The Muscarinic Acetylcholine Receptor in Dermal Papilla Cells Regulates Hair Growth

TL;DR

This study shows that the muscarinic acetylcholine receptor in skin cells regulates hair growth through specific signaling pathways.

Contribution

The study identifies the M4 muscarinic receptor as a key regulator of hair growth via Wnt/β-catenin signaling in dermal papilla cells.

Findings

Inhibiting AChE or activating mAChR promotes hair growth in cultured cells and mouse skin.

Activation of Wnt/β-catenin signaling occurs through PI3K/AKT and ERK pathways upon mAChR stimulation.

Treatment with bethanechol increases hair shaft elongation in mouse vibrissae.

Abstract

The role of cholinergic system in hair biology is poorly understood. In M4 muscarinic receptor (mAChR) knockout mice, the hair follicles have a prolonged telogen phase and fail to produce hair shafts. Here, we reported that hair growth was regulated by cholinergic signalling via mAChRs. Dermal papilla cells expressed different cholinergic biomarkers. Inhibiting AChE or activating mAChR in dermal papilla cells, cultured vibrissae and mouse skin epidermis promoted the hair growth. In cultured papilla cells treated with bethanechol, an agonist of mAChR, an activation of Wnt/β-catenin signalling was illustrated by various indicative biomarkers, including phosphorylation of GSK-3β and mRNA expression of various molecules for Wnt/β-catenin signalling. Activation of Wnt/β-catenin signalling was mediated by PI3K/AKT and ERK signalling upon the stimulation of bethanechol. Moreover, the bethanechol-mediated signalling was blocked by specific antagonist of M4 mAChR. In addition, an increase in hair shaft elongation was observed in mouse vibrissae upon the treatment of bethanechol, suggesting the cholinergic role in hair growth.

Figures

Full text

1. Introduction

In non-neuronal cells, the cholinergic system, composing acetylcholine (ACh) synthesizing enzymes, transporters, receptors, and degrading enzymes, exhibits a wide range of biological roles on different organs and pathologies [1,2,3,4]. Several lines of evidence suggest that the dysregulation of cholinergic system in different tissues is leading to diverse diseases, having significant implications for human health. For an example in skin epidermis, the role of cholinergic molecules has been proposed to be involved in different skin functions [5,6]. A synapse-like interplay between melanocyte and keratinocyte, named as “skin synapse”, has been proposed [7,8,9]. The direct involvement of cholinergic molecules is to regulate skin melanogenesis in both melanocyte and keratinocyte. Under the exposure to UV, the stimulated keratinocyte releases ACh acting on acetylcholine receptor (AChR) localized on melanocyte to trigger the synthesis and release of melanin. In parallel, the patients suffering from atopic dermatitis showed over 14-fold higher levels of ACh in epidermis and dermis, as compared to normal levels [10,11].

In mammalian skin, hair follicle is an organ that spans both the epidermal and dermal layers of the skin. The hair follicle, containing over 20 different cell types, is responsible for hair growth via a complex and dynamic process [12]. The hair growth is tightly regulated by the follicle in a cyclic manner [13], consisting of three distinct phases, namely: (i) anagen (the active hair growth phase); (ii) catagen (the regression phase); and (iii) telogen (the resting phase): the phases of hair follicle are characterized by unique morphological and molecular features [14,15]. To achieve a well-balanced cycle between hair growth and regression, a range of growth inhibiting and/or promoting signals are involved in regulating different phases of hair follicle. Among those growth promoting signals, Wnt, fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF), and insulin-like growth factor 1 (IGF-1) have been identified as key players in promoting hair growth [16,17,18,19,20,21]. Activation of Wnt/β-catenin signalling triggers the growth of hair follicles by stimulating different gene expressions at the anagen phase [22,23,24,25]. Dermal papilla cell (DPC), a group of specialized mesenchymal cells clustering at the base of hair follicle, is a crucial player in regulating hair growth [26]. DPC secretes a range of paracrine and autocrine factors, including Wnt, R-spondin, FGF, and Noggin, acting on follicular bulge stem cells to initiate the anagen phase [27,28,29].

The role of cholinergic system in hair biology remains an interesting topic with limited finding, having only a few reports exploring the relationship of the complex process. Several lines of evidence support the notion of ACh playing role in hair biology. Alzheimer’s patients taking oral reversible and competitive acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) inhibitor, rivastigmine, exhibited symptoms of hypertrichosis and hair re-pigmentation [30,31,32]. In DPCs treated with AChE inhibitor, nor-galantamine, a stimulation of anagen’s activating signalling has been identified [33,34,35,36]. In addition, the knockout mice of M4 muscarinic receptor (M4 mAChR) showed a defect in hair growth: the hair follicles had a significantly prolonged telogen phase and it failed to produce pigmented hair shafts [37,38]. These findings provide supporting evidence that the cholinergic system may play roles in the multifaceted processes of hair growth. Here, we hypothesize that the cholinergic signalling in hair follicles can regulate hair growth. The expression profiles of cholinergic molecules in DPCs and hair follicles were determined. In addition, the release of ACh, induced by solar light, from DPC was demonstrated, which thereafter activated the AChRs localized on DPCs; the cholinergic activation triggered the Wnt/β-catenin signalling to stimulate hair growth. These findings shed light on the role of cholinergic system in hair growth.

2. Results

2.1. Characterization of Cholinergic Molecules in DPC

DPC was employed here as the major cell type in testing the induction of hair growth. The cultures showed a rapid growth from day 2 to 6 after plating: the maximal cell number was at day 7 (Figure S1A). DPC cultures contained mainly the enzymatic activity of AChE instead of BChE, as revealed by the effect of different specific inhibitors, i.e., BW284c51 as the AChE inhibitor and iso-OMPA as the BChE inhibitor (Figure S1B). The expressions of cholinergic molecules in cultured DPCs were identified, showing changes during the growth. Specifically, the protein expression of AChE decreased, while BChE increased as the culture time progressed (Figure 1A). This indicated a potential regulatory mechanism governing the expressions of these two enzymes.

The globular tetramer (G_4_) form of AChE is composed of four AChE catalytic subunits. With the help of proline-rich membrane anchor, G_4_ form of AChE attaches to cell membrane and hydrolyses ACh, extracellularly. We suspected that AChE in DPC has similar function in hair follicles. The form of AChE in DPC culture was determined by a sucrose density gradient, showing that the G_4_ AChE was the major form (Figure 1B). Notably, the mRNA expression of BChE was higher than that of AChE by over 30% (Figure S1C). In contrast, the enzymatic activity of AChE was ~ 6 times higher than that of BChE (Figure 1C). The hydrolysing rate of ACh was much faster by AChE than BChE, and therefore, AChE should still play a dominant role in breaking down ACh here. The mRNA expressions of AChE, BChE, and choline acetyltransferase (ChAT) in different cell lines were compared by qPCR. The expression of AChE mRNA of DPC was much lower than that of HaCaT and SH-SY5Y cells (Figure S1D). In addition, the identification of subunits for nAChR and mAChR, as well as other cholinergic molecules in DPC, was done by specific primers in PCR (Figures S2 and S3).

2.2. DPC as a Source of ACh

ChAT is an enzyme known for its pivotal role in the synthesis of ACh. While vesicular ACh transporter (VAChT) is a crucial transporter responsible for sequestration of ACh into secretory organelles, i.e., synaptic vesicle in neurons. To confirm the presence of cholinergic molecules in DPC, immunofluorescent staining was performed. In the presence of Triton X-100, the expressions of ChAT, VAChT and synaptophysin (SYP) in DPCs were recognized by using anti-ChAT, anti-SYP and anti-VAChT antibodies (Figure 1D). Dermal papilla is the regulating centre in hair follicle, which secretes growth factors to stimulate the development of hair follicle. Here, we hypothesized that DPC could be the source of ACh in activating cholinergic system in the hair follicle. To support this hypothesis, the light sensitive protein should be present in DPC. Opsin has been proposed to sense light by the skin surface, causing ACh release from skin epidermal cells [7]. The mRNAs encoding OPN1, OPN2, and OPN3 were identified in DPC, while the expressions of other opsins were not observed. The expression profile of these opsins peaked at one day after culture (Figure 2A). To reveal the possible release of ACh from DPC, KCl was employed to depolarize the cell membrane [39]. Starting from 25 mM KCl, an increase in ACh in the conditioned medium, released by DPC culture, was detected by an increase of ~2 folds (Figure 2B). The treatment with KCl in DPCs resulted in the release of ACh over time, reaching a maximum at 15 min of the challenge (Figure 2C). In addition, the exposure of solar light to DPCs was able to cause the release of ACh in a time-dependent manner, with a maximum at 5 min of the exposure (Figure 2D). EDTA, a Ca^2+^ chelator, was able to block the light-mediated release of ACh, suggesting the involvement of Ca^2+^ in the ACh release. In line with this notion, the amount of ACh in P7 anagen was ~5 times higher than that of the P23 telogen in skin tissues (Figure 2E). Thus, the level of ACh in mouse skin varied during the development of hair cycle. In addition, the pTOPflash plasmid containing TCF-binding sequence for β-catenin TCF complex was used to quantify the activation of Wnt/β-catenin signalling, a key signal during hair growth. The exposure of solar light in pTOPflash-transfected DPC was able to increase the luciferase activity in a time-dependent manner (Figure 2F).

2.3. Cholinergic Signalling in Hair Growth

The hair cycle in C57BL/6 mice is highly synchronized during the first postnatal weeks, which allows reliable selection of specific phases simply based on the animal’s age [14]. Specifically, postnatal days 5 to 12 correspond to anagen VI (the active growth phase with fully developed follicles and ongoing hair shaft elongation), days 17 to 19 correspond to catagen (the regression phase), and days 20 to 23 correspond to telogen (the resting phase) [14]. The immunostaining analysis of cryosections of mouse skin showed the localization of AChE and BChE in dermal papilla, indicated by SOX2 and ALP staining, at anagen phase of hair growth cycle, which thereafter disappeared in catagen and telogen phases (Figure 3A,B and Figure S4). Thus, AChE and BChE might function during the anagen phase. Because of high expressions of these enzymes, the dermal papilla could be the centre of cholinergic signalling in hair follicle. To verify the membrane localization, immunostaining was performed without Triton X-100 to preserve plasma membrane integrity, resulting in cell surface restricted fluorescence. As expected, without the use of Triton, AChE and BChE were recognized and localized on the DPC membrane (Figure 3C).

To support the role of cholinergic signalling in hair growth, the downstream targets of Wnt signalling, i.e., axin-related protein, (AXIN2), β-catenin, insulin-like growth factor-1 (IGF-1), and ALP, were measured in the presence of BW284c51, an AChE inhibitor, for 48 h in cultured DPCs: the mRNA levels of AXIN-2, IGF-1, and ALP were increased significantly in dose-dependent manners (Figure 4A). Moreover, the growth biomarker of DPC, the mRNA encoding ALP, was increased significantly by a maximum of ~2.5 folds. The mRNA level of IGF-1, a hair growth promoting factor, was increased by over 3 folds upon BW284c51 treatment. The recombinant Wnt3a served as a positive control here, as an agonist in activating the downstream genes of Wnt (Figure 4A). On the other hand, the role of Wnt/β-catenin signalling in regulating the cholinergic molecules was determined. Valproic acid (VPA), an inhibitor of GSK-3β, activates Wnt/β-catenin signalling in cultured DPC. The mRNA expression of AChE was decreased by ~80% under the treatment of VPA and BChE mRNA was increased by ~3 folds (Figure 4B). The mRNA expression of ChAT was relatively unchanged. Bt_2_-cAMP was served as a control in regulating AChE, BChE and ChAT expressions.

To determine the role of AChE inhibitor in hair growth, mouse vibrissae ex vivo cultures were used. The isolated vibrissae were treated with different concentrations of BW284c51 for 72 h. As compared to the control, the length of hair, generated from the follicle being treated with BW284c51, was increased significantly in a dose-dependent manner (Figure 4C). The isolated vibrissae treated with 100 mM BW284c51 showed an increase in hair length by ~60%, as compared to the control. The concentration of 5 mM VPA has shown to promote hair growth in DPC in mice and humans by activating the Wnt/β-catenin signalling [40,41]. VPA served as a positive control. The effect of VPA showed an induction of hair length by ~45%, less than that of the AChE inhibitor (Figure 4C).

2.4. The Cholinergic Drugs in Hair Growth

M4 muscarinic acetylcholine receptor (M4 mAChR) was expressed in hair follicle and DPC, as recognized by immunohistochemical staining of its antibody (Figure 5A). M4 mAChR was co-localized with SOX2 (a marker of dermal papilla), indicating that M4 mAChR was expressed predominantly at the dermal papilla. The mRNA expressions of M4 mAChR and ALP were decreased in the first 48 h of culture and increased thereafter (Figure 5B). The mRNA expression of ChAT remained relatively unchanged during the growth period. Bethanechol is a non-specific mAChR agonist. Here, bethanechol was applied in pTOPflash-transfected DPC, the luciferase activity was markedly induced in a dose-dependent manner after the treatment: the maximal induction was at ~5 mM bethanechol (Figure 5C). CHIR-99021 (C91) and VPA are known GSK-3β inhibitor serving as positive controls.

Pirenzepine (a M1 selective antagonist), gallamine (a M2 selective antagonist), 4-DAMP (a M3 selective antagonist), tropicamide (a M4 selective antagonist), PD102807 (a potent M4 selective antagonist), and atropine (an antagonist of muscarinic receptors) were used here to identify the receptor specificity in responding to challenge of bethanechol. In pTOPflash-transfected DPC, the treatment with pirenzepine, gallamine, tropicamide and PD102807 did not induce luciferase activity, only 4-DAMP was able to increase the luciferase activity by ~30% (Figure 5D). In the bethanechol-treated transfected DPC, the luciferase activity was increased by ~50% (Figure 5E). The pre-treatments with tropicamide, PD102807 and atropine were able to inhibit markedly the bethanechol-induced luciferase activity, suggesting the involvement of M4 mAChR.

In addition, bethanechol was applied in cultured DPCs. The mRNA levels of the Wnt-induced downstream genes, i.e., AXIN2, LEF-1, β-catenin, IGF-1, and ALP were revealed by RT-PCR after the treatment. The results indicated that the levels of mRNAs encoding AXIN2, LEF-1, and β-catenin were significantly increased in a dose-dependent manner upon treatment with bethanechol (Figure 5F). In parallel, the growth markers of DPC, e.g., IGF-1 and ALP were increased. In line with the notion, the treatment of bethanechol in cultured hair follicles was able to induce hair growth in a dose-dependent manner: the maximal induction by over 80% was at 1 mM bethanechol (Figure 5G). VPA served as a positive control.

GSK-3β plays a crucial role in regulating Wnt signalling. GSK-3β phosphorylates β-catenin, leading to its destabilization and subsequent degradation, thereby maintaining a low level of β-catenin in the cytosol and nucleus. Treating bethanechol in DPC was able to induce GSK-3β phosphorylation at 5 min of incubation, while the pre-treatment of atropine or tropicamide was able to block the induction of bethanechol (Figure 6A). Additionally, the bethanechol-induced GSK-3β phosphorylation occurred in a transient manner, having a maximal phosphorylation at ~5 min (Figure 6B). In contrast, VPA induced the phosphorylation at a rather substantial manner. In order to illustrate the role of M4 mAChR in Wnt signalling, the siRNA knock out of the receptor was performed. The incorporation of siRNA of M4 mAChR in DPCs was able to reduce the receptor expression to ~60% of the control transfected with negative control siRNA (Figure 6C). In parallel, the protein expression of M4 mAChR in the siRNA transfection was reduced by ~80% (Figure S1E). In the siRNA-transfected cultures, the levels of mRNAs encoding AXIN2, LEF-1, β-catenin, and IGF-1, as induced by applied bethanechol, were significantly decreased in dose-dependent manners upon an increasing dose of siRNA-M4 mAChR (Figure 6C). Moreover, the treatment of bethanechol, BW284c51, and C91 (a positive control) increased the localization of β-catenin in the nucleus by ~50% in the β-catenin immunofluorescence experiment (Figure S5). The results suggested that M4 mAChR could be a regulator of Wnt downstream signalling in DPC for hair growth.

To test the cholinergic system being involved in hair growth, the dorsal back of depilated mice was treated every day with the vehicle (negative control), 0.2 mM BW284c51, 25 mM bethanechol, and 500 mM VPA (a positive control). Here, the treatments of BW284c51, bethanechol, and VPA exhibited an accelerated re-entry of anagen, as determined by hair growth in surface view and H&E cross-section (Figure 7A,B). By histologic analysis, the dermis thickness and the hair weight, and the number of hair follicles were significantly increased under the treatments of BW284c51, bethanechol, and VPA (Figure 7C–E). By immunostaining analysis, the expression level of Ki67, a marker of cell proliferation in DPC, was significantly increased, at least by over 150%, in the hair matrix of follicles treated with BW284c51, bethanechol or VPA (Figure 7F,G).

2.5. Downstream Signalling of M4 mAChR in Hair Growth

The M4 mAChR couples to G_i/o_ protein leading to the inhibition of cAMP production by inhibiting adenylyl cyclase [42]. The treatment of forskolin activates adenylyl cyclase to increase the endogenous cAMP level [43]. The treatment of DPC, transfected with pGloSensor-22F^TM^ plasmid (a live-cell assay for cAMP signalling coupling to G protein) or pCRE-Luc (a construct for cAMP activation), with forskolin led to increased luminescence or luciferase signal, and which was dose-dependently inhibited by bethanechol (Figure 8A,C). The pre-treatment with tropicamide was able to block the inhibition mediated by bethanechol, which further confirmed the involvement of M4 mAChR in DPC (Figure 8B,D).

The interaction between M4 mAChR and Wnt signalling is poorly studied. We hypothesized that M4 mAChR alone can bypass the Frizzled and LRP5/6 receptors to activate the Wnt signalling in DPC. In pTOPflash-transfected DPC, the application of Wnt3a, or bethanechol, induced the promoter activity (Figure 9A). The activation triggered by Wnt3a was fully blocked by DKK-1 (an inhibitor of Wnt signalling binding to the LRP6 co-receptor) but not by tropicamide. As expected, tropicamide application blocked the activation by bethanechol but not by DKK-1. This implied that there was no interference from the two antagonists and/or signalings. The co-treatment of Wnt3a and bethanechol increased the luciferase activity by ~100%, higher than the single drug challenge. With the pre-treatment of DKK-1, or tropicamide, in pTOPflash-transfected DPC, the luciferase activity was inhibited partially in the co-treatment of Wnt3a together with bethanechol (Figure 9A). Similar results were further illustrated in pTOPflash-transfected PC12 cells, a neuronal cell line (Figure S1F). The co-treatment of Wnt3a and bethanechol induced the phosphorylation of GSK-3β by ~3 fold. The pre-treatment of DKK-1, or tropicamide, inhibited partially the phosphorylation of GSK-3β triggered by the co-treatment (Figure 9B), suggesting the two signalling events should be independent. This result was further illustrated in ex vivo cultures of mouse vibrissae. Both Wnt3a and bethanechol stimulated hair growth. The co-treatment of Wnt3a and bethanechol increased the hair shaft length by ~80%. As expected, the pre-treatment of DKK-1 or tropicamide partially inhibited the growth triggered by the co-treatment (Figure 9C).

To investigate the relationship between mAChR and Wnt signalling pathways, different inhibitors, including H89 (PKA inhibitor), U73312 (PLCβ inhibitor), LY294002 (PI3K inhibitor), BIS (PKC inhibitor), and U0126 (ERK/MEK inhibitor), were pre-treated to the cultured DPC, to block the activation triggered by bethanechol. Treatments of LY294002 and U0126 were able to inhibit the effect of bethanechol, significantly (Figure 9D). U73312 and BIS showed minor inhibition. The treatment of LY294002 or U0126 alone showed no effect to DPC (Figure S1G). Given that the PI3K/AKT pathway is well-documented to directly intersect with canonical Wnt signalling through AKT-dependent phosphorylation and functional inactivation of GSK-3β (thereby promoting β-catenin stabilization and transcriptional activity), we have chosen to focus the subsequent mechanistic studies on the PI3K/AKT pathway. Here, the treatment of bethanechol in DPCs induced the phosphorylation of AKT in a time-dependent manner and peaked at 5 min; this effect was fully inhibited by pre-treatment with LY294002 (Figure 9E,F). Moreover, the bethanechol-induced GSK-3β phosphorylation was inhibited by the pre-treatment of LY294002 (Figure 9G). This further verified that AKT could be the mediator between the signalling of the M4 mAChR and Wnt pathways.

3. Discussion

For the first time, a detailed mechanistic action of cholinergic system in regulating hair growth has been proposed. Under the proposed hypothesis, ACh is being released from the solar light-stimulated DPC, and which thereafter activates the M4 mAChR, as well as the Wnt signalling pathway, at the DPC (Figure 10). This stimulatory effect of Wnt signalling by M4 mAChR in promoting hair growth is mediated via the PI3K/AKT and the ERK/MEK signalling pathways. Several lines of evidence support this notion: (i) the expression and localization of cholinergic molecules in DPC; (ii) induction of ACh release by solar light; (iii) activation of M4 mAChR triggering Wnt signalling and hair growth; (iv) blockage of hair growth by mAChR antagonist; and (v) regulation of cholinergic enzyme by Wnt agonist. Our findings support the relationship of hair pigmentation and hypertrichosis in the possible therapy of having intake of cholinesterase inhibitors. For an example in Alzheimer’s patients, 24 out of 62 patients reported hair darkening in the occipital region after using cholinesterase inhibitor for at least 6 months [30]. In another example, a Caucasian 80-year-old male who received AChE/BChE inhibitor, rivastigmine, for one month was diagnosed with acquired localized hypertrichosis on both forearms [31]. These findings highlight the potential regulation of hair growth by cholinergic system. Having this hypothesis, the cholinergic molecules in hair follicles may be a potential therapeutic target for alopecia. In addition, several AChE inhibitors are currently approved by the US Food and Drug Administration for the treatment of Alzheimer’s disease, and in general these drugs are safe to use [44,45]. To fully comprehend the potential of these inhibitors as an alopecia treatment, additional preclinical and clinical studies are required. Our studies raise the possibility that the cholinergic drug can be a new niche to cure alopecia.

Wnt signalling plays an important role in growth and development of hair follicles. Here, we have demonstrated that the activation of Wnt signal in DPC could be triggered by M4 mAChR. The relationship between cholinergic signalling and Wnt signalling has been reported in various cell types, but the detailed mechanism is not elucidated [46,47,48]. Huperzine A, a selective AChE inhibitor, has been shown to inhibit GSK3α/β activity and enhance the level of β-catenin in transgenic mouse model of Alzheimer’s disease, as well as in cultured SH-SY5Y cells [49]. In parallel, the application of Donepezil (an Alzheimer’s drug) in scopolamine-induced mice showed similar inhibition on the activity of GSK3α/β [50]. Nor-galanthamine, an alkaloid and a potent AChE inhibitor, triggered the anagen-activating signalling pathways, including the ERK1/2, AKT, and β-catenin pathway, in cultured DPC [33]. These findings are consistent with our results that the inhibition of AChE in DPC can promote hair growth via activation of AChR and subsequently the Wnt signalling.

Our study demonstrates that the activation of M4 mAChR can trigger the PI3K/Akt pathway, leading to the phosphorylation and inhibition of GSK-3β, which may facilitate Wnt/β-catenin signalling by stabilizing β-catenin. This observation aligns with emerging evidence of crosstalk between diverse signalling pathways and the Wnt pathway through GSK-3 modulation. For instance in the insulin signalling pathway, insulin binds to its receptor tyrosine kinase and activates PI3K, resulting in Akt-mediated GSK-3 inhibition, a process primarily linked to glycogen synthesis but also capable of stabilizing β-catenin in specific contexts [51]. Similarly, neurotrophic factors, such as brain-derived neurotrophic factor, engage TrkB receptors to stimulate PI3K/Akt signalling, phosphorylating and inhibiting GSK-3, potentially influencing β-catenin-dependent transcription in neuronal cells [52]. Furthermore, prostaglandin E2 binding to its receptor activates PI3K/Akt, leading to GSK-3β phosphorylation and subsequent β-catenin pathway activation [53]. These examples highlight a broader paradigm wherein non-Frizzled receptors, including GPCRs and RTKs, converge on GSK-3 regulation as to modulate Wnt signaling. This underscored the significance of our findings on the muscarinic receptor in revealing alternative mechanisms of the Wnt pathway activation.

ACh was first identified in skin at 1962 [54]. The functionalities of ACh in skin have been proposed, and one of them is playing a role in skin pigmentation [7,55]. Keratinocyte and melanocyte form a so called “skin synapse” in the skin epidermis. Keratinocyte is able to release ACh, triggered by light-activated opsin, and that acts on mAChR of melanocyte, as such to regulate the production and release of melanin [8,9]. Similar to the situation in skin, the hair DPC could release ACh, possibly triggered by solar light-activated opsins [56,57]. In contrast, the released ACh is acting as an autocrine to trigger the responses on DPC itself. However, the exact sub-types of opsins in directing this function are not known, but at least the expressions of opsins 1–3 peaked in the first day after DPC culture. In addition, the level of ACh has been shown to be higher in the anagen phase than in the telogen phase of hair follicle, which is consistent with the hypothesis that the opsin is interacting with cholinergic molecules via solar light in regulating hair growth.

Although various mAChRs, including M1 to M5, were identified in the dermal papilla, here, we propose that M4 mAChR could be the primary target to mediate the cholinergic signalling in hair growth. In line to this notion, M3 and M4 mAChR have been reported to involve in keratinocyte migration and hair growth cycling. The M4 mAChR knock-out mice displayed retardation of hair follicle morphogenesis [37] and similarly a decreased keratinocyte migration was observed upon transfection with siRNA of M4 mAChR [58]. However, the reciprocal effect was revealed in keratinocyte transfected with siRNA-M3 mAChR and M3 mAChR KO mice [58]. In line to these reports, an inhibition of M3 mAChR using 4-DAMP was able to stimulate the Wnt signalling in cultured DPCs, as shown here. Conversely, the inhibition of M4 mAChR using tropicamide or M4 mAChR targeted siRNA can inhibit the Wnt signalling, even in the present of bethanechol. Moreover, with the pretreatment of the non-specific antagonist atropine in DPC, Wnt signalling was inhibited even with the induction of bethanechol. Thus, the inhibitory effect of M4 mAChR on Wnt signalling outweighed the stimulatory effect of M3 mAChR in the presence of atropine. This suggests that the role of M4 mAChR may be prioritized over M3 mAChR.

4. Materials and Methods

4.1. Cell Cultures and Chemicals

Immortalized hair follicle DPCs—hTERT (T0501) were obtained from Applied Biological Materials (Richmond, BC, Canada). SH-SY5Y cells (a human neuroblastoma line), human keratinocytes (HaCaT) and rat pheochromocytoma PC12 cells were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). DPC, SH-SY5Y, and HaCaT cells were cultured in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum (FBS) and 1% (v/v) penicillin/streptomycin (stock as 10,000 U and 10,000 mg/mL) in 5% CO_2_ at 37 °C. PC12 cells were cultured in DMEM supplemented with 6% fetal bovine serum (FBS), 6% horse serum, and 1% penicillin/streptomycin (10,000 U/mL and 10,000 μg/mL). All culture reagents were purchased from Thermo Fisher Scientific (Waltham, MA, USA). The siRNA-M4 was custom synthesized by GenePharma (Shanghai, China). The target sequences for the human CHRM4 (NM_000741) gene was 5′-CCT CTA CAC CGT GTA CAT C-3′ as reported previously [58]. The sequence of negative control siRNA was 5′-UUC UCC GAA CGU GUC ACG UTT-3′.

4.2. DNA Transfection and Luciferase Reporter Assay

The plasmid having TCF/LEF-firefly luciferase reporter (pTOPflash) (Cat: 21-170) with two repeats, each containing three copies of the TCF-binding site upstream of thymidine kinase minimal promoter, was purchased from Upstate Biotechnology (Erie, PA, USA). The pGloSensor-22F cAMP plasmid (Cat: E1290) was purchased from Promega (Madison, WI, USA). GloSensor™ luciferase contains cAMP binding domain fused to a circularly permuted form of firefly luciferase. The pCRE-Luc (Cat: 219076) contains CRE promoter sequences with a luciferase reporter, purchased from Stratagene (La Jolla, CA, USA). The siRNA-M4, 5′-CCT CTA CAC CGT GTA CAT C-3′ (corresponding to amino acids 259–279) and negative control siRNA, 5′-UUC UCC GAA CGU GUC ACG UTT-3′, was custom synthesized by GenePharm, as reported previously [58]. The reporter constructs were then transfected using human hair follicle dermal papilla cell (HFDPC) Avalanche^®^ Transfection Reagent (EZ Biosystems, College Park, MD, USA) according to the given instruction. In short, DPC was seeded into 6-well plate incubated overnight in 5% CO_2_ at 37 °C in a humidified environment without the use of penicillin/streptomycin in the culture medium. After 3 h incubation with the transfected mixture, the medium was aspirated and replaced by DMEM supplement with 10% FBS and different drug treatment. For the transfection of pGloSensor-22F cAMP plasmid and siRNA-M4, the transfection mixture was incubated overnight. After 24 h, the cells were collected for the luciferase reporter assay. The luciferase assay was performed using Pierce^TM^ firefly luciferase kit (Thermo Fisher Scientific). The luminescent reaction was quantified in a GloMax^®^ 96 microplate luminometer (Thermo Fisher Scientific) and normalized to the total protein. For the pGloSensor-22F cAMP assay, the pre-incubated medium was replaced by fresh medium containing 500 µM of 3-isobutyl-1-methylxanthine and treated with or without drug for 15 min before the addition of 10 µM forskolin. After 15 min of forskolin treatment, the culture was lysed, followed by the luciferase assay.

4.3. SDS-PAGE and Western Blot Analysis

Cells were lysed in whole cell lysis buffer and shaken for 30 min at 4 °C followed by centrifugation at 12,000× g at 4 °C for 10 min. The supernatants were collected and protein concentrations were determined using Bradford protein assay (Bio-Rad Laboratories, Hercules, CA, USA). The aliquots normalized to 40 μg of protein were applied to 8% sodium dodecyl sulfate (SDS)-polyacrylamide gels and then transferred to nitrocellulose membranes. The membranes were blocked with 5% BSA in Tris-buffered saline with 0.1% Tween-20 for 2 h. After blocking, the membranes were incubated at 4 °C overnight with specific primary antibodies including anti-AChE at 1:1000 (ab253201, Abcam, Cambridge, UK), anti-BChE at 1:1000 (SC-377403, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-α-tubulin at 1:1000 (DM1A, Cell Signalling Technology, Danvers, MA, USA), anti-P-GSK-3β at 1:1000 (#9336, Cell Signalling Technology), anti-GSK-3β at 1:1000 (SC-377213, Santa Cruz Biotechnology), anti-P-AKT at 1:1000 (#9271, Cell Signalling Technology), anti-T-AKT at 1:1000 (#9272, Cell Signalling Technology), anti-M4 mAChR at 1:1000 (ab189432, Abcam), followed by incubation with Horseradish peroxidase secondary antibodies (Sigma-Aldrich, St Louis, MO, USA) at 25 °C for 1 h. The immune-reactive proteins were detected using enhanced chemiluminescence (ECL) Western blot detection kit (Thermo Fisher Scientific). The intensities of the bands were quantified using ChemiDoc Imaging System (Bio-Rad Laboratories). The intensities of protein bands were in the non-saturating range of calibration curves.

4.4. Sucrose Density Gradient Analysis

Different AChE forms were separated by sucrose density gradient analysis according to previous method [59,60]. In brief, a continuous 5–20% sucrose gradient in lysis buffer containing 10 mM HEPES, pH 7.5, 1 mM EDTA, 1 mM EGTA, 0.2% Triton X-100, and 150 mM NaCl was prepared in a 12 mL polyallomer ultracentrifugation tube. Two hundred μL (1 μg/μL) cell lysates mixed with sedimentation markers, including alkaline phosphatase (6.1 S) and β-galactosidase (16 S), were loaded onto the gradients followed by centrifugation as 38,000 rpm in SW41 Ti Rotor (Beckman, Palo Alto, CA, USA) at 4 °C for 16 h. Approximately 48 fractions were collected for determination of AChE activity by Ellman assay. In Ellman assay, 0.1 mM tetra-iso-propylpyrophosphoramide (iso-OMPA; Sigma-Aldrich), 0.625 mM acetylthiocholine (ATCh; Sigma-Aldrich), and 0.5 mM 5,5-dithiobis-2-nitrobenzoic acid (DTNB) in 80 mM Na_2_HPO_4_, pH 7.4, was added to the sample. The mixture was incubated at room temperature for 30 min, and AChE activities were measured at 405 nm using Multiskan™ FC microplate photometer. The enzyme activity was expressed as absorbance units/min/g of protein.

4.5. Real-Time PCR

Total RNA was extracted from cultured DPCs using RNAzol™ Reagent (Sigma-Aldrich) and 3 μg of RNA was reverse transcribed using PrimeScript™ RT reagent kit (Takara, San Jose, CA, USA), according to manufacturer’s instruction. Template cDNAs were subjected to RT-PCR using the following specific primers provided in Figure S6. Real-time PCR was performed in LightCycler 480 (Roche Molecular Biochemical, Mannheim, Germany) using KAPA SYBR FAST qPCR kits in accordance with the manufacturer’s instruction. The 2^−∆∆Ct^ method was used to calculate the relative expression levels.

4.6. Vibrissae Culture

The ex vivo mouse vibrissae were carefully isolated from the upper lip pad of 4-week-old C57BL/6 male mice and maintained the intact follicular unit, including the dermal papilla, hair matrix, inner or outer root sheath, and bulb region, while removing the surrounding dermal connective tissue sheath and adjacent interfollicular skin. The isolated vibrissae were cultured in Williams E medium (Sigma-Aldrich), supplemented with 1% (v/v) penicillin/streptomycin solution, under 5% CO_2_ at 37 °C. In each experiment, follicles were collected from a single mouse and assigned to individual treatment groups. Each individual experiment involved four hair follicles per treatment group. The data presented in Figure 5G were derived from n = 4 independent experiments, while those in Figure 4C and Figure 9C are from n = 3 independent experiments. This resulted in a total of at least 12–16 vibrissae being analyzed per treatment condition across the replicate experiments. Vibrissae were incubated in various conditioned medium for 3 days, and the increase in length of hair was measured from days 0 to 3.

4.7. Animals and In Vivo Hair Growth Test

Six-week-old male C57BL/6 mice were chosen for the experiment and allowed to adapt to their new environment for 1 week. The hairs on the backs of seven-week-old mice, whose hair follicles were in the telogen phase, were depilated by depilatory cream (Veet, Parsippany, NJ, USA). Various drugs were applied topically daily for up to 14 days, with four mice in each group. All reagents used for the hair re-growth test were dissolved in a vehicle composed of 100% ethanol and mixed with equal amount of cream. The skin was harvested and subjected to haematoxylin and eosin (H&E) staining according to the protocol of H&E staining kit (Yuanye Biology, Shanghai, China) or immuno-fluorescent staining. The skin section was examined by Zeiss Celldiscoverer 7 automated microscope (Zeiss, Jena, Germany) and analyzed by ImageJ software (version 1.54).

4.8. Skin Cryosection Preparation

The dorsal skin from C57BL/6 mice were depilated to remove hair and collected after being sacrificed. Skin tissue was fixed with 4% paraformaldehyde for 1 h at room temperature and then perfused with 30% sucrose solution in PBS. Afterward, the skin was embedded in OCT and frozen at −80 °C overnight followed by cutting into 30 μm section with Thermo CryoStar NX 70 Cryostat (Thermo Fisher Scientific). The section was preserved at −20 °C for the following experiments.

4.9. Immuno-Fluorescent Staining

Cultured cells were grown on glass coverslips in 35 mm culture dishes. After PBS wash, cells were fixed with 4% paraformaldehyde for 15 min. Cells were incubated with or without 0.1% Triton X-100 in PBS for 10 min and then blocked by 5% BSA for 1 h. Cultures were stained with primary antibodies for 16 h at 4 °C, followed by Alexa 488/555/647-conjugated secondary antibodies (Sigma-Aldrich) for 2 h. The following antibodies were used: anti-AChE at 1:100 (Santa Cruz), anti-BChE at 1:100 (R&D System, Minneapolis, MN), anti-M4 mAChR at 1:100 (Abcam Ltd.), anti-ChAT at 1:100 (Abcam Ltd.), anti-SYP at 1:100 (Santa Cruz), anti-SOX2 at 1:100 (Abcam Ltd.), anti-Ki67 at 1:100 (Abcam Ltd.), and anti- β-catenin at 1:100 (Santa Cruz). For the immunofluorescence analysis of β-catenin localization in DPC, DPC was treated with or without Bch, BW284c51, C91 or VPA for 4 h before fixation. Samples were mounted with ProLong™ Gold antifade mountant with or without DAPI (Thermo Fisher Scientific). Samples were then examined by Zeiss Celldiscoverer 7 automated microscope (Zeiss).

4.10. Measurement of ACh

The amount of ACh in conditioned medium or skin tissue from wild-type mice was measured using acetylcholine ELISA Kit (Colorimetric) (Abcam Ltd.) and LC–MS/MS. In brief, cells were seeded and cultured in a 60 mm culture dish. After 48 h, the culture medium was replaced by DMEM without FBS supplied with 20 µM BW284c51 and 20 µM iso-OMPA for 30 min. The culture medium was replaced by Ringer’s buffer (7.2 g/L NaCl, 0.37 g/L KCl, 0.17 g/L CaCl_2_, and pH 7.4) with 20 µM BW284c51 and 20 µM iso-OMPA, then cells were exposed under the solar light simulator machine employed with a Xenon arc lamp (Newport Corporation, Irvine, CA, USA) or treated with KCl. The lamp produced a broad-spectrum output covering 200–2500 nm (UV through visible to near-infrared), with the fused silica lens ensuring high transmittance in the UV range. Specifically, the irradiance delivered in our setup was approximately 0.5 mW/m^2^ for UVA, 50 mW/m^2^ for UVB, and 50 mW/m^2^ for UVC. The Ringer’s buffer containing ACh, released by cells, was collected, heated at 55 °C for 5 min to deactivate cholinesterase and concentrated with freeze drying, followed by ACh assay. In brief, the amount of choline in each sample was determined using the probe for choline before and after adding AChE. The amount of ACh was acquired by subtracting the amount of choline before adding AChE from that after adding AChE. For the ACh amount in the skin tissue, the mouse skin tissue was heated at 55 °C for 5 min to deactivate cholinesterase immediately after dissection in lysis buffer containing 10 mM HEPES, pH 7.5, 1 mM EDTA, 1 mM EGTA, 0.2% Triton X-100, and 150 mM NaCl. The samples were then homogenized, vortex and centrifuge to obtain the extract. For the LC-MS/MS analysis, LC-MS/MS analysis was conducted by an Agilent 6410B triple quadrupole mass spectrometer (QQQ-MS/MS) with ESI source (Agilent Technologies, Santa Clara, CA, USA) coupled with Agilent 1290 Infinity Binary Liquid Chromatography Systems. Chromatographic separation was achieved on a PC HILIC (150 mm × 2 mm, 3 µm, Osaka Soda, Osaka, Japan) at 35 °C. For conditioned medium, after freeze drying, the solid obtained was redissolved in 100% acetonitrile before injection into the mass spectrometer. For tissue sample, an equal amount of acetonitrile was added to the sample to precipitate the protein before injecting into the mass spectrometer. The mobile phase consisted of water containing 20 mM ammonium formate and 0.2% formic acid (A) and acetonitrile containing 0.1% formic acid (B) at a flow rate of 0.4 mL/min. The gradient elution was programmed as follows: 0–2 min, 2% A; 2–2.01 min, 2–25% A; 2.01–4 min, 25% A; 4–5 min, 25–50% A; 5–8 min, 50% A; 8–10 min, 50–2% A, and 10–13 min, 2% A.

4.11. Statistical Analysis

Each result is represented as the mean ± SEM, calculated from independent replicates. Comparisons of the means for untreated control cells and treated cells were analysed using Student’s t-test. Significant values were represented as * p < 0.05, ** p < 0.01, *** p < 0.001. Post hoc power analyses were conducted using based on observed effect sizes and sample sizes, confirming adequate power (>80%) for detecting the reported differences in the majority of significant comparisons.