AKT3 deficiency in M2 macrophages impairs cutaneous wound healing by disrupting tissue remodeling

AKT signaling and M2 macrophage-guided tissue repair are key factors in cutaneous wound healing. A delay in this process threatens human health worldwide. However, the role of AKT3 in delayed cutaneous wound healing is largely unknown. In this study, histological staining and transcriptomics demonstrated that prolonged tissue remodeling delayed wound healing. This delay was accompanied by defects in AKT3, collagen alpha-1(I) chain (COL1A1), and collagen alpha-1(XI) chain (COL11A1) expression and AKT signaling. The defect in AKT3 expression was M2 macrophage-specific, and decreased AKT3 protein levels were observed in CD68/CD206-positive macrophages from delayed wound tissue. Downregulation of AKT3 in M2 macrophages did not influence cell polarization but impaired collagen organization by inhibiting COL1A1 and COL11A1 expression in human skin fibroblasts (HSFs). Moreover, a co-culture model revealed that the downregulation of AKT3 in the human monocytic cell line (THP-1)-derived M2 macrophages impaired HSF proliferation and migration. Finally, cutaneous wound healing in AKT3-/- mice was much slower than that of AKT3+/+ mice, and F4/80 macrophages from the AKT3-/- mice had an impaired ability to promote wound healing. Thus, the downregulation of AKT3 in M2 macrophages prolonged tissue remodeling and delayed cutaneous wound healing.


INTRODUCTION
Chronic wound healing occurs when a wound cannot restore the anatomical and functional integrity of the skin through a normal, orderly, and timely repair process, resulting in delayed wound healing [1]. The prevalence of various chronic diseases (diabetes) increases each year, leading to a higher incidence of associated chronic wounds. Although chronic refractory wounds are not immediately lifethreatening, the delay in wound healing (e.g., months or years) has serious effects on a patient's recovery from the primary disease and quality of life. It is also a significant burden on a patient's family, both financially and as caregivers [2]. Cutaneous wound healing includes inflammation, tissue regeneration, and tissue remodeling that involves a complex orchestration of resident stem cells, immune cells, cytokines, and the extracellular matrix (ECM) [3,4]. However, imbalances or defects in these processes can perturb the delicate equilibrium of cells and signaling pathways that are necessary for complete tissue repair. These defects can result in chronic wounds and fibrotic scars that impair normal tissue function, leading to organ failure and death [5]. Tissue remodeling was considered to be the final step of cutaneous wound AGING healing and also closely related to well wound closure [6]. ECM components provide a "scaffold" for different cell types involved in tissue remodeling that are essential for the tissue repair process [7].
Complete wound healing and functional restoration of damaged skin in adults remain a significant challenge. Recent studies have demonstrated that the microenvironment, which consists of diverse immune cells, is crucial for human cutaneous wound healing, especially for tissue remodeling [8]. The immune system is an active component of tissue repair and regeneration. At the initiation and terminal stages of wound healing, the injured skin requires the activation of an immune response, which is characterized by the massive recruitment of immune cells [9]. Among the various immune cell types, macrophages, particularly those of the M2 phenotype (i.e., M2 macrophages), play a predominant role in tissue remodeling [10]. It is noteworthy that an alternative macrophage phenotype (M2 macrophage) is educated by the microenvironment at the injury site [11][12][13][14]. Recent studies showed that M2 macrophages encourage constructive tissue remodeling due to their capacity to remodel the ECM and synthesize multiple cytokines and growth factors [5]. A lack of M2 macrophages during tissue remodeling leads to delayed wound healing [5]. Although tissue repair is orchestrated by numerous cell types, macrophages are involved at all stages of the wound repair response and, thus, have emerged as potentially important therapeutic targets [10,15,16]. Although the importance of macrophages in cutaneous wound healing is clear, the specific molecular mechanism underlying the role of M2 macrophages in this process is unknown.
Signal transduction and activation are crucial for mammalian biological processes, such as growth, metabolism, angiogenesis, and wound healing [17]. AKT signaling plays a critical role in wound healing. Disrupted AKT signaling prolongs corneal epithelial wound healing by inhibiting epithelial cell proliferation [18]. The AKT pathway is also important for regulating macrophage survival, migration, and proliferation and orchestrating the responses of macrophages to different metabolic and inflammatory signals [19,20]. AKT, also known as protein kinase B, is a key component of the PI3K/AKT signaling pathway. There are three AKT isoforms (i.e., AKT1, AKT2, and AKT3), which are responsible for different biological processes [21]. There is growing evidence that these isoforms are not redundant and have partially opposing effects [22]. Among the AKT isoforms, AKT3 is responsible for homeostasis and is frequently targeted by microRNA post-transcriptionally to attenuate peripheral nerve injury [23,24]. Genetic ablation of AKT3 in macrophages promotes foam cell formation and atherosclerosis in mice [25]. We hypothesized that AKT3 might affect cutaneous wound healing through AKT signaling and AKT3-mediated M2 macrophage reprogramming.
In the current study, we found that both prolonged tissue remodeling and downregulated AKT3 expression occurred during delayed cutaneous wound healing. M2 macrophages derived from delayed wound tissue lacked AKT3 and were incapable of promoting human skin fibroblast (HSF) proliferation and migration. Genetic ablation of AKT3 in mice delayed cutaneous wound healing, specifically at the tissue remodeling stage.

Extracellular
matrix remodeling and reepithelialization was weaker in delayed wound tissue Extracellular matrix components (e.g., fibronectin, elastin, and collagen) are essential for wound repair [26]. The restoration of tissue integrity is the result of neutrophils, monocytes, macrophages, fibroblasts, endothelial cells, and keratinocytes and the scaffold provided by the ECM [7,27]. We found that tissue from cutaneous wounds with delayed healing had more inflammatory cells compared to tissue from wounds with a normal rate of healing ( Figure 1A-a). Masson and EVG staining demonstrated that delayed wound tissue had significantly less assembled collagenous, muscular, and elastic fibers ( Figure 1A-b). Reepithelialization is another essential process in wound healing [28]. IHC staining with the re-epithelialization marker CK5 revealed that re-epithelialization was not complete in the delayed wound tissue, and PCNA expression on the delayed wound tissue surface was decreased compared to normal wound tissue ( Figure  1B-a, b). There was also an increased number of apoptotic cells in the delayed wound tissue compared to the normal wound tissue ( Figure 1B-c).
To explore the mechanism of delayed wound healing and re-epithelialization, RNA sequencing with clustering analysis (heatmap and volcano) identified a total of 1792 downregulated genes and 1570 upregulated genes (Figure 2A, 2B). Gene ontology (GO) and KEGG pathway analysis showed that the top 20 enrichment functions included ECM organization, interaction, and cell adhesion ( Figure  2C, 2D). These results demonstrated that the assembly of ECM-associated collagenous, muscular, and elastic fibers was weakened, and the reepithelization process was slowed in tissues with delayed wound healing. AGING AKT3, COL1A1, and COL11A1 levels were downregulated in delayed wound tissue To explore the molecular mechanism of impaired tissue remodeling, we evaluated three KEGG pathways associated with tissue remodeling, namely PI3K-AKT signaling, ECM-receptor interaction, and focal adhesion [7,20]. These three KEGG pathways accounted for a total of 35 changed genes ( Figure 3A). Based on the Venn data for the GO analysis, AKT3, COL1A1 (collagen type I alpha 1 chain), and COL11A1 (collagen type XI alpha 1 chain) were significantly enriched ( Figure 3B). These data were consistent with the KEGG pathway analysis. Gene Set Enrichment Analysis (GSEA) was also used to analyze the tissue remodelingassociated gene set ( Figure 3C a-d).
COL1A1 and COL11A1 are components of the ECM and play a pivotal role in the tissue remodeling phase of cutaneous wound healing [29,30]. We evaluated AKT3, COL1A1, and COL11A1 expression in both normal and delayed wound tissue and found that expression of all three genes was decreased in the delayed wound tissue ( Figure 3D, 3E, Supplementary Figure 2A). Activation of AKT signaling was also diminished in the delayed wound tissue, as indicated by weak phosphorylation at S472 of AKT3 ( Figure 3F, Supplementary Figure 2B). These results suggested that AKT3, COL1A1, AGING and COL11A1 were downregulated, and AKT3-related AKT signaling was deficient in delayed wound tissue.

AKT3 deficiency in M2 macrophages caused downregulation of COL1A1 and COL11A1
M2 macrophages are crucial for cutaneous wound healing and healing of wounds in other organs [10,28]. AKT3 deficiency in M2 macrophages is responsible for cholesterol metabolism which was closely corelated to wound repair [25,31]. We evaluated whether the downregulation of AKT3 was M2 macrophage-specific and responsible for the changes observed in COL1A1 and COL11A1 expression in the delayed wound tissue. GSEA revealed that PI3K-AKT signaling and phagosome-related genes (characteristic of macrophages) were negatively enriched in the delayed wound tissue ( Figure 4A). The PI3K-AKT signaling and phagosomeassociated heatmaps revealed that AKT3, which was one of the top 10 changed genes, was decreased in both the PI3K-AKT3 signaling and phagosome categories for the delayed wound tissue ( Figure 4B). AGING Immunofluorescence staining for CD68, CD206, and AKT3 demonstrated that AKT3 was downregulated in CD68/CD206-positive M2 macrophages, and there was less infiltration of this macrophage population into the delayed wound tissue ( Figure 4C). qRT-PCR confirmed that the sorted CD68/CD206-positive cells were M2 macrophages with decreased AKT3 mRNA levels ( Figure 4D). In addition, total AKT3 and phosphorylated AKT3 S472 protein levels were lower in the delayed wound tissue compared to the normal wound tissue ( Figure 4E, Supplementary Figure 2C). Furthermore, decreased COL1A1 and COL11A1 protein levels were observed in the CD68-positive macrophages ( Figure 4F a-b). These results suggested that AKT3 expression was specifically altered in M2 macrophages, leading to reduced infiltration into the wound and decreased COL1A1 and COL11A1 expression.

AKT3 knockdown in M2 macrophages impaired HSF proliferation and migration in ex vivo coculture model
To further investigate the influence of AKT3 deficiency on M2 macrophages, we established a co-culture model of THP-1-derived M2 macrophages and HSFs using a transwell non-contact co-culture system ( Figure 5A). QRT-PCR confirmed the differentiation of THP-1 cells into M2 macrophages (Supplementary Figure 1A). AKT3 knockdown with shAKT3 in M2 macrophages was verified by western blotting. Phosphorylation of AKT3 S472 was also reduced in these M2 macrophages ( Figure 5B).

AGING
In the ex vivo co-cultures, the M2 macrophages derived from the THP-1 cells significantly increased HSF proliferation, which was partially abolished by AKT3 knockdown (Figure 5C, 5E). In contrast, M2 Delayed macrophages failed to promote HSF proliferation compared to M2 Normal macrophages ( Figure 5D, 5F). Cell migration and scratch wound healing are considered to be the effector of wound healing [34]. In this study, we tested the effect of M2 macrophage coculture on the migration of HSFs in vitro. THP-1derived M2 macrophages significantly increased HSF migration, which was partially abrogated by AKT3 knockdown in the M2 macrophages ( Figure 5G). Not surprisingly, M2 Delayed macrophages did not promote HSF migration compared to M2 Normal macrophages ( Figure 5H). Thus, the elimination of AKT3 expression in M2 macrophages impaired the proliferation and migration of co-cultured HSFs.
The presence of the THP-1-derived M2 macrophages dramatically increased COL1A1 and COL11A1 expression in the HSF cells; however, this effect was abolished by AKT3 knockdown in the M2 macrophages ( Figure 5I, Supplementary Figure 2D). The patientderived M2 macrophages had a similar effect on the HSFs. Lower COL1A1 and COL11A1 expression levels were observed in the HSFs exposed to M2 Delayed macrophages, which had decreased AKT3 levels compared to the M2 Normal macrophages ( Figure 5J).

AKT3 knockout impeded cutaneous wound healing in vivo
To study the functional role of Akt3 in cutaneous wound healing in vivo, we used CRISPR/Cas9 technology to genetically ablate the AKT3 gene in C57BL/6 mice ( Figure 6A). AKT3 expression and AKT3 S472 phosphorylation were significantly reduced in the AKT3 -/mice compared to the AKT3 +/+ mice ( Figure 6B). Interestingly, loss of AKT3 did not alter mouse body weight (Supplementary Figure 1B). Therefore, we hypothesized that AKT3 might be involved in the regulation of cutaneous wound healing through its effects on M2 macrophage function. Although M2 macrophages are active in all stages of cutaneous wound healing, the most infiltration by these macrophages into the wound occurs during the tissue remodeling phase. We observed that on day 0 postinjury, the wound lesion area was similar between the AKT3 -/and AKT3 +/+ mice. By day 7 post-injury, the lesion areas of the AKT3 -/mice were significantly larger than those of the AKT3 +/+ mice. Not surprisingly, the lesions of the AKT3 +/+ mice were almost healed by day 14 post-injury. In contrast, the lesions of the AKT3 -/mice were almost the same size as they were on day 7 ( Figure 6C). Histological analysis of the wound tissue showed that the wound structure in the AKT3 +/+ mice demonstrated greater integrity and tightness compared to that of the AKT3 -/mice. In addition, collagenous, muscular, and elastic fibers were more abundant in the wound lesion area of the AKT3 +/+ mice compared to that observed for the AKT3 -/mice ( Figure 6D). AKT3 knockout also affected ECM deposition by inhibiting COL1A1 and COL11A1 expression (Supplementary Figure 1C, 1D(a)). Moreover, AKT3 knockout decreased M2 macrophage infiltration into the cutaneous wound site ( Figure 6E, Supplementary Figure  1D(b)). These results suggested that loss of AKT3 delayed cutaneous wound healing by disrupting the tissue remodeling process.
During the tissue remodeling phase of wound healing, we observed that both TGF-β and IL-10 were expressed on days 7 and 14 post-injury, and the levels of expression were lower in the AKT3 -/mice ( Figure 7A). AKT3 expression levels in these F4/80/CD206-positive M2 macrophages were also verified by western blotting ( Figure 7B). Previous research demonstrated keratinocyte meditated epidermal proliferation was crucial for cutaneous wound healing and was positively regulated by Erk/Akt signaling pathway [33]. To further investigate M2 macrophage to proliferation and migration of mouse JB6 cells, we co-cultured these mouse-derived F4/80/CD206 positive M2 macrophages with the JB6 murine epidermal cell line, results demonstrated that M2 macrophages derived from the wound tissue of AKT3 -/mice lost their ability to promote cell proliferation and migration ( Figure 7C-7E). COL1A1 and COL11A1 expression levels were also decreased in the JB6 cells co-cultured with M2 macrophages from the AKT3 -/mice ( Figure 7F). Overall, our study demonstrated that decreased M2 macrophage infiltration and impaired function were the underlying causes of delayed cutaneous wound healing. AKT3 deficiency in M2 macrophages appeared to be responsible for this abnormal M2 macrophage infiltration and function.

DISCUSSION
Restoration of cutaneous integrity after an injury is of vital importance. Intact skin provides the first barrier against invading microbes and pathogens. Loss of the integrity of large portions of the skin as a result of injury or illness can lead to major disabilities and even death [34]. Cutaneous wound healing consists of three main phases (i.e., inflammation, regeneration, and remodeling), which are tightly linked [35]. Delayed cutaneous wound healing is frequently encountered with the large wound lesions of patients with diabetes, vascular disease, or dermatosis [36,37]. The present AGING study consisted of male patients with injury to a large portion of their legs. We observed the absence of skin integrity in the delayed wound tissue that was accompanied by reduced numbers of collagenous, muscular, and elastic fibers. It is well known that reepithelialization of a wound plays a crucial role in maintaining cutaneous integrity [38,39]. We observed slowed re-epithelialization in the delayed wound tissue. We also noticed that this wound tissue was incapable of proliferating compared to normal wound tissue. TUNEL staining suggested that there was excessive inflammation in the delayed wound tissue, which is consistent with data demonstrating that a chronic wound inflammatory response can contribute to delayed wound healing [40]. Our current results suggested that a disruption of the tissue remodeling phase may be the main cause of the delayed cutaneous wound healing.

AGING
We used next-generation sequencing to further investigate the molecular mechanisms involved in the impaired tissue remodeling observed with delayed cutaneous wound healing. We found that the ECM organization, ECM-receptor interactions, and cell adhesion were enriched. In particular, transcriptome analysis demonstrated that the expression of AKT3, COL1A1, and COL11A1 in delayed wound tissue was downregulated in three tissue remodeling-related processes, including PI3K-AKT signaling, focal adhesion, and ECM-receptor interaction. These changes were confirmed by IHC and western blotting. Based on these results, we asked whether the changes observed in AKT3, COL1A1, and COL11A1 expression were truly linked or just coincidental.
M2 macrophages are indispensable for cutaneous wound healing, especially during the tissue remodeling phase [10]. A deficiency in AKT3-related PI3K-AKT signaling can suppress the polarization of these macrophages [41]. In the present study, GSEA revealed that PI3K-AKT signaling and phagosome-related genes were negatively enriched in the delayed wound tissue, with AKT3 being downregulated in both gene sets. We observed reduced CD68/CD206-positive M2 macrophage infiltration in the delayed wound tissue, which was accompanied by decreased AKT3 expression in the M2 macrophage population and concomitant decreases in COL1A1 and COL11A1 expression in this tissue. Studies have shown that HSFs are the main sources of collagen in the skin [42]. Our current findings suggest that an AKT3 deficiency leads to the inability of M2 macrophages to induce COL1A1 and COL11A1 expression in cutaneous wounds (i.e., in HSFs). Indeed, AKT3 knockdown abolished M2 macrophage-induced increases in HSF proliferation and migration and COL1A1 and COL11A1 expression in our co-culture system.

AGING
In the present study, we failed to connect AKT3 directly to the expression of COL1A1 and COL11A1 in HSFs. Understanding the mechanism underlying this phenomenon requires further investigation. AKT3 is a vital component of the PI3K-AKT signaling pathway, which is crucial for androgen receptor regulation [43,44]. However, the androgen receptor is negatively correlated with cutaneous and prostatic wound healing [45,46]. Currently, we cannot exclude the influence of androgen on cutaneous wound healing in our study because the tissue samples that we analyzed were mainly obtained from middle-aged men. In addition, we need to analyze more wound tissue samples along with clinical data to clarify the relationship between low AKT3 levels and delayed cutaneous wound healing.
AKTs are a family of protein kinase B involved in cellular biology. Although they share some properties, it is clear that they each have distinct functions. AKT1 can enhance apoptosis and cell growth in mice [47], whereas AKT2 deficient mice have a diabetic phenotype and are resistant to insulin therapy [48]. In contrast, loss of AKT3 in mice induces foam cell formation and atherosclerosis [25]. Currently, very little is known about the role of AKT3 in delayed cutaneous wound healing. In the present study, we uncovered the potential of AKT3 deficiency to impair M2 macrophage infiltration and function in delayed cutaneous wound healing. Through the generation of an AKT3 knockout mouse model, we have identified a potential target for accelerating cutaneous wound healing in the absence of significant morbidity or death. Whether AKT3 plays a critical role in other M2 macrophage-associated biological processes is currently unknown. Future studies will evaluate the potential contribution of other AKT isoforms (AKT1 and AKT2) to delayed cutaneous wound healing. The classical and extensively activated PI3K/AKT/mTOR signaling pathway should also be included in the future research to clarify its function in cutaneous wound healing. Our transcriptome data suggest that these isoforms are unlikely to be involved in this process. Extensive studies have determined that AKT3 plays a role in the activation of the DNA repair pathway [49], suggesting that the involvement of AKT3 in tissue injury repair is potentially wide-ranging and not limited to cutaneous wounds.

Patients and cell culture
The present study was approved by the ethics committee of Shanghai General Hospital. Crural wound tissue samples (n=13) were obtained from patients of Shanghai General Hospital. All samples were collected with informed consent. The inclusion criteria were as follows: male; 40 to 60 years of age; no diabetes; no vascular disease of the lower extremities; no nerve injury. Patients whose wound area showed no apparent reduction after standard treatment for one week were assigned to the delayed wound group. Patients that did not demonstrate delayed wound healing were assigned to the normal wound group. HSFs and JB6 cells were obtained from the Chinese Academy of Sciences Committee on Type Culture Collection Cell Bank (Shanghai, China). The HSFs and JB6 cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM; Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS) and maintained in 5% CO2 at 37°C. The human monocytic cell line THP-1 was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in RPMI 1640 medium with 10% FBS at 37°C and 5% CO2. THP-1 cells were differentiated into M2 macrophages, as previously described [28]. Briefly, THP-1 cells were incubated with 100 nmol PMA for 30 h to induce differentiation into macrophages. After washing three times, the adherent macrophages were treated with IL-4 (20 ng/mL) and IL-13 (20 ng/mL) to induce the CD68+/CD206+ M2 phenotype. For macrophage sorting, wound tissue samples were digested at 37°C. The samples were incubated with fluorescein isothiocyanate (FITC)-CD68 and phycoerythrin (PE)-CD206 or FITC-F4/80 antibodies for 40 min. After filtering through a 70-µm nylon cell strainer, the samples were analyzed using a Sony SH800 flow cytometer (Sony Biotechnology, Japan). Ten thousand events were collected using a forward scatter threshold of 50,000 (5%). Debris was excluded according to the FSC/SSC (forward scatter/side scatter) dot plot.
The establishment of in vitro macrophage co-culture model was followed the method which has been previously described [50]. THP-1 induced M2 macrophages and wound tissue sorted macrophages were seeded into Transwell chambers (Corning, New York, USA). The inserts with macrophages were put into the HSF or JB-6 cells, forming in vitro co-culture models to be applied in the following studies. was performed on tissue sections using antibodies against CK5, PCNA, AKT3, COL1A1, and COL11A1. Immunofluorescence (IF) was carried out with antibodies against CD68, CD206, F4/80, COL1A1, and COL11A1 and observed using fluorescence microscopy. To be specific, formalinfixed, paraffin-embedded tissue samples were cut into 4-μm-thick sections. 0.01 M citrate buffer (pH 6.0) was used for antigen retrieval which performed in a pressure cooker for 3 min. After blocked in 1% BSA for 1.5 h at room temperature, tissue sections were incubated with primary antibodies at 4°C. Then, for IF, the samples were incubated with secondary antibodies conjugated with Alexa fluor for 1 h at room temperature, counterstained with 4′,6diamidino-2-phenylindole dihydrochloride (DAPI) to detect nuclei and visualized using fluorescence microscopy. For IHC, the secondary antibody was diluted to 1:750 for recognizing primary antibodies. The staining for IHC was visualized using the VECTASTAIN ABC peroxidase system and peroxidase substrate DAB kit.
Following TruSeqTM RNA sample preparation Kit from Illumina (San Diego, CA), RNA-seq transcriptome library was prepared using 10μg of total RNA. The raw paired end reads were trimmed and quality controlled by SeqPrep and Sickle with default parameters. Then clean reads were separately aligned to reference genome with orientation mode using TopHat software. Differential expression genes between two different samples was calculated according to the fragments per kilobase of exon per million mapped reads (FRKM) method. RSEM was used to quantify gene abundances. R statistical package software EdgeR was utilized for differential expression analysis. In addition, functionalenrichment analysis including GO and KEGG were performed to identify which DEGs were significantly enriched in GO terms and metabolic pathways at Bonferroni-corrected P-value ≤0.05 compared with the whole-transcriptome background. GO functional enrichment and KEGG pathway analysis were carried out by Goatools and KOBAS.

RNA isolation and quantitative reverse transcription polymerase chain reaction (qRT-PCR)
Total RNA was extracted using Trizol (Invitrogen, Grand Island, NY). First-strand cDNA was synthesized using SuperScript III reverse transcriptase (Invitrogen, USA). mRNA expression was analyzed by qRT-PCR using PowerUp™ SYBR® Green Master Mix (Thermo Scientific, Waltham, MA, USA). Relative mRNA expression levels were quantified using the 2 -ΔΔCt method with β-actin as the internal control. The specific primers used for qRT-PCR are listed in Supplementary Table 1. Cell proliferation was also examined using the Cell-Light™ EdU DNA Cell Proliferation Kit (Ribobio, Guangzhou, China). EdU (5-ethynyl-2'-deoxyuridine) was added to the cells for 2 h. EdU incorporation was analyzed by fluorescence microscopy.

Cell migration assay
HSFs were seeded in the chambers of transwell inserts, and the bottom chambers were filled with normal control medium or conditioned medium. After 24 h, the cells remaining on the top side of the filter membrane were wiped off gently with a cotton swab. The cells that migrated to the lower surface were fixed with 10% buffered formalin, stained with 0.1% crystal violet for 10 min at room temperature, and counted using an inverted microscope (Leica, Wetzlar, Germany).

Animal experiments
AKT3 knockout mice (AKT3 -/-, C57BL/6 background) were generated by Model Organisms (Shanghai, China). Briefly, exons 3 to 5 were deleted from the AKT3-203 transcript using CRISPR/Cas9 technology. The sequence of the guide RNA was 3′-TGAACCAGTCAG ACTGAAGA TGG-5′. The AKT3 knockout vector was injected into ES cells, and then the homologous recombination (HR) of ES cells was selected. The HR ES cells were injected into the blastocysts to generate chimera mice. The chimera mice were crossed with AKT3 wild-type mice (AKT3 +/+ , C57BL/6 background) to obtain the AKT3 -/mouse strain. The cutaneous wound model was established as previously described [51]. Briefly, mice were anesthetized with ketamine/ xylazine (100 mg·kg -1 , i.p.), and the back skin was shaved and disinfected with an ethanol solution. Fullthickness excisional wounds were created using a 10mm biopsy punch and left without cover during the wound healing process.

Statistical analysis
Data are presented as the mean ± SEM. Statistical analysis was performed using SPSS 17.0 software (SPSS Inc., Chicago, IL, USA). Groups were compared using the two-tailed Student's t-test or one way ANOVA. *P < 0.05, **P<0.01, ***P<0.001 was considered statistically significant. Each experiment was performed in triplicate.

CONFLICTS OF INTEREST
The authors declare that they have no conflict of interest.

FUNDING
There was no funding in this study.