Abstract

   Fibroblast growth factor 21 (Fgf21) is a key regulator of hair follicle
   development and cycling. Although microRNAs (miRNAs) are involved in
   this process, the specific mechanisms through which Fgf21 regulates
   hair follicle growth remain unclear. This study investigates the role
   of Fgf21 and its associated miRNAs in hair follicle growth and
   development. Using CRISPR/Cas9, we generated Fgf21 knockout mice
   (Fgf21^−/^−), which exhibited a delayed transition from the telogen to
   anagen phases compared to wild-type (WT) mice. miRNA sequencing
   identified differentially expressed miRNAs in Fgf21^−/^− mice, with
   dual-luciferase assays confirming that miR-134-5p directly targets
   vascular endothelial zinc finger 1 (Vezf1) and miR-136-5p targets
   mitogen-activated protein kinase kinase kinase 1 (Map3k1). Real-time
   qPCR analysis revealed that Vezf1 and Map3k1 expression was higher in
   Fgf21^−/^− mice than in WT mice during catagen, but lower during
   telogen. These findings indicate that Fgf21 plays a critical role in
   regulating hair follicle growth and may modulate Vezf1 and Map3k1
   expression through miRNAs. This study provides novel insights into the
   molecular regulation of hair follicle growth and suggests potential
   therapeutic strategies for hair follicle-related disorders.

1. Introduction

   Hair follicles are essential skin appendages that arise from the
   interactions between cells of different origins, including epithelial,
   mesenchymal, and neurodermal cells [[34]1]. Hair follicles serve as the
   foundation for hair growth and provide structural support and
   protection to the skin [[35]2]. Hair follicle formation is a complex
   process involving the coordinated actions of surrounding tissues and
   cells and is regulated by multiple signaling pathways and transcription
   factors [[36]3].

   In recent years, the molecular mechanisms underlying hair follicle
   development have gained considerable attention. The fibroblast growth
   factor (FGF) family plays a crucial regulatory role in this process.
   The FGF family comprises a large group of growth factors that regulate
   cell proliferation, apoptosis, migration, and differentiation
   [[37]4,[38]5]. Dysregulation of FGF family genes significantly affects
   hair follicle growth and development. For instance, Fgf18 is considered
   a key regulator of the hair follicle growth cycle, as Fgf18 knockout
   mice exhibit shortened telogen [[39]6], whereas Fgf7 overactivation
   markedly inhibits hair follicle growth [[40]7]. Fgf20 is essential for
   early dermal condensate formation [[41]8,[42]9], whereas Fgf9 and Fgfr2
   IIIb are involved in hair follicle regulation [[43]10,[44]11].
   Interactions between Fgf and Wnt signaling in the mesenchymal niche
   regulate the hair follicle cycle clock in mice [[45]12], and Fgfr2
   negatively regulates sonic hedgehog (Shh) during hair follicle
   development [[46]13]. Members of the FGF and FGFR families regulate the
   entire hair follicle growth cycle in mice, with different genes
   expressed at various stages. During anagen, Fgf21 and its receptor
   Fgfr2 are expressed; during catagen, Fgfr1 and Fgfr4 are expressed; and
   during telogen, Fgf1, Fgf2, Fgf13, Fgf18, and Fgf20 are expressed
   [[47]14]. These findings highlight the critical role of the FGF family
   in hair follicle growth and development.

   Fgf21 was first isolated in 2000 by Nishimura et al. [[48]15] and
   belongs to the FGF19 subfamily, which also includes Fgf19, Fgf21, and
   Fgf23 [[49]16]. While Fgf21 has been recognized as a key regulator of
   hair follicle growth and development, its role in the miRNA regulatory
   network, particularly in the context of hair follicle growth and cell
   cycle regulation, remains underexplored. Therefore, elucidating the
   specific mechanisms by which Fgf21 regulates hair follicle growth and
   development, as well as its role in hair follicle cell cycle
   regulation, is crucial for understanding hair follicle biology and may
   provide new therapeutic strategies for related disorders.

   MicroRNA sequencing technology has revealed the significant role of
   miRNAs in hair follicle growth, development, and transition between
   hair follicle cycles [[50]17]. To investigate the role of Fgf21 in hair
   follicle growth and cycle regulation, our laboratory used CRISPR/Cas9
   to develop a Fgf21 knockout mouse model. High-throughput miRNA
   sequencing found 93 differentially expressed miRNAs, with 10 selected
   for their potential roles in hair follicle regulation; Vezf1 and Map3k1
   were confirmed as direct targets of miR-134-5p and miR-136-5p,
   respectively, and their expression within follicles was altered by
   Fgf21 during the first natural growth cycle of mouse hair follicles.
   These findings indicate that Fgf21 knockout induces changes in the
   miRNA expression profile and regulates Vezf1 and Map3k1 expression,
   thereby influencing hair follicle growth and development.

2. Materials and Methods

2.1. Animal Models and Sample Collection

   This study used FVB strain, an outbred strain derived from Swiss mice
   and developed by the National Institutes of Health from the Friend
   Virus Susceptibility 1b (FV1b) inbred strain. We established a Fgf21
   knockout mouse model using CRISPR/Cas9 technology. The mouse model was
   bred through intercrosses and reciprocal crosses of heterozygous and
   homozygous mice. All experimental animals were housed in a controlled
   environment with a temperature of 22–24 °C, humidity of 45–65%, and
   free access to food and water. Genomic DNA was extracted from mouse
   tail tips using a genomic DNA extraction kit (Tiangen Bio, Beijing,
   China) to confirm successful Fgf21^−/^− model establishment. Genotyping
   was performed via PCR using the following primers:
   Fgf21-F:5′-GTTCCTGCCAAGTGTGTCAAATA-3′;
   Fgf21-R:5′-GACCACAGAAAGAGACTCACCG-3′. The reaction procedure was as
   follows: initial denaturation at 94 °C for 5 min, followed by 35 cycles
   of denaturation at 94 °C for 30 s, annealing at 56 °C for 30 s, and
   extension at 72 °C for 50 s. Finally, the reaction was held at 72 °C
   for 10 min and then cooled to 16 °C for 1 h. The genotype
   identification results are listed in [51]Figure S1 in the Supplementary
   Materials.

   Dorsal skin tissues of 8-week-old FVB Fgf21^−/^− and WT mice were
   collected for miRNA sequencing. All animal experiments were conducted
   in strict accordance with the Guidelines for the Management and Use of
   Laboratory Animals established by the Laboratory Animal Research Center
   of Inner Mongolia University (SYXK2020-0006). The animal experiments in
   this study were approved and authorized by the Animal Ethics Committee
   of Inner Mongolia University (IMU-MOUSE-2020-047). These measures
   ensured scientific rigor and ethical compliance, providing a reliable
   model for investigating the role of Fgf21 in the hair follicle growth
   cycle.

2.2. Preparation of Hair Follicle Cycle-Synchronized Mouse Model

   To assess the role of Fgf21 in the hair follicle growth cycle, we
   synchronized the hair follicle cycle in 8-week-old FVB mice using a
   physical depilation method. After inducing anesthesia, a mixture of
   rosin and paraffin was applied to the dorsal skin to induce depilation.
   Subsequently, we compared the hair growth rates between WT and
   Fgf21^−/^− mice. Hair growth was monitored and photographed at 1-, 3-,
   5-, 7-, 9-, 11-, 13-, 15-, 17-, 19-, and 21-days post-depilation to
   document the process in detail.

   8-week-old mice are in the telogen phase of the hair follicle cycle
   [[52]18]. After depilation, hair follicles synchronously transition to
   the anagen phase [[53]19]. We use the model to study the transition
   from telogen to anagen in mice. For target gene validation in the hair
   follicle cycle, naturally cycling mice are selected. On postnatal day 6
   (P6), hair follicles are in the anagen phase, characterized by active
   cell proliferation and hair shaft formation, while P5 is regarded as
   early anagen due to the continuity of follicular development [[54]20].
   P15 and P17 mice are in the catagen phase, which lasts until P19,
   followed by the resting telogen phase, with P20 being in early telogen
   [[55]20]. Based on this information, skin tissue from P5 mice
   represents anagen, P16 represents catagen, and P20 represents telogen.

2.3. Real-Time Quantitative PCR (RT-qPCR)

   In the RT-qPCR experiment, total RNA was extracted from fresh frozen
   samples. After collection, the samples were stored in liquid nitrogen
   tanks and RNA extraction was performed once all samples were collected.
   Total RNA was extracted from the dorsal skin tissues of 8-week-old WT
   and Fgf21^−/^− mice using the RNAiso Plus reagent kit (Takara Bio Inc.,
   Tokyo, Japan). The extracted RNA was reverse transcribed into cDNA
   using the PrimeScript Reverse Transcriptase Kit (Takara Bio Inc.,
   Tokyo, Japan). Reverse transcription was performed with the GoScript
   Reverse Transcription System (Promega Corporation, Madison, WI, USA) on
   a C1000 Thermal Cycler (Bio-Rad Laboratories Inc., Hercules, CA, USA).
   RT-qPCR reactions were conducted on the C1000 Thermal Cycler using the
   GoTaq qPCR Master Mix (Promega Corporation, Madison, WI, USA). GAPDH
   and U6 were selected as internal reference genes to ensure the accuracy
   and comparability of the data. The primer sequences used for RT-qPCR
   are listed in [56]Table S1 in the Supplementary Materials. Relative
   gene expression levels were calculated using the 2^−ΔCT method, based
   on the difference between the CT values of the target genes and the
   reference genes.

2.4. miRNA Library Construction and High-Throughput Sequencing

   miRNA sequencing libraries were constructed from the total RNA
   extracted from the fresh dorsal skin tissues of 8-week-old FVB mice
   using the TruSeq Small RNA Sample Preparation Kit (Illumina Inc., San
   Diego, CA, USA). PCR products of 140–160 bp were recovered using 6%
   polyacrylamide gel electrophoresis for library preparation. Single-end
   high-throughput sequencing was performed on an Illumina HiSeq 2000/2500
   platform (Illumina Inc., San Diego, CA, USA).

2.5. Data Quality Control and miRNA Identification

   The quality of raw miRNA sequencing data was assessed using Illumina
   FastQC (Illumina Inc., San Diego, CA, USA), and high-quality sequences
   meeting the Q30 standard were selected for further studies. These
   sequences were aligned to the miRBase 21.0 database (Memorial
   Sloan-Kettering Cancer Center, New York, NY, USA) using Bowtie (Johns
   Hopkins University, Baltimore, MD, USA) to identify and quantify known
   miRNAs. Sequences that did not match the known miRNAs were further
   aligned with the reference genome. The secondary structure of the
   successfully aligned sequences was predicted using RNAfold (University
   of Vienna, Vienna, Austria) to identify and analyze the structural
   features of the miRNAs.

2.6. Differential Expression Analysis of miRNAs

   The expression levels of miRNAs were analyzed using RSEM (Harvard
   Medical School, Boston, MA, USA) based on sequencing data, with
   expression values normalized to transcripts per million (TPM).
   Differential expression analysis of the miRNAs was performed using
   DESeq2 (Bioconductor, Heidelberg, Germany). DESeq2 employs a negative
   binomial distribution model to evaluate the statistical significance of
   miRNA expression differences while accounting for sample variability.
   The analysis was performed using default parameters, a significance
   level of p-value < 0.05, and a fold-change cutoff of 2 for identifying
   upregulated and downregulated miRNAs. The Benjamani–Hochberg method was
   applied for multiple testing correction to enhance accuracy.

2.7. Target Gene Enrichment Analysis and Network Construction

   To identify the target genes of differentially expressed miRNAs
   associated with hair follicle growth and development, we employed three
   prediction tools: TargetScan 8.0 (Whitehead Institute for Biomedical
   Research, Cambridge, MA, USA), miRanda 2.066 (Memorial Sloan-Kettering
   Cancer Center, New York, NY, USA), and miRDB 6.0 (University of
   Illinois, Chicago, IL, USA). Potential target genes were identified by
   intersecting the prediction results from these tools. Subsequently, GO
   and KEGG pathway enrichment analyses were performed on candidate target
   genes to elucidate their potential functions and associated biological
   pathways. In both GO and KEGG analyses, a significance threshold of p <
   0.05 was set to identify significantly enriched terms. Cytoscape
   software 3.7.1 (Cytoscape Consortium, San Diego, CA, USA) was used to
   construct miRNA-GO/KEGG-target gene networks, enabling visualization of
   relationships among miRNAs, their target genes, and associated
   pathways.

2.8. Validation of Target Relationships

   To validate the targeting relationships between miR-134-5p and
   miR-136-5p and their target genes, we used the dual-luciferase reporter
   assay ([57]Figure 1). The backbone vector was amplified in LB broth
   (AMP+) at 37 °C for 12 h. The bacterial culture was processed to remove
   debris, and the vector was purified and digested with XhoI and XbaI
   (New England Biolabs, Ipswich, MA, USA). The vector was linearized via
   1% agarose gel electrophoresis. Target binding sites for miR-134-5p and
   miR-136-5p within the target genes’ 3′UTR were identified using
   TargetScan, and primers were designed ([58]Supplementary Table S1). The
   target fragments were synthesized, annealed into double-stranded DNA,
   and ligated into the linearized vector. The ligation product was
   transformed into competent cells, cultured, and plated on LB agar
   (AMP+). Selected colonies were cultured, and plasmid DNA was extracted
   and sequenced to verify the construct. The confirmed recombinant
   plasmid was propagated in E·coli for subsequent transfection
   experiments.

Figure 1.

   [59]Figure 1
   [60]Open in a new tab

   Diagram of the dual-luciferase reporter assay.

   The 293T cells were transfected in four groups: miR-134/136-5p mimics
   or NC mimics (RiboBio, Guangzhou, China) with WT or MUT target gene
   vectors. Cells (2 × 10^5 per well) were seeded in 24-well plates and
   transfected at ~95% confluence. Transfection complexes were formed by
   mixing Solution A (50 µL Opti-MEM (Thermo Fisher Scientific, Waltham,
   MA, USA) with 2 µL Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA))
   and Solution B (50 µL Opti-MEM with 1.25 µL mimics/NC and 300 ng
   vector) at a 1:1 ratio, incubating this for 5 min at room temperature,
   and diluting it to 500 µL with Opti-MEM and adding this to the cells.
   After 6 h incubation at 37 °C with 5% CO[2], the medium was replaced.
   For detection, cells were washed three times with 1× DPBS (Thermo
   Fisher Scientific, Waltham, MA, USA), lysed with 100 µL 1× PLB
   (Promega, Madison, WI, USA) for 15 min at 145 rpm, then centrifuged at
   4 °C and 12,000 rpm for 30 s. Luciferase activity was measured by
   mixing 10 µL supernatant with 50 µL Luciferase Assay Reagent (Promega,
   Madison, WI, USA) to read RLU1, followed by adding 50 µL Stop Reagent
   (Promega, Madison, WI, USA) to read RLU2. The ratio (RLU1/RLU2) was
   recorded.

2.9. Morphological Observation

   The skin tissues used in H&E staining experiments were obtained from
   fresh samples. Dorsal skin tissues were cut into 1 cm^2 pieces, rinsed
   with PBS, and immediately immersed in 4% paraformaldehyde for 24 h for
   fixation. After fixation, the samples were dehydrated using a graded
   ethanol series, cleared in xylene, and embedded in paraffin. Thin
   sections were then prepared and stained with H&E. Morphological
   observations were conducted using a microscope.

2.10. Immunofluorescence Staining

   The skin tissues used in the immunofluorescence staining experiments
   were also obtained from fresh samples, which were rinsed with PBS and
   fixed in paraformaldehyde for 24 h before experimentation. Tissue
   sections were dewaxed in xylene and rehydrated using a graded ethanol
   series. These sections were subjected to heat-induced epitope retrieval
   in 0.01 M sodium citrate buffer to expose antigens. After antigen
   retrieval, the sections were blocked with blocking solution at room
   temperature for 1 h and then incubated with primary antibodies
   overnight at 4 °C. The primary antibodies used were anti-BCL2 (Bioss,
   Beijing, China, bsm-1174R, dilution 1:400), anti-KI67 (Bioss, Beijing,
   China, bsm-52455R, dilution 1:400), anti-VEZF1 (Bioss, Beijing, China,
   bs-12791R, dilution 1:400), and anti-MAP3K1 (Proteintech, Rosemont, IL,
   USA, 19970-1-AP, dilution 1:400). After the incubation period, sections
   were uniformly covered with the secondary antibody and incubated for 1
   h in the dark, followed by three washes with PBS. Nuclei were
   counterstained with DAPI (Sigma, Saint Louis, MO, USA, D8417, dilution
   1:1000) for 8–10 min at room temperature in the dark, with three
   additional washes with PBS. These sections were then mounted and
   examined to assess BCL2, KI67, MAP3K1, and VEZF1 expression in various
   hair follicle structures.

2.11. Statistical Analysis

   Each experiment was conducted independently in triplicate. Data
   analysis was performed using the GraphPad Prism 8 software (GraphPad
   Software, San Diego, CA, USA). Student’s t-test was used to compare
   differences between the two groups. When evaluating differences among
   multiple groups, one-way analysis of variance (ANOVA) was performed. A
   p-value of <0.05 was considered statistically significant.

3. Results

3.1. Fgf21 Knockout Delays the Transition from Telogen to Anagen in Hair
Follicles

   During telogen, the skin of friend virus B-type (FVB) mice typically
   appears pink, gradually turns pale pink, and begins to grow white hair
   at the onset of anagen. To investigate the role of Fgf21 in hair
   follicle cycle transition, we synchronized the hair follicle cycles of
   8-week-old Fgf21^−/^− and WT mice through depilation. Immediately after
   depilation, the skin of all the mice appeared pink, indicating that
   they were in the telogen phase, with no significant differences
   observed between the two groups. On day 7 post-depilation, both
   Fgf21^−/^− and WT mice retained pink skin, confirming that they
   remained in the telogen phase. On day 9, while the skin of Fgf21^−/^−
   mice remained pink, that of WT mice began to turn pale pink, indicating
   the initial entry into the anagen phase. By day 11, Fgf21^−/^− mice
   showed no significant changes, whereas WT mice began to grow fine
   vellus hair. On day 13, Fgf21^−/^− mice gradually turned pale pink,
   marking the initial entry into the anagen phase, whereas WT mice had
   already developed longer hair covering the skin ([61]Figure 2A).

Figure 2.

   [62]Figure 2
   [63]Open in a new tab

   Effects of Fgf21 knockout on hair follicle cycle transition. (A)
   Time-series images showing the dorsal skin and hair development in WT
   and Fgf21^−/^− mice from day 1 (P1) to 21 (P21) post-depilation. (B)
   H&E staining illustrating hair follicle structures in WT and Fgf21^−/^−
   mice on day 7 (7 d); day 9 (9 d); day 11 (11 d); and day 13 (13 d)
   post-depilation. Scale bar = 1000 μm.

   To further investigate the impact of Fgf21 knockout on hair follicle
   cycle transition, we performed hematoxylin and eosin (H&E) staining on
   depilated skin tissues of Fgf21^−/^− and WT mice. The results were
   consistent with phenotypic observations. On day 7 post-depilation, both
   groups exhibited typical in the telogen phase features with shortened
   hair follicles and tightly packed cell layers between the inner and
   outer root sheaths, showing no significant differences. By day 9, WT
   mice began to show increased hair follicle length, enlarged dermal
   papilla, and the thickened dermis, indicating a transition to anagen,
   whereas Fgf21^−/^− mice remained in telogen with shorter hair follicles
   and shrunken dermal papilla. By day 11, most hair follicles in WT mice
   had extended downward close to the adipose layer, with increased dermal
   thickness and dermal papilla enlargement, confirming their entry into
   the anagen phase. In contrast, Fgf21^−/^− mice only began to extend
   their follicles into the dermis and indicated a transition to the
   anagen phase. By day 13, both groups exhibited increased follicle
   length and dermal thickening, confirming their eventual progression
   into the anagen phase. These results indicate that Fgf21 knockout
   delays the transition from telogen to anagen in hair follicles
   ([64]Figure 2B).

3.2. Expression Analysis and Functional Annotation of miRNAs

   To investigate the impact of Fgf21 knockout on hair follicle growth and
   development, we performed miRNA sequencing of the dorsal skin tissues
   of Fgf21^−/^− and WT mice, each with three biological replicates per
   group. After quality control and sequence alignment, we identified 937
   miRNAs in the WT group and 952 miRNAs in the Fgf21^−/^− group.
   Principal component analysis (PCA) revealed distinct differences in
   miRNA expression between the experimental and control groups, whereas
   the intra-group samples exhibited high reproducibility, indicating the
   consistency of experimental operations and reliability of the results
   ([65]Figure 3A). Violin plots also illustrated the distribution of
   miRNA expression levels in each sample ([66]Figure 3B).

Figure 3.

   [67]Figure 3
   [68]Open in a new tab

   Expression analysis and functional annotation of miRNAs. (A) PCA plot
   of the distribution of miRNAs expression in dorsal skin tissues of WT
   and Fgf21^−/^− mice. (B) Violin plots of the distribution of miRNAs
   expression in six different samples of WT and Fgf21^−/^− mice. (C) Venn
   diagram comparing miRNAs expression between WT and Fgf21^−/^− mice. A
   total of 776 miRNAs were commonly expressed between the two groups,
   with 176 miRNAs specifically expressed in Fgf21^−/^− mice and 161
   miRNAs specifically expressed in WT mice. (D) Volcano plot showing the
   differentially expressed miRNAs between the two groups. The plot
   highlights the 66 upregulated and 27 downregulated miRNAs. (E)
   Hierarchical clustering heatmap of DEmiRNAs, showing the expression
   patterns of differentially expressed miRNAs in the dorsal skin tissues
   of WT and Fgf21^−/^− mice. Color gradations indicate relative changes
   in miRNAs expression levels, with yellow indicating upregulation and
   blue indicating downregulation. (F) GO functional annotation of target
   genes of DEmiRNAs. The bubble chart displays the top 20 most
   significantly enriched GO terms. (G) KEGG pathway enrichment analysis
   of target genes of DEmiRNAs. The bubble chart displays the top 20 most
   significantly enriched KEGG pathways.

   A Venn diagram analysis of miRNAs from Fgf21^−/^− and WT mice revealed
   that 776 miRNAs were commonly expressed between the two groups, 176
   miRNAs were specifically expressed in the Fgf21^−/^− group, and 161
   miRNAs were specifically expressed in the WT group ([69]Figure 3C). A
   comparative analysis of expression levels between the two groups
   identified 93 differentially expressed miRNAs, of which 66 were
   upregulated and 27 were downregulated. Volcano plots and hierarchical
   clustering heat maps also showed the distribution and clustering of
   such differentially expressed miRNAs ([70]Figure 3D,E).

   To further elucidate the functions of these differentially expressed
   miRNAs (DEmiRNAs), we performed Gene Ontology (GO) and Kyoto
   Encyclopedia of Genes and Genomes (KEGG) functional annotation and
   enrichment analysis of their target gene sets using miRanda 2.066
   prediction software. GO analysis revealed that the host genes of the
   DEmiRNAs were primarily enriched in functions such as protein binding,
   DNA binding, nucleotide binding, multicellular organismal development,
   positive regulation of RNA polymerase II transcription, and protein
   phosphorylation ([71]Figure 3F). KEGG analysis revealed that the target
   genes of DEmiRNAs were associated with several key signaling pathways
   during hair follicle development, including the MAPK signaling pathway,
   Ras signaling pathway, and Wnt signaling pathway ([72]Figure 3G). These
   results suggest that miRNAs are likely involved in the molecular
   pathways of hair follicle development by regulating their target gene
   expression.

3.3. Clustering Analysis and Expression Level Detection of Hair
Follicle-Related miRNAs

   To further explore the effects of Fgf21 knockout on the expression of
   hair follicle-related miRNAs, we identified 10 miRNAs associated with
   hair follicle growth and development and performed clustering analysis
   on their expression levels ([73]Figure 4A). The clustering heatmap
   revealed distinct miRNAs expression patterns between the two groups.
   Specifically, miR-423-5p and miR-186-5p were downregulated in
   Fgf21^−/^− mice, whereas miR-409-5p, miR-127-3p, miR-134-5p,
   miR-381-3p, miR-152-3p, miR-434-5p, miR-369-3p, and miR-136-5p were
   upregulated. These results indicate that there are significant
   differences in miRNA expression levels between the dorsal skin tissues
   of WT and Fgf21^−/^− mice.

Figure 4.

   [74]Figure 4
   [75]Open in a new tab

   Effects of Fgf21 knockout on the expression of hair follicle-related
   miRNAs and validation. (A) Hierarchical clustering heatmap of the
   expression patterns of 10 hair follicle-related miRNAs in the dorsal
   skin tissues of WT and Fgf21^−/^− mice. Color changes denote relative
   changes in miRNA expression levels, with upregulation indicated by red
   and downregulation indicated by blue. (B–K) RT-qPCR results indicating
   the relative expression levels of specific miRNAs in the dorsal skin
   tissues of WT and Fgf21^−/^− mice. RT-qPCR results were consistent with
   the sequencing data trends. * p < 0.05, ** p < 0.01, *** p < 0.005.
   Each group included three mice, and experiments were repeated three
   times independently.

   To validate the accuracy of the miRNA sequencing results, we performed
   RT-qPCR analysis and selected miRNAs using U6 as an endogenous
   reference miRNA. The RT-qPCR results showed that the expression levels
   of miR-409-5p, miR-127-3p, miR-134-5p, miR-381-3p, miR-152-3p,
   miR-434-5p, miR-369-3p, and miR-136-5p were upregulated in Fgf21^−/^−
   mice, whereas those of miR-186-5p and miR-423-5p were downregulated
   ([76]Figure 4B–K). These findings are consistent with the trends
   observed in the sequencing data. These results further confirmed the
   role of Fgf21 in modulating the expression of hair follicle-related
   miRNAs.

3.4. Clustering Analysis and Network Construction of Target Genes

   To elucidate the molecular mechanisms by which Fgf21 knockout
   influences hair follicle growth and development, we identified target
   genes for the 10 hair follicle-related miRNAs and performed GO and KEGG
   functional enrichment analysis. GO functional enrichment analysis
   revealed that the target genes are primarily involved in biological
   processes including regulation of transcription from the RNA polymerase
   II promoter, protein phosphorylation, and regulation of transcription
   from the DNA template. Subcellular localization analysis showed that
   the target genes were predominantly located in the cytoplasm, nucleus,
   and plasma membrane. In terms of molecular function, the target genes
   were primarily involved in binding to proteins, ATP, and DNA, as well
   as with protease activation ([77]Figure 5A). KEGG pathway enrichment
   analysis indicated that the target genes were enriched in pathways
   related to signal transduction, cancer, and nervous, endocrine, and
   immune systems ([78]Figure 5B).

Figure 5.

   [79]Figure 5
   [80]Open in a new tab

   Clustering analysis and network construction of target genes. (A) GO
   functional analysis of target genes. (B) KEGG pathway enrichment of
   target genes. (C) miRNAs–GO–target gene network showing interactions
   between miRNAs and their target genes within GO terms. Green nodes
   represent target genes, purple nodes represent GO terms, and yellow
   nodes represent miRNAs. Lines indicate regulatory relationships between
   miRNAs and target genes, as well as associations between target genes
   and GO terms. (D) miRNAs–KEGG–target gene network showing interactions
   between miRNAs and their target genes within KEGG pathways. Green nodes
   represent target genes, purple nodes represent KEGG pathways, and
   yellow nodes represent miRNAs. Lines indicate regulatory relationships
   between miRNAs and target genes, as well as associations between target
   genes and KEGG pathways.

   By combining literature reports, we selected hair follicle-related
   target genes for GO and KEGG enrichment analyses. GO analysis revealed
   that these target genes are primarily enriched in biological processes
   such as protein binding, homologous protein binding, and the regulation
   of transcription from RNA polymerase II promoter, as well as in
   cellular components like the cytoplasm and nucleus. Consequently, we
   constructed a miRNAs–GO–target gene network association diagram
   ([81]Figure 5C). KEGG analysis showed that these target genes were
   significantly enriched in hair follicle development-related pathways,
   including the Wnt signaling pathway, SNARE interactions in vesicular
   transport, ABC transporters, protein processing in the endoplasmic
   reticulum, the Notch signaling pathway, ubiquitin-mediated proteolysis,
   and the MAPK signaling pathway. We also constructed a
   miRNAs–KEGG–target gene network association diagram ([82]Figure 5D).
   These network diagrams highlight the potential molecular processes
   through which Fgf21 may regulate hair follicle growth and development
   through miRNAs, providing new insights into the molecular basis of this
   process. Thus, miRNAs may serve as a potential therapeutic target for
   follicle-related diseases.

3.5. Functional Implications of miRNAs in Hair Follicle Development

   To identify the potential target genes influenced by Fgf21 knockout, we
   used three prediction tools—TargetScan, miRanda, and miRDB—to predict
   target genes for the 10 hair follicle-related miRNAs ([83]Figure 6A).
   By comparing the prediction results from these tools, we identified
   miR-134-5p and miR-136-5p as the miRNAs with the highest number of
   predicted target genes among the three software programs. Specifically,
   miR-134-5p had 39 predicted target genes, while miR-136-5p had 90. To
   further validate these findings, we constructed Venn diagrams to
   visualize the intersection of these target gene predictions for further
   experimental analysis ([84]Figure 6B).

Figure 6.

   [85]Figure 6
   [86]Open in a new tab

   Target gene prediction and dual-luciferase reporter assay. (A) Bar
   chart shows the number of target genes for 10 miRNAs predicted by
   TargetScan, miRanda, and miRDB. (B) Venn diagram showing the predicted
   target genes of miR-134-5p and miR-136-5p, intersected from TargetScan,
   miRanda, and miRDB. Numbers indicate the number of target genes
   predicted by each software and their overlaps. (C) Bar chart showing
   the dual-luciferase reporter assay results for miR-134-5p and its
   potential target genes Banf1, Golm1, Tspan18, Zmat2, Antxr1, and Vezf1.
   Blue bars represent the NC mimics group and red bars represent the
   miR-134-5p mimics group. (D) Bar chart showing the dual-luciferase
   reporter assay results for Vezf1 3′UTR WT co-transfected with
   miR-134-5p mimics and Vezf1 3′UTR WT co-transfected with NC mimics.
   Blue bars represent the NC mimics group and red bars represent the
   miR-134-5p mimics group. (E) Bar chart showing the dual-luciferase
   reporter assay results for Vezf1 3′UTR MUT co-transfected with
   miR-134-5p mimics and Vezf1 3′UTR MUT co-transfected with NC mimics.
   Blue bars represent the NC mimics group and red bars represent the
   miR-134-5p mimics group. (F) Bar chart showing the dual-luciferase
   reporter assay results for miR-136-5p and its potential target genes
   Sumo2, Braf, Cbx4, Ppp2r2a, Yes1, Pcdh19, Sema4c, and Map3k1. Blue bars
   represent the NC mimics group and red bars represent the miR-136-5p
   mimics group. (G) Bar chart showing the dual-luciferase reporter assay
   results for Map3k1 3′UTR WT co-transfected with miR-136-5p mimics and
   Map3k1 3′UTR WT co-transfected with NC mimics. Blue bars represent the
   NC mimics group and red bars represent the miR-136-5p mimics group. (H)
   Bar chart showing the dual-luciferase reporter assay results for Map3k1
   3′UTR MUT co-transfected with miR-136-5p mimics and Map3k1 3′UTR MUT
   co-transfected with NC mimics. Blue bars represent the NC mimics group
   and red bars represent the miR-136-5p mimics group. ** p < 0.01, *** p
   < 0.005. Each experiment was independently repeated three times.

   For miR-134-5p, we selected six high-scoring target genes associated
   with hair follicle development from the 39 predicted targets for
   dual-luciferase reporter assays. The results showed that Banf1, Golm1,
   Tspan18, Zmat2, and Antxr1 did not significantly interact with
   miR-134-5p ([87]Figure 6C). However, Vezf1 showed a significant
   reduction in luciferase activity when co-transfected with miR-134-5p
   mimics compared to the negative control (NC) mimics ([88]Figure 6D).
   Mutation assays further confirmed that Vezf1 is a target gene of
   miR-134-5p, as the mutant Vezf1 vector showed no significant changes in
   luciferase activity when co-transfected with miR-134-5p mimics and NC
   mimics ([89]Figure 6E).

   For miR-136-5p, we selected eight target genes that not only scored
   highly in the bioinformatics prediction but also have relevance to hair
   follicle growth. The results showed that Sumo2, Braf, Cbx4, Ppp2r2a,
   Yes1, Pcdh19, and Sema4c did not show significant interactions with
   miR-136-5p ([90]Figure 6F). In contrast, Map3k1 exhibited a significant
   reduction in luciferase activity when co-transfected with miR-136-5p
   mimics compared to NC mimics ([91]Figure 6G). Mutation assays further
   confirmed that Map3k1 is a target gene of miR-136-5p, as the mutant
   Map3k1 vector did not show significant changes in luciferase activity
   when co-transfected with miR-136-5p mimics and NC mimics ([92]Figure
   6H).

3.6. Fgf21 Regulates Expression and Localization of Vezf1 and Map3k1 in the
Hair Follicle Growth Cycle

   To characterize the hair follicle growth cycle, we examined the dorsal
   skin tissues of mice on postnatal day 5 (P5), 16 (P16), and 20 (P20).
   We observed the distribution of hair follicles with enlarged dermal
   papilla regions at P5 within the dermis and subcutaneous tissue, which
   is a feature of the anagen phase. At P16, the hair follicles extended
   through the epidermis, dermis, and subcutaneous tissue, with hair
   shafts reaching the epidermal surface, indicative of the catagen phase.
   Hair follicles were reduced in size and in a resting state at P20,
   which is a characteristic of the telogen phase. Therefore, we,
   respectively, define P5, P16, and P20 as the anagen, catagen, and
   telogen phases. These observations were consistent with the findings of
   Müller-Röver et al. [[93]21] ([94]Figure 7A).

Figure 7.

   [95]Figure 7
   [96]Open in a new tab

   Identification of the hair follicle growth cycle and the effects of
   Fgf21 on target gene expression. (A) Histological staining (H&E)
   showing the hair follicle structures of dorsal skin from P5, P16, and
   P20 mice. Scale bar = 200 μm. (B) Immunofluorescence staining showing
   the localization of BCL2 (green) in hair follicles of P5, P16, and P20
   mice, with DAPI (blue) used for nuclear staining. Scale bar = 100 μm.
   (C) Immunofluorescence staining showing the localization of Ki67
   (green) in hair follicles of P5, P16, and P20 mice, with DAPI (blue)
   used for nuclear staining. Scale bar = 100 μm. (D) Bar chart showing
   the quantification of BCL2 and KI67 fluorescence intensity in WT mice
   at P5, P16, and P20. (E) Immunofluorescence staining showing the
   localization of VEZF1 (green) in hair follicles of P5, P16, and P20 WT
   and Fgf21^−/^− mice, with DAPI (blue) used for nuclear staining. Scale
   bar = 100 μm. (F) Bar chart showing the quantification of VEZF1
   fluorescence intensity in WT and Fgf21^−/^− mice at P5, P16, and P20.
   (G) Immunofluorescence staining showing the expression of MAP3K1
   (green) in hair follicles of P5, P16, and P20 WT and Fgf21^−/^− mice,
   with DAPI (blue) used for nuclear staining. Scale bar = 100 μm. (H) Bar
   chart showing the quantification of MAP3K1 fluorescence intensity in WT
   and Fgf21^−/^− mice at P5, P16, and P20. (I) RT-qPCR analysis of Vezf1
   expression in dorsal skin tissues from P5, P16, and P20 mice. (J)
   RT-qPCR analysis of Map3k1 expression in dorsal skin tissues from P5,
   P16, and P20 mice. P5, P16, and P20 correspond to postnatal days 5, 16,
   and 20 in mice, representing the anagen phase, catagen phase, and
   telogen phase of the hair follicle cycle, respectively. White arrows
   have been included to highlight protein expression sites throughout the
   hair follicle cycle. Each group included three mice, and the
   experiments were independently repeated three times.

   Immunofluorescence staining was performed to detect the expression of
   the apoptosis marker BCL2 and the proliferation marker KI67, with
   quantification via ImageJ 1.45. During the anagen phase, KI67 exhibited
   the highest expression levels in hair follicle stem cells and
   throughout the hair follicle, while BCL2 was significantly expressed in
   the outer root sheath. In the catagen phase, KI67 expression was
   attenuated, and BCL2 expression was highest in both the outer and inner
   root sheaths. During the telogen phase, the expression levels of both
   KI67 and BCL2 were low ([97]Figure 7B–D).

   The expression levels and protein localization of Vezf1 and Map3k1
   during the natural growth cycle of hair follicles in WT and Fgf21^−/^−
   mice were examined using immunofluorescence staining and quantified via
   Image J. During the anagen phase, VEZF1 and MAP3K1 were expressed in
   the outer root sheath and inner root sheath of hair follicles. During
   the catagen phase, VEZF1 was expressed in the hair shaft, inner root
   sheath, and outer root sheath of hair follicles in Fgf21^−/^− mice,
   with higher expression levels than WT mice. MAP3K1 was expressed in the
   outer and inner root sheaths of both WT and Fgf21^−/^− mice. During the
   telogen phase, VEZF1 was predominantly expressed in the hair shaft of
   WT mice, whereas MAP3K1 was mainly expressed in the outer root sheath
   and the hair shaft of WT mice ([98]Figure 7E–H).

   RT-qPCR analysis revealed that during the anagen phase, Vezf1 and
   Map3k1 expression levels were not significantly different between WT
   and Fgf21^−/^− mice (p > 0.05). However, in the catagen phase, Vezf1
   expression in Fgf21^−/^− mice was 2.27 times higher than in WT mice,
   while Map3k1 expression level was 8.3 times higher (p < 0.05). During
   the telogen phase, Vezf1 expression levels in Fgf21^−/^− mice were 0.51
   times lower than those in WT mice ([99]Figure 7I,J), while the Map3k1
   expression level was 0.5 times lower (p < 0.005). These results
   indicate that Fgf21 knockout alters the expression patterns, including
   the localization and expression levels, of VEZF1 and MAP3K1 during the
   hair follicle growth cycle, suggesting that Fgf21 may regulate hair
   follicle growth and cycle transition by modulating the expression of
   these genes.

4. Discussion

   The hair follicle growth cycle, comprising the anagen, catagen, and
   telogen phases, is a complex and dynamic physiological process
   regulated by multiple signaling pathways. The FGF family plays crucial
   roles in the regulation of hair follicle growth and development.
   Specifically, the interaction between FGF and Wnt signaling pathways in
   the dermal papilla regulates the duration of the anagen phase through
   positive and negative feedback loops [[100]12]. Fgf7, secreted by
   dermal papilla cells (DPCs), promotes the proliferation of DPCs and
   hair follicle stem cells via the Wnt signaling pathway and induces hair
   follicle stem cell (HFSC) differentiation [[101]22]. Additionally, Fgf1
   and Fgf2 promote hair growth, while Fgf5 induces catagen formation and
   reduces hair length. Fgf12 is expressed in the outer root sheath cells
   during anagen, and its knockdown delays the transition from telogen to
   anagen in mice and reduces the length of whisker hair follicle, whereas
   overexpression of Fgf12 increases the migration of the outer root
   sheath cells [[102]23]. Fgf21, a relatively new member of the FGF
   family, produces a secreted protein [[103]24]; however, its role in the
   hair follicle growth cycle has not yet been fully explored. In this
   study, we generated a Fgf21^−/^− mouse model using the CRISPR/Cas9
   system and performed high-throughput miRNA sequencing on the dorsal
   skin tissues of WT and Fgf21^−/^− mice, identifying 93 differentially
   expressed miRNAs. We further selected 10 hair follicle-related miRNAs
   (miR-409-5p, miR-127-3p, miR-134-5p, miR-381-3p, miR-152-3p,
   miR-434-5p, miR-369-3p, miR-136-5p, miR-186-5p, and miR-423-5p) and
   validated their expression using RT-qPCR; the results were consistent
   with the sequencing data. Through synchronized hair follicle cycle
   experiments, we observed a delayed transition from telogen to anagen in
   Fgf21^−/^− mice, indicating a crucial role for Fgf21 in the regulation
   of the hair follicle growth cycle. Liu et al. demonstrated that,
   compared to WT mice, Fgf21^−/^− mice exhibit a slower hair regrowth
   rate, a significant reduction in hair shaft density, and markedly
   decreased expression levels of Erk and Akt [[104]25]. Our findings
   reveal the role of Fgf21 in the hair follicle growth cycle, though the
   specific mechanisms require further investigation.

   Vezf1, a zinc-finger transcription factor, is specifically expressed in
   vascular endothelial cells and plays a critical role in vascular
   development [[105]26]. Liang et al. demonstrated that reduced
   expression of Vezf1 can disrupt the expression of genes related to
   vascular endothelial cells in human induced pluripotent stem cells
   (iPSCs) [[106]27]. Consequently, iPSCs regulated by miR-495 exhibited
   upregulated Vezf1 expression, thereby promoting the proliferation of
   vascular endothelial cells [[107]27]. Kuschnerus et al. found that
   miR-483-3p reduces Vezf1 expression, leading to increased apoptosis of
   vascular endothelial cells and diminished regenerative capacity
   mediated by M2 macrophages [[108]28]. These findings suggest that Vezf1
   may be regulated by multiple miRNAs, forming a complex regulatory
   network. In this study, we identified a targeting relationship between
   miR-134-5p and Vezf1, suggesting that miR-134-5p may influence Vezf1
   expression in a manner distinct from miR-495 and miR-483-3p, thereby
   affecting the proliferation and regenerative capacity of vascular
   endothelial cells. Fgf21 regulates Vezf1 expression during the hair
   follicle cycle. Research shows Vezf1 can indirectly influence the VEGF
   pathway by regulating VEGF receptor expression. For example, Vezf1,
   with Etv2, activates Flt1 transcription [[109]29]. Flt1, a major VEGF
   receptor, when upregulated, boosts VEGF signaling, promoting
   angiogenesis [[110]30]. Vascular endothelial growth factor (VEGF), also
   known as the “vascular permeability factor”, is overexpressed in the
   outer root sheath keratinocytes of hair follicles, promoting the
   formation of perifollicular vasculature and enhancing hair regeneration
   [[111]31]. Our study shows that Fgf21 plays a key role in regulating
   the hair follicle cycle. It likely modulates metabolic and
   proliferative processes via the FGF family, suggesting it may share
   functional similarities with VEGF in promoting hair follicle
   development. We observed that VEZF1 protein was present in the outer
   root sheath of WT mice during telogen, but absent in Fgf21^−/^− mice.
   Since Vezf1 is essential for vascular development in mice [[112]32], it
   may function similarly to VEGF by enhancing blood supply, thereby
   promoting hair follicle growth and development. Miyashita et al. showed
   that Vezf1 regulates the proliferation, formation, and migration of
   endothelial cells during vascular development [[113]33]. Hair
   follicles, as skin appendages, rely on the vascular network of the
   dermal papilla to supply nutrients. In this study, we found that VEZF1
   was expressed in the dermal papilla of WT mice during anagen and
   telogen, whereas no expression was detected in Fgf21^−/^− mice. This
   suggests that Fgf21 deficiency may lead to the absence of Vezf1,
   thereby affecting proliferation and reconstruction of the vascular
   network in the dermal papilla. This, in turn, reduces the nutrient
   supply from the blood to the hair follicle, delaying the transition of
   hair follicles from telogen to anagen.

   Map3k1, a member of the mitogen-activated protein kinase kinase kinase
   (MAP3K) family, plays a crucial role in cell proliferation and
   apoptosis. Research by Andrew Parker et al. has shown that Map3k1
   regulates keratinocyte migration and actin stress fiber formation via
   the JNK signaling pathway, while also independently modulating cell
   proliferation and apoptosis through this pathway [[114]34]. In our
   study, we observed a significant accumulation of MAP3K1 protein in the
   outer and inner root sheaths of hair follicles during the catagen phase
   in both WT and Fgf21^−/^− mice. Based on Parker’s findings, we
   speculated that Map3k1 regulates apoptosis in hair follicle cells via
   the JNK signaling pathway during catagen. Epithelial–mesenchymal
   transition (EMT) is a key cellular process. Chang et al. demonstrated
   that Map3k1 is a target gene of miR-145-5p and that its overexpression
   negatively regulates miR-145-5p expression through the JNK signaling
   pathway, thereby inhibiting EMT in non-small cell lung cancer cells
   [[115]35]. In this study, we validated the targeting relationship
   between miR-136-5p and Map3k1, suggesting that miR-136-5p may
   negatively regulate Map3k1 through the JNK signaling pathway, thereby
   affecting hair follicle growth and development. But it is uncertain
   whether this process depends on the regulation of the Fgf21. Sennett et
   al. demonstrated that the EMT between dermal cells and epithelial stem
   cells influences hair follicle induction and formation [[116]1]. During
   the anagen phase of the hair follicle growth cycle, the interaction
   between mesenchymal dermal papilla cells and epithelial matrix cells
   induces hair shaft formation [[117]1]. The hair shaft is formed by
   rapidly proliferating and differentiating matrix cells that further
   differentiate and migrate to form the inner root sheath, thereby
   determining the shape of the hair [[118]36]. In our study, the MAP3K1
   protein was detected in the hair bulb, outer root sheath, and inner
   root sheath of hair follicles during the anagen phase in both WT and
   Fgf21^−/^− mice. However, during telogen, MAP3K1 expression levels were
   reduced in Fgf21^−/^− mice. This suggests that Fgf21 deficiency may
   lead to decreased Map3k1 expression, thereby affecting the
   differentiation and migration of matrix cells and ultimately
   influencing the transition from telogen to anagen phase in hair
   follicles.

   In summary, our data show that Fgf21 gene regulation leads to
   significant changes in the expression levels and protein localization
   of its target genes, Vezf1 and Map3k1, within hair follicles. However,
   the exact mechanisms by which Vezf1 and Map3k1 regulate hair follicle
   development are still unknown. Consequently, our next study will focus
   on investigating the specific functions of Vezf1 and Map3k1 in hair
   follicles.

5. Conclusions

   Fgf21 knockout delays the transition from telogen to anagen in hair
   follicles. Dual-luciferase assays confirm that miR-134-5p directly
   targets Vezf1 and miR-136-5p targets Map3k1. We found that the
   expression levels and protein localization of the target genes Vezf1
   and Map3k1 in hair follicles were significantly altered by Fgf21
   treatment. The findings demonstrate that Fgf21 modulates miRNA
   expression profiles to regulate hair follicle development, providing
   theoretical insights for hair follicle biology research and the
   clinical management of hair growth disorders ([119]Figure 8).

Figure 8.

   [120]Figure 8
   [121]Open in a new tab

   Model for regulation of hair follicle cycle by Fgf21-related miRNAs.

Acknowledgments