Previously, we screened the signature genes of ovine DPCs using single-cell transcriptome sequencing (32). However, the conservation and the validity of these marker genes are still elusive. To understand the conservation of these marker genes, we compared the transcriptome data of DPCs from different species. As a result, 15 marker genes are conserved among sheep, mice, humans and goats, 73 marker genes are conserved among sheep and other two species, while 226 genes are specific for sheep (Figure 1A; Supplementary Table S2). Subsequently, we verified some conserved marker genes through immunofluorescence staining on sheep skin paraffin sections and cultured DPCs. The results showed that CRABP2 was expressed in ovine DP and outer root sheath, IGFBP3 was expressed in DP, matrix and outer root sheath, CRABP1 and IGF1 were expressed in DP and matrix, SOX18 and PDGFRA were specifically expressed in DP, while VIM was specifically expressed in dermal cells as a positive control (33). In addition, CORIN, a known marker gene in mice (34) while not screened in sheep, expressed in dermal sheath rather than DP (Figure 1B), SOX2 and CD133, the well-known marker genes in mice (7), were not detected in sheep. Besides, we further verified the expression of these markers in cultured DPCs through immunofluorescence. The results showed that CRABP1, CRABP2, PDGFRA and α-SMA were widely expressed in cultured DPCs, in which α-SMA is a positive control as a certified marker (35). IGFBP3 was highly expressed in agglutinative cells, which was similar to VERSICAN (Figure 1C). It is worth noting that PDGFRA was expressed in the cell membrane, which could be used as the cell surface marker. These results together implied that SOX18 and PDGFRA were novel marker genes of ovine DPCs and IGFBP3 was a novel marker gene for aggregated DPCs.

Cell aggregation is a critical characteristic of DPCs, which is typically lost after a period of culture ex vivo in humans and mice (27). In this study, we detected the ability of cell aggregation from early passaged ovine DPCs to multiple passaged DPCs. In contrast, we found ovine DPCs consistently aggregated for at least 35 passages, in which the aggregated cells will eventually form three-dimensional spheroids projecting up from the culture substrate (Figure 2A). For those cells < 30 passages, the DPCs presented obvious aggregation after about 7 days, However, for those cells with more than 30 passages, they took a long time to form aggregation (Figure 2B). Although, the aggregate size had no significant difference between early passaged DPCs and multiple passaged DPCs. These results suggested that ovine DPCs exhibit particularly robust and stable aggregation, despite the ability of cell aggregation being decreased till 30 passages.

Alkaline phosphatase is expressed in DP in vivo and is a marker of follicle-inducing activity (19). Previous study demonstrated that early passaged DPCs have alkaline phosphatase activity, while the alkaline phosphatase activity will lose after a period of culture (36). In contrast, we found ovine DPCs have alkaline phosphatase activity from early passaged ovine DPCs to multiple passaged DPCs (Figure 2C). It should be noted that alkaline phosphatase staining was observed in most ovine DPCs composing aggregates, while was observed in a small portion of DPCs before confluence (Figure 2D). These results suggested that the process of cell agglutination may affect alkaline phosphatase activity.

In addition, the cell viability results showed that there was no significant difference between P3 and P20 of DPCs, while the cell viability significantly decreased in P35 compared with P3 or P20 (P < 0.01) (Figure 2E). This might be a potential reason for the declined cell aggregation ability.

Subsequently, we detected the expression changes of marker genes from early passaged ovine DPCs to multiple passaged DPCs using qPCR. As a result, the expression of SOX18, HHIP, WIF1, and FGFR1 have no significant differences between early passaged ovine DPCs and multiple passaged DPCs. Of note, the expression of CRABP2, SFRP2, and VCAN were dramatically decreased in multiple passaged DPCs compared with early passaged ovine DPCs, while IGFBP3, FGF7, and PDGFRA were dramatically increased in multiple passaged DPCs compared with early passaged ovine DPCs (Figure 3A). Furtherly, immunofluorescence results of CRABP2, PDGFRA, and SOX18 confirmed the above results (Figure 3B).

Meanwhile, considering the important role of the Wnt signaling in hair follicle development and maintaining DPC hair-inducing activity (28), we further quantitated the mRNA expression level of CTNNB1, the transcriptional coactivator of Wnt signaling, using qPCR. The result showed that CTNNB1 was downregulated in multiple passaged DPCs compared with early passaged ovine DPCs (Figure 4A). Additionally, with immunofluorescence, we found the cells positive for nuclear β-catenin decreased in multiple passaged DPCs compared with early passaged ovine DPCs, which suggested that Wnt signaling decreased after a period of culture ex vivo (Figure 4B).

As Wnt signaling decreased after a period of culture ex vivo accompanied by decreased cell aggregation ability, we speculated Wnt signaling plays an important role in maintaining the hair-inducing ability of ovine DPCs. To verify it, we detected the influence of Wnt signaling on the intrinsic properties of DPCs using Wnt/β-catenin agonist and antagonist. The Wnt/β-catenin agonist and antagonist worked well as verified by β-catenin immunofluorescence (Supplementary Figure S1). As a result, ovine DPCs lost the ability of cell aggregation when treated with Wnt/β-catenin antagonist ICG001 and IWP2 (Figure 5A). Meanwhile, the cells positive for alkaline phosphatase staining declined when treated with Wnt/β-catenin antagonist IWP2, while increased when treated with Wnt/β-catenin agonist SKL2001 (Figure 5B). Besides, gene expression analysis of some DP signature genes was conducted. The mRNA level VCAN and CRABP2 were decreased in the DPCs treated with Wnt/β-catenin antagonist ICG001 compared with the control group. The mRNA level of SOX18 had no significant difference between DPCs treated with Wnt/β-catenin antagonist ICG001 and the control group. It is worth noting that FGF7, an intercellular signaling factor known to act on follicular keratinocytes (37), was increased in the DPCs treated with Wnt/β-catenin antagonist. IGFBP3 transcripts, which encode a secreted inhibitor of IGF signaling (38), were increased in the DPCs treated with Wnt/β-catenin antagonist (Figure 5C). These results suggested that Wnt/β-catenin plays an important role in maintaining the intrinsic properties of DPCs through regulating FGF and IGF signaling.

Skin tissues of Hu sheep were obtained from Suzhou sheep breeding farm (Suzhou, Jiangsu, China) for isolating DPCs and making tissue sections. Briefly, three healthy lambs (3 days old, 1 male and 2 female) were anesthetic through hypodermic injection of 2% lidocaine hydrochloride (10 mg), then skin samples ~1 cm² and 2 mm deep were harvested from the body side of lambs. The skin samples were divided into two parts and stored in Dulbecco's modified essential medium (DMEM)/F12 medium (Thermo Fisher Scientific, MA, USA) with penicillin-streptomycin (Thermo Fisher Scientific, MA, USA) or 4% paraformaldehyde (Solarbio, Beijing, China) for subsequent analysis. The wound was sutured. All the experimental procedures and sheep used in this study received prior approval from the Experimental Animal Management Committee of Yangzhou University (No: NSFC2020-NFY-1).

Previously, we had conducted single-cell transcriptome sequencing on Hu sheep lambskin from curly wool lambs and straight wool lambs (19). Based on other related published literature in humans, mice and cashmere goats (32, 33, 50), the transcriptome data of DPCs from different species were compared to analyze the conservation of marker genes in DPCs. Conserved marker genes among different species and sheep-specific marker genes were identified. The results were shown through a Venn diagram (https://bioinfogp.cnb.csic.es/tools/venny/index.html).

Immunofluorescence staining on the skin paraffin section and cultured DPCs were performed as previously described (32, 51). Briefly, skin samples fixed with paraformaldehyde, as mentioned in Section 4.1, were dehydrated in gradient ethanol series. The samples were then rinsed with xylene for 30 min and further embedded in paraffin. Embedded samples were serially cut into longitudinal sections at 5 μm thickness using a microtome (Leica RM2255, Nussloch, Germany). Before staining, slides were deparaffinized in xylene for 30 min and then treated in 0.01 M sodium citrate buffer at 96°C (10 min) for antigen recovery. Then the slides were blocked using PBS supplemented with 10% goat serum and 3% bovine serum albumin (Merck KGaA, Darmstadt, Germany) at room temperature for 40 min. Ovine DPCs grown on coverslips were fixed with 4% paraformaldehyde for 15 min. They were then permeabilized in 0.5% Triton X-100 PBS solution for 15 min, and incubated in block solution (PBS with 0.5% triton X-100 and 10% goat serum) for >30 min at room temperature. Primary antibodies against target proteins were then incubated with the skin paraffin section or cultured DPCs samples at 4°C overnight. Fluorescently-labeled secondary antibodies were used to specifically bind to primary antibodies at 37°C for 30 min. The primary antibodies and secondary antibodies used in this study are listed in Table 1. Hoechst33342 (Beyotime Biotechnology, Shanghai, China) was used for nuclei staining and the slides were finally mounted with Vecatshield mounting media (Vector, USA). Each primary antibody was conducted on three biological replicates. Fluorescent pictures were taken under a Fluorescence Inversion Microscope System (Nikon, Tokyo, Japan).

Primary DPCs were isolated via micro-dissecting dermal papillae and explanting them in a culture medium as previously described (32). The DPCs were cultured in adherent culture dishes with DMEM/F12 media supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific, MA, USA), penicillin (100 U/ml), and streptomycin (100 mg/ml) at 37°C with 5% CO₂. DPCs migrated from original and compact DP at 3–4 days after seeding, which were named P0. Cells were passaged every 2 days after P1. DPCs are well known to aggregate in culture, which is likely to relate to hair-inducing activity (3). The DPCs with agglutinate capability will eventually form three-dimensional spheroids projecting up from the culture substrate. We detected DPC aggregation for different passaged DPCs. The images were acquired by Fluorescence Inversion Microscope System (Nikon, Tokyo, Japan).

Before staining, the DPCs were fixed with 4% formaldehyde for 30 min. The alkaline phosphatase activity of DPCs was detected by using BCIP/NBT Alkaline Phosphatase Color Development Kit (Beyotime, Shanghai, China) following the manufacturer's instructions. Each group set 3 replicates. Finally, the images were acquired by Fluorescence Inversion Microscope System (Nikon, Tokyo, Japan).

Total RNA was extracted from DPCs using Trizol reagent (Thermo Fisher Scientific, MA, USA) following the manufacturer's instructions. The first-strand cDNA was obtained using a PrimeScript™ RT reagent Kit with gDNA Eraser (TAKARA, Beijing, China), and then was subjected to quantification of the mRNA with GAPDH as an endogenous control on the Bio-Rad CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad, CA, USA). qRT-PCR reaction system and amplification procedure were performed as previously described (51). Each group included at least 3 samples, and all reactions were performed in triplicate for each sample. The primers and annealing temperatures for genes were listed in Supplementary Table S1.

Wnt/β-catenin agonist SKL2001 (MedChemExpress, Shanghai, China) and antagonist IWP2 (Beyotime Biotechnology, Shanghai, China) as well as ICG001 (Beyotime Biotechnology, Shanghai, China) were used to study the effect of Wnt/β-catenin on DPCs. SKL2001 (52), a small molecule that controls the biological activity of Wnt, was added to DPCs (P4 and P20) at a final concentration of 20 μM based on the preliminary experiment. IWP-2 (53), an inhibitor of Wnt processing and secretion, was added to DPCs (P4 and P20) at a final concentration of 20 μM and ICG-001 (54), an inhibitor of β-catenin/TCF mediated transcription, was added to DPCs (P4 and P20) at a final concentration of 10 μM based on the preliminary experiment.

SPSS 20.0 was used for statistical analysis, the student's t-test was utilized to calculate the statistical significance between the two groups and One-Way ANOVA was used for many groups. Composite data are shown as the mean ± standard error. A P-value of ≤ 0.05 was considered significant.