Uhrf1 governs the proliferation and differentiation of muscle satellite cells.

DNA methylation is an essential form of epigenetic regulation responsible for cellular identity. In muscle stem cells, termed satellite cells, DNA methylation patterns are tightly regulated during differentiation. However, it is unclear how these DNA methylation patterns affect the function of satellite cells. We demonstrate that a key epigenetic regulator, ubiquitin like with PHD and RING finger domains 1 (Uhrf1), is activated in proliferating myogenic cells but not expressed in quiescent satellite cells or differentiated myogenic cells in mice. Ablation of Uhrf1 in mouse satellite cells impairs their proliferation and differentiation, leading to failed muscle regeneration. Uhrf1-deficient myogenic cells exhibited aberrant upregulation of transcripts, including Sox9, with the reduction of DNA methylation level of their promoter and enhancer region. These findings show that Uhrf1 is a critical epigenetic regulator of proliferation and differentiation in satellite cells, by controlling cell-type-specific gene expression via maintenance of DNA methylation.

Introduction

Adult skeletal muscle has a great capacity for regeneration, which is mediated by muscle satellite (stem) cells that express the Pax7 gene (Seale et al., 2000). Genetic ablation of Pax7-expressing satellite cells in mice led to a drastic loss of muscle tissue and loss of muscle regeneration (Lepper et al., 2011; McCarthy et al., 2011; Murphy et al., 2011; Sambasivan et al., 2011). During regeneration, satellite cells are rapidly activated to proliferate and give rise to myoblasts that fuse to form new skeletal muscle fibers. Although it has been reported that many intrinsic and extrinsic cell signals contribute to the maintenance, activation, and differentiation of satellite cells, little information is available on the role of epigenetic regulation, especially maintenance of DNA methylation patterns, in satellite cells during muscle regeneration.

DNA methylation is critical for regulating gene transcription. Several studies reported that DNA methylation patterns are altered during the proliferation and differentiation of mouse and human myogenic cells (Carrió et al., 2015; Miyata et al., 2015; Tsumagari et al., 2013). A family of DNA methyltransferases (DNMT) mediates DNA methylation by enzymatic activities including de novo methylation (catalyzed by DNMT3a and DNMT3b) or maintenance of methylation (by DNMT1). Deleting DNMT3a in satellite cells led to loss of proliferation with increased expression of the Cdkn1c (Naito et al., 2016) or the Gdf5 gene (Hatazawa et al., 2018). Further, we reported that deletion of Dnmt1 in satellite cells impaired muscle regeneration in vivo, resulting in a reduced number of myogenic cells in vitro (Iio et al., 2021). Thus, identification of epigenetic regulator is important to understand the mechanisms by which DNA methylation governs the proliferation and differentiation of satellite cells.

Ubiquitin like with PHD and RING finger domains 1 (Uhrf1; also known as Np95 in mice and ICB90 in humans), acts as an epigenetic regulator; it is essential for maintenance of DNA methylation by recruiting DNMT1 to hemi-methylated DNA sites (Bostick et al., 2007; Sharif et al., 2007), and it also can interact with DNMT3a and DNMT3b (Meilinger et al., 2009). Consistently, UHRF1 is upregulated in proliferating cells, such as neural stem cells and basal stem cells in the airway, and downregulated during differentiation or quiescence (Ramesh et al., 2016; Xiang et al., 2017). Previous reports have shown that ablation of Uhrf1 in different cell types, including embryonic stem cells (Sharif et al., 2007), T cells (Obata et al., 2014), hematopoietic stem cells (Zhao et al., 2017), chondrocytes (Yamashita et al., 2018), and adult neural stem cells (Blanchart et al., 2018), leads to DNA hypomethylation, resulting in abrogation of proliferation and/or differentiation. However, no information is available on the functions of Uhrf1 in satellite cells in muscle regeneration. Here, we investigated the expression pattern and function of Uhrf1 within satellite cells and their progeny in vivo and in vitro by using satellite cell-specific Uhrf1 knockout mice.

Results

Uhrf1 is active in myogenic cells after muscle injury and during in vitro culture

To determine whether and when Uhrf1 is expressed in muscles, the tibialis anterior (TA) muscles were injured by cardiotoxin (CTX) injection; uninjured and injured TA muscles were collected on different days postinjury (dpi). During muscle regeneration, the Uhrf1 transcript level peaked at 4 dpi and then decreased by 14 dpi (Figure S1A). To determine which myogenic cells contain active Uhrf1, immunofluorescence staining of Uhrf1, Pax7, and MyoD was performed in TA muscle sections (Figure 1A). In uninjured TA muscles, almost none of the Pax7+ satellite cells expressed Uhrf1 (Figure 1B). Although only 21% of MyoD+ cells were Uhrf1+ at 2 dpi, about 70% of MyoD+ cells expressed Uhrf1 at 3 dpi, when the number of myogenic cells peaked (Ogawa et al., 2014). At 5 dpi, when the number of Pax7+ cells peaked during normal regeneration postinjury (Murphy et al., 2011), 11% of Pax7+ cells expressed Uhrf1, and this percentage decreased to 5.8% at 14 dpi (Figure 1B), which was compatible with the Uhrf1 transcript level in injured TA muscles. More EdU+ cells were found in Pax7+Uhrf1+ cells than Pax7+Uhrf1- cells at 5 dpi (Figure S1B), indicating Uhrf1 is expressed in proliferating myogenic cells. In addition, Uhrf1 expression gradually decreased after induction of differentiation in C2C12 cells (Figures S1C–S1E).

Figure 1

Uhrf1 is active in myogenic cells after muscle injury and during in vitro culture

(A) Immunostaining of Pax7, MyoD, and Uhrf1 during muscle regeneration. Arrows show Uhrf1+Pax7+ or Uhrf1+MyoD + cells.

(B) Quantification of Uhrf1+ cells among Pax7+ or MyoD+ cells in uninjured and injured TA muscles collected at 2, 3, 5, and 14 dpi.

(C) Time course of myogenic cells cultured in vitro.

(D) Immunostaining of Uhrf1, Pax7, Ki67, Myogenin, and Myh1 after 1, 3, 4, and 7 days of culture.

(E) Quantification of Uhrf1+ cells in vitro.

(F) Quantification of proliferating (Ki67+) cells in vitro.

(G) Quantification of Ki67+ cells among Uhrf1+ cells after 3 or 4 days of culture.

(H) Quantification of Myogenin+ cells among Uhrf1+ cells after 4 days of culture.

(I) Quantification of Myh1+ cells in Uhrf1+ fibers after 7 days of culture. For all graphs, data were pooled from three or four independent experiments and are expressed as the mean with 95% CI. Scale bars: 20 μm. See also Figure S1

To further examine the expression pattern of Uhrf1, satellite cells were isolated by FACS and cultured for up to 7 days in vitro (Figure 1C). Almost all cells were positive for Pax7 on day 1, but none of these cells expressed Uhrf1 (Figure 1D), which supported the in vivo results in intact muscles. The majority of myogenic cells (78%) were positive for Uhrf1 on day 3, and this percentage gradually decreased during differentiation (Figure 1E). The expression pattern of Ki67+ proliferation cells was similar to that of Uhrf1+ cells (Figure 1F). Whereas almost all Uhrf1+ cells expressed Ki67 on day 3 and 4 (Figure 1G), only 10% of Uhrf1+ cells expressed Myogenin on day 4 (Figure 1H), and almost no Uhrf1+ cells expressed Myh1 on day 7 (Figure 1I). These observations indicate that Uhrf1 is not expressed in satellite cells during quiescence but is transiently expressed in proliferating myogenic cells before being downregulated during differentiation.

Loss of Uhrf1 expression in satellite cells impairs muscle regeneration

To determine whether Uhrf1 expression in satellite cells and their progeny is necessary for muscle regeneration, we selectively deleted this gene using Pax7 (Lepper et al., 2009) and Uhrf1 (Skarnes et al., 2011) mice. The efficiency of Uhrf1 ablation in satellite cells and their progeny was measured by isolating satellite cells by FACS from hindlimb muscles of Pax7;Uhrf1 mice following administration of oil (control mice), Pax7 mice following administration of tamoxifen (TMX, Cre control mice), or Pax7;Uhrf1 mice following administration of TMX (mutant mice) for 5 consecutive days (Figures S2A and S2B). Deletion was highly efficient, as the Uhrf1 transcript and protein level was 90% lower in mutant mice than Cre control myogenic cells (Figures S2C and S2D). Together, these experiments indicate that Uhrf1 expression was effectively abolished in satellite cells and their progeny from Pax7;Uhrf1 mice.

To investigate the effects of Uhrf1 expression loss in satellite cells and their progeny on intact muscle and during muscle regeneration, the TA muscles of controls and mutant mice were injured by CTX and harvested at 3 and 5 dpi (Figure 2A). Because Pax7 mice treated with TMX (Cre control mice) exhibit a mild muscle regeneration phenotype (Lahmann et al., 2021; Mademtzoglou et al., 2018; Sakai et al., 2020), Cre control mice were included in the experiments for regeneration. There was no change in muscle mass or the number of Pax7+ cells in uninjured TAs of each mouse (Figures S2E–S2G), because quiescent satellite cells originally lack the expression of Uhrf1 (Figure 1D). At 3 dpi, there was a significant decrease in the number of MyoD+ cells in mutant mice (Figures 2B and 2C), indicating a decrease in myogenic cells. The regenerating fibers, characterized by expression of embryonic myosin heavy chain (Myh) 3, were detected at 5 dpi in control mice (Figure 2D), similar to a previous report (Murphy et al., 2011). Although there was no difference in the muscle mass (Figure 2E), the Myh3 level was decreased in Cre control mice compared with the control, whereas the proportion of Myh3+ regenerating fibers was severely reduced in mutant mice (Figure 2F). These results indicate that loss of Uhrf1 affects the regeneration of new fibers. In addition, there was a significant decrease in the number and proliferation of Pax7+ cells (Figures 2G–2I) at 5 dpi in mutant mice compared with control and Cre control mice.

Figure 2

Loss of Uhrf1 expression in satellite cells impairs muscle regeneration

(A) Experimental design for TMX treatment followed by CTX injection and sample harvest.

(B) Immunostaining of MyoD in TA muscles at 3 dpi.

(C) Quantification of MyoD+ cells at 3 dpi (n = 5 mice/condition).

(D) Immunostaining of Myh3 in TA muscles at 5 dpi.

(E) Mass of TA muscles at 5 dpi (n = 5 mice/condition).

(F) The average area of Myh3+ regenerating myofibers in TA muscle cross-sections at 5 dpi (n = 3 or 5 mice/condition).

(G) Immunostaining of Pax7 in TA muscle cross-sections at 5 dpi.

(H) Quantification of Pax7+ cells at 5 dpi (n = 6 mice/condition).

(I) Quantification of Pax7+EdU+ cells at 5 dpi (n = 6 mice/condition).

(J) Whole-mount TA at 28 dpi.

(K) Mass of TA muscles at 28 dpi (n = 4 or 5 mice/condition).

(L) The average total area of TA muscle cross-sections at 28 dpi (n = 3 or 4 mice/condition).

(M) Immunostaining of Pax7 in TA muscle cross-sections at 28 dpi.

(N and O) The total and average number of Pax7+ cells in TA muscle cross-sections at 28 dpi (n = 4 mice/condition). Statistical significance was determined by Welch's ANOVA with Dunnett's T3 multiple comparison test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Scale bars: 50 μm in (B, G, and M), 500 μm in (D), and 5 mm in (J). See also Figure S2

To determine whether the observed phenotype in the early phase of the muscle regeneration was sustained during the process, injured TA muscles were harvested at 28 dpi (Figure 2A), at which time regeneration is largely complete (Murphy et al., 2011). Although there was no difference in the TA muscle mass between control and Cre control muscle according to macroscopic observations (Figure 2J), the mutant mice exhibited significant muscle mass loss in the TA muscles (Figure 2K). There was a significant decrease in the total CSA in mutant mice (Figures 2L and S2H) but no difference between the control and Cre control mice. Also, TA sections contained a significantly lower number of Pax7+ cells in mutant mice compared with control mice (Figures 2M and 2N), and this significant decrease remained even after normalization to the total CSA (Figure 2O). In summary, loss of Uhrf1 expression in satellite cells significantly impaired their proliferation and muscle regeneration.

Uhrf1 deficiency in satellite cells disturbs proliferation and differentiation of myogenic cells in vitro

To investigate the mechanisms of satellite cell impairment in Pax7;Uhrf1 mice, single myofibers were isolated from the extensor digitorum longus muscle of control and mutant mice and cultured for up to 4 days (Figure S3A). On days 0 and 2, the number of Pax7+ cells per myofiber was comparable between the mutant and control mice (Figure S3B). In contrast, the total number of myogenic cells and the proportion of Pax7-MyoD+ myoblasts on day 4 were decreased in the myofibers of mutant mice compared with control mice (Figure S3B).

To further characterize the impaired proliferation and differentiation observed in culture, satellite cells from control, Cre control, and mutant mice were characterized in vitro (Figure 3A). Uhrf1 protein was expressed in 65% of myogenic cells cultured for 3 days from both control and Cre mice but in 5.2% of those from mutant mice (Figures S3C and S3D), demonstrating efficient deletion of Uhrf1 allele in isolated myogenic cells. In EdU labeling experiments to assess proliferation, although Cre control myogenic cells reduced proliferation compared with control mice, only 24% of myogenic cells were positive for EdU in mutant mice, compared with approximately 40% in Cre control mice, on day 3 (Figures 3B and 3C). Furthermore, the number of Myogenin+ cells was significantly decreased in mutant mice on day 4 (Figures 3D and 3E). In addition, myotube formation, observed by Myh1 staining on day 7, was also compromised in the mutant mouse cells (Figures 3F and 3G). Taken together, these in vitro findings indicate that Uhrf1 is required in myogenic cells for proliferation and the loss of Uhrf1 leads to impairments of differentiation.

Figure 3

Uhrf1 deficiency in satellite cells disturbs proliferation and differentiation of myogenic cells in vitro

(A) Experimental design for TMX followed by FACS and immunostaining after 3, 4, and 7 days in culture.

(B) EdU detection in myogenic cells after 3 days in culture.

(C) Quantification of EdU+ cells after 3 days in culture (n = 4 mice/condition).

(D) Immunostaining of Myogenin after 4 days in culture.

(E) Quantification of Myogenin+ cells after 4 days in culture (n = 5 mice/condition).

(F) Immunostaining of Myh1 after 7 days in culture.

(G) Fusion index after 7 days in culture (n = 3 mice/condition). Statistical significance was determined by Welch’s ANOVA with Dunnett's T3 multiple comparison test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. Scale bars: 100 μm. See also Figure S3

Loss of Uhrf1 in satellite cells alters DNA methylation patterns and subsequently transcriptional programs

To identify the genes affected by Uhrf1 deficiency in satellite cells and their progeny, RNA-Seq was performed in cultured myogenic cells obtained from control, Cre control, and mutant mice (Figure 4A). Because a heatmap of the expression profiles showed a clear distinction between the control and Cre control myogenic cells (Figure S4A), we focused on the comparison between Cre control and mutant myogenic cells for further analysis (Figure 4B). We detected 3,644 upregulated and 3,672 downregulated genes in the mutant compared with Cre control myogenic cells. According to Gene Ontology (GO) enrichment analysis, the downregulated genes showed significant enrichment of the biological process terms related to cell proliferation and cell cycle (Figure 4C), consistent with the impairment of these processes observed in the mutant cells. The genes that were upregulated in mutant myogenic cells showed enrichment of myogenesis-independent GO terms such as “positive regulation of smooth muscle cell migration” and “blood vessel development.” Because no caspase in the critical apoptotic signaling pathway was upregulated in the mutant cells (Table S1), the activation of apoptosis may not be a main cause of the phenotype of the mutant cells. In addition, gene sets associated with transposable elements (TEs), such as “regulation of transposition,” was also enriched among the upregulated genes (Figure 4C). As loss of DNA methylation due to Uhrf1 deficiency in embryonic stem cells (Sharif et al., 2007) or neural stem cells (Ramesh et al., 2016) leads to activation of TEs, we thus examined the expression of TEs using RNA-Seq data by STAR program (v2.7.6a) (Dobin et al., 2013) with dedicated parameters and algorithms for TE analysis (see STAR Methods). We found upregulation of intracisternal A particles (IAPs) in mutant compared with Cre control myogenic cells (Figure 4D), which is in line with the results from Uhrf1-deficient embryonic stem cells (Bostick et al., 2007; Sharif et al., 2007).

Figure 4

Loss of Uhrf1 in satellite cells alters DNA methylation patterns and subsequently transcriptional programs

(A) Experimental design for RNA-Seq and MBD2-Seq.

(B) Heatmap of expression similarity between Cre control and mutant samples.

(C) Bar graph of the Gene Ontology (GO) terms enriched among the upregulated and downregulated genes in mutant myogenic cells.

(D) Heatmap showing the RNA-Seq Z score of top 10 differentially expressed retrotransposon families.

(E) Distribution of Uhrf1-mediated methylated DNA with the given intervals and scores, with the genomic features determined using a cis-regulatory element annotation system.

(F) Venn diagram of the upregulated genes associated with Uhrf1-mediated methylated DNA loci.

(G) Bar graph of the GO terms enriched among the upregulated genes associated with Uhrf1-mediated methylated DNA loci.

(H) Methylated DNA signals around the Sox9 loci in control and mutant myogenic cells visualized using IGV.

(I) Experimental design for the siRNA treatments.

(J) Quantification of EdU+ Sox9-knockdown myogenic cells (n = 4 mice/condition). Statistical significance was determined by Welch's ANOVA with Dunnett's T3 multiple comparison test. ∗p < 0.05, ∗∗∗p < 0.001. See also Figure S4.

To determine whether the genes upregulated by Uhrf1 deficiency were specific to myogenic cells, we compared the upregulated genes in chondrocytes (Yamashita et al., 2018) and hematopoietic stem cells lacking Uhrf1 (Zhao et al., 2017) with those in myogenic cells (Figure S4B). Among the 3,644 genes upregulated in Uhrf1 deficient-myogenic cells, 436 overlapped with the 1,122 genes upregulated in Uhrf1-knockout chondrocytes. On the other hand, only 82 genes overlapped with the 626 genes upregulated in hematopoietic stem cells lacking Uhrf1. These results indicate that Uhrf1 deficiency induces upregulation of partially overlapped common genes regardless of cell types, but cell-specific knockout of Uhrf1 may lead to upregulation of cell-specific genes by disruption of cell-type-specific maintenance DNA methylation mechanisms.

We next performed methyl-CpG-binding domain protein-2-enriched genome sequencing (MBD2-Seq) in the control and mutant cultured myogenic cells (Figure 4A). The higher GC count per read in methylated DNA compared with unmethylated DNA (Figure S4C) and the distinct methylation patterns observed at the H19 and Kcnq1ot1 loci, which are imprinted genes (Figure S4D), indicated proper enrichment of methylated DNA by MBD2. To identify DNA loci methylated by Uhrf1, peak calling was performed using MACS2. We identified 45,582 peaks, and these loci were enriched in promoters, downstream regions, and gene bodies of the genome (Figure 4E). To identify Uhrf1 target genes in myogenic cells, we analyzed the genes upregulated by Uhrf1 deficiency that overlapped with the Uhrf1-mediated DNA methylated peaks. We found 196 genes with Uhrf1-mediated peaks within -2000 bp from the transcriptional start site (Figure 4F). GO enrichment analysis showed that myogenesis-independent gene sets, such as “embryonic organ morphogenesis,” “smooth muscle tissue development,” and “ossification,” were enriched in these genes (Figures 4G and S4F).

Among these genes, Sox9, which is a key transcription factor in chondrocytes differentiation and skeletal development (Bi et al., 1999), was associated with all of the top six GO categories. We also detected reduced DNA methylation levels in the regions upstream of the Sox9 loci in mutant myogenic cells (Figure 4H). This region is annotated as proximal enhancer by the ENCODE (Consortium et al., 2020) and overlapped with H3K4me1 peaks, indicating the primed enhancer, in MC3T3-E1 osteoblastic cell line expressing Sox9 (Figure S4E) (Oki et al., 2018). To explore whether upregulation of Sox9 is responsible for the impaired proliferation and differentiation of Uhrf1-deficient myogenic cells, we treated mutant myogenic cells with siRNA targeting Sox9 (Figure 4I). Knockdown of Sox9 (Figure S4G) significantly rescued the impaired proliferation of the mutant myogenic cells (Figures 4J and S4H). However, Sox9 knockdown was not sufficient to rescue the impaired differentiation of the mutant myogenic cells (Figures S4I and S4J). Taken together, Uhrf1-mediated DNA methylation in satellite cells and their progeny controls specific gene expression in a cell-type-specific manner and negatively regulates the aberrant expression of genes that can induce the differentiation toward other cell type, such as Sox9, for physiological proliferation and/or differentiation of satellite cells.

Discussion

In this study, we identified Uhrf1 as a critical regulator of skeletal muscle satellite cells. Notably, Uhrf1 is not expressed during quiescence but is upregulated during proliferation and downregulated during differentiation in satellite cells. Our findings agree with previous reports showing that Uhrf1 is necessary for maintaining the DNA methylation profile during DNA replication in proliferating cells and downregulated during differentiation or quiescence other than satellite cells (Ramesh et al., 2016; Xiang et al., 2017). As Uhrf1 overexpression was reported to contribute to cancer initiation and progression (Mudbhary et al., 2014), tight regulation of Uhrf1 expression in stem cells appears to be critical for self-renewal and differentiation.

Further, ablation of Uhrf1 in satellite cells led to impaired muscle regeneration, similar to Dnmt1- (Iio et al., 2021) or Dnmt3a-deficient mice (Naito et al., 2016). The marked loss of EdU incorporation indicated a remarkable reduction in the number of Uhrf1-deficient myogenic cells in vitro and in vivo at 28 dpi. Therefore, the impairment of muscle regeneration of mutant mice could result from decreased numbers of myogenic cells.

In addition, we report that loss of Uhrf1 in satellite cells alters transcriptional programs caused by alteration of genome-wide DNA methylation patterns. Consistent with the impaired proliferation of Uhrf1-deficient myogenic cells, we found that many cell-cycle-related genes were downregulated. We propose that Uhrf1 deficiency in mutant myogenic cells results directly in hypomethylation and abnormal upregulation of transcriptions. Indeed, the GO term enriched among the upregulated genes by hypomethylation in Uhrf1-deficient myogenic cells was myogenesis independent such as “embryonic organ morphogenesis” and “smooth muscle tissue development.” As satellite cells have a potential to adopt a pericyte phenotype following stimulation of the Notch and PDGF signaling pathways (Gerli et al., 2019), loss of Uhrf1 in satellite cells might trigger partial transdifferentiation into vascular-like cells. We found that TEs, especially IAPs, were strongly upregulated in Uhrf1-deficient myogenic cells. This phenotype is comparable with that of Uhrf1-deficient cortical neural stem cells showing strong activation of IAPs (Ramesh et al., 2016). Because long-terminal repeats in IAPs could change gene regulation at a given locus (Morgan et al., 1999), further studies are needed to clarify whether transposition of IAPs into new genomic loci would contribute to the phenotype of mutant satellite cells.

Finally, we found the aberrant expression of Sox9, which is a transcription factor that is essential for chondrocyte differentiation and cartilage formation, in mutant myogenic cells with the reduction of DNA methylation at its proximal enhancer. We showed that suppression of the increased Sox9 in Uhrf1-deficienct myogenic cells partially rescued their impaired proliferation. These results are congruent with the report showing that Sox9 overexpression in chondrocytic cells suppress their cell cycle (Panda et al., 2001). The fact that Sox9-expressing chondrocytes and skeletal muscle cells share a developmental origin as somites (Lagha et al., 2008) might have influenced the aberrant expression of Sox9 in Uhrf1-deficienct myogenic cells. Thus, we propose that Uhrf1 epigenetically governs, at least in part, appropriate gene expression and suppression in satellite cells for their ideal proliferation and differentiation.

In conclusion, the present study points Uhrf1 as a regulator of self-renewal and differentiation of satellite cells via genome-wide DNA methylation patterning.

Limitations of the study

Knockdown of Sox9 partly rescued the impaired proliferation but not the differentiation of Uhrf1-deficient myogenic cells. Therefore, other factors or mechanisms, including not only the transcripts regulated by Uhrf1 but also the chromatin state such as histone modifications, can also regulate the behavior of mutant myogenic cells. Considering the enrichment of GO terms such as “DNA repair” in downregulated genes in mutant myogenic cells, the genome instability seems to be one of the mechanisms for explaining the phenotype of mutant myogenic cells. In addition, we found a number of phenotypes in Uhrf1-deficient myogenic cells overlapped with Dnmt1-deficient ones (Iio et al., 2021). Further efforts to perform the integrated analysis of omics data, including histone modifications, of Uhrf1- and Dnmt1-deficient myogenic cells may help understand the different roles between Uhrf1 and Dnmt1 for regulating satellite cells.

STAR★Methods

Key resources table

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Yuuki Imai (y-imai@m.ehime-u.ac.jp).

Materials availability

This study did not generate new unique reagents.

Experimental model and subject details

Animals

All animals were maintained and used according to a protocol approved by the Animal Experiment Committee of Ehime University, Japan. C57BL/6JJcl mice were purchased from CLEA Japan. The Pax7 (Lepper et al., 2009); Jackson Laboratories, stock no: 012476) and Uhrf1 (B6Dnk;B6N-Uhrf1Ieg (Skarnes et al., 2011); mouse strains have been described previously. Uhrf1 mice were crossed with Pax7 mice to generate Pax7;Uhrf1 mice. All mice were housed in a specific pathogen-free facility under climate-controlled conditions and a 12-h light/dark cycle and were provided water and a standard diet ad libitum. All male animals were evaluated at age 7–12 weeks old. At 7 weeks of age, tamoxifen (150 μL, 20 mg/mL) (Sigma, cat# T5648) dissolved in corn oil (Sigma, cat# C8267) was injected intraperitoneally for 5 consecutive days to induce Cre-mediated recombination. After 1 week of tamoxifen injection, injury was induced by injecting 100 μL of 10 μM cardiotoxin (CTX) solution (Latoxan, cat# L8102) into the left tibialis anterior (TA) muscles under anesthesia using isoflurane. The right TA served as the intact control.

Cell lines

The C2C12 cell line (female, ATCC, CRL-1772) was culture in DMEM (Gibco, cat# 10569-010) supplemented with 10% FBS (Gibco, cat# 10099-141) and antibiotic–antimycotic (Gibco, cat# 15240-062) in a 37°C incubator with 5% CO2. The medium was replaced with differentiation medium (DMEM supplemented with 2% horse serum) to induce myotube formation.

Primary cell cultures

Isolated skeletal muscle-derived mononuclear myogenic cells from male mice were resuspended in DMEM/F12 (Gibco, cat# 10565-018) supplemented with 20% FBS, 2% Ultroser G (Pall, cat# 15950-017), and antibiotic–antimycotic in a 37°C incubator with 5% CO2. Cells at 5000 or 10000/cm2 were cultured on an 8-well chamber slide (Thermo Scientific, cat# 177445) or in a 6-well plate coated with 1 mg/ml Matrigel (Corning, cat# 354230). Please see the following sections for more information on isolation methods.

Method details

Immunofluorescence and microscopy

For immunofluorescence (IF) staining, muscles were frozen in liquid nitrogen-chilled isopentane, and 10 μm cryosections were prepared. For Pax7 and Uhrf1 staining, sections were fixed in 4% paraformaldehyde (PFA) in PBS for 5 min and then boiled in antigen retrieval buffer (10 mM sodium citrate, pH 6) for 10 min. For Myh3 staining, tissue sections were fixed for 10 min in acetone at −20 °C before boiling, blocked in 5% goat serum (Gibco, cat# 16210-064) in PBS for 60 min, and incubated with the primary antibodies overnight at 4°C. The antibodies used for IF staining were anti-Pax7 (DSHB, cat# PAX7, RRID: AB_528428; 1/5 dilution; or Invitrogen, cat# PA1-117, RRID: AB_2539886; 1/100 dilution), anti-MyoD (Abcam, cat# ab133627, RRID: AB_2890928; 1/100 dilution), anti-Uhrf1 (Santa Cruz Biotechnology, cat# sc-373750, RRID: AB_10947236; 1/50 dilution), anti-Myh3 (DSHB, cat# F1.652, RRID: AB_528358; 1/20 dilution), Alexa fluor 568 goat anti-mouse IgG1 (Invitrogen, cat# A-21124; RRID: AB_2535766; 1/1000 dilution), and Alexa fluor 488 goat anti-rabbit IgG (Invitrogen, cat# A-11008; RRID:AB_143165; 1/1000 dilution).

For IF staining in cells, cells were fixed in 4% PFA in PBS for 10 min, followed by permeabilization using 0.5% Triton X-100 in PBS for 10 min. Cells were blocked using 5% goat serum in PBS. The following primary antibodies were used: anti-Uhrf1 (Santa Cruz Biotechnology, cat# sc-373750, RRID: AB_10947236; 1/100 dilution; or MBL, cat# D289-3 RRID: AB_10597260; 1/100 dilution), anti-Ki67 (Novus, cat# NB500-170, RRID: AB_10001977; 1/100 dilution), anti-Myogenin (Santa Cruz Biotechnology, cat# sc-12732, RRID: AB_627980; 1/100 dilution), and anti-Myh1 (DSHB, cat# MF20, RRID: AB_2147781; 1/10 dilution). The signals were visualized using the following fluorochrome-coupled secondary antibodies: Alexa fluor 568 goat anti-mouse IgG1a, Alexa fluor 488 goat anti-rabbit IgG, and Alexa fluor 488 goat anti-rat IgG (Invitrogen, cat# A-11006, RRID: AB_2534074; 1/1000 dilution). Nuclei were counterstained with DAPI.

Stained cells or tissues were photographed using BIOREVO (Keyence) and quantified using Fiji (v2.3.0) (https://imagej.net/Fiji). The numbers of Uhrf1+, Pax7+, Ki67+, Myogenin+, and EdU+ cells in vitro and the area of Myh3+ fibers in vivo were determined using the Fiji Analyze Particles function. For total muscle CSA quantification, the whole area of transverse TA muscle sections was measured by laminin staining. The numbers of Uhrf1+, Pax7+, MyoD+, and EdU+ cells in vivo were determined by manual counting using the Adobe Photoshop count tool.

Isolation of skeletal muscle-derived mononuclear myogenic cells

Isolation of mononuclear myogenic cells from hindlimb muscles was performed as described (Liu et al., 2015) with minor modifications. Briefly, muscles were chopped in cold PBS and digested using 800 U/mL collagenase type 2 (Worthington, cat# LS004177) in F10 medium (Gibco, cat# 11550-043) supplemented with horse serum (Gibco, cat# 26050-088) and antibiotic–antimycotic (Gibco, cat# 15240-062) for 60 min followed by another digestion using 1000 U/mL collagenase type 2 and 11 U/mL dispase (Gibco, cat# 17105-041) for 30 min. The supernatants were filtered through a 40 μm cell strainer (Falcon, cat# 352340) and centrifuged to yield a cell suspension. The following antibodies were used: CD31-FITC (BioLegend, cat# 102406; RRID: AB_312901; 1/100 dilution), CD45-FITC (BioLegend, cat# 103108, RRID: AB_312973; 1/100 dilution), Sca1-APC (BioLegend, cat# 122512, RRID: AB_756197; 1/100 dilution), and Vcam1-PE (Invitrogen, cat# 12-1061-82, RRID: AB_2572573; 1/100 dilution). Cells were isolated using the FACS Aria II (Becton, Dickinson and Company) and analyzed by FlowJo v10.1r7 (BD Life Sciences).

Western blotting

Cultured myogenic cells were washed with PBS and dissolved in RIPA buffer (FUJIFILM, cat# 182–02451) containing protease inhibitors (nacalai tesque, cat# 25,955-11). Whole-cell extracts were separated by SDS-PAGE and transferred to polyvinylidene fluoride membranes. The membranes were blocked with 5% skim milk in TBS with Tween-20 (PBST), followed by incubation with an anti-Uhrf1 antibody (Santa Cruz Biotechnology, cat# sc-373750, RRID: AB_10947236; 1/500 dilution) or anti-Gapdh antibody (Cell Signaling Technology, cat# 5174, RRID: AB_10622025; 1/1000 dilution) overnight at 4 °C. After washing with PBST, the membranes were incubated with an anti-mouse (Promega, cat# W4021, RRID: AB_430834; 1:5000) or anti-rabbit immunoglobulin/HRP secondary antibody (Agilent, cat# P0448, RRID: AB_2617138; 1:5000) for 60 min at room temperature. Immunoreactive signals were detected using ECL Prime Western Blotting Detection Reagents (Amersham, cat# RPN2232) on the Image Quant LAS 4000 system (GE Healthcare).

siRNA transfection

Cells were transfected with non-targeting siRNA (Dharmacon, cat# D-001810-10-05) or siRNA targeting Sox9 (Dharmacon, cat# L-059108-01-0005) using Lipofectamine RNAiMAX (Thermo Scientific, cat# 13778-075) following the manufacturer's protocol.

EdU labelling and detection

For in vivo labelling, EdU (5-ethynyl-2′-deoxyuridine, Invitrogen, cat# A10044) was dissolved in PBS at 0.5 μg/μL and injected intraperitoneally into mice at 5 μg/g body weight 24 h before harvesting. For in vitro labelling, the cells were incubated with EdU (10 μM) for 3 h. EdU was detected using Click-iT EdU Imaging Kits (Invitrogen, cat# C10340) with Alexa Flour 647.

RNA isolation and quantitative real-time PCR

Total RNA was extracted from 3 day cultured myogenic cells using the RNeasy Plus Micro Kit (Qiagen, cat# 74034) following the manufacturer's protocol. cDNA was synthesized from total RNA using PrimeScript (Takara, cat# RR036A). qPCR was performed in duplicate samples using TB Green Premix Ex Taq II (Takara, cat# RR820S) and Thermal Cycler Dice (Takara, cat# TP850). The primer sequences are listed in Table S2.

Isolation, culture, and immunostaining of single myofibers

Single myofibers were isolated from extensor digitorum longus (EDL) muscles as described previously (Rosenblatt et al., 1995). Briefly, EDL muscles were mildly digested with 0.2% collagenase type 2 in DMEM for 90 min at 37°C. The isolated fibers were cultured in DMEM supplemented with 30% FBS, 1% chicken embryo extract (US Biological, cat# C3999), and 10 ng/mL bFGF (Wako, cat# 062-06661). Immunostaining was performed following a previous protocol (Shinin et al., 2009). The isolated fibers were fixed in 4% PFA in PBS for 10 min and then blocked using 5% goat serum in PBS for 30 min. The antibodies used for IF staining were anti-Pax7 (DSHB, cat# PAX7, RRID: AB_528428; 1/5 dilution) and anti-MyoD (Abcam, cat# ab133627, RRID: AB_2890928; 1/100 dilution).

RNA-Seq and data analysis

The integrity of isolated RNA was verified using the Agilent 2100 Bioanalyzer with RNA 6000 Pico Kit (Agilent, cat# 5067-1513). RNA-Seq libraries were prepared using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (NEB, cat# E7760S) with the NEBNext rRNA Depletion Kit (NEB, cat# E6310L) and NEBNext Multiplex Oligos for Illumina (NEB, cat# E7335S) according to the manufacturer's instructions. The constructed libraries were verified using the Agilent 2100 Bioanalyzer with DNA 1000 kit (Agilent, cat# 5067-1504) and quantified using KAPA Library Quantification Kits (KAPA Bio, cat# KK4824) and the 7500 Real-Time PCR System (ABI, cat# 7500-01). The libraries were sequenced as 50 or 150 bp paired-end reads using the HiSeq 1500 or NovaSeq 6000 Sequencing System (Illumina), respectively. Raw reads were trimmed using TrimGalore (v0.6.6; http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) and mapped to the reference genome (GRCm38, PRI) using HISAT2 (v2.2.1) (Kim et al., 2019). The option ‘--trim3 99’ was used to compare reads of different lengths. Expression level was quantified using featureCounts (v2.0.1) (Liao et al., 2014) with GRCm38.102 annotation. The differentially expressed genes were extracted by DESeq2 (v1.30.1) (Love et al., 2014). GO enrichment analysis was performed using Metascape (Zhou et al., 2019).

For transposable element (TE) analysis, raw reads were trimmed using TrimGalore and mapped to the reference genome (GRCm38, PRI, https://www.gencodegenes.org/mouse/) using STAR (v2.7.6a; Dobin et al., 2013) with the PRI annotation. For the mapping, the STAR option ‘--runMode alignReads --runThreadN 4 outMultimapperOrder Random --outSAMtype BAM SortedByCoordinate --outSAMmultNmax 1 --outFilterMultimapNmax 1000 --outFilterMismatchNmax 3 --winAnchorMultimapNmax 1000 --alignIntronMax 1 --alignMatesGapMax 350 --alignEndsType EndToEnd’ was used, allowing a higher number of multi-mapped reads (Teissandier et al., 2019). TE families were quantified using featureCounts (v2.0.1; -M -T 1 -s 0 -p) and TE annotation (http://labshare.cshl.edu/shares/mhammelllab/www-data/TEtranscripts/TE_GTF/). The differentially expressed TEs were extracted by DESeq2.

MBD2-Seq and analysis of the sequencing data

Genomic DNA was isolated and purified using the NucleoSpin Tissue Kit (MACHEREY-NAGEL, cat# 740952). Purified DNA was sheared using the Covaris S220 ultrasonicator with microTUBE AFA (Covaris, cat# 520045). Methylated and non-methylated DNA fragments were separated using the EpiXplore Methylated DNA Enrichment Kit (Clontech, cat# PT5034-2). Libraries were generated using the DNA SMART ChIP-Seq Kit (Takara, cat# 634865). The libraries were selected from 250 to 350-bp fragments obtained by E-Gel 2% SizeSelect electrophoresis (Invitrogen, cat# G661012). Size-selected libraries were verified using the Agilent 2100 Bioanalyzer with the High Sensitivity DNA kit (Agilent, cat# 5067-4626) and quantified using KAPA Library Quantification Kits (KAPA Bio, cat# KK4824) and the 7500 Real-Time PCR System (ABI, cat# 7500-01). The libraries were sequenced using NovaSeq 6000 with 150 bp paired end reads. Raw reads were down-sampled to 50 million reads for each condition to avoid the bias due to different sequencing depths. The reads were trimmed using TrimGalore and were mapped using HISAT2. Peaks were obtained by MACS2 (v2.2.6) (Zhang et al., 2008) using cells derived from TMX-treated mice as the ‘control,’ with the following options: --slocal 0 --llocal 0 -q 0.01. ChIPpeakAnno (v3.24.2) (Zhu et al., 2010) was used to annotate and visualize the peaks. Methylated DNA signals were visualized using Integrative Genomics Viewer (IGV) (Robinson et al., 2011). We downloaded the Proximal enhancer-like (pELS) bed file for mouse from the SCREEN portal (Consortium et al., 2020) (https://screen.encodeproject.org). H3K4me1 peaks of MC3T3-E1 was downloaded from ChIP-Atlas web tools (Oki et al., 2018) (http://chip-atlas.org).

Quantification and statistical analysis

The scatterplot with means are used to show values. In a bar chart, values are expressed as means ± SD (Figure S3B) or 95% CI (Figures 1 and S1B). The exact values of n are indicated in the corresponding figure legends. Two-sample test for equality of proportions was performed by Test of Equal or Given Proportions. Welch's t-test was used to compare variable between two groups. Welch's ANOVA with Dunnett's T3 multiple comparison test was used to compare variables among three or more. Data were analyzed using Prism 9 (GraphPad Software) or R (v4.0.5) (https://www.r-project.org). Statistical significance was set at p < 0.05.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-Pax7	DSHB	cat# PAX7; RRID: AB_528428
Rabbit polyclonal anti-Pax7	Invitrogen	cat# PA1-117; RRID: AB_2539886
Rabbit monoclonal anti-MyoD	Abcam	cat# ab133627; RRID: AB_2890928
Mouse monoclonal anti-Myogenin	Santa Cruz Biotechnology	cat# sc-12732; RRID: AB_627980
Mouse monoclonal anti-Uhrf1	Santa Cruz Biotechnology	cat# sc-373750; RRID: AB_10947236
Rat monoclonal anti-Uhrf1	MBL	cat# D289-3; RRID: AB_10597260
Mouse monoclonal anti-Myh1	DSHB	cat# MF20; RRID: AB_2147781
Mouse monoclonal anti-Myh3	DSHB	cat# F1.652; RRID: AB_528358
Rabbit polyclonal anti-Ki67	Novus	cat# NB500-170; RRID: AB_10001977
Rabbit monoclonal anti-Gapdh	Cell Signaling Technology	cat# 5174; RRID: AB_10622025
Rat monoclonal anti-CD31-FITC	BioLegend	cat# 102406; RRID: AB_312901
Rat monoclonal anti-CD45-FITC	BioLegend	cat# 103108; RRID: AB_312973
Rat monoclonal anti-Sca1-APC	BioLegend	cat# 122512; RRID: AB_756197
Rat monoclonal anti-Vcam1-PE	Invitrogen	cat# 12-1061-82; RRID: AB_2572573
Alexa fluor 568 goat anti-mouse IgG1	Invitrogen	cat# A-21124; RRID: AB_2535766
Alexa fluor 488 goat anti-rabbit IgG	Invitrogen	cat# A-11008; RRID:AB_143165
Alexa fluor 488 goat anti-rat IgG	Invitrogen	cat# A-11006; RRID: AB_2534074
anti-mouse immunoglobulin/HRP	Promega	cat# W4021; RRID: AB_430834
anti-rabbit immunoglobulin/HRP	Agilent	cat# P0448; RRID: AB_2617138
Chemicals, peptides, and recombinant proteins
Cardiotoxin	Latoxan	cat# L8102
Tamoxifen	Sigma	cat# T5648
Corn oil	Sigma	cat# C8267
Collagenase type 2	Worthington	cat# LS004177
Dispase	Gibco	cat# 17105-041
1% chicken embryo extract	US Biological	cat# C3999
bFGF	Wako	cat# 062-06661
EdU	Invitrogen	cat# A10044
Critical commercial assays
ECL Prime Western Blotting Detection Reagents	Amersham	cat# RPN2232
Lipofectamine RNAiMAX	Thermo Scientific	cat# 13778-075
Click-iT EdU Imaging Kits	Invitrogen	cat# C10340
RNeasy Plus Micro Kit	Qiagen	cat# 74034
PrimeScript	Takara	cat# RR036A
TB Green Premix Ex Taq II	Takara	cat# RR820S
Agilent 2100 Bioanalyzer with RNA 6000 Pico Kit	Agilent	cat# 5067-1513
NEBNext Ultra II Directional RNA Library Prep Kit for Illumina	NEB	cat# E7760S
NEBNext rRNA Depletion Kit	NEB	cat# E6310L
NEBNext Multiplex Oligos for Illumina	NEB	cat# E7335S
Agilent 2100 Bioanalyzer with DNA 1000 kit	Agilent	cat# 5067-1504
KAPA Library Quantification Kits	KAPA Bio	cat# KK4824
NucleoSpin Tissue Kit	MACHEREY-NAGEL	cat# 740952
EpiXplore Methylated DNA Enrichment Kit	Clontech	cat# PT5034-2
DNA SMART ChIP-Seq Kit	Takara	cat# 634865
E-Gel 2% SizeSelect electrophoresis	Invitrogen	cat# G661012
Agilent 2100 Bioanalyzer with the High Sensitivity DNA kit	Agilent	cat# 5067-4626
Deposited data
Raw and analyzed data	This paper	GEO: GSE169193
Mouse reference genome GRCm38, PRI	GENCODE	https://www.gencodegenes.org/mouse/
Proximal enhancer-like (pELS) bed file	Consortium et al., 2020	https://screen.encodeproject.org
H3K4me1 peaks of MC3T3-E1	Oki et al., 2018	http://chip-atlas.org
Experimental models: Cell lines
Mouse: C2C12 cells	ATCC	CRL-1772
Experimental models: Organisms/strains
Mouse: Pax7CE/+: B6;129-Pax7tm2.1(cre/ERT2)Fan/J	The Jackson Laboratory	JAX: 012476
Mouse: Uhrf1fl/fl: B6Dnk; B6N-Uhrf1<tm1a(EUCOMM)Wtsi>/Ieg	The EUCOMM program	MGI:4434100
Oligonucleotides
non-targeting siRNA	Dharmacon	cat# D-001810-10-05
siRNA targeting Sox9	Dharmacon	cat# L-059108-01-0005
Primers for Uhrf1, see Table S2	This paper	NA
Primers for MyoD1, see Table S2	This paper	NA
Primers for Myog, see Table S2	This paper	NA
Primers for Sox9, see Table S2	This paper	NA
Software and algorithms
Fiji (v2.3.0)	Schneider et al., 2012	https://imagej.net/Fiji
TrimGalore (v0.6.6)	https://github.com/FelixKrueger/TrimGalore	http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/
HISAT2 (v2.2.1)	Kim et al., 2019	http://daehwankimlab.github.io/hisat2/
DESeq2 (v1.30.1)	Love et al., 2014	https://bioconductor.org/packages/release/bioc/html/DESeq2.html
Metascape	Zhou et al., 2019	http://metascape.org
STAR (v2.7.6a)	Dobin et al., 2013	https://github.com/alexdobin/STAR
featureCounts (v2.0.1)	Liao et al., 2014	http://subread.sourceforge.net
MACS2 (v2.2.6)	Zhang et al., 2008	https://github.com/macs3-project/MACS/tree/release_v2.2.6
ChIPpeakAnno (v3.24.2)	Zhu et al., 2010	https://bioconductor.org/packages/release/bioc/html/ChIPpeakAnno.html
Prism (v9.3.1)	GraphPad	https://www.graphpad.com/scientific-software/prism/
R (v4.0.5)	R Core Team 2020	https://www.r-project.org/
Integrative Genomics Viewer (v2.11.4)	Robinson et al., 2011	https://igv.org
FlowJo (v10.1r7)	BD Life Sciences	https://www.flowjo.com