A Novel Regulatory Axis, CHD1L-MicroRNA 486-Matrix Metalloproteinase 2, Controls Spermatogonial Stem Cell Properties

Spermatogonial stem cells (SSCs) are unipotent germ cells that are at the foundation of spermatogenesis and male fertility. However, the underlying molecular mechanisms governing SSC stemness and growth properties remain elusive.

ABSTRACT

Spermatogonial stem cells (SSCs) are unipotent germ cells that are at the foundation of spermatogenesis and male fertility. However, the underlying molecular mechanisms governing SSC stemness and growth properties remain elusive. We have recently identified chromodomain helicase/ATPase DNA binding protein 1-like (Chd1l) as a novel regulator for SSC survival and self-renewal, but how these functions are controlled by Chd1l remains to be resolved. Here, we applied high-throughput small RNA sequencing to uncover the microRNA (miRNA) expression profiles controlled by Chd1l and showed that the expression levels of 124 miRNA transcripts were differentially regulated by Chd1l in SSCs. KEGG pathway analysis shows that the miRNAs that are differentially expressed upon Chd1l repression are significantly enriched in the pathways associated with stem cell pluripotency and proliferation. As a proof of concept, we demonstrate that one of the most highly upregulated miRNAs, miR-486, controls SSC stemness gene expression and growth properties. The matrix metalloproteinase 2 (MMP2) gene has been identified as a novel miR-486 target gene in the context of SSC stemness gene regulation and growth properties. Data from cotransfection experiments showed that Chd1l, miR-486, and MMP2 work in concert in regulating SSC stemness gene expression and growth properties. Finally, our data also revealed that MMP2 regulates SSC stemness gene expression and growth properties through activating β-catenin signaling by cleaving N-cadherin and increasing β-catenin nuclear translocation. Our data demonstrate that Chd1l–miR-486–MMP2 is a novel regulatory axis governing SSC stemness gene expression and growth properties, offering a novel therapeutic opportunity for treating male infertility.

INTRODUCTION

The balance between spermatogonial stem cell (SSC) self-renewing divisions and differentiating divisions maintains the SSC pool and meets the proliferative demand of the testis to produce millions of sperm each day (1). Therefore, elucidating the signal pathways controlling SSC self-renewal and stemness is critical for us to understand spermatogenesis and treat male infertility. Although both epigenetic (2) and transcriptional (3, 4) mechanisms, several signaling pathways (5, 6), and an increasing list of regulators (e.g., epigenome-related regulators [2], small noncoding RNAs [2, 7], and multiple transcription factors [3, 4, 6, 8]) have been proposed as regulators of SSC self-renewal and stemness, the exact mechanisms remain largely unidentified.

The chromodomain helicase/ATPase DNA binding protein 1-like gene (CHD1L gene) was originally isolated from hepatocellular carcinoma tissue and identified as an oncogene in our previous study (9). A functional study using CHD1L transgenic mice has confirmed its role in tumorigenesis (10), and epidemiological studies have suggested that CHD1L is either associated with cancer-inducing processes or can be used as a predictive biomarker for hepatocellular carcinoma (11–13) and several other cancers, such as ovarian carcinoma (14), bladder cancer (15), colorectal carcinoma (16), breast cancer (17), lung adenocarcinoma (18), glioma (19), and gastric cancer (20). CHD1L has also been defined as an ATP-dependent chromatin-remodeling enzyme that plays an important role in preserving chromosome integrity and promoting DNA repair (21). Interestingly, other studies have reported that CHD1L is required for early embryo development in mice (22) and early-stage cellular reprogramming during murine induced pluripotent stem (iPS) cell generation (23). Importantly, we have recently reported that CHD1L promotes neural differentiation from human embryonic stem cells (24) and is required for SSC survival (25). The exact underlying mechanisms through which CHD1L controls SSC survival and stemness, however, remain to be elucidated.

Of great interest to us is the growing evidence that supports a role for a novel class of gene regulators, microRNAs (miRNAs), in regulating stem cell pluripotency, self-renewal, and differentiation (26–29). miRNAs are short, noncoding RNA molecules (∼21 to 23 nucleotides) that suppress gene expression at the posttranscriptional or translational level by degrading target mRNA or blocking mRNA translation. Several stage-specific miRNAs, such as miR-221, miR-203, and miR-34b-5p, have been suggested to play a functional role in spermatogenesis (30). miR-21, regulated by the SSC self-renewal-related transcription factor ETV5, plays a role in Thy1-positive (Thy1⁺) SSC self-renewal (31). miR-221 and -222 play a critical role in maintaining the undifferentiated state of mammalian male germ cells by repressing the expression of Kit receptor (32). Other studies have also suggested a role for miR-34c, miR-20, and miR-106a in SSC differentiation (33) and self-renewal (34).

Combined, the findings from the aforementioned studies point to a novel role for CHD1L and miRNAs in SSC self-renewal and spermatogenesis; however, the links between CHD1L and miRNAs in the context of SSC self-renewal and spermatogenesis are missing. In this study, through small RNA high-throughput sequencing analysis, we found that stem cell pluripotency and proliferation were two of the most regulated pathways upon CHD1L inhibition. As a proof of concept, we have also identified CHD1L–miR-486–matrix metalloproteinase 2 (MMP2) as a novel regulatory axis governing SSC stemness gene expression and growth properties. Our results offer a possible therapeutic basis for treating male infertility by modulating SSC stemness and self-renewal.

RESULTS

miRNA expression profile in mSSCs with Chd1l depletion.

We have recently reported that CHD1L is required for SSC survival and self-renewal (25). To explore the underlying molecular mechanism through which CHD1L regulates SSC functions, freshly isolated THY1⁺ mouse SSCs (mSSCs) were infected with small hairpin RNA (shRNA) lentivirus against Chd1l (sh-Chd1l) or a scrambled, nontargeted shRNA (sh-NT) generated in our previous study (25). Real-time quantitative PCR (RT-qPCR) analysis confirmed that Chd1l gene expression in mSSCs was successfully downregulated (Fig. 1A). To identify the potential miRNAs regulated by Chd1l in SSCs, small RNAs isolated from SSCs treated with control (sh-NT) or Chd1l gene knockdown (sh-Chd1l) shRNA were subjected to high-throughput small RNA sequencing. We obtained approximately 14.6 to 16.7 million effective reads in different samples and mapped reads with lengths of 18 to 23 nucleotides (nt) to the genome using CLC Genomics Workbench 6.0. Approximately 80% of the reads were perfectly mapped to the reference genome sequence, and the small transcripts identified were then classified into several different miRNA categories according to their annotations. After applying strict criteria (P < 0.05; false discovery rate [FDR] < 0.05; log₂ fold change > 1.5 or < −0.5), we observed that 124 miRNAs were differently expressed in SSCs upon Chd1l knockdown; of these, 34 and 90 miRNAs were up- or downregulated, respectively. The hierarchical clustering heat map shown in Fig. 1B depicts the hierarchical clustering of the top 20 up- or downregulated miRNAs. Five well-established miRNA target prediction databases (Target Scan, miRanda, mi RDB, miRWalk, and PicTar5) were utilized to predict the target genes of the differently expressed miRNAs, and the possible functions investigated using gene function annotation methods. A gene ontology (GO) analysis was conducted to understand the functional roles of the target genes, and an enrichment analysis based on the hypergeometric test was used to test whether a GO term was statistically enriched for a given set of genes. Only genes predicted as the target gene of a given miRNA by four or five of the above-mentioned computational algorithms were included in the GO and KEGG pathway analyses. We observed that several pathways involved in cancer, such as pathways in cancer, proteoglycans in cancer, acute myeloid leukemia, phosphoinositide 3-kinase (PI3K)–Akt, focal adhesion, Wnt signaling, and pluripotency, were significantly enriched in the testes upon Chd1l depletion (Fig. 1C), which further supports a role for Chd1l in various cancers, as aforementioned, and in regulation of SSC stemness. Fifteen miRNAs (10 upregulated and 5 downregulated) were randomly selected for validation by RT-qPCR, and the data shown in Fig. 1D revealed that the expression levels of 14 of 15 miRNAs examined in this study were confirmed by RT-qPCR, suggesting that these 14 miRNAs are regulated by Chd1l in SSCs. Importantly, we found that miR-486 is the miRNA most upregulated by Chd1l in SSCs in our RT-qPCR validation analysis (Fig. 1D). Therefore, as a proof of concept, miR-486 was chosen for the following analysis to further explore the molecular mechanism underlying SSC functions regulated by Chd1l.

FIG 1.

miRNAs regulated by CHD1L in mouse spermatogonial stem cells (mSSCs). Thy1⁺ mSSCs were infected with scrambled, nontargeted (sh-NT) or Chd1l gene-specific (sh-Chd1l) shRNA lentivirus. After 48 h, total RNAs, including small RNAs, were harvested and subjected to real-time quantitative PCR (RT-qPCR) and high-throughput miRNA sequencing (Sequencing) analysis, respectively. (A) RT-qPCR analysis of Chd1l gene expression. Data presented here are mean values ± standard errors of the means (SEM) from three independent experiments (n = 3). *, P < 0.05 (versus sh-NT). (B) Hierarchical cluster and heat map to show changes in miRNAs in CHD1L knockdown SSCs (sh-Chd1l) compared to expression levels in control SSCs (sh-NT). Red and green represent increase and decrease of expression, respectively. Each row represents the results for a single miRNA in small RNA sequencing analysis. The 20 miRNAs most up- or downregulated after CHD1L knockdown are shown. (C) KEGG pathways associated with the target genes of the miRNAs most regulated by CHD1L. (D) miRNA expression profile and validation. The miRNAs most up- or downregulated by CHD1L in mSSCs were selected and validated by standard RT-qPCR. Data shown here are mean values ± SEM from three independent experiments (n = 3), which are presented as log₂ fold changes with the miRNA expression level in control mSSCs set as 0. *, P < 0.05 and FDR < 0.05 (for sequencing); #, P < 0.05 (for RT-qPCR).

miR-486 is transcriptionally regulated in SSCs by Chd1l.

Transcriptional modulation and regulation of miRNA biogenesis are two widely known mechanisms controlling miRNA activity. Here, we sought to determine whether either of these mechanisms was responsible for the regulation of miR-486 by Chd1l in SSCs. We found that miR-486 was dramatically downregulated by Chd1l overexpression, whereas both miR-486 and its primary (pri-miR-486) transcript were significantly upregulated in response to Chd1l knockdown (Fig. 2A and B), suggesting that miR-486 is transcriptionally regulated by Chd1l in SSCs. Such a notion was further supported by our RNA decay data (Fig. 2C). Two previous studies have suggested that miR-486, an miRNA that is enriched in muscle, is controlled by an alternative promoter within intron 40 of the Ankyrin-1 gene and that the transcription of miR-486 is directly controlled by serum response factor (SRF), its coactivator myocardin-related transcription factor A (MRTF-A), and MyoD (35), as well as myostatin (also known as growth and differentiation factor 8) (36). We wondered if CHD1L regulated miR-486 transcription in SSCs through a similar mechanism. To investigate this, we generated five miR-486 promoters as described in the previous studies (35, 36). Data from our luciferase activity analysis using these five promoters showed that the luciferase activities of these five promoters were not significantly regulated by Chd1l knockdown (Fig. 2D and E), implying that Chd1l’s transcriptional regulation of miR-486 expression in SSCs is independent of these reported promoters.

FIG 2.

CHD1L regulates miR-486 in mouse spermatogonial stem cells (mSSCs) through a transcriptional mechanism. (A) C18-4 cells (mouse spermatogonial stem cell line) with Chd1l overexpression had downregulated miR-486. C18-4 cells were transfected with control (pcDNA3.1) or Chd1l overexpression (pcDNA3.1-Chd1l) plasmid. (B) Both mature and primary (pri-miR-486) miR-486 transcripts were upregulated in CHD1L knockdown SSCs. C18-4 cells were infected with scrambled, nontargeted (sh-NT) or Chd1l gene-specific (sh-Chd1l) shRNA lentivirus. (A and B) After 48 h, total RNAs, including small RNAs, were harvested and subjected to RT-qPCR analysis. (C) CHD1L regulates pri-miR-486 in mSSCs through a transcriptional mechanism. C18-4 cells infected with sh-NT or sh-Chd1l lentivirus were treated with an inhibitor of transcription (actinomycin D [ActD]; 1 µg/ml) for the indicated times. The result for pri-miR-486 at 0 h for both groups was set as 1.0 (arbitrary unit), and values at other time points calculated accordingly. (D and E) Neither the embryonic and adult skeletal muscle promoter nor the adult cardiac and skeletal muscle promoter for the Ank1.5 gene is required for CHD1L-mediated miR-486 expression in mSSCs. C18-4 cells infected with sh-NT or sh-Chd1l lentivirus were transfected with the indicated reporters. Luciferase activities were measured at 48 h posttransfection. The data presented here are representative mean values ± SEM from three independent experiments (n = 3). *, P < 0.05 (versus sh-NT).

miR-486 controls SSC stemness gene expression and growth properties.

The above-described data show that miR-486 expression is upregulated by Chd1l knockdown in SSCs, and our previous study reported that CHD1L is required for SSC survival and self-renewal (25); we speculated, therefore, that miR-486 plays a role in these processes. To this aim, we transfected C18-4 cells with a control or miR-418 mimic and subjected them to various analyses in order to determine whether miR-486 is required for SSC functions. RT-qPCR data showed that miR-486 expression in C18-4 cells was significantly upregulated by the miR-486 mimic (Fig. 3A). Consequently, decreased expression of all four SSC stemness genes (the Oct4, Sox2, Nanog, and Plzf genes) was observed in C18-4 cells transfected with the miR-486 mimic (Fig. 3B and C), suggesting that miR-486 controls SSC stemness. Moreover, miR-486 overexpression also significantly reduced SSC growth, as demonstrated by the results of MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] assays (Fig. 3D) and phosphorylated histone H3 (pH3) immunofluorescence staining (Fig. 3E). Furthermore, data from our apoptosis analysis showed an increase of cell apoptosis in the C18-4 cells overexpressing miR-486 (Fig. 3F), and the cell cycle data revealed that miR-486 overexpression significantly increased SSC arrest at G₀/G₁ phase (Fig. 3G). These data suggest that miR-486 mediates SSC growth through controlling cell apoptosis and cell cycle processes.

FIG 3.

miR-486 controls SSC stemness gene expression and growth properties. (A to C) SSC stemness gene expression levels were downregulated by miR-486. C18-4 cells were transfected with miRNA negative control (Ctrl mimic) or miR-486 mimic, respectively. After 48 h, total RNAs, including small RNAs, and proteins were harvested and subjected to RT-qPCR (A and B) and Western blot analysis (C), respectively. (D) miR-486 inhibits SSC growth. C18-4 cells transfected with miRNA negative control (Ctrl mimic) or miR-486 mimic were subjected to cell proliferation analysis using an MTS assay kit. OD, optical density. (E) Immunofluorescence analysis of pH3-positive mouse SSCs. (F) miR-486 increases SSC apoptosis. C18-4 cells transfected with miRNA negative control (Ctrl mimic) or miR-486 mimic were harvested and subjected to apoptosis assay. Cells that stained positive for annexin V (x axis) but negative for propidium iodide (y axis) were considered apoptotic cells. (G) miR-486 promotes SSC cell cycle arrest at G₀/G₁ phase. C18-4 cells transfected with miRNA negative control (Ctrl mimic) or miR-486 mimic were harvested and subjected to cell cycle analysis. The data presented here are representative images or mean values ± SEM from three independent experiments (n = 3). *, P < 0.05 (versus Ctrl mimic).

Importantly, an opposite effect was also observed in terms of SSC stemness, growth, and apoptosis when the endogenous expression level of miR-486 in C18-4 cells was inhibited by an miR-486 shRNA lentivirus (Fig. 4). The above-described data collectively confirm a role for miR-486 in the regulation of SSC stemness gene expression and growth properties.

FIG 4.

The stemness marker expression and growth properties of SSCs are regulated by miR-486. C18-4 cells were infected with control (sh-NT) or miR-486 (sh-miR-486) lentivirus. Cells were harvested and subjected to various analyses as indicated. (A to C) SSC stemness marker expression levels were upregulated by miR-486 inhibition. (D) miR-486 suppression increased SSC growth. (E) SSC proliferation (pH3-positive cells) was increased by miR-486 knockdown. (F) miR-486 knockdown decreased SSC apoptosis. Cells that stained positive for annexin V (x axis) but negative for propidium iodide (y axis) were considered apoptotic cells. The data presented here are representative images or mean values ± SEM from three independent experiments (n = 3). *, P < 0.05 (versus sh-NT).

MMP2 is a novel target of miR-486 in SSCs.

To identify the possible target genes of miR-486 in the context of SSC functions, additional high-throughput mRNA sequencing analysis was conducted with the freshly isolated THY1⁺ mSSCs infected with Chd1l-targeting (sh-Chd1l) or scrambled, nontargeted (sh-NT) shRNA lentivirus in a separate study (unpublished data). As mentioned above, increased expression of miR-486 was observed in SSCs with Chd1l knockdown; thus, inhibiting Chd1l gene expression using its shRNA mimics the results of miR-486 overexpression. One could assume that the genes downregulated by Chd1l knockdown (or miR-486 overexpression) in SSCs could be the candidate target genes of miR-486. Compared to the results for control cells, we observed significantly decreased expression for 316 genes in the cells with Chd1l knockdown. Data from our GO and KEGG pathway analysis of the genes significantly downregulated by CHD1L knockdown in SSCs revealed that the top two biological pathways modulated by CHD1L (or miR-486) in SSCs are extracellular matrix (ECM)-receptor interaction and focal adhesion (Fig. 5A). Additionally, miRWalk was enlisted to predict possible target genes for miR-486 in order to identify downstream targets of miR-486 that are responsible for SSC functions. One or more conserved miR-486 binding sites were found within 1,449 genes. Among them, only five genes (the Mmp2, Fbln2, Gstm7, Rhod, and Itm2a genes) were significantly downregulated by Chd1l knockdown (or increased expression of miR-486) in our high-throughput mRNA sequencing analysis (Fig. 5B), inferring that these five genes are possible target genes of miR-486 in SSCs. MMP2 was chosen for further validation in this study because of its prominent role in ECM-receptor interaction and focal adhesion, as well as having a reported role in follicle stem cell proliferation (37). As expected, we found that MMP2 protein was highly expressed in the immature mouse testis, while its expression was decreased in the mature mouse testis (Fig. 5C and D). Importantly, MMP2 was significantly downregulated in SSCs by Chd1l knockdown at both the mRNA and protein level (Fig. 5E and F). Interestingly, MMP2 protein, but not its mRNA expression level, was dramatically inhibited by the miR-486 mimic (Fig. 5G and H), while an opposite effect was observed with miR-486 inhibition (Fig. 5I and J), indicating that MMP2 is translationally repressed by miR-486.

FIG 5.

MMP2 is the novel target of miR-486 in SSCs. (A) The GO and KEGG pathway analyses of the mRNAs significantly downregulated by CHD1L knockdown in SSCs. Thy1⁺ mSSCs were infected with scrambled, nontargeted (sh-NT) or Chd1l gene-specific (sh-Chd1l) shRNA lentivirus. After 48 h, total RNAs were harvested and subjected to high-throughput mRNA sequencing analysis. The genes that displayed significant downregulation by Chd1l knockdown in SSCs were analyzed using DAVID. The top 10 KEGG pathways are presented here. (B) Venn diagram illustrating the downregulated genes observed in mRNA sequencing (RNA-seq) analysis and the miR-486 target genes predicted by miRWalk. (C) Immunohistochemistry analysis of MMP2 protein expression in mouse testis. dpp, days postpartum. (D) Western blot analysis of Chd1l and MMP2 protein expression in mouse testis. (E and F) MMP2 expression was suppressed in SSCs with Chd1l knockdown. (G to J) The MMP2 protein level, but not the mRNA level, was modulated by miR-486 in SSCs. (K) Schematic diagram illustrating the miR-486 sequence and the wild-type (WT) and mutated (MUT) miR-486 binding sites within the 3′UTR of Mmp2 gene reporters. (L and M) miR-486 binding site was required for miR-486-mediated Mmp2 gene repression. Control miRNA mimic (Ctrl mimic) or miR-486 mimic was cotransfected into C18-4 cells with the wild-type reporter (pmiR-mmp2-WT) or the reporter containing the mutated miR-486 binding site (pmiR-Mmp2-Mut). Transfected cells were cultured for 48 h before luciferase activity assay. The data presented here are representative images or mean values ± SEM from three independent experiments (n = 3). *, P < 0.05 (versus sh-NT or Ctrl mimic).

To determine whether MMP2 is regulated by miR-486 through a direct or indirect mechanism, the MMP2 3′ untranslated region (UTR) harboring the miR-486 binding site (Fig. 5K) was cloned into a pmiRNA reporter. Our data show that the activity of luciferase from the MMP2 3′UTR reporter was significantly inhibited by miR-486 overexpression (Fig. 5L), and this effect was abolished once the miR-486 binding site within the MMP2 3′UTR was mutated (Fig. 5M). Taken together, our data demonstrate that MMP2 is an authentic target gene of miR-486 and that its expression is negatively controlled by miR-486 expression in SSCs.

Mmp2 inhibition reproduces the effects of miR-486 overexpression on SSC stemness gene expression and growth properties.

To study the functional implication of Mmp2 in SSC functions, C18-4 cells were transfected with a scrambled control or either of two independent Mmp2-specific small interfering RNAs (siRNAs). RT-qPCR data showed that the endogenous Mmp2 gene expression was successfully knocked down by both Mmp2-specific siRNAs (Fig. 6A). Consequently, we found that the expression of all four SSC stemness genes was significantly downregulated by Mmp2-specific siRNAs (Fig. 6B and C). Similarly, decreased levels of cell mitosis and meiosis (Fig. 6D and E) and a higher proportion of apoptotic cells (Fig. 6F) were observed in these SSCs. As expected, Mmp2 gene knockdown resulted in SSCs arresting at G₀/G₁ cell cycle phase (Fig. 6G). Altogether, our data prove that MMP2 suppression in SSCs recapitulates the effects of miR-486 overexpression on SSC stemness gene expression and growth properties.

FIG 6.

Mmp2 inhibition reproduces the effects of miR-486 overexpression on SSC stemness gene expression and growth properties. (A to C) SSC stemness marker expression was suppressed by Mmp2 knockdown. C18-4 cells were transfected with control (si-NT) or one of two Mmp2-specific [si-mmp2 (#1 & #2)] siRNAs. After 48 h, total RNAs and proteins were harvested and subjected to RT-qPCR (A and B) and Western blot analysis (C), respectively. (D and E) Mmp2 knockdown decreased SSC growth (MTS assay) (D) and proliferation (pH3-positive cells) (E). (F) Mmp2 knockdown increased SSC apoptosis. Cells that stained positive for annexin V (x axis) but negative for propidium iodide (y axis) were considered apoptotic cells. (G) Mmp2 knockdown increased SSC G₀/G₁ cell cycle arrest. The data presented here are representative images or mean values ± SEM from three independent experiments (n = 3). *, P < 0.05 (versus si-NT).

miR-486 repression is required for CHD1L-mediated SSC stemness gene expression and growth properties.

To further understand the functional significance of miR-486 in CHD1L-mediated SSC functions, control and Chd1l knockdown C18-4 cells were infected with either a control nontargeted or an miR-486 shRNA lentivirus. RT-qPCR data showed that both Chd1l and miR-486 were successfully downregulated in C18-4 cells by Chd1l and miR-486 shRNA, respectively, and the expression level of miR-486 was significantly increased by Chd1l knockdown in control C18-4 cells, while such promotive effect was abolished in miR-486 knockdown C18-4 cells (Fig. 7A, left and middle). We also found a decreased expression level of the Mmp2 gene in C18-4 cells with Chd1l knockdown, while no such change was observed in cells with miR-486 inhibition (Fig. 7A, right). As expected, Mmp2 protein expression was dramatically downregulated and upregulated in cells infected with Chd1l and miR-486 shRNA lentivirus, respectively, but such effects disappeared in cells infected with both shRNAs (Fig. 7B). Importantly, we found that while Chd1l knockdown (Fig. 7C, 2nd bar) and miR-486 inhibition (Fig. 7C, 3rd bar) alone decreased and increased SSC stemness gene expression, respectively, miR-486 inhibition in Chd1l knockdown SSCs (Fig. 7C, 4th bar) restored gene expression levels (Fig. 7C). Such effects were also observed with the protein expression of the Nanog, Plzf, Oct4, and C-Myc genes (Fig. 7D). A similar phenomenon was observed in our MTS assay (Fig. 7E) and pH3 immunofluorescence staining (Fig. 7F), while an opposite effect was found in cell apoptosis (Fig. 7G) and cell cycle analysis (Fig. 7H). Taking these results together, we demonstrate that miR-486 is at least partially responsible for CHD1L-mediated SSC stemness gene expression and growth properties.

FIG 7.

miR-486 repression is required for CHD1L-mediated SSC stemness gene expression and growth properties. C18-4 cells infected with control (sh-NT) or Chd11 (sh-Chd1l) lentivirus were coinfected with control (sh-NT) or miR-486 (sh-miR-486) lentivirus as indicated. Infected cells were cultured for 48 h and subjected to RT-qPCR (A and C), Western blot analysis (B and D), MTS assay (E), pH3 immunofluorescence staining (F), apoptosis assay (G), or cell cycle assay (H). (A) RT-qPCR detection of Chd1l, miR-486, and Mmp2 gene expression levels. (B) Western blot detection of MMP2 protein expression in the infected cells. (C and D) miR-486 inhibition partially rescued the expression levels of the SSC stemness markers examined. (E and F) miR-486 repression partially released the inhibitory effects of Chd1l knockdown on SSC growth (E) and proliferation (F). (G) miR-486 knockdown almost blunted the apoptotic effects induced by Chd1l inhibition in SSCs. (H) The promotive effect of Chd1l knockdown on SSC cell cycle arrest at G₀/G₁ phase was almost abolished by miR-486 inhibition. The data presented here are representative images or mean values ± SEM from three independent experiments (n = 3). *, P < 0.05 (versus sh-NT); ǂ, P < 0.05 (sh-Chd1l/sh-miR-486 versus sh-Chd1l).

miR-486 regulates SSC stemness gene expression and growth properties through modulating MMP2 protein expression.

Since our data show that the MMP2 gene is a novel gene target of miR-486, and the expression of MMP2 was negatively regulated by miR-486, we further attempted to delineate the precise relationship between miR-486 and MMP2 in the context of SSC functions by conducting coinhibitory experiments using miR-486 shRNA or MMP2 siRNA and the respective negative control (nontargeted shRNA or control siRNA) in C18-4 cells.

Gene expression data showed that neither miR-486 knockdown nor Mmp2 inhibition could affect Chd1l gene expression (Fig. 8A, left), suggesting a higher position for Chd1l in the signaling cascade examined in this study. Similarly, the miR-486 and Mmp2 gene expression profiles (Fig. 8A, middle and right) and the MMP2 protein expression patterns (Fig. 8B) in C18-4 cells transfected with the vectors indicated in the figure further support the idea that Mmp2 is a downstream target of miR-486. As expected, the expression levels of all four SSC stemness genes (the Oct4, Sox2, Nanog, and Plzf genes) were significantly upregulated by miR-486 inhibition, while such gene inductions no longer occurred in SSCs when both miR-486 and Mmp2 expression were suppressed (Fig. 8C). Moreover, we also observed a similar phenomenon with the protein expression levels of Nanog, Plzf, Oct4, and C-Myc (Fig. 8D). Data from MTS assays (Fig. 8E) and pH3 immunofluorescence staining (Fig. 8F) showed that miR-486 inhibition significantly promoted SSC proliferation, mitosis, and meiosis, while such promotive effects were abolished when the endogenous Mmp2 gene expression was also inhibited. Conversely, SSC apoptosis (Fig. 8G) and cell cycle arrest were significantly inhibited by miR-486 knockdown, and such inhibitory effects were eliminated by cotransfection of C18-4 cells with miR-486 shRNA and Mmp2 siRNA (Fig. 8H). Collectively, our data demonstrate that miR-486 regulates SSC stemness gene expression and growth properties by modulating MMP2 protein expression.

FIG 8.

miR-486 regulates SSC stemness gene expression and growth properties through modulating MMP2 protein expression. C18-4 cells infected with control (sh-NT) or miR-486 (sh-miR-486) lentivirus were cotransfected with control (si-NT) or Mmp2 (si-mmp2) siRNA as indicated. Transfected cells were cultured for 48 h and subjected to RT-qPCR (A and C), Western blot analysis (B and D), MTS assay (E), pH3 immunofluorescence staining (F), apoptosis assay (G), or cell cycle assay (H). (A) RT-qPCR detection of Chd1l, miR-486, and Mmp2 gene expression. (B) Western blot detection of MMP2 protein expression in the infected cells. (C and D) Stemness gene induction caused by miR-486 knockdown in SSCs was normalized by Mmp2 inhibition. (E and F) Mmp2 repression abrogated the promotive effects of miR-486 inhibition on SSC growth (E) and proliferation (F). (G and H) The inhibitory effects of miR-486 knockdown on SSC apoptosis and cell cycle arrest were completely reversed by Mmp2 gene inhibition. The data presented here are representative images or mean values ± SEM from three independent experiments (n = 3). *, P < 0.05 (versus sh-NT/si-NT); ǂ, P < 0.05 (sh-miR-486/si-mmp2 versus sh-miR-486).

MMP2 regulates SSC stemness gene expression and growth through activation of β-catenin signaling by cleaving N-cadherin.

To further elucidate the molecular mechanism underlying MMP2 regulation of SSC stemness gene expression and functions, C18-4 cells were incubated with exogenous activated MMP2 protein and subjected to various assays. The data showed that the mRNA and protein expression levels of all the SSC stemness markers examined in this study were upregulated by MMP2 in a time-dependent manner (Fig. 9A and B). MMP2 incubation also significantly increased SSC proliferation (Fig. 9C). Since our RNA sequencing analysis suggested the Wnt/β-catenin signal is one of the pathways most modulated/enriched by Chd1l, we wondered if MMP2 incubation could activate this pathway. Indeed, data from our TopFlash/FopFlash luciferase activity assays confirmed such a mechanism (Fig. 9D). Moreover, immunofluorescent staining showed an accumulation of β-catenin within nuclei upon MMP2 incubation (Fig. 9E). Finally and importantly, we observed typical cell plasma membrane and cell junctional locations for N-cadherin in control SSCs, while it was mainly localized within the cytoplasm in SSCs treated with MMP2 (Fig. 9E), implying a role for MMP2 in dislocating N-cadherin protein from the cell membrane and junctions by cleaving/degrading this molecule. This was confirmed through in vitro digestion assays, which showed that the recombinant N-cadherin protein was degraded into several fragments by activated MMP2 (Fig. 9F).

FIG 9.

MMP2 regulates SSC stemness gene expression and growth through activating β-catenin signaling by cleaving N-cadherin. (A and B) MMP2 increased SSC stemness gene expression. C18-4 cells were treated with 10 ng/ml recombinant MMP2 for the indicated times. Total RNAs and proteins were harvested and subjected to RT-qPCR (A) and Western blot analysis (B), respectively. (C) MMP2 promoted SSC growth. C18-4 cells were incubated with vehicle control or 10 ng/ml recombinant MMP2 and subjected to cell proliferation analysis using an MTS assay kit. (D) Wnt/β-catenin/TCF/LEF signaling was activated by MMP2 in SSCs. C18-4 cells were transfected with either M50 Super 8x TOPFlash (plasmid 12456; Addgene) or M51 Super 8x FOPFlash (plasmid 12457p; Addgene) reporter plasmid for 44 h, followed by incubation with vehicle control or 10 ng/ml recombinant MMP2 for an additional 4 h. Cell lysates were harvested and subjected to a standard luciferase activity assay. (E) Immunofluorescence staining revealed the cellular locations of β-catenin and N-cadherin. (F) MMP2 cleaves N-cadherin. Recombinant N-cadherin (1,000 ng per reaction mixture volume) (ab112268; Abcam) was incubated with activated MMP2 (100 ng per reaction mixture volume) (ab174022; Abcam) in assay buffer for 4 h, followed by Western blot analysis with an antibody against N-cadherin (ab76057; Abcam). FL, full length recombinant N-cadherin protein. The data presented here are representative images or mean values ± SEM from three independent experiments (n = 3). *, P < 0.05 (versus 0 h or vehicle).

DISCUSSION

SSCs not only maintain normal spermatogenesis but also sustain fertility by fine-tuning their self-renewal and differentiation, which are tightly regulated by a combination of intrinsic (from the SSCs) and extrinsic (from the niche and supporting cells) signaling (4, 38). A crucial role for Chd1l in SSC survival and self-renewal has recently been appreciated in this field (25), but the underlying mechanism is unknown. Through our combined efforts, we have now been able to delineate a novel underlying mechanism for spermatogenesis. Specifically, we demonstrate for the first time that miR-486 and its target gene MMP2 are at least partially responsible for CHD1L-mediated SSC stemness gene expression and growth properties.

Using small RNA sequencing and annotation with miRbase (version 20, June 2014), we identified 124 miRNAs that were significantly differentially expressed in SSCs upon Chd1l depletion. Further bioinformatic analysis indicated that the most enriched biological pathways for Chd1L-modulated miRNAs are associated with cancer processes, stem cell pluripotency gene regulation, cell proliferation and differentiation, PI3K-Akt signaling, and Wnt signaling, all of which have been implicated in cancer processes, stem cell pluripotency, and differentiation. Such implications were further supported by the reported roles for the top 20 most up- and downregulated miRNAs in cancer (e.g., miR-874 [39, 40], miR-709 [41], miR-668 [42], and miR-770 [43]) and stem cell biology (miR-127 [44, 45], miR-183 [46], and miR-184 [47–49]).

Since our previous study reported that CHD1L is required for SSC survival and self-renewal (25), we assumed these miRNAs may play a role in SSC functions and spermatogenesis. Among them, miR-486 was chosen for further investigation in this study due to the following facts: (i) miR-486 was confirmed as the most upregulated miRNA in SSCs with Chd1l knockdown in our validation analysis, (ii) miR-486 has been proven to play a dual role in normal erythropoiesis and chronic myeloid leukemia (CML) by modulating both CD34-positive hematopoietic stem cell and CML progenitor cell growth, survival, and drug sensitivity (50–52), as well as by controlling the hematopoietic cell lineage fate decision (3, 52, 53), and (iii) miR-486 also plays a role in mesenchymal stem cell survival (54) and replicative senescence (55). In considering the prominent role for CHD1L in SSC survival and self-renewal, we reasoned that miR-486 could be the downstream gene target of Chd1l and that miR-486 should also play a role in SSC functions. Indeed, data from miR-486 overexpression and knockdown experiments showed that miR-486 suppresses SSC stemness gene expression and inhibits SSC growth but promotes SSC apoptosis and cell cycle arrest at G₀/G₁ phase. Importantly, through our coinhibition experiments, we demonstrate that miR-486 is the downstream gene target of Chd1l and that modulation of miR-486 expression is the underlying mechanism of CHD1L-mediated SSC stemness and growth properties.

One of the interesting findings in our current study is the identification of miR-486 as the functional downstream target of Chd1l in SSC biology. We have shown that, in SSCs, miR-486 is negatively regulated by CHD1L through a transcriptional repression mechanism, which is at odds with the well-established function of CHD1L as an ATP-dependent chromatin remodeling enzyme and a transcriptional activator in gene regulation. The CHD1L protein is comprised of two helicase domains, a C-terminal Macro domain, and a nuclear localization sequence (56). Through a high-density oligonucleotide microarray analysis, it has been shown that CHD1L is highly expressed in early erythroid cells, CD34-positive cells, endothelial cells, dendritic cells, and some leukemic cells (57). Unsurprisingly, CHD1L protein is mainly localized in the nucleus due to its nuclear localization sequence. It has been suggested that CHD1L confers DNA-binding capability and functions as a transcription factor (56). A few direct transcriptional targets of CHD1L, such as ARHGEF9 (Rho guanine nucleotide exchange factor 9) (11), TCTP (translationally controlled tumor protein) (13), and SPOCK1 (sparc/osteonectin, cwcv, and kazal-like domains proteoglycan 1) (58) have been reported. CHD1L transcriptionally regulates the aforementioned downstream genes through direct binding to their respective target gene promoters. Through our small RNA sequencing analysis, we observed that at least 34 miRNAs were significantly upregulated upon Chd1l knockdown. Taking these results together with our finding that miR-486 is negatively regulated by Chd1l through a transcriptional repression mechanism, we therefore speculated that CHD1L can also function as a transcriptional repressor and that its role in gene regulation is gene specific and cell context dependent.

Evidence has suggested that miR-486 expression is controlled through a cis regulatory element within its host gene Ankyrin-1 by several transcriptional factors, such as SRF, MRTF-A, MyoD, and myostatin (35, 36). Other transcriptional factors, Myb (52) or GATA1 (51), have also been reported to play a regulatory role in miR-486 transcription during the hematopoietic cell lineage fate decision. We speculated that CHD1L regulates miR-486 expression in SSCs through a similar mechanism; however, we failed to confirm such a possibility. In this study, we observed no evidence in our promoter luciferase activity assays to support the idea that the reported cis regulatory elements within the Ankyrin-1 gene are responsible for CHD1L-mediated miR-486 transcription, indicating that (i) none of the identified transcriptional factors (e.g., SRF, MRTF-A, MyoD, myostatin, Myb, or GATA1) that directly bind to the cis regulatory elements is regulated/affected by CHD1L in SSCs, (ii) CHD1L mediates miR-486 gene expression in an indirect manner, and (iii) the genomic locations of the putative regulatory element(s) for CHD1L are beyond the regions examined within the Ankyrin-1 gene and CHD1L remotely regulates miR-486 transcription. To confirm this, newly developed methods, such as 3C and its derivatives (4C, 5C, and HiC) (59), that can detect all the chromosomal regions that physically interact with promoters and CHD1L should be used in our future studies.

Another interesting finding in this study is that MMP2 has been identified as a novel target gene of miR-486 in the context of SSC functions. Several lines of evidence have been sought out to support the idea that MMP2 is a bona fide miR-486 target gene in SSCs: first, MMP2 gene expression was significantly downregulated in SSCs upon Chd1l depletion, which resulted in miR-486 upregulation. Second, the MMP2 protein expression level, but not the RNA level, was adversely regulated by miR-486, as demonstrated in miR-486 overexpression and inhibition experiments. Third, miR-486 overexpression dramatically downregulated MMP2 3′UTR luciferase activity, but such downregulation was completely abolished when the miR-486 binding sites within the MMP2 3′UTR were mutated. However, it should be noted that both the mRNA and protein levels of MMP2 were downregulated by CHD1L knockdown, but only the MMP2 protein expression level was modulated by miR-486 overexpression and inhibition, inferring that there are additional mechanisms by which CHD1L operates to regulate MMP2 gene expression. It came as no surprise for us to observe such a discrepancy in terms of MMP2 gene regulation by CHD1L and miR-486, due to the well-known facts that a given miRNA can target multiple genes and a given mRNA can be targeted by many miRNAs. Another 33 miRNAs besides miR-486 were also upregulated by Chd1l knockdown in our sequencing analysis. Therefore, other miRNAs upregulated by Chd1l, such as miR-668, might also target the MMP2 gene and negatively regulate the MMP2 gene transcript, which warrants further investigation in future studies.

Interestingly, we observed that MMP2 inhibition can recapitulate the effects of miR-486 overexpression on SSC stemness gene expression and other functions, including cell growth, mitosis, meiosis, apoptosis, and cell cycle arrest, indicating a role for a Chd1l–miR-486–Mmp2 axis in SSC functions. In order to definitively clarify the importance of the Chd1l–miR-486–Mmp2 axis in SSCs, we demonstrated that coinhibiting miR-486 and Mmp2 gene expression in SSCs partially reversed the effects of miR-486 inhibition on SSC gene expression and functions.

Finally, we demonstrate that MMP2 regulates SSC stemness gene expression and growth by activating β-catenin signaling through cleaving N-cadherin. Emerging evidence supports a prominent role for MMP2 in regulating embryonic stem cell and other stem/progenitor cell functions. The underlying mechanisms of MMP2-mediated stem cell functions and differentiation have been attributed mainly to its protease activity on a variety of matrix proteins, adhesion molecules, and cell membrane proteins, as well as growth factors and cytokines (60). In particular, MMP2-mediated cleavage of the division abnormally delayed (Dally)-like protein Dlp (37) and laminin-111 (61) have been proposed as the main mechanisms underlying Wingless-mediated follicle stem cell proliferation in the Drosophila melanogaster ovary and early differentiation of embryonic stem cells, respectively. In the latter study, the authors proposed a new mechanism by which an MMP2-cleaved laminin fragment regulates the expression of E-cadherin, MMP2, and MMP9 and promotes stem cell differentiation. In our current study, we demonstrate that MMP2 cleaves the transmembrane molecule N-cadherin and disrupts the interaction of N-cadherin with β-cadherin at the SSC membrane, which in turn results in the activation of Wnt/β-cadherin signaling by increasing the nuclear translocation of β-cadherin, finally leading to SSC proliferation and stemness gene expression.

As a whole, we have identified a CHD1L–miR-486–MMP2 regulatory axis as a new molecular mechanism through which CHD1L regulates SSC stemness and growth properties. Our findings provide novel insights into normal spermatogenesis and may offer a therapeutic option for male infertility.

MATERIALS AND METHODS

Small RNA sequencing.

Purified total RNAs and small RNAs were analyzed using the Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA). Small RNA libraries were prepared from purified RNAs with the TruSeq small RNA sample preparation kit (Thermo Fisher) and quantified using the Qubit 2.0 fluorometer (Thermo Fisher). The samples were subjected to cluster generation and sequencing using a IIx (Illumina, San Diego, CA) in a single-read 1 × 36-nt multiplex format.

Data processing of miRNA sequencing.

The raw sequence files generated underwent quality control analysis using FASTQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Before aligning sequences to a reference genome, the sequencing adaptors (Illumina) used in the sequencing library preparation were trimmed using Fastx (fastx_toolkit 0.0.13.2). In order to generate sRNAbench sequence alignments, CLC Genomics Workbench 6.0 was used. The alignment against Sanger miRBase (version 20) was performed with the following parameters: seed length for alignment = 20, minimum read count = 2, bowtie alignment type = n, allowed number of mismatches = 0, minimum read length = 15, and maximum number of multiple mappings = 10. miRNA expression among the several samples was normalized to the read-per-million (RPM) value as computed by sRNAbench. Heat maps of expressed miRNAs were generated using tMev version 4.9 on RPM-normalized values.

Prediction of miRNA targets and functional analysis of target genes.

Computationally predicted mRNA targets of the top 20 up- and downregulated miRNAs were identified using five well-established miRNA target prediction databases (Target Scan, miRanda, mi RDB, miRWalk, and PicTar5). We performed gene ontology (GO) and KEGG pathway analyses in order to investigate the downstream genes of CHD1L using DAVID Bioinformatics Resources 6.8 (https://david.ncifcrf.gov/home.jsp). Ingenuity Pathway Analysis (IPA) core analysis was used, and only statistically significant data were considered (FDR ≤ 0.05, Benjamini-Hochberg for biological function, and P ≤ 0.05 for pathways analysis).

SSC isolation, culture, and CHD1L knockdown.

Mouse THY1⁺ SSCs were isolated and maintained on mitomycin C-treated (Sigma) mouse embryonic fibroblast (MEF) feeder cells as described in our previous study (25). Mouse SSC line C18-4 was obtained from Shaorong Gao (School of Life Sciences and Technology of Tongji University, Shanghai, China) and cultured in our laboratory using the culture medium described in our previous study (25). MEFs were isolated from BALB/c mouse embryos and cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) supplemented with 10% (vol/vol) fetal bovine serum (FBS) at 37°C and 5% CO₂. The procedures for nontargeted and CHD1L shRNA lentivirus generation and CHD1L knockdown in SSCs were similar to those used in our previous study (25).

RNA isolation and RT-qPCR analysis of pri-miRNAs, miRNAs, and mRNAs.

Total RNAs for mRNA quantitation were isolated from SSCs using TRIzol reagent (Qiagen, CA), and small RNAs were purified using the miRNA isolation kit (Qiagen) according to the manufacturer’s instructions. To remove genomic DNA, all the samples were treated with RNase-free DNase I according to the manufacturer’s protocol (Thermo Fisher). RNA quantity and quality were determined using a NanoDrop spectrophotometer. Total RNAs were reverse transcribed using the PrimeScript RT reagent kit (TaKaRa Bio, Shiga, Japan) following the manufacturer’s instructions, and real-time quantitative PCR (RT-qPCR) detection of mRNA expression was performed using SYBR premix Ex Taq (TaKaRa) with the StepOne plus real-time PCR system (ABI). TaqMan pri-miR assays (Applied Biosystems) were used to detect the mature miRNAs and pri-miR-486 according to the manufacturer’s instructions. The expression levels of miRNAs and mRNAs were normalized against those of the endogenous snRNA U6 and GAPDH, respectively. Relative quantification of gene expression was performed using the comparative cycle threshold (C_T) method (2^–ΔΔCT). The results were expressed as fold changes against their respective control treatments. All RT-qPCR reactions were performed in triplicate. All the primers are listed in Table 1.

TABLE 1.

Primers used in this study

Gene	Primer sequence (5′–3′)	Application
Gapdh F	GCCTCAAGATCAGCAAT	RT-qPCR
Gapdh R	AGGTCCACCACTGACACGTT	RT-qPCR
Chd1l F	GACCTGAGTTTGGGTGATG	RT-qPCR
Chd1l R	CGGATAAGTAGTTCGGTA	RT-qPCR
Oct4 F	GGCGTTCTCTTTGGAAAGGTGTTC	RT-qPCR
Oct4 R	CTCGAACCACATCCTTCTCT	RT-qPCR
Nanog F	GCTGAGATGCCTCACACGGAG	RT-qPCR
Nanog R	TCTGTTTCTTGACTGGGACCTTGTC	RT-qPCR
Mmp2 F	CCCCATGTGTCTTCCCCTTC	RT-qPCR
Mmp2 R	GTCAGTATCAGCATCGGGGG	RT-qPCR
Plzf F	GTGCCTCGCCATACCAGTGTAC	RT-qPCR
Plzf R	CCCCTTTTCTTTTCTGTTATTCTTTT	RT-qPCR
c-Myc F	GGTGGGTCGTCGAGTGCTAG	RT-qPCR
c-Myc R	AGTGGTTACCGCCTTGTTGTTA	RT-qPCR
Mmp2 3′UTR (WT) F	CGAGCTCGGCCCTATCATCTTCATCGC	pmiR-Mmp2
Mmp2 3′UTR (WT) R	GCTCTAGATATTTATACTTGTTTGCCATTT	pmiR-Mmp2
Mmp2 3′UTR (MUT) F	CCTGGCATGGGGCAGTGGGCTGACGAT GTGGCCAAGGAAATC	pmiR-Mmp2
Mmp2 3′UTR (MUT) R	GATTTCCTTGGCCACATCGTCAGCCCA CTGCCCCATGCCAGG	pmiR-Mmp2
U6 F	CTCGCTTCGGCAGCACA	RT-qPCR
U6 R	AACGCTTCACGAATTTGCGT	RT-qPCR
miR-486 F	ACACTCCAGCTGGGCGGGGCAGCTCAGTA	RT-qPCR
miR-486 R	TGGTGTCGTGGAGTCG	RT-qPCR
miR-486 shRNA	TCCTGTACTGAGCTGCCCCGAG	miR-486 sh
1,080-bp upstream region F	GGGGTACCCCCTTGTAACTCCCAGTGGCTG	pGL3-miR-486
1,080-bp upstream region R	GAAGATCTGATGAAGGTCCACATCCTCCT	pGL3-miR-486
Intron F	GGGGTACCGTAACACCAGGCAAGCGGCAG	pGL3-miR-486
Intron R	GAAGATCTCTGGAGTTAGAAGAAGGGAAA	pGL3-miR-486
Distal region F	GGGGTACCCCCTTGTAACTCCCAGTGGCTG	pGL3-miR-486
Distal region R	GAAGATCTGAGAGGCTCACAGCACAGGA	pGL3-miR-486
Middle region F	GGGGTACCCAACCCAGCCTGAGGAGAGC	pGL3-miR-486
Middle region R	GAAGATCTCTTGGTCTCAGGAAACACAT	pGL3-miR-486
Proximal region F	GGGGTACCATACAGCACCTGCTGTCTCC	pGL3-miR-486
Proximal region R	GAAGATCTGATGAAGGTCCACATCCTCCT	pGL3-miR-486

Cell proliferation assay.

Cell viability was evaluated by measuring the absorbance of MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] (Promega). Two thousand C18-4 cells were seeded in each well of a 96-well plate and cultured overnight, followed by various treatments as indicated in the individual figure legends. The viable cells were counted and evaluated by MTS assay according to the manufacturer’s protocol. The absorbance value of each experimental well was read at 490 nm on a spectrophotometer and normalized by subtracting the mean of blank-well values. The experiments were performed in sextuplicate.

Immunofluorescent staining.

The cultured C18-4 cells were fixed and incubated with primary antibodies against the phosphorylated histone H3 (pH3) (Abcam), β-catenin (Abcam), or N-cadherin (Abcam) at 4°C overnight, followed by incubation with Alexa Fluor-conjugated secondary antibody. Cell nuclei were labeled with DAPI (4,6-diamidino-2-phenylindole dihydrochloride hydrate) (Vector Laboratories, CA). Cells were examined using fluorescence microscopy (Leica, Wetzlar, Germany), and images were taken randomly. pH3-positive cells and total nuclei were counted from 10 randomly selected fields under a 200× objective, and the percentage of pH3-positive nuclei was calculated. The experiment was performed in triplicate.

Immunohistochemical analysis of mouse testes.

Testes harvested from newly born (5 dpp) or adult (56 dpp) mice were processed for cryosectioning. Microscopic sections of the mouse testes were air dried and fixed in cold acetone at 4°C, followed by incubation with a peroxidase blocking solution (3% H₂O₂) and then 10% goat serum (Dako). Thereafter, the sections were incubated with a primary antibody against MMP2 (ab37150; Abcam) or rabbit IgG at 4°C overnight. The sections were then incubated with goat anti-rabbit IgG-conjugated horseradish peroxidase and then with 3,3′-diaminobenzidine (Dako). After counterstaining with hematoxylin and mounting, the slides were examined using an Olympus BX61 microscope and images taken using a BX-PMTVC camera.

Western blotting.

Protein concentrations of C18-4 cells or mouse testis samples were measured using the Micro bicinchoninic acid (BCA) assay (Pierce, Milwaukee, WI). The extracted proteins (25 μg) were loaded on 12% SDS-polyacrylamide gels and then transferred onto polyvinylidene difluoride (PVDF) membranes (Hybond P). The nonspecific binding was blocked with phosphate-buffered saline (PBS) buffer containing 0.1% Tween 20, 2% BSA, and 5% nonfat dry milk, and the blots were then incubated with anti-mouse CHD1L (Abcam), anti-rabbit Mmp2 (Abcam), or anti-mouse β-actin (Sigma). After washing with 0.1% Tween 20 in PBS, the blots were incubated with horseradish peroxidase-conjugated anti-mouse or anti-rabbit IgG (Santa Cruz) at room temperature for 1 h. Visualization was developed using the ECL plus Western blotting detection kit (Amersham, Helsinki, Finland). Quantification of blot intensities was performed using Image software (Bio-Rad, Hercules, CA) according to the developer’s protocol. β-Actin was immunoblotted and visualized as a loading control.

Flow cytometry analysis for apoptosis and cell cycle.

Cells with various treatments were collected and labeled with annexin V-fluorescein isothiocyanate and propidium iodide (PI), provided in the annexin V-Fluos staining kit (BD Biosciences, San Jose, CA), according to the manufacturer’s instructions. Labeled cells were subjected to flow cytometry analysis. Cells labeled with annexin V but not PI were counted as apoptotic cells. For cell cycle analysis, cells with different treatments were collected and fixed with 70% ice–ethanol overnight at 4°C. After being spun down, the cell pellets were gently loosened and subsequently stained with PI-RNase buffer (BD Biosciences, San Jose, CA) according to the manufacturer’s instructions. The data were collected and analyzed on a BD FACSCalibur flow cytometer with the FlowJo software. The experiments were performed in triplicate and repeated three times.

miR-486 promoter, MMP2 3′UTR reporter, and mutation of miR-486 binding site within MMP2 3′UTR reporter.

The miR-486 promoter DNAs (the 1,080-bp upstream region and the intron of the Ank1.5 gene) were amplified from mouse genomic DNA by PCR using primers reported in the previous study (36) and shown in Table 1. Amplified DNA fragments were cloned into the pGL3-basic vector (Promega, UK), and the constructs designated pGL3-miR-486_1080 bp and _Intron, as indicated in Fig. 2. Similarly, the distal, middle, and proximal regions of the 1,080-bp upstream region of the Ank1.5 gene were PCR amplified using primers reported in the previous study (36) and shown in Table 1 and then cloned into the pGL3-basic vector to generate the pGL3-miR-486_1080bp-Distal, _1080bp-Middle, and _1080bp-Proximal constructs, respectively.

Reporter vectors harboring 3′ untranslated region (3′UTR) sequences of the murine Mmp2 gene were created using cDNA from SSCs. The 3′UTR of the murine Mmp2 gene was amplified by PCR with primers shown in Table 1 and cloned into the pmirGLO dual-luciferase miRNA target expression vector (Promega), designated pmiR-Mmp2-WT. The miR-486 binding site mutation was introduced into pmiR-Mmp2-WT by using the QuikChange site-directed mutagenesis kit (Agilent Technologies) according to the manufacturer’s instructions and the construct designated pmiR-Mmp2-Mut. All plasmids were verified by DNA sequencing.

Transient-transfection and luciferase assays.

Luciferase assays for the miR-486 promoter or Mmp2 3′UTR reporters were conducted as in previous studies (26, 29, 62–65). Briefly, for gene promoter activity assays, SSCs were cotransfected with vectors expressing individual miR-486 promoters (pGL3-miR-486__1080 bp, _Intron, _1080 bp-Distal, _1080 bp-Middle, and _1080 bp-Proximal; 0.15 μg/2.5 × 10⁴ cells) and pRenilla (15 ng/2.5 × 10⁴ cells) using Lipofectamine 2000 according to the manufacturer’s instructions. Transfected cells were cultured for 48 h prior to analysis. The dual-luciferase reporter assay system was used for detecting luciferase and Renilla activities according to the protocol provided in the system. The relative luciferase unit (RLU) was defined as the ratio of luciferase activity versus Renilla activity with that of the control set as 1.0.

For Mmp2 3′UTR reporter activity assays, SSCs were cotransfected with individual reporter genes (pmiR-Mmp2-WT or pmiR-Mmp2-Mut; 0.15 μg/2.5 × 10⁴ cells) and the control or miR-486 mimic (25 nM) using Lipofectamine 2000 according to the manufacturer’s instructions. The pRL-TK vector (0.02 μg/2.5 × 10⁴ cells; Promega) was included in all transfection assays as an internal control. Luciferase and Renilla activities were detected 48 h after transfection using a standard protocol. The relative luciferase unit (RLU) was defined as the ratio of luciferase activity versus Renilla activity with that of the control set as 1.0.

miRNA mimics and siRNA transfection.

The miR-486 mimic was obtained from Ribobio (Guanzhou China). A scrambled control mimic with no homology to the human or mouse genome was used as the control. C18-4 cells were transfected with the miRNA mimics in 6-well plates at a final concentration of 20 nM using Lipofectamine RNAiMax (Invitrogen, Carlsbad, CA). Similarly, the siRNA against Mmp2 or the scrambled sequence (Qiagen, CA) was transfected to C18-4 cells at a final concentration of 50 nM.

miR-486 shRNA generation and infection.

The shRNA sequences targeting miR-486 were designed online (Invitrogen RNAi design and ordering; ThermoFisher Scientific) and are listed in Table 1. A scrambled sequence that shared no homology with mammalian genomes was used as the control. The pGC-LV, pHelper 1.0, and pHelper 2.0 packaging system and the 293T cell line were used for the production of pseudoviral particles using procedures similar to those described in our previous study (25). C18-4 cells were infected with shRNA-expressing lentiviral particles at a ratio of 20 particles to 1 cell for 12 h, and the infection efficiency was evaluated by green fluorescent protein (GFP) expression. The infected cells were maintained in SSC culture medium. RT-qPCR analysis was performed to evaluate the knockdown effect for miR-486 in SSCs.

Statistical analysis.

Statistical analysis was performed using GraphPad Prism 5. Numbers (n) in the figure legends refer to the number of independent experiments. The two-tailed unpaired Student t test and one- (or two-)way analysis of variance (ANOVA) with the Dunnett (or Bonferroni) post hoc test were applied for comparisons between two or multiple groups, respectively. A P value of <0.05 was considered statistically significant.