which keeps them separate from Epi-SCs and fibroblasts. 63 In the first 8 hours after separation, Epi-SCs readily adhered to the bottom of the well 52 and did not generate spherical colonies under SKPs first culture conditions described in the methods section.
Secondary clonality of the spheres expressing SKPs markers was subsequently formed from dissociated primary spheres, acquisition of which was confirmed according to a growing body of evidence from other 3D colony-forming systems (eg, methylcellulose or matrigel).
Following this period, premature adherence has typically resulted in a reduced SKPs yield, owing to the fact that the SKPs in the adherent sphere will immediately differentiate into dermal fibroblasts and gradually lose their stemness. 52 Therefore, some studies utilized  76 One remarkable determinant of specific triggers of corresponding signals in the EGFR system is its differential spatiotemporal expression in response to a given stimulus, which is observed in various physiological phenomena related to AREG. For example, AREG is the predominant EGFR ligand upregulated in skin mimicking cutaneous injury models, where heparin-binding EGF upregulation was followed by that of AREG in promoting wound healing through a potent stimulus of keratinocyte proliferation in skin homeostasis. [77][78][79] Additionally, AREG was identified as the ascendant EGFR ligand upregulated during blastocyst implantation, 80,81 and mammary gland, neuronal, bone and trophoblast development. [82][83][84] Furthermore, AREG does not result in EGFR degradation in contrast to EGF, an effective trigger of EGFR degradation, 85 but rather targets EGFR towards a recycling pathway 86 and promotes EGFR accumulation. 87 These findings F I G U R E 5 Amphiregulin promotes SKP stemness in vitro through the PI3K and MAPK pathways. Evaluation of PI3K (wortmannin, 100 nmol/L) and MAPK (PD98059, 50 μmol/L) inhibition alone or together in the presence of AREG (80 ng/mL) or not in vitro for 3 d. Changes in SKP stemness were evaluated according to (A) changes in cell number, (B) Ki-67 staining, and (C) AP staining. #P < .05 vs AREG treatment; *P < .05 vs control. S, significant difference; NS, not significant F I G U R E 6 Amphiregulin induces the telogen-to-anagen transition in vivo through the PI3K and MAPK pathways. Mice were randomly divided into eight groups, and AREG (0.5 mg/kg body weight, every 48 h), Wortmannin (0.5 mg/kg body weight, every 48 h), and PD98059 (3 mg/kg body weight, every 48 h) were subcutaneously injected into the dorsal skin of 56-d-old C57 mice, whereas controls received PBS. At 10-d post-injection, (A) the dorsal skin was photographed under a stereoscopic microscope to assess changes in telogen-to-anagen transition, and (B) HFs were evaluated by IF for Ki-67 + cells suggest AREG as a promising target for promoting EGFR signalling in skin restoration; however, no study had previously demonstrated the influence of AREG on dermis-derived cells, such as SKPs.
Therefore, the present study determined whether AREG improves SKP stemness and the associated mechanisms. We employed RNA-seq rather than cDNA microarrays, because RNA-seq allows assessment of the presence and quantity of RNA transcripts from whole-genome samples, whereas cDNA microarrays only identify finite allele variants designed into the microarrays. Additionally, RNA-seq shows higher resolution for detecting changes in gene expression relative to cDNA microarrays. 88 Following acquisition of RNA-seq results, GO 89 and KEGG 90 enrichment analyses were performed to identify enriched metabolic pathways and/or biological processes, which subsequently identified involvement of the PI3K and MAPK pathways.
Both the PI3K and MAPK pathways played a critical role in multiple cellular responses elicited by AREG signalling, which agreed with our previous findings reporting involvement of the PI3K pathway in wound-induced HF telogen-to-anagen transition 91 and a vital pathway for epidermal and dermal cell communication, which is indispensable for HF regeneration. 54 Moreover, the PI3K pathway is involved in tissue regeneration, as highlighted by the reported decline in the long-term regeneration capability of hematopoietic stem cells in Akt knockout mice, 92 whereas phosphatase and tensin homolog knockout in Lgr5 + HF stem cells enhanced HF regeneration after wounding. 54 Importantly, the PI3K pathway inhibits senescence and promotes the self-renewal of human SKPs in vitro. 93 Regarding the MAPK pathway, EGFR stimulation provokes MAPK in a multistep process, which results in transcription factor translocation to the nucleus and alters gene expression to promote growth, proliferation and/or differentiation 94 and described by the effect of lysophosphatidic acid on fibroblasts. 95 Moreover, previous studies suggest that the MAPK pathway cooperates with PI3K and Rac1 signalling to induce DNA synthesis. 96,97 These findings, as well as those of the present study, suggest a critical role for AREG in HF regeneration and its potential efficacy as a therapeutic approach for de novo HF regeneration.
This study has limitations. First, we were unable to elucidate the precise mechanism associated with the AREG-mediated activity; therefore, further studies are necessary to determine the transcription factors responsible for the SKP-specific expression profile observed following co-culture with the Epi-SCs and ASCs. Second, there is overwhelming evidence supporting a role for AREG in tumour development, as AREG is upregulated in numerous neoplasms, including head, neck, lung, breast, stomach, liver, colon, prostate, bladder and skin tumours, 73 and capable of self-sufficient growth and survival signals. 72 Moreover, functional studies indicate that AREG can perform as a pro-oncogenic factor, affecting most cancer hallmarks. 98 Therefore, a gap exists between experimental and clinical findings regarding AREG-related roles, and further investigations are required to clarify its efficacy for future clinical applications.
Furthermore, caution is needed and safety should be confirmed when extrapolating experimental results to a clinical setting, including the long-term treatment effects and further in vitro studies in human cells and in vivo experiments in larger animals, which are required to confirm the approaches to inducing SKP stemness in HFs as promising yet underexplored lines of research.

| CON CLUS IONS
We applied a 3D co-culture model and determined that AREG promoted SKP stemness by enhancing both cell proliferation and hairinducing capacity through the PI3K and MAPK pathways, thereby, providing insight into a promising strategy of AREG for de novo hair regeneration in treating alopecia.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest regarding this study.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data included in this study are available upon request to the corresponding author.