The Zhongwei goat is a famous native Chinese goat breed used for high quality suede. The principal production area of the Zhongwei goat is located in Xiangshan in Zhongwei City, Ningxia Hui Autonomous Region. The wool of a Zhongwei one-month lamb is approximately 7 cm long with highly curved patterns. However, from two to three months after birth, the delicate flower pattern disappears gradually due to a reduction in the bending of the hair strands and the looseness of the wool strands, which remarkably affects the economic value of the coat. However, the molecular mechanism underlying this short-term change in hair shape is still unknown.
Auber suggested that the bending direction of the hair ball structure at the bottom of the wool follicles is opposite to that of the hair structures at the top [1
]. For uniformity and curvature characteristics, wool bending may be a feature that is determined during embryonic development. This is mainly due to the consistency of the depth of fetal hair follicles in the skin and whether the growth direction is uniform among diverse populations that can be attributed to the asymmetric [2
]. Previous studies have revealed that hair bending is caused by an irregular growth and bending of the hair follicles in the epidermis, the asymmetric distribution of proteins producing different biological forces, which squeeze the hair into a variety of shapes [3
The regulatory mechanism underlying the hair curl trait formation is still elusive [4
]. The molecular mechanisms underlying hair follicle progression, growth and hair curl are similar in mammals [8
]. These research results provide a theoretical basis to illuminate the molecular mechanism underlying wool curl formation. In recent years, the study of hair fiber curl formation using mice as a model has accelerated our understanding on the genetic mechanism underlying hair curl traits. Classic signal pathways, such as the Wnt
, and Shh
pathways, have been found to be related to hair bending [9
]. Andl et al. reported that Dicer and miRNAs are expressed in mouse skin tissue, suggesting that miRNAs play major roles in the development of skin and hair follicles [13
]. Previous studies have shown that miRNAs have two characteristics in regulating gene expression: temporal specificity and tissue specificity, which determine the direction of cell differentiation [14
]. Similarly, during the growth of the epidermis and hair follicles, the miRNA expression profile also exhibits these properties, i.e., some miRNAs are only expressed during a specific period or in a specific tissue. For example, the expression level of miR-184 gradually increases during the growth period of sheep hair follicles, and the expression level of miR-205 increases and then decreases from anagen to telogen, with the highest expression in catagen [15
]. Kang et al. screened candidate genes determining wool growth and functional clusters closely related to this process in Chinese Tan sheep at 1 month after birth (curly wool) and 48 months after birth (straight wool) by transcriptase sequencing [16
]. These results shed light on our experimental design. Nissimov et al. [17
] proposed the hypothesis of multiple dermal papillary (MPC) transitions in which hair follicle growth is divided by the base of the hair bulb. It was believed that hair bending was caused by a partial separation of the dermal papilla at the bottom of the hair follicle. This phenomenon produces multiple hair fibers that are wrapped around the stratum corneum with different discrete hair shaft tips as the axis. In addition, this phenomenon is related to the degree of separation of hair papilla cells; the more independent the structure of each subunit is, the greater the difference in the predetermined center, which gives the subunits different growth speeds, resulting in the generation of bending. Therefore, we chose caprine dermal papilla cells as the model to explore miRNAs function in hair follicle development in vitro.
In recent years, there have been several studies related to wool shape and wool quality identification [18
]; however, these studies mainly focused on genetic differences among different populations with divergent morphological wool types. Few studies have concentrated on the molecular mechanism underlying the dynamic changes of wool morphology at different development stages. Here, we identified the expression profile of Zhongwei goats at two different ages (45 and 108 days) by high-throughput sequencing. To construct a network related to hair development, we analyzed and tested potential miRNA/mRNA interactions. We sought to investigate the roles of miR-26a and miR-130a in hair development. The aim of this study was to provide a theoretical basis for the identification of the potential miRNAs involved in the wool growth of Zhongwei goats and to further discover the regulatory mechanism underlying the hair follicle development in animals.
Recent studies have demonstrated that miRNAs are widely involved in the occurrence and periodic growth of hair follicles [22
]. These miRNAs, numerous regulatory factors and signal pathways constitute an extremely complex network regulation system in various types of hair follicle cells. Furthermore, the regulation of hair follicle morphogenesis and periodic development is achieved [26
]. The shape, type and color of hair are determined not just during embryogenesis, but also repeatedly during each hair growth cycle [27
]. To screen the miRNAs involved in the regulation of hair growth and the formation of striation patterns, a high-throughput sequencing of skin tissue from Zhongwei goats at 45 days and 108 days old was performed. We observed 28 differentially expressed miRNAs in the goat skin at the two developmental stages and verified 12 of these miRNAs by RT-qPCR. These miRNAs may be important factors influencing the changes in the hair bending phenotype. Some of them have been found to be related to the regulation of hair follicle development in previous reports. Zhang et al. identified multiple miRNAs in duck skin, including miR-10a and miR-451, which may regulate genes involved in the Wnt/β-catenin
signal pathways [28
]. Gao et al. detected 14 miRNAs participating in the regulation of the development of wool, including miR-143, miR-10a and let-7, among which miR-10a was differentially expressed between the large corrugated and small corrugated skins of Hu sheep [29
]. Wen et al. screened 159 miRNAs expressed in skin and ear tissues, of which 105 were strongly conserved, and the miR-30 family was highly expressed in adult goat and sheep [30
]. In our study, we detected some novel, significantly differentially expressed miRNAs that were not previously reported to be associated with hair bending or hair follicle development. From the RT-qPCR results, it can be seen that the difference between miR-130a and miR-26a is extremely significant compared to other miRNAs. Therefore, we chose miR-130 and miR-26a to further examine their function and effects on the DPC proliferation. The latest study found that miR-130a expression was increased in hypertrophic scar tissue and derived primary fibroblasts, which was accompanied by the upregulation of collagen 1/3 and α-SMA
expression. The miR-130a/CYLD/Akt
pathway may serve as a novel entry point for future skin fibrosis research [31
]. Icli suggested that inhibiting miR-26a increased the mRNA level of its target gene SMAD1
in the ECs at nine days post-wounding in diabetic mice. In addition, high glucose levels reduced the activity of the SMAD1
3′-UTR. Diabetic dermal wounds treated with LNA-anti-miR-26a had an increased expression of ID1
, a downstream modulator of SMAD1
, and a decreased expression of the cell cycle inhibitor p27 [32
]. So far, great progress has been made on understanding the molecular mechanism underlying hair follicle regulation in human and model animals like mice and rats. The highly conserved characteristics of miRNAs are greatly convenient for understanding the relevant molecular mechanisms in animals.
MicroRNAs mainly exert their biological functions through interactions with target genes. In vivo, different gene products coordinate with each other to perform biological functions. An annotation analysis of the pathways for target genes of differentially expressed miRNAs can help further interpret the functions of miRNA. In the present study, pathway annotations indicate that DE miRNAs’ target genes are significantly enriched in the TGF-β/SMAD
], and MAPK
], which are known to be involved in regulating hair follicle development, and miR-130a and miR-26a were among our key DE miRNAs of interest. Given the dramatic expression in the two development stages, miR-130 and miR-26a are selected as candidate miRNAs. They were predicted to have 1211 target genes and 1135 target genes, respectively. To further investigate the regulatory relationship between miRNAs and mRNAs in the hair development process, we integrated a miRNA-seq and transcriptional data from our laboratory. A miRNA target gene regulatory network associated with hair bending was constructed. Notably, the target genes of both miR-130a and miR-26a were enriched in the TGF-β/SMAD
signal pathway. Studies have found that TGF-β
plays an important regulatory role in cell growth and differentiation as well as hair follicle development and formation. SMADs
are known as the special signal factors in the TGF-β
signal pathway. When the corresponding receptor on the surface of the cell membrane binds to TGF-β
are responsible for transmitting the TGF
signal from the receptor to the nucleus [39
]. Previous studies have demonstrated that SMADs
can regulate the development, periodic growth and pigmentation of hair follicles [40
]. SMAD proteins can be roughly divided into three classes: receptor-activated SMADs, co-type SMADs and inhibitory SMADs. The receptor-activated SMADs (r-SMAD) include SMAD1
, and SMAD8
, which can specifically bind to the BMP protein, and SMAD2
, which are the active substrates of the T-R1 receptor. The co-type SMADs (Co-SMAD), SMAD4, act as a co-required protein in the TGF-1
pathway. The inhibitory SMADs (I-SMAD) include SMAD6
, which inhibit SMAD
phosphorylation and prevent signal transduction to the nucleus [42
]. The regulatory networks of these two miRNA assemblies may contribute to the study of the mechanism of hair bending in Zhongwei goats.
Dermal papilla cells are mesenchymal cells that not only regulate the growth and development of hair follicles, but are also considered as a reservoir of pluripotent stem cells [43
]. To determine the roles of miRNAs in regulating hair follicle development, we successfully isolated DPCs in vitro and provided a model to explore the miRNA potential functions. Dermal papilla cells are a group of cells located at the base of hair follicles that differentiate from dermal stromal cells, which have key regulatory effects on the morphogenesis and differentiation direction of hair follicles. Since Jahoda et al. [44
] successfully isolated and cultured rat tentacled nipple cells, in vitro isolation and culture techniques for dermal papilla cells from other mammals (such as humans, pigs, sheep, and rabbits) have also been established [45
]. Alpha-SMA is a specific marker of hair dermal papilla cells cultured in vitro, which is mainly used to distinguish these cells from fibroblasts [49
]. VIM is a dermal hair follicle-derived cell marker for the identification of dermal-derived cells and epidermal-derived cells [50
]. In this study, the expression of α-SMA and VIM was detected by immunofluorescence staining and western blotting of isolated cultured cells, indicating the successful isolation and acquisition of goat hair dermal papilla cells. This isolation of DPCs lays a foundation for subsequent gene function verification at the cellular level.
There are numerous reports on miR-130 and miR-26a in cancer and tumorigenesis, but research on hair follicles is still lacking. We first detected the effects of miR-130a and miR-26a on the proliferation of dermal papilla cells. The results showed that miR-130a could inhibit the proliferation of sheep dermal papilla cells, while miR-26a could promote the proliferation of these cells. Additionally, miR-130a has been shown to promote collagen secretion, myofibroblast transformation and cell proliferation by targeting CYLD and enhancing Akt activity [31
]. It can be speculated that the same specific miRNA has different effects on different hair follicle cells. This may be due to differences in the periods of hair follicles, or it might be that the miRNA can play different roles in the regulation of hair follicles in different species. Next, we detected the changes in the expression levels of several SMAD
genes in the TGF-β/SMAD
pathway, which provided a molecular basis for further study on the occurrence and developmental mechanism of goat hair follicles. As expected, the results suggested that miR-130a and miR-26a can regulate the mRNA and protein expression level of the SMAD
family. These miRNAs regulate the growth and development of hair follicles. Research has shown that the inhibition of miR-26a promotes angiogenesis and dermal wound healing in mice by increasing the endothelial SMAD1
]. Our bioinformatic predictions suggested that SMAD1
might be a target gene of miR-26a. Notably, the mRNA level of SMAD1
does not change after the overexpression or inhibition of miR-26a. We further confirmed that miR-26a could reduce the protein expression of SMAD1
. Therefore, we conclude that miR-26 plays a role after transcription.
4. Materials and Methods
4.1. Animals and Samples
All animal experimental procedures were approved by the Ministry of Agriculture of the People’s Republic of China and Institute of Animal Science, Chinese Academy of Agricultural Sciences and were performed according to the guidelines for the care and use of experimental animals established by this ministry. Before collecting samples, we obtained the permission of Ningxia Zhongwei Goat Conservation Farm. Three Zhongwei goats located at a Zhongwei goat breeding farm in Ningxia, China, were randomly selected and had no relationship with each. We collected skin tissue from the goats at two time points: specifically, three 1 cm2 pieces of skin tissue were collected from the scapula using a sterilized scalpel. Some samples were immediately stored in RNAlater (Thermo Fisher Scientific, NY, USA) and kept at −80 °C until further processing. Other samples were quickly stored in a 4% paraformaldehyde fixative to prepare the paraffin sections. All wounds were treated with Yunnan Baiyao Powder to stop bleeding (China Yunnan Baiyao Group Co., Ltd., Kunming, China).
Ethical approval for animal survival was provided by the animal ethics committee of the Institute of Animal Science, Chinese Academy of Agricultural Sciences (IAS-CAAS) with the following reference number: IASCAAS-AE-03, on 1 September 2014.
4.2. Total RNA Extraction and Small RNA Library Preparation
Total RNA was extracted according to the RNeasy Plus Universal Mini Kit method. To ensure the accuracy of the data obtained, a nanodrop was used to detect whether the purity, concentration, and nucleic acid absorption peak of the isolated RNA were normal, and an Agilent 2100 was used to accurately evaluate the RNA integrity. Starting with the total RNA, a connector was added to each end of small RNA transcripts, and cDNA was synthesized by reverse transcription. Subsequently, after PCR amplification, we separated the target DNA fragments by PAGE. A cDNA library was obtained by gelatinization recovery.
4.3. Small RNA Sequencing and Read Processing
The libraries were sequenced on an Illumina HiSeq 2500 platform, and 50 bp single-end reads were generated. We conducted a series of data quality assessments, including an inspection of error rates, removal of low-quality reads, and length selection. Next, we analyzed the length distribution statistics. The peak of the length distribution helps determine the small RNA species. We used Bowtie [51
] to compare the screened reads to a reference sequence (ftp://ftp.ncbi.nlm.nih.gov/genomes/refseq/vertebrate_mammalian/Capra_hircus/latest_assembly_versions/GCF_001704415.1_ARS1
) and analyzed the distribution of the reads along the reference genome. To remove the tags that corresponded to low-complexity sequences (including rRNA, tRNA, snRNA and snoRNA), we compared the clean reads with the GtRNAdb database [52
], Rfam database [53
], Repbase database [54
] and NCBI database [55
]. Candidate miRNAs were obtained after deduplication. To identify miRNAs conserved among species, we compared the reads obtained by sequencing with those in other mammalian miRBase 22.0 databases (1 to 2 base mismatches were allowed during the process). The expression of the known and new miRNAs in each sample were counted, and the expressions were normalized with TPM
A differential expression analysis of two groups of data was performed using DESeq [56
]. The differential miRNA screening conditions were FDR < 0.01 and |fold change| > 1.5.
4.4. The Sequencing Results for Small RNA Were Verified by RT-qPCR
Differentially expressed miRNAs were randomly selected for RT-qPCR. The U6 gene was used as a reference, and primers were designed and synthesized by Shanghai Bioengineering Co., Ltd. (Shanghai, China). The total RNA isolated from the above-described 6 samples (3 from each 45-day-old and 108-day-old lamb) was subjected to reverse transcription cDNA synthesis. Reverse transcription cDNA synthesis was carried out in accordance with the Takara reverse transcription kit instructions. The products were stored at −20 °C until use. The reverse transcription products were used as templates for quantitative RT-qPCR. RT-qPCR was performed with reference to Takara real-time quantitative kit specifications: reaction system (20 μL); cDNA (2 μL), upstream and downstream primers (0.8 μL (Table S1
)), SYBR (10 μL), ddH2
O (6.4 μL). The test results were converted into relative expression data using the 2−ΔΔCt
method for relative quantitative analysis. A t
test was performed, and p
< 0.05 was defined as a significant difference. The reaction system was placed on an ABI quantitative PCR instrument for reaction. At least three samples were included for each time point, and all reactions for each sample were repeated three times.
4.5. Target Gene Prediction and Functional Notation of Differentially Expressed miRNAs
To better understand the biological functions of the identified miRNAs, we used RNAhybrid [57
] and miRanda [58
] software to predict the miRNA target genes. Gene Ontology (GO; http://www.geneontology.org/
) is an international standard classification system for gene function. The Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.kegg.jp/kegg/
) is a system for analyzing gene function and genomic information databases. We screened target genes of differentially expressed miRNAs for experimental purposes, and enrichments in the distribution of the target genes were assessed by a GO analysis. Additionally, the gene-enriched pathways of the target genes of the differentially expressed miRNAs were detected by a KEGG analysis to elucidate the expression pattern in each sample in the experiment by evaluating the functions of the target genes.
4.6. Differentially Expressed miRNA-Target Gene Interaction Network Analysis
Prior to a differential gene expression analysis for each sequenced library, a differential expression analysis of two samples was performed using IDEG6. The p value was adjusted using the q value. An FDR < 0.01 and |foldchange| ≥ 1.5 were set as the threshold for significant differential expression. MicroRNAs associated with the research purposes were selected from validated differentially expressed miRNAs. The predicted miRNA target genes were combined with transcriptomic analysis data to construct an mRNA−miRNA network. Transcriptome analysis data have been published, and the NCBI accession number is PRJNA555706. The software Cytoscape 3.4.0 was used to graphically visualize the network [59
4.7. Dermal Papilla Cell Separation and Culture
Approximately 1 cm2 of scapular skin tissue was removed by aseptic surgery, placed in DMEM containing two antibodies, and stored in an ice box. The harvested skin tissue was immersed in a petri dish containing 75% alcohol for 1 min and was rapidly washed 4–5 times with PBS. The tissue pieces were cut into 1 mm2 pieces using a sterile surgical blade, and these small tissue pieces were then collected in a petri dish containing 0.25% neutral protease and were incubated at 37 °C for approximately 2 h. Under a dissecting microscope, the dermal layer and subcutaneous tissue layer were gently separated with high-precision forceps to expose hair follicle bulges. The bulges were collected in a 1.5 mL centrifuge tube (including 1 mL of complete medium). We centrifuged the tube at 1500 rpm for 5 min and discarded the waste; the pellet in the tube was the size of a soybean. Then, 400 µL of type IV collagenase was added, and the precipitate was thoroughly mixed. After 15 min, the digestion was terminated, and the mixture was centrifuged. The cell mixture was passed through a 150 µm mesh cell sieve and seeded in a cell culture flask for cultivation.
4.8. Dermal Papilla Cell Identification
Fourth-passage cells in good growth conditions were seeded in petri dishes and cultured in an incubator at 37 °C for 48 h. After the cell density reached 70~80%, the cultivation was terminated. The expression of α-smooth muscle actin (α-SMA) and vimentin (VIM), a specific marker of hair papilla cells cultured in vitro, were detected by immunofluorescence staining. The total protein was extracted from hair dermal papilla cells in vitro, and the expression of specific markers was detected by Western blotting.
4.9. Overexpression and Inhibition of miR-26a and miR-130a
miRNA mimics and inhibitors were formulated in 20 µM stock solutions for transfection experiments according to miRNA product instructions (Shanghai Jiama Biotechnology Co., Ltd., Shanghai, China). The mimic and inhibitor sequences are showed in Table 2
. Cells were transfected with miR-26a/miR-130a mimics or inhibitors according to the instructions for Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA). The transfection groups are as follows: miR-26a mimic/miR-26a mimic-NC, miR-26a inhibitor/miR-26a inhibitor-NC, miR-130a mimic/miR-130a mimic-NC, miR-130a inhibitor/miR-130a inhibitor-NC. We extracted RNA from hair papillary cells, and RT-qPCR was performed to detect the overexpression and inhibition effect. Methods and reagents are shown in Part 4.4 of this section. The primers for RT-qPCR are shown in Table S2
4.10. Cell Proliferation Assay
Forty-eight hours after transfection, dermal papilla cells were digested with 0.25% trypsin. After 5 min, the digestion was terminated, and the medium was centrifuged. The medium was added to a cell suspension, which was seeded in a 96-well plate at 100 µL/well (approximately 1 × 104 cells) and cultured in an incubator at 37 °C in saturated humidity and 5% CO2. The experiment included a control group (no treatment) and an experimental group, and each group comprised 6 parallel wells. The old medium was discarded after 0, 12, 24, 36, or 48 h of incubation, and the cells were washed twice with PBS. Next, we added 10 µL of CCK-8 solution to each well (avoiding the generation of air bubbles) and continued to incubate the plate for 2 h in a 37 °C incubator. Finally, the OD values were measured at 450 nm by an enzyme-labelling instrument. The relative proliferation level of hair dermal papilla cells is represented by the OD value.
4.11. RNA Extraction, Reverse Transcription and RT-qPCR
Three days after the dermal papilla cells were transfected, a total RNA extraction was performed using the miRNeasy Mini Kit (QIAGEN, Dusseldorf, Germany). Reverse transcription cDNA synthesis was carried out in accordance with the Takara reverse transcription kit instructions. A quantitative analysis was performed using SYBR Premix Ex TaqTM II (Taiwan TaKaRa Bio., Taibei, China) and an ABI quantitative PCR instrument. The primers used for the analysis are listed in the attached table. The experimental results were analyzed by the 2−ΔΔCt
method. A t test was performed, and p
< 0.05 was defined as a significant difference. Three samples were included for each time point, and all the reactions for each sample were repeated three times. The primers for RT-qPCR are shown in Table S3
4.12. Western Blot Analysis
Based on previous data analysis and experimental results, we selected SMAD2 and SMAD6 in the TGF-β/SMAD signal pathway for protein level validation. We extracted the total protein from dermal papilla cells 3 d after transfection. Western blot analysis mainly included the following steps: preparation of a colloid, transfer to a membrane, blocking, binding of a specific antibody to the corresponding target protein, and visualization. The primary antibodies used were murine anti-β-actin (1:1000 (Abcam, Cambridge, UK)), murine anti-SMAD2 (1:1000 (Abcam, Cambridge, UK)), and rabbit anti-SMAD6 (1:1000 (Abcam, Cambridge, UK)). The antibody was diluted in a 5% BSA blocking solution (Beijing Solable Technology Co., Ltd., Beijing, China).
4.13. Luciferase Reporter Gene Assay
According to the dual-luciferase reporter product instructions, (Shanghai Jiama Biotechnology Co., Ltd.), the recombinant mutant vector SMAD1-miR26a-MUT and the wild-type vector SMAD1-miR26a-WT were cotransfected with the miRNA mimic miR-26a mimetic or miR-26a mimic-NC. Forty-eight hours after cotransfection, the activities of the Renilla luciferase and firefly luciferase were measured by a dual-luciferase reporter assay kit (Promega, Madison, WI, USA). At least 3 samples were included for each time point, and all reactions for each sample were repeated three times.
4.14. Statistical Analysis
All data are presented as the mean ± S.D. based on at least three replicates for each treatment. The data were analyzed by performing one-sample t tests.
4.15. Availability of Data and Materials
This manuscript contains previously published data. The sample information for transcriptome sequencing data were submitted to the NCBI BioProject section under accession number PRJNA555706.