Micro-RNA-31 controls hair cycle-associated changes in gene expression programs of the skin and hair follicle

Abstract

The hair follicle is a cyclic biological system that progresses through stages of growth, regression, and quiescence, which involves dynamic changes in a program of gene regulation. Micro-RNAs (miRNAs) are critically important for the control of gene expression and silencing. Here, we show that global miRNA expression in the skin markedly changes during distinct stages of the hair cycle in mice. Furthermore, we show that expression of miR-31 markedly increases during anagen and decreases during catagen and telogen. Administration of antisense miR-31 inhibitor into mouse skin during the early- and midanagen phases of the hair cycle results in accelerated anagen development, and altered differentiation of hair matrix keratinocytes and hair shaft formation. Microarray, qRT-PCR and Western blot analyses revealed that miR-31 negatively regulates expression of Fgf10, the components of Wnt and BMP signaling pathways Sclerostin and BAMBI, and Dlx3 transcription factor, as well as selected keratin genes, both in vitro and in vivo. Using luciferase reporter assay, we show that Krt16, Krt17, Dlx3, and Fgf10 serve as direct miR-31 targets. Thus, by targeting a number of growth regulatory molecules and cytoskeletal proteins, miR-31 is involved in establishing an optimal balance of gene expression in the hair follicle required for its proper growth and hair fiber formation.—Mardaryev, A. N., Ahmed, M. I., Vlahov, N. V., Fessing, M. Y., Gill, J. H., Sharov, A. A., and Botchkareva, N. V. Micro-RNA-31 controls hair cycle-associated changes in gene expression programs of the skin and hair follicle.

Micro-RNAs (miRNAs) are a class of small (∼22 nt) naturally occurring noncoding RNAs that act as the repressors of gene activity in animals and plants (1–2). MiRNAs negatively regulate gene expression by base pairing of 5′-end (i.e., nt 2–8, the “seed” region) with the 3′ untranslated regions (3′ UTRs) of target messenger RNAs (mRNAs). MiRNAs regulate gene expression post-transcriptionally by interacting with target mRNA in a sequence-specific manner to either impair its stability, translation, or both. This post-transcriptional regulation serves as an important mechanism controlling the expression of many protein-coding genes (1).

In humans, only ∼850 mature miRNAs target more than one-third of protein-encoding mRNAs, as determined by computational prediction analysis (3–5). MiRNAs and their mRNA targets appear to represent remarkably diverse regulatory networks, such that single miRNA can bind to and regulate the expression of up to 200 different mRNA targets, and/or conversely, several different miRNAs can bind to and cooperatively control the expression of a single mRNA target (6–7). These findings indicate that this class of noncoding RNA (ncRNA) molecules constitute a new layer in the regulatory mechanisms controlling gene expression programs in living organisms (8).

MiRNAs have crucial roles in the control of tissue development, differentiation, organogenesis, stem cell activity, growth control, and apoptosis (9–16). Recent findings suggest that miRNAs regulate the protein output of the transcriptome via modulating the levels of their target mRNAs and fine tuning distinct gene expression programs (3, 17). Furthermore, some miRNAs may also target another miRNAs and antagonize their effects on gene expression, thus maintaining the levels of a target protein in the cell (18).

The hair follicle (HF) is a cyclic biological system that progress through stages of growth, regression, and quiescence, each characterized by unique patterns of gene activation and silencing. The period of relative quiescence of the hair cycle (telogen) is characterized by silencing of many genes involved in the control of hair-specific keratinocyte differentiation and by minimal signaling exchange between the epithelial and mesenchymal HF compartments (19–22). The HF transition from the resting to the growth phase (anagen) results in a formation of new hair shaft, and is accompanied by the activation of a large number of signaling pathways controlling the expression of genes encoding hair-specific keratins, components of the inner root sheath, and follicular pigmentation apparatus (23–25). Involution of the HF (catagen) and cessation of hair production is accompanied by silencing of the majority of the genes, which were active during anagen and, in turn, by activation of the gene expression programs to induce apoptosis or promote survival in selected cell populations of the HF (26–28).

During the past few years, it was shown that miRNAs play important roles in the control of skin and HF development. Mice carrying a keratinocyte-specific Dicer deletion, which eliminates the production of all miRNAs in a tissue-specific manner, are characterized by severe alterations in HF morphogenesis, formation of large germ-like cysts, and hyperproliferation of the epidermis (9, 11). A recent study revealed a role of miR-203 in the regulation of epidermal keratinocyte differentiation, at least in part, by direct repression of the p63 expression (29–30). Moreover, miR-200b and miR-196a have been implicated in the control of HF development as potential targets for Wnt signaling pathway, since their expression was reduced in the skin of transgenic mice overexpressing Wnt inhibitor Dkk1 in the epidermis (9).

However, the role of miRNAs in hair cycle-associated changes in the gene expression programs in the skin and the HF remains to be explored. In this study, we examine the global changes in the expression of miRNAs in murine skin during the hair cycle, and define a role for miR-31 in the control of complex program of gene expression during hair cycle.

MATERIALS AND METHODS

Animals and tissue collection

All animal works were performed under the license of the University of Bradford (Bradford, UK) and the Institutional Animal Care and Use Committee protocol of Boston University (Boston, MA, USA). Skin samples were collected from neonatal C57BL/6 mice at postnatal days 12–23 (P12–P23), as well as from 8- to 10-wk-old adult mice. Skin was frozen in liquid nitrogen and embedded, as described elsewhere (31). To induce hair cycle, depilation of the back skin was performed, as described previously (32). Skin was harvested at the telogen stage of the hair cycle (unmanipulated skin), as well as at anagen II [3 days postdepilation (dpd)], anagen IV (5 dpd), anagen VI (8–12 dpd), and catagen (16–19 dpd), using ≥5 mice/time point.

RNA extraction, microarray, and real-time PCR

Total RNA was isolated from snap-frozen tissue samples using miRNeasy kit (Qiagen, Crawley, UK). For miRNA microarray analysis, 5 μg of RNA was isolated from the snap-frozen full-thickness mouse skin collected at P12-P23. MiRNA microarray analysis was performed by LC Sciences, (Houston, TX, USA). mRNA microarray analysis was performed by Mogene Co. (St. Louis, MO, USA) using 41K Whole Mouse Genome 60-mer oligo-microarray (Agilent Technologies, Santa Clara, CA, USA).

Expression of miR-31 was determined using TaqMan real-time PCR Assay (Applied Biosystems, Foster City, CA, USA) and MyiQ single-color real-time PCR detection system (Bio-Rad, Hemel Hempstead, UK) under the following cycling conditions: 95°C for 10 min, followed by 40 cycles of: 95°C for 15 s, and 60°C for 60 s. Differences between samples and controls were calculated using the Genex database software (Bio-Rad) based on the C_t (^ΔΔC_t) equitation method and normalized to the corresponding small nucleolar RNA 202 (SnoRNA) values. Data were pooled, means ± se were calculated, and statistical analysis was performed using unpaired Student's t test.

Quantitative RT-PCR for mRNA was performed with iQ SYBR Green Supermix (Bio-Rad), using 10 ng cDNA and 1 μM primers. PCR primers were designed with Beacon Designer software (Premier Biosoft International, Palo Alto, CA, USA; Table 1). Amplification was performed at the following conditions: 95°C for 5 min, followed by 40 cycles of denaturation (95°C for 15 s), annealing (30 s at temperature experimentally determined for each primer pairs), and elongation (72°C for 15 s). For each gene of interest, qRT-PCR was performed in triplicate. Data analysis was performed as described above.

Table 1.

PCR primers

Accession number	Sequence definition	Sense/antisense primers
NM_026505	BMP and activin membrane-bound inhibitor (BAMBI)	TCCTGTATCTGTTTCCTTCCTGAG ACTGATGGTGGTGACTGTGTAG
NM_010055	Distal-less homeobox 3 (DLX3)	CCAAATCCACTCCTCTCTG GTCTTGCCTGGTCTATCTC
NM_008002	Fibroblast growth factor 10 (FGF10)	CCACCATGCTGAAGTGTGTTAG TTTGAGGATTAGGAGGAGGGAAG
NM_016958	Keratin 14 (Krt1–14)	CCACCTTTCATCTTCCCAATTCTC GTGCGGATCTGGCGGTTG
NM_008470	Keratin 1–16 (Krt1–16)	AATATCCACTCCTCCTCAC GTTGAACCTTGCTCCTTG
NM_010663	Keratin 1–17 (Krt1–17)	ACCTGACTCAGTACAAGCC CCTTAACGGGTGGTCTGG
NM_024449	Sclerostin (Sost)	CGGACCTATACAGGACAAG TAGCCCAACATCACACTC

In situ hybridization

For miR-31 detection on tissue sections, cryosections (10 μm) were fixed in 4% paraformaldehyde for 10 min at room temperature. After acetylation in triethanolamine buffer (4.5 mM triethanolamine, 6 N NCl, and 3 mM acetic anhydride) for 10 min and premobilization (1% Triton X-100/1× DEPC-treated PBS) for 30 min, slides were hybridized with 2.5 pmol DIG-labeled miR-31 probe (Exiqon, Copenhagen, Denmark) diluted in hybridization buffer (50% formamide DI, 2× SCC, 1% dextran sulfate, and 0.4 mg/ml t-RNA) for 16–18 h at 50°C overnight. Slides subsequently were washed in 2× SCC (10 min, 4 times, 67°C), 0.1× SCC (60 min, 67°C), 0.2× SCC (10 min, RT). Immunodetection of miR-31 was performed with sheep alkaline phosphatase conjugated anti-DIG antibody (1:5000; Roche, Mannheim, Germany) followed by a staining reaction with NBT/BCIP solution (Roche) for 16–18 h at room temperature.

Pharmacological experiments and morphometric analyses

Synthetic miR-31 inhibitor (anti-miR-31) or miRIDIAN inhibitor negative control-1 (Dharmacon, Chicago, IL, USA) was administered subcutaneously to back skin of 8-wk-old C57BL/6 mice in concentration 20 μM using atelocollagen for their local and sustained delivery, as described previously (33–34). In the first experiment, anti-miR-31 treatment was performed on dpd 1, 2, 3, and 4, and skin samples were collected on dpd 5. In the second experiment, anti-miR-31 was administrated on dpd 4, 5, 6, and 7, and skin was harvested on dpd 8. In each experiment, ≥4 or 5 mice/time point were used for analyses in both experimental and control groups. Collected samples were processed for histological and morphometric analyses, which were performed using a bright-field microscope, DS-C1 digital camera, and ACT-2U image analysis software (Nikon, Tokyo, Japan).

Cell culture

Primary mouse epidermal keratinocytes (PMEKs) were prepared from newborn mice at P2 or P3, as described previously (35). PMEKs were grown in EMEM calcium-free medium (Lonza, Slough, UK) supplemented with 0.05 mM calcium, at 33°C, 8% CO₂ (Scientific Laboratory Suppliers, Hessle, UK) until 60–70% confluent. PMEKs were transfected with 200 nM synthetic miR-31 inhibitor (anti-miR-31), miR-31 mimic or miRNA negative controls (Dharmacon), using Lipofectamine RNAiMax (Invitrogen, Paisley, UK), according to the manufacturer's protocols. Cells were harvested 24 h after transfection and used for further analyses.

Flow cytometry analysis

Primary mouse epidermal keratinocytes were transfected with 200 nM anti-miR-31, miR-31 mimic, or miRNA negative controls (Dharmacon) as described above. After 24 h, the medium was removed, and cells were washed twice with PBS, and then trypsinized. Pellets were then fixed in 70% ethanol in PBS (at −20°C for 30 min). The cell suspension was centrifuged at 2000 rpm and resuspended in PBS containing propidium iodide (400 μg/ml; Sigma, St. Louis, MO, USA) and RNase A (10 mg/ml; Invitrogen) at 37°C for 30 min. Vials were placed on ice before analysis. Flow cytometry analyses were performed using a FACS-Calibur flow cytometer (BD Biosciences; San Jose, CA, USA). Data obtained were analyzed using the CellQuest software (BD Biosciences).

Western blot analysis

Proteins were extracted from snap-frozen skin samples or cultured cells with lysis buffer, as described previously (36). Five micrograms of protein were processed for Western blot analysis as described previously, followed by incubation with primary antibodies (Table 2) overnight at 4°C. Horseradish peroxidase-tagged IgG antibody was used as secondary antibody (1:5000; Thermo Scientific, Colchester, UK). Antibody binding was visualized with an enhanced chemiluminescence's system (SuperSignal West Pico Kit; Thermo Scientific) and autoradiographed with X-ray film (CL-Xposure Film, Thermo Scientific). Densitometric analysis was performed using Total Lab v1.10 software (Biogenetic Services, Brookings, SD, USA).

Table 2.

Primary antibodies for Western blotting

Antigen	Host	Dilution	Manufacturer
Actin	Mouse	1:2000	Abcam (Cambridge, UK)
FGF10	Goat	1:2000	Abcam
K14	Mouse	1:3000	Biomeda (Foster City, CA, USA)
K16	Rabbit	1:3000	Abcam
K17	Rabbit	1:3000	Abcam
Sost	Rabbit	1:50	Abcam

Immunohistochemistry

For immunohistochemical analyses, formalin-fixed cryostat sections (10 μm) of mouse back skin were used. The cryosections were incubated with primary antisera against Keratin 16 (Epitomics, Burlingame, CA, USA; diluted 1:50), Keratin 17 (Abcam, Cambridge, UK; diluted 1:5000) and Dlx3 (Santa Cruz Biotechnology, Santa Cruz, CA, USA; diluted 1:100) overnight at room temperature, followed by application of corresponding Cy³-labeled antibody (Invitrogen; diluted 1:200) for 45 min at 37°C. Incubation steps were interspersed by four washes with phosphate buffer-saline (PBS, 5 min each). Image preparation and analysis were performed using a fluorescent microscope in combination with DS-C1 digital camera and ACT-2U image analysis software (Nikon).

Luciferase reporter assay

HaCaT cells were grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with heat-inactivated 10% FBS in an atmosphere of 5% CO₂ at 37°C, until 60–70% confluent. 3′UTR fragments containing miR-31 putative target sites were amplified from mouse genomic DNA using forward and reverse primers containing XhoI and NotI restriction sequences, respectively. The following primers were used: Krt14, GGGCTCGAGAGATCCGCACCAAGGTCAT and TTTGCGGCCGCGCAACTCAGAAAAAGAAGC; Krt16, GGGCTCGAGGTCCATCCTCAAGGAGCAAG and TTTGCGGCCGCCCAAAAAGCTTTATTAGCCTACC; Krt17, GGGCTCGAGGCTGCAGAGAGGCAGCTTCCCT and TTTGCGGCCGCCGACGTCTCTCCGTCGAAGGGA; Fgf10, GGGCTCGAGTGACGATCCAAACATAGAAG and TTTGCGGCCGCGCTTTCCAGTAAATGCTTG; Dlx3 1, AATCTCCCTCCCCTTGCTTA and ACACTCTGGCTCCCATTTTG; Sost, GGGCTCGAGTTTCTACACAACAGTTTAAGG and TTTGCGGCCGCATTAACAATGCCTCTGGTC.

These amplified fragments were cloned at XhoI and NotI sites downstream of CV40 promoter-driven Renilla luciferase cassette in pCHECK2 (Promega, Madison, WI, USA). For dual luciferase assay, these constructs (200 ng) were cotransfected with 200 nM miR-31 mimic or negative control mimic (Dharmacon) into HaCaT cells using 0.5 μl Lipofectamine 2000 (Invitrogen) in 96-well plates. At 24 h after transfection, the relative luciferase activities were determined using Dual-Glo Luciferase Assay System (Promega).

RESULTS

Hair cycle-associated changes in the miRNA signatures in the skin

To identify the changes in the miRNA expression in the skin and HF during distinct stages of the hair cycle in mice (i.e., during anagen, the period of active growth and hair production; catagen, the stage of apoptosis-driven involution; and telogen, the period of relative resting), global microarray analysis of miRNAs was performed as described previously (37). Total RNA was isolated from the skin of neonatal mice at P12–P15, i.e., when HFs complete morphogenesis and actively produce hair (anagen-like stage), as well as at P16–P17 and P20–P23 (i.e., during the catagen and telogen stages, respectively) (31).

Global miRNA expression profiling revealed the substantial hair cycle-associated changes in the expression of a large number of miRNAs in the skin (Fig. 1A, B). In particular, expression of 219 of 568 miRNAs analyzed showed significant (P<0.01) differences between the distinct hair cycle stages (Table 3). Among these miRNAs, the largest proportion showed significant changes in expression between the anagen and telogen skin, while a relatively lower number of miRNAs displayed differences in expression between anagen and catagen, as well as between the catagen and telogen stages of the hair cycle (Fig. 1B, C).

Figure 1.

Global miRNA expression profile in mouse skin at distinct hair cycle stages. A) Heat map represents differentially expressed miRNAs between distinct stages of the hair cycle. Color map is used to visualize the difference in expression. B) Differences in the expression of miRNAs between distinct hair cycle stages. C) Heat map represents a cluster of miRNAs with significantly (P<0.01) up-regulated expression in anagen, compared to catagen and telogen skin.

Table 3.

Differentially expressed miRNAs in distinct hair cycle stages

MiRNA	Anagen	Catagen	Telogen	P value
mmu-let-7a	34,080	37,282	28,616	4.99E-05
mmu-let-7b	26,775	28,245	21,593	3.97E-03
mmu-let-7c	30,679	34,524	25,817	1.35E-04
mmu-let-7d	26,886	26,308	21,365	2.89E-03
mmu-let-7e	12,254	7,611	6,136	2.32E-03
mmu-let-7f	29,618	27,453	22,636	5.85E-04
mmu-let-7g	6,762	8,690	10,316	1.03E-04
mmu-let-7i	8,921	12,732	12,953	1.04E-06
mmu-miR-1	10,182	16,405	28,682	2.09E-08
mmu-miR-100	226	259	633	1.11E-16
mmu-miR-103	2,484	2,053	884	8.74E-09
mmu-miR-106a	308	322	193	1.88E-02
mmu-miR-107	2,235	1,703	793	1.40E-08
mmu-miR-10a	643	450	313	2.66E-04
mmu-miR-10b	1,871	1,807	2,449	1.29E-03
mmu-miR-125a-5p	2,408	2,538	3,562	2.22E-05
mmu-miR-125b-5p	9,991	12,985	16,895	1.02E-10
mmu-miR-126-3p	2,239	5,403	7,564	8.03E-13
mmu-miR-127	1,386	1,974	2,282	1.02E-03
mmu-miR-128a	275	287	453	1.02E-08
mmu-miR-130a	180	282	171	1.17E-06
mmu-miR-133a	1,450	2,493	4,025	1.11E-16
mmu-miR-133b	1,214	2,089	3,611	0.00E+00
mmu-miR-140*	669	341	205	2.78E-15
mmu-miR-143	941	2,705	3,587	4.02E-13
mmu-miR-145	5,257	3,950	3,341	1.18E-07
mmu-miR-146a	78	169	678	4.44E-16
mmu-miR-148a	250	854	2,796	2.48E-12
mmu-miR-150	159	215	581	1.66E-10
mmu-miR-151-5p	1,382	970	924	2.87E-06
mmu-miR-152	873	1,359	3,092	0.00E+00
mmu-miR-155	618	45	25	0.00E+00
mmu-miR-15a	875	649	989	6.91E-04
mmu-miR-15b	6,140	2,928	2,384	2.81E-09
mmu-miR-16	14,023	11,139	9,687	2.58E-02
mmu-miR-17	1,780	2,100	877	2.34E-04
mmu-miR-181a	1,253	2,207	1,267	4.99E-07
mmu-miR-182	942	666	501	1.07E-06
mmu-miR-183	1,855	935	482	2.32E-13
mmu-miR-191	3,454	3,012	1,780	2.00E-15
mmu-miR-195	4,159	3,828	6,763	0.00E+00
mmu-miR-199a-3p	2,141	6,754	14,750	0.00E+00
mmu-miR-199a-5p	207	363	641	1.55E-13
mmu-miR-19b	195	296	117	5.05E-06
mmu-miR-200a	144	842	809	4.34E-13
mmu-miR-200b	8,130	4,745	2,703	1.73E-10
mmu-miR-200c	11,342	7,148	3,421	0.00E+00
mmu-miR-203	57,573	59,271	32,593	4.44E-16
mmu-miR-205	13,774	21,404	14,430	3.63E-04
mmu-miR-20a	2,290	2,601	1,246	3.66E-04
mmu-miR-21	4,219	4,299	9,121	1.72E-06
mmu-miR-214	4,088	3,488	4,671	1.15E-04
mmu-miR-22	231	484	535	4.38E-05
mmu-miR-221	297	411	317	8.37E-06
mmu-miR-222	442	435	297	1.55E-04
mmu-miR-223	473	25	166	4.29E-09
mmu-miR-24	4,791	9,292	10,421	3.17E-07
mmu-miR-25	3,795	2,616	1,680	3.85E-14
mmu-miR-26a	17,633	26,562	26,411	1.30E-07
mmu-miR-26b	3,627	3,544	6,210	7.88E-05
mmu-miR-27a	1,744	5,366	7,887	1.13E-11
mmu-miR-27b	2,699	8,388	9,501	1.86E-12
mmu-miR-295*	5	62	4	1.37E-03
mmu-miR-29a	309	776	2,669	0.00E+00
mmu-miR-30a	445	1,226	2,074	6.16E-13
mmu-miR-30b	6,269	3,227	3,908	1.05E-04
mmu-miR-30c	7,827	4,044	3,683	1.61E-07
mmu-miR-30d	943	1,205	1,717	4.73E-12
mmu-miR-30e	95	244	514	2.11E-13
mmu-miR-31	1,136	683	68	0.00E+00
mmu-miR-320	5,704	4,424	3,131	1.93E-10
mmu-miR-322	101	115	388	5.57E-10
mmu-miR-335–5p	59	184	237	1.08E-02
mmu-miR-361	1,292	711	621	1.44E-10
mmu-miR-377	98	119	36	5.35E-05
mmu-miR-378	215	462	511	1.21E-14
mmu-miR-379	1,042	849	1,439	1.17E-10
mmu-miR-382	395	235	371	9.26E-05
mmu-miR-423-5p	688	413	345	1.77E-06
mmu-miR-429	2,107	1,808	1,089	4.88E-07
mmu-miR-434-3p	296	465	991	8.07E-13
mmu-miR-466f-3p	364	2,133	232	8.58E-04
mmu-miR-466g	79	734	49	1.01E-02
mmu-miR-467a*	31	187	20	2.28E-03
mmu-miR-467b*	47	372	22	9.44E-04
mmu-miR-486	249	294	667	5.49E-12
mmu-miR-669c	302	191	98	1.48E-04
mmu-miR-674	359	502	243	0.00E+00
mmu-miR-676	425	310	250	4.49E-09
mmu-miR-689	189	313	1,389	0.00E+00
mmu-miR-690	8,249	2,060	1,247	1.68E-08
mmu-miR-705	5,692	3,432	4,784	5.89E-05
mmu-miR-709	41,199	27,175	29,106	2.83E-11
mmu-miR-720	871	2,393	399	2.27E-11
mmu-miR-762	6,228	4,776	5,657	4.81E-02
mmu-miR-805	503	98	118	4.53E-07
mmu-miR-92a	5,013	3,545	2,200	9.70E-14
mmu-miR-92b	1,488	850	681	2.96E-05
mmu-miR-93	572	848	378	4.18E-04
mmu-miR-98	2,000	540	715	3.10E-04
mmu-miR-99a	602	624	989	9.17E-07
mmu-miR-99b	1,094	1,206	704	9.70E-06

In comparison to telogen and catagen, the anagen phase of the hair cycle is characterized by most dramatic changes in expression of a large number of genes in the HF and in other structures of the skin (21, 24), which implicates a role for miRNAs in regulation of their expression. For further analyses of the roles of miRNAs in the control of hair cycle-associated gene expression programs in the skin, we selected miR-31, as its expression showed the most remarkable changes between the anagen and catagen/telogen stages of the hair cycle (Table 3).

miR-31 expression in the skin is markedly increased during anagen and decreased in catagen and telogen

Microarray data validated by qRT-PCR showed very high levels of miR-31 transcripts in neonatal skin during the anagen-like stage at P12, while during catagen (P17–P19), its expression progressively decreased and remained at very low levels during telogen (P20–P23; Fig. 2A). Similar changes in the miR-31 levels were observed in adult skin during the depilation-induced hair cycle: miR-31 expression progressively increased during HF transition from telogen to anagen and reached maximum at late anagen stage (dpd 12), followed by rapid decrease during catagen (dpd 16–19) (Fig. 2B).

Figure 2.

Spatiotemporal expression of miR-31 during hair cycle. A) Detection of miR-31 in neonatal skin by qRT-PCR: it is expressed maximally during the anagen-like stage (P12), while its expression progressively decreased during catagen (P17–P19) and remained at very low levels during telogen (P20–P23). B) Detection of miR-31 in the skin during depilation-induced hair cycle: miR-31 expression progressively increased during HF transition from telogen to anagen and reached a maximum at late anagen stage (d 12 after depilation), followed by rapid decrease during catagen (d 16–19 after depilation). C–F) Representative photomicrographs of in situ hybridization for miR-31 in HF at different hair cycle stages. C) Lack of miR-31 expression in telogen. D) miR-31 expression in the epidermis (large arrow) and in the growing hair bulb of midanagen HF (arrowhead). E) Prominent miR-31 expression in the hair matrix (arrowhead), outer and inner root sheaths (small and large arrows), and its lower expression in the dermal papilla in late anagen HFs (asterisk). F) Low miR-31 expression in the epithelium of catagen HF (arrow). **P < 0.01.

Consistently with microarray and qRT-PCR data, in situ hybridization for miR-31 showed lack of its expression in telogen skin (Fig. 2C). miR-31 expression appeared first in the epidermis and in the growing hair bulb of midanagen HF on dpd 5 (Fig. 2D). In late anagen HFs (dpd 12), miR-31 was prominently expressed in the hair matrix, as well as in the outer and inner root sheaths, while relatively lower expression was seen in the dermal papilla, and lack of expression was seen in the dermis (Fig. 2E). During catagen, miR-31 expression progressively decreased (midcatagen; Fig. 2F) and disappeared completely from the entire skin during telogen (data not shown). These data suggested a role for miR-31 in the control of gene expression programs in the HF and epidermal keratinocytes during the anagen phase of the hair cycle.

Inhibition of miR-31 activity in the skin accelerates anagen development and alters hair shaft formation

To explore the role of miR-31 in the control of hair cycle, a synthetic inhibitor designed to specifically bind to and block the miR-31 activity was administered into the back skin of 10- to 11-wk-old mice at different time-points of the depilation-induced hair cycle. According to the experimental approaches established previously (38), the efficiency of anti-miR-31 in inhibiting the miR-31 activity was assessed by analyzing the expression of miR-31 in the treated and control samples using TaqMan miRNA assay and qRT-PCR. These experiments showed significant decrease of the miR-31 expression in the skin treated with miR-31 inhibitors vs. the control (data not shown).

In the first experiment, anti-miR-31 was administered daily into mouse back skin during the early anagen (d 1–4 of the depilation-induced hair cycle), and skin was harvested at d 5 after the beginning of the experiment (Fig. 3A). Inhibition of miR-31 activity during the early anagen resulted in acceleration of anagen progression compared to the control (Fig. 3B–D). In mice treated with anti-miR-31, significantly more HFs were found in anagen IV stage (P<0.05), characterizing by larger and more pigmented hair bulbs, whereas the majority of HFs in the control skin reached only anagen III phase of the hair cycle (Fig. 3B–D). Acceleration of anagen development in mice treated with anti-miR-31 vs. the controls was also associated with a significant (P<0.05), morphologically recognizable, increase in the skin thickness (Fig. 3E), used as an additional well-established parameter of the hair cycle progression in mice (39).

Figure 3.

Inhibition of miR-31 accelerates early anagen development and alters HF morphology. A–E) Hair cycle was induced by depilation in the back skin of 10-wk-old C57BL/6 mice; anti-miR-31 or vehicle control were administered daily subcutaneously at dpd 1–4. Skin was harvested at dpd 5 (A). C–D) Representative skin examples of control (B) and anti-miR-31-treated mice (C) at d 5; sections were stained for the detection of endogenous alkaline phosphatase activity to visualize the morphology of dermal papilla as an important indicator of the defined stages in HF cycle. D) Percentage of HFs in defined stages of anagen was evaluated in cryostat sections of the skin of control or anti-miR-31-treated mice by quantitative histomorphometry using established morphological criteria (32); there was a significant increase in the percentage of HFs in anagen IV stage in anti-miR-31-treated skin, compared with the control. E) Skin thickness after anti-miR-31 treatment is significantly enhanced, compared to the control. F–I) Hair cycle was induced by depilation in the back skin of 10-wk-old C57BL/6 mice; anti-miR-31 or vehicle control were administered daily subcutaneously at dpd 4–7. Skin was harvested at dpd 8 (F). H) Representative skin examples of control and anti-miR-31-treated mice (G–I) at d 8. G) Representative skin example of control skin showing HFs with anagen VI morphology. H) Microphotograph of HFs that received anti-miR-31 treatment, depicting hair shaft deformation. I) Hyperplastic and deformed outer root sheath of the HF after anti-miR-31 treatment. *P < 0.05.

In the second experiment, mice were treated with anti-miR-31 during the midanagen phase of the hair cycle (dpd 4–7), and skin samples were collected on d 8 of the experiment (Fig. 3F). Administration of anti-miR-31 during midanagen did not cause any significant changes in the rate of hair cycle progression, and HFs in both experimental and control skin reached anagen VI stage. However, HFs treated with anti-miR-31 were characterized by larger hair bulbs, altered hair shaft structure with irregular distribution of melanin, and hyperplastic and deformed outer root sheath compared to the control (Fig. 3G–I).

These results were consistent with data demonstrating high expression levels of miR-31 in anagen HFs (Fig. 2B) and suggested a complex role for miR-31 in the control of anagen development and hair production: during early anagen, miR-31 serve as a negative regulator anagen progression, while during mid- to late-anagen, miR-31 is involved in the control of keratinocyte differentiation in the hair matrix and hair shaft formation.

Modulation of the miR-31 activity induces complex changes in gene expression program in primary keratinocytes

To further explore mechanisms underlying the effects of miR-31 on hair cycle, primary mouse epidermal keratinocytes were transfected either with anti-miR-31 or with miR-31 mimic to inhibit or enhance miR-31 activity, respectively. Transfection efficiency was assessed by qRT-PCR, which revealed significant decrease and increase in the miR-31 levels in the keratinocytes after treatment with anti-miR-31 or miR-31 mimic, respectively, vs. the controls (data not shown).

To assess whether miR-31 shows any effects on cell proliferation, keratinocytes were stained by propidium iodide and FACS analysis was performed. FACS quantification did not reveal any significant changes in cell cycle progression as a result of either miR-31 inhibition or activation, suggesting that miR-31 is not involved in the regulation of keratinocyte proliferation (Fig. 4A).

Figure 4.

Complex changes in gene expression program in primary keratinocytes and in the skin due to miR-31 inhibition. A) No changes in cell cycle progression through the different phases as a result of either miR-31 inhibition or activation were detected by FACS analysis. B) Microarray analysis of the global gene expression in keratinocytes transfected with anti-miR-31 or control cells: functional assignments of the genes with altered expression due to anti-miR-31 treatment. C) Elevated expression of FGF10, BAMBI, and SOST transcripts in anti-miR-31-treated skin vs. the control was detected in the extracts of the full-thickness skin by qRT-PCR. D) Western blotting assay of the extracts of the full-thickness skin either treated with anti-miR-31 or with vehicle control to detect FGF10 and SOST proteins: up-regulation of FGF10 and SOST proteins was detected in anti-miR-31-treated samples vs. the control confirmed by the densitometry. E) Detection of Dlx3 transcripts by qRT-PCR showed its up-regulation in anti-miR-31-treated skin vs. the control. F, G) Analysis of Dlx3 expression by immunofluorescence. In the control, Dlx3 is expressed in the hair matrix (large arrow), inner root sheath (small arrow), and in the hair shaft (asterisk) (F); In anti-miR-31 skin, in addition to the hair matrix (large arrow), inner root sheath (small arrow), and in the hair shaft (asterisk), Dlx3 is present in the outer root sheath (arrowhead) (G). *P < 0.05; **P < 0.01.

However, microarray analysis of the global gene expression in keratinocytes treated with anti-miR-31 or control cells revealed that miR-31 inhibition resulted in 2-fold or higher changes in expression of 419 genes that encode distinct adhesion molecules, components of the cytoskeleton, metabolic enzymes and growth factors/receptors/signaling molecules involved in the control of cell fate decision and differentiation (Fig. 4B; Supplemental Table 1). Functional assignments of the genes the expression of which were altered in keratinocytes after anti-miR-31 treatment, revealed that 53% of them represent distinct signaling and growth regulatory molecules or transcription factors, thus suggesting a potential involvement of miR-31 in the control of keratinocyte responsiveness to growth factor stimulation/inhibition during anagen development and hair cycle-associated tissue remodeling.

miR-31 regulates expression of the distinct components of the FGF, BMP, and Wnt signaling pathways and the Dlx3 transcription factor in the skin

Expression of the selected genes whose expression in keratinocytes was changed after anti-miR-31 treatment (Supplemental Table 1) and which are implicated in the control of hair cycle and keratinocyte differentiation (fibroblast growth factor 10 (Fgf10), bone morphogenetic protein, and activin membrane-bound inhibitor (Bambi), Sclerostin (Sost), and distal-less homeobox 3 (Dlx3) (21–22, 40) was further examined in mouse back skin treated with anti-miR-31 (Fig. 3). Consistent with microarray data, the anti-miR-31 treatment resulted in increased expression of transcripts for Fgf10, Bmp pathway inhibitor Bambi, Wnt, and Bmp signaling antagonist Sost (Fig. 4C). Western blot analysis of the full-thickness skin samples obtained after anti-miR-31 treatment in vivo also showed that Fgf10 and Sost protein levels were increased compared to the controls (Fig. 4D). Expression of the Dlx3 transcription factor was also increased in primary keratinocytes treated with anti-miR-31 (Fig. 4E). By immunofluorescence analysis of the skin treated with anti-miR-31 in vivo, ectopic Dlx3 expression in the HF outer root sheath and increase of the Dlx3 expression in the hair matrix, inner root sheath, and hair shaft were seen (Fig. 4G) vs. the controls (Fig. 4F). These data suggest that the effects of the anti-miR-31 on anagen progression and hair shaft formation (Fig. 3) were executed, at least in part, via modulation of the activity of the Fgf, Bmp, and Wnt pathways and expression of the Dlx3 transcription factor in the HFs.

miR-31 controls the expression of keratin 14, 16, and 17 (Krt14, Krt16, and Krt17) in the HF

In addition to the changes in expression of genes that encode signaling/regulatory molecules and transcription factors implicated in hair cycle control, microarray analysis revealed that expression of several keratin genes has also been affected after inhibition of the miR-31 activity (Fig. 5, Supplemental Table 1). Transcripts for Krt14, Krt16, and Krt17 were increased in the keratinocytes treated with anti-miR-31, compared to the controls (Fig. 5A). Treatment of the keratinocytes with anti-miR-31 also resulted in increase in the levels of K14, K16, but not K17 proteins (Fig. 5B). Moreover, administration of anti-miR-31 into mouse back skin also resulted in up-regulation of K14, K16, and K17 expression. Elevated levels of Krt14, Krt16, and Krt17 transcripts and corresponding proteins were detected in the extracts of the full-thickness skin treated with anti-miR-31 vs. the controls, as determined by qRT-PCR and Western blot analysis, respectively (Fig. 5C, D).

Figure 5.

miR-31 regulates expression of keratins in the primary epidermal keratinocytes and in mouse skin. A, B) K14, K16, and K17 expression in the primary keratinocytes transfected with either anti-miR-31 or control oligonucleotides: qRT-PCR analysis showing a significant up-regulation of K14, K16, and K17 mRNA levels in the keratinocytes transfected with anti-miR-31 (A); Western blot analysis demonstrating increased levels of K14, K16, and K17 proteins after transfection with anti-miR-31 confirmed by the densitometric analysis (B). C, D) Expression of K14, K16, and K17 in skin treated either with anti-miR-31 or with vehicle control: a significant up-regulation of K14, K16, and K17 mRNA in skin that received anti-miR-31 treatment vs. control was detected by qRT-PCR (C). Western blot analysis shows elevated K14, K16, and K17 protein levels after anti-miR-31 treatment, confirmed by the densitometric analysis (D). E, F) K17 detection by immunofluorescence in the control and anti-miR-31 treated skin, respectively. Increase in K17 expression in the unilateral disk (small arrow) and in the opposite side of the hair matrix (large arrow) in the treatment vs. the control. Prominent K17 expression in the hair shaft (arrowhead) and in the thickened outer root sheath (asterisk) in anti-miR-31 treated skin (F) compared to the hair shaft (arrowhead) and outer root sheath (asterisk) in the control (E). G, H) K16 immunofluorescence in the HF: increased immunoreactivity of K16 in the outer root sheath and companion layer of the HF after anti-miR-31 treatment (H, small arrows and large arrow, respectively), compared to the controls (G). *P < 0.05; **P < 0.01.

K16 and K17 expression by immonofluorescence showed marked alterations in the HFs treated with anti-miR-31 vs. the control (Fig. 5E, F). Consistent with data reported previously (41–42), K17 immunoreactivity was detected in the unilateral cluster of hair matrix cells, as well as in the outer root sheath and in the medulla of the hair shaft of control HFs (Fig. 5E). However, anti-miR-31 administration resulted in the increase of K17 expression in the unilateral disk and in the opposite side of the hair matrix, as well as in the hair shaft and thickened outer root sheath vs. the controls (Fig. 5F). K16 expression also increased in the outer root sheath and companion layer of the HF after anti-miR-31 treatment compared to the controls (Fig. 5G, H).

Krt16, Krt17, Dlx3, and Fgf10 are direct targets of miR-31

Bioinformatic analysis using RNA22 algorithm, a method for identifying miRNA binding sites and their corresponding heteroduplexes (43), revealed that Fgf10, Sost, Dlx3, Krt14, Krt16, and Krt17 carry several putative miR-31 binding sites (Fig. 6A). To validate whether miR-31 directly regulates expression of these genes, the effects of miR-31 on the relevant 3′UTR constructs were tested using a luciferase reporter assay. We judged the repression of average luciferase activity ≥30% as a significant effect of miRNA on gene expression (43).

Figure 6.

Krt16, Krt17, Dlx3, and FGF10 are primary targets of miR-31. A) Predicted interactions between miR-31 and Krt14, Krt16, Krt17, Dlx3, and Fgf10 mRNA. Alignment of mouse sequences in the 3′-UTR of Krt14, Krt16, Krt17, Dlx3, and Fgf10 mRNA. Representation is limited to the region around the miR-31 complementary site. B) Cotransfection of HaCaT cells with miR-31 mimic and the Krt16 3′UTR, Krt17 3′UTR, and Dlx3 3′UTR 3′UTR constructs encompassing putative target sites caused >50% reduction in normalized luciferase activity, compared to the corresponding controls. About 30% suppression in the luciferase activity has also been observed after cotransfection of the cells with miR-31 mimic and FGF10 3′UTR reporter construct, compared to the control. miR-31 mimic did not cause significant reporter inhibition with constructs containing segments of the 3′UTR of Krt14 mRNA.

Cotransfection of HaCaT cells with miR-31 mimic and the Krt16 3′UTR reporter construct caused >50% reduction in luciferase activity, compared to the corresponding control (Fig. 6B). A single-mismatched hexamer found within the 3′UTR of Krt17 mRNA decreased luciferase activity by 60% when it was cotransfected with miR-31 mimic (Fig. 6B). Substantial suppression in the luciferase activity has also been observed after cotransfection of the cells with miR-31 mimic and Dlx3 3′UTR reporter construct, compared to the control (Fig. 6B). A putative site within the 3′UTR of Fgf10 mRNA introduced into the luciferase reporter construct has also demonstrated sensitivity to miR-31 in the cotransfection assay (Fig. 6B). However, the reporter constructs containing segments of the 3′UTR of Krt14 or Sost mRNAs have not demonstrated significant sensitivity to miR-31 in the cotransfection experiments (Fig. 6B, data not shown). Collectively, these data indicate that Krt16, Krt17, Dlx3, and Fgf10, but not Krt14 or Sost transcripts might represent genuine targets of miR-31.

DISCUSSION

HF cycling is a unique biological phenomenon that is accompanied by the profound changes in the skin and HF microanatomy and in pigmentation, as well as by the remodeling of the cutaneous vascular and nervous apparatus (21, 25, 44–47). We show here that each stage of the hair cycle is characterized by the distinct patterns of the miRNA profiles in the skin (Fig. 1), thus suggesting miRNAs as an important regulatory layer in the complex program controlling hair cycle-associated changes of gene expression in the skin and the HF.

Remarkable fluctuations in the expression levels of distinct miRNAs during hair cycle suggest that these miRNAs may be involved in at least 3 distinct mechanisms of the control of gene expression proposed for miRNAs previously: 1) silencing of the distinct gene expression programs during the HF transition between hair cycle stages (i.e., switch between the telogen-associated growth inhibitory programs and hair growth-promoting programs during the telogen-anagen transition or between the hair growth-associated programs to proapoptotic programs during transition of the HFs from anagen to catagen); 2) establishing the optimal levels of the transcripts to prevent their overexpression, thus modulating a balance of activity of distinct signaling pathways (i.e., limiting an excessive activation of procarcinogenic signaling pathways, such as Wnt or Hedgehog, during active HF growth and hair production); and 3) control of the expression of the other miRNAs to limit their inhibitory effects on the levels of the target transcripts (i.e., similar to the antagonistic interactions between the miR-184 and miR-205 described previously (18).

Although the function of the distinct miRNAs or their clusters in the control of HF cycling remains to be further defined, we provide here the evidence that miR-31 plays an important role in the control of anagen-associated gene expression programs in the HF. We show that miR-31 expression is markedly increased in the skin during anagen phase of the hair cycle, and its expression is seen in both the epidermal and follicular epithelial skin compartments, as well as in the dermal papilla (Fig. 2). Furthermore, we demonstrate that inhibition of the miR-31 activity in the skin results in anagen acceleration and in alterations in the hair shaft formation and outer root sheath morphology (Fig. 3).

However, we observed that changes in miR-31 expression levels using either anti-miR-31 or miR-31 mimic do not affect proliferation in vitro (Fig. 4). Our findings are consistent with recently reported data showing that overexpression of miR-31 in aggressive breast cancer cells does not affect their proliferation when compared to normal mammary epithelial cells (48). This suggests that elevated expression of miR-31 in cells with high proliferative activity is possibly required to prevent alterations in growth factor signaling leading to tumor initiation or progression.

We show here that the inhibitory effects of miR-31 on anagen development are likely to be realized, at least in part, via modulation of the activity of the Fgf, Wnt, and Bmp signaling pathways. By analyzing the global gene expression changes in primary keratinocytes treated with anti-miR-31, we demonstrate that miR-31 is involved in regulating the expression of the several components of these pathways, such as the FGF receptor ligand Fgf10, the Wnt and Bmp inhibitor Sost, and the Bmp antagonist Bambi (Fig. 4). Moreover, Fgf10 might serve as one of the direct targets of miR-31 (Fig. 6).

Fgf and Wnt pathways promote the HF telogen-anagen transition via providing the stimulatory signals to the HF stem cells and/or their progenies residing in the HF bulge and secondary hair germ, while Bmp signaling operates as an anagen inhibitor antagonizing the activity of the Fgf and Wnt pathways during telogen (49–52). Sost is capable of antagonizing the Wnt and Bmp pathways in osteoblasts and adipocytes; however, it can also directly bind Noggin, a potent Bmp antagonist, stimulator of telogen-anagen transition and modulator of hair shaft pigmentation (23, 49, 53–54).

Thus, by regulating the expression of the Fgf10, Bambi, and Sost miR-31 may be involved in the fine-tuning the activity of the Fgf, Wnt, and Bmp pathways in distinct subpopulations of hair progenitor cells and in modulating the effects of these pathways on anagen progression and hair shaft pigmentation associated with it.

However, in addition to the effects of miR-31 on early stages of anagen development, we also demonstrate here that inhibition of the miR-31 activity during mid-/late anagen results in alterations in the hair shaft formation accompanied by its thickening, irregular distribution of the pigment, as well as by the hyperplasia of the outer root sheath (Fig. 3). Fgf, Wnt, and Bmp signaling pathways are closely involved in the control of keratinocyte differentiation and hair shaft formation, and genetic alterations in the activity of these pathways results in the distinct abnormalities of the hair shaft formation and pigmentation (20, 22–23).

We provide here evidence that, in addition, to the modulatory effects on the activity of the Fgf, Wnt, and Bmp pathways, miR-31 is also involved in the control of hair shaft formation, at least in part, via regulating the expression of the Dlx3 transcription factor. Expression of the Dlx3 in the skin samples treated with anti-miR-31 significantly up-regulated and, in addition to the expression in the hair matrix, inner root sheath and hair shaft, Dlx3 is ectopically expressed in the outer root sheath (Fig. 4). The luciferase reporter assay confirmed that expression of Dlx3 is directly regulated by miR-31 through its 3′UTR (Fig. 6).

Dlx3 transcription factor plays an important role in the control of the hair matrix keratinocyte differentiation toward the hair shaft and inner root sheath cell lineages (55). Genetic Dlx3 ablation in mice results in failure of proper formation of the hair shaft and inner root sheath leading to complete alopecia (55). Dlx3 also serves as a downstream of the Wnt and Bmp signaling pathways and as an upstream regulator of the Hoxc13 and Gata3 transcription factors that are essential components of the transcription programs controlling the formation of the hair shaft and inner root sheath, respectively (56–59). Thus, elevated Dlx3 expression and its ectopic appearance in the outer root sheath in response to miR-31 inhibition could contribute to the abnormalities in the hair shaft and outer root sheath morphology seen in our study (Fig. 3).

In addition to the changes observed in the expression of distinct signaling molecules, we demonstrate here that miR-31 regulates the expression of the keratins 14, 16, and 17, the essential components of the keratinocyte cytoskeleton. Alterations in the levels of miR-31 induced by miR-31 synthetic inhibitor resulted in up-regulation of the expression of these keratins both in vitro and in vivo (Fig. 5). Moreover, luciferase reporter assay revealed that expression of K16 and K17 are directly regulated by miR-31 through their 3′UTR, whereas K14 are most likely an indirect target of miR-31 (Fig. 6).

We found that alterations of the hair shaft structure and outer root sheath hyperplasia after inhibition of the miR-31 activity in the skin are accompanied by up-regulation of the expression of K17 in the hair matrix keratinocytes, hair shaft medulla, and outer root sheath, as well as by increase of the K16 expression in the companion layer of the outer root sheath (Fig. 5). In addition to the maintenance of keratinocyte cytoskeleton, K17 is capable of influencing keratinocyte growth and size by regulating protein synthesis (60). This suggests that K17 may contribute, at least in part, to the acceleration of anagen development and hyperplastic changes in the outer root sheath in the HF treated by miR-31.

K16 and K17 are expressed in the medulla of the central portion of the hair shaft (42, 61). K16 overexpression under control of the K14 promoter leads to the formation of curly hairs and hyperplasia of the outer root sheath (62). Genetic ablation of the keratin 17 results in hair shaft fragility and strain-dependent alopecia (63), while intradermal administration of chimeric oligonucleotides carrying the mutated α-helical K17 domain leads to marked alterations of the hair shaft morphology and shape (64). These data suggest that the elevated levels of the K16 and K17 may, at least in part, contribute to the alterations in the hair shaft structure seen after anti-miR-31 treatment.

Taken together, our data demonstrate a previously unrecognized role of miR-31 in complex regulation of the gene expression programs that control anagen progression and hair shaft formation in the HF. By targeting a number of growth regulatory molecules, transcription factors and cytoskeletal proteins, miR-31 is involved in establishing an optimal balance of gene expression in the HF required for its proper growth and hair fiber formation. Although many aspects of the miR-31-dependent effects on the HF cycling remain to be clarified, these data will help in further establishing molecular signaling networks that control organ regeneration and raise the possibility of exploring the role of miR-31 in pathobiology of distinct clinical conditions with impaired skin regeneration and hair growth.