Defective plasticity in dermatomyositis patients muscle stem cells is associated with sustained intrinsic inflammatory signaling and disruption of the histone H3.3 chromatin loading pathway

Abstract

Skeletal muscle regeneration is driven by muscle stem cells (MuSCs), which proliferate, differentiate, and fuse to reform myofibers and restore muscle function. This myogenesis process is driven both by intrinsic MuSC properties and extrinsic cues. While coordinated inflammatory signals are necessary for healthy regeneration, chronic inflammation participates in various pathologies affecting the skeletal muscle. In the idiopathic inflammatory myopathy dermatomyositis (DM), MuSCs exhibit impaired myogenesis in vitro, indicating that they may have acquired intrinsic defects, contributing to the disease and providing a mechanism for sustained patient muscle weakness despite efficient anti-inflammatory treatments. Here, we investigated the transcriptomic regulation of DM-derived MuSCs, with a focus on the H3.3 histone variant that regulates myogenesis progression. DM-derived MuSCs were unable to effectively execute the myogenic transcriptional program during in vitro differentiation. They exhibited an activated canonical tumor necrosis factor (TNF)-⍺ signaling. They also showed reduced expression of H3.3 and its chaperone genes, coupled with a decrease in H3.3 deposition across the entire genome, and particularly at myogenic regulatory factor loci. The loss of H3.3 combined with elevated TNF-⍺ signaling was associated with a failure of DM-derived MuSCs to achieve myogenesis, suggesting a mechanistic link between epigenetic dysregulation and defective muscle regeneration in humans.

Graphical Abstract

Graphical Abstract.

Introduction

Adult muscle stem cells (MuSCs) sustain skeletal muscle regeneration after myofiber damage. They implement adult myogenesis encompassing proliferation, differentiation into myocytes, and fusion to form new functional myofibers. Myogenesis is an intrinsic property of MuSCs that is tightly regulated by their environment [1]. Muscle regeneration is associated with an inflammatory response where a pro-inflammatory environment initially stimulates MuSC proliferation and represses early differentiation; while at later stages, an anti-inflammatory environment promotes differentiation and fusion of MuSCs [2]. Sustained inflammation is linked to muscle wasting and MuSC dysfunction in aging [3]. However, the impact of chronic inflammation on MuSC biology has not been thoroughly investigated.

Dermatomyositis (DM) is a rare disease belonging to the family of idiopathic inflammatory myopathies. It is characterized by high levels of type I interferon (IFN-I), which correlates with the severity of the disease [4, 5]. While anti-inflammatory treatments such as glucocorticoid administration improve patient’s health, many of them experience muscle weakness [6, 7]. This autoimmune disease is characterized by permanent myofiber damage, associated with a strong inflammatory environment due to the presence of a huge leukocyte infiltrate within the skeletal muscle. We have previously shown that MuSCs from DM patient’s muscle (DM-MuSC) proliferate and differentiate much less than healthy control (HC)-MuSCs [8]. Those intrinsic defects of myogenesis likely participate in the disease by impairing muscle regeneration. DM-MuSCs offers the opportunity to investigate the intrinsic effects of long-term inflammation on MuSCs.

Cell identity is maintained by several layers of epigenetic regulations. These mechanisms preserve cell specificity over time while allowing a degree of plasticity to respond to environmental cues. Controlled deposition and recycling of histone variants is such a mechanism. Studies in various mouse cell types have shown that the replication-independent deposition of the H3.3 histone variant is crucial for safeguarding cell identity [9–11] and is necessary for the proper regulation of myogenic regulatory factor (MRF) transcription and myogenesis [12]. Also, chromodomain-helicase-DNA-binding protein 2 (CHD2)-dependent deposition of H3.3 allows proper myogenic differentiation, while the repression of either CHD2 or H3.3 leads to the loss of MRF expression and to a defective myogenic differentiation [13, 14]. Finaly, the loss of H3.3 chaperone protein histone cell cycle regulator (HIRA) leads to the downregulated expression of skeletal muscle genes and to the upregulated expression of genes from other lineages [12]. Despite compelling evidence from mouse models highlighting the importance of H3.3 deposition in regulating myogenesis and maintaining cell identity, these mechanisms remain largely uncharacterized in human systems. DM offers a unique pathological context to explore the impact of chronic inflammation on H3.3 dynamics and its consequences for human MuSCs function. Notably, H3.3 deposition is known to be altered in inflammatory contexts. In senescent cells, both HIRA and H3.3 relocalize to promoters of active genes in line with transcriptional changes [15], and during viral infection or interferon stimulation, increased HIRA recruitment enhances H3.3 deposition at interferon-stimulated genes (ISGs) [16, 17]. These findings suggest that chronic inflammatory signals in DM may similarly reshape H3.3 dynamics, thereby contributing to defective myogenesis.

In this study, the transcriptomic profiles of human primary DM-MuSCs and HC-MuSCs were analyzed throughout the myogenesis process, which can be recapitulated in vitro. DM-MuSCs displayed markedly reduced transcriptomic plasticity from proliferation to differentiation. The role of IFN-I signaling in this defect was found to be limited. In contrast, several inflammatory pathways, particularly those associated with Tumor Necrosis Factor (TNF)-α, were found to be overexpressed in DM-MuSCs. At the chromatin level, a widespread reduction in H3.3 deposition was observed, and overexpression of H3.3 partially rescued the DM-MuSC phenotype. Additionally, treatment of HC-MuSCs with TNF-α recapitulated both the myogenesis defects and the global decrease in H3.3 deposition seen in DM-MuSCs.

Materials and methods

Ethics statement

All procedures of the study were in accordance with the ethical standards of the Helsinki Declaration. Written informed consent was obtained from all patients (or legal representatives) for the use of their biopsy for research purposes. The whole procedure was handled by the Hospices Civils de Lyon cell bank according to institutionally approved protocols (agreement CRB-HCL Codecoh, AC2013-1867). The laboratory agreement from French Ministry of research for the use of human cells is Codecoh, DC-2022-5055.

Patient cells selection

Human skeletal MuSCs derived from DM patients were selected from a previous study [8]. Standard protocol approvals, registrations, patient consents, and case selection methods are detailed in [8]. Rapidly, patients without immuno-suppressive and cytotoxic drugs administration before muscle biopsy were selected. In the previous study, histologic staining and immunohistochemistry of muscle sections were performed to evaluate the histologic DM severity score. In the current study, we chose four patients with high severity on this score, which together with other clinical data indicate severe muscle harm. A summary of relevant patients’ information and clinical data can be found in Supplementary Table S6. More detailed information can be found in [8]. MuSCs derived from the four DM patients were obtained from the Hospices Civils de Lyon cell bank (Cardiobiotec, CRB-HCL AC2013-1867). All experiments involving DM-MuSCs were done with those cells. MuSCs from HCs paired on sex and age were obtained from surgical waste of patients undergoing programmed knee or hip surgery.

Cell culture

MuSCs derived from HCs were directly extracted by mechanical and enzymatic digestion from fresh muscle biopsies [8]. After an expansion period, they were frozen and stored in liquid nitrogen until further use. After thawing, MuSCs were purified at the first passage of primary culture using anti-CD56 coated magnetic beads (Miltenyi Biotech, 130-050-401) to obtain pure myogenic cell cultures as described previously [8]. Cell purity was checked by flow cytometry (Miltenyi Biotech, MACSQuant Analyzer) using anti-CD56 antibodies (APC-conjugated anti-CD56 antibodies, BD Pharmingen, #555518) and isotypic control (BD Pharmingen, #555751).

MuSCs were cultured in complete medium (Skeletal Muscle Cell Growth Medium, #23060 Promocell, supplemented with Mix C-39365, Promocell), 10% of heat-inactivated Fetal Bovine Serum (FBS, Sigma–Aldrich), and 1% of Penicillin/streptomycin (Gibco, 14190169). Cells were passaged upon reaching 80% confluency using 0.05% trypsin–ethylenediaminetetraacetic acid (EDTA) (Gibco, 25300-054) digestion, and then seeded back onto new supports.

Proliferation assay

Analysis of proliferation of MuSCs was performed as follows: MuSCs were seeded at 3000 cells/cm² on labteks (Dutscher, 177437). The next day, medium was changed with fresh medium. Twenty-four hours later, 50% of the medium was replaced with fresh medium. Proliferation was evaluated at day 2, using 5-ethynyl-29-deoxyuridine (EdU) incorporated for 6 h (kit EdU Clickit Alexa Fluor 488, Invitrogen). At the end of the proliferation assay, cells were washed with phosphate-buffered saline (PBS) twice and fixed with formaldehyde 4%. Fixed cells were permeabilized cells in PBS containing 0.5% Triton X-100. For proliferation assay, EdU was revealed for 30 min according to the manufacturer instructions (kit EdU Clickit Alexa Fluor 488, Invitrogen). Cells were washed twice in PBS and blocking was performed with 1-h incubation in bovine serum albumin (BSA) 4% (diluted in PBS, Fisher scientific, BP9701). Incubation with MKI67 primary antibody (Abcam, ab15580) diluted in BSA 4% was performed overnight. Secondary antibodies were then incubated for 40 min at 37°C (see Supplementary Table S7 for antibodies used). After washing several times with PBS the samples, they were incubated with streptavidin dylight (Vector laboratories, SA-5549) for 20 min at 37°C. Cells were washed thoroughly with PBS, and cells were briefly incubated in Hoechst 33342 (Sigma–Aldrich, 14533) for nuclei visualization. Slides were mounted using Fluoromount G (Invitrogen, 15586276).

Differentiation Assay

Analysis of differentiation capacities of MuSCs was performed as follows: cells were seeded in labtek at 1500 cells/cm² in growth medium. The next day, medium was changed with differentiation medium with or without cytokines consisting of skeletal muscle cell basal medium (Promocell, C-23260) supplemented with 10 μg/ml insulin (Sigma–Aldrich, I2643), FBS 0.5%, and 1% of penicillin/streptomycin. Fifty per cent of the medium was replaced with fresh medium every day for 3 days. Proliferation was evaluated at day 3 by using 5-ethynyl-29-deoxyuridine (EdU) incorporated for 6 h (kit EdU Clickit Alexa Fluor 488; Invitrogen). At the end of the differentiation assay, cells were washed with PBS twice and fixed with formaldehyde 4%. Fixed cells were permeabilized cells in PBS containing 0.5% Triton X-100. For proliferation assay, EdU was revealed for 30 min according to the manufacturer instructions (kit EdU Clickit Alexa Fluor 488, Invitrogen). Cells were washed twice in PBS and blocking was performed with 1-h incubation in BSA 4% (diluted in PBS, Fisher scientific, BP9701). Incubation with anti-MYOG (BD Pharmingen, 556358) and anti-DES (Abcam, ab32362) primary antibodies diluted in BSA 4% was performed overnight. Secondary antibodies were then incubated for 40 min at 37°C (see Supplementary Table S7 for antibodies used). After washing several times with PBS the samples, they were incubated with streptavidin dylight (Vector laboratories, SA-5549) for 20 min at 37°C. Cells were washed thoroughly with PBS, and cells were briefly incubated in Hoechst 33342 (Sigma–Aldrich, 14533) for nuclei visualization. Slides were mounted using Fluoromount G (Invitrogen, 15586276).

High-density differentiation Assay

Analysis of differentiation capacities of MuSCs was performed as follows: cells were seeded in labtek at 10 000 cells/cm² in growth medium. The next day, medium was changed with differentiation medium consisting of skeletal muscle cell basal medium (Promocell, C-23260) supplemented with 10 μg/ml insulin (Sigma–Aldrich, I2643), FBS 0.5%, and 1% of penicillin/streptomycin. Fifty per cent of the medium was replaced with fresh medium every day for 3 days. At the end of the high-density differentiation assay, cells were washed with PBS twice and fixed with formaldehyde 4%. Fixed cells were permeabilized cells in PBS containing 0.5% Triton X-100. Cells were washed twice in PBS and blocking was performed with 1-h incubation in BSA 4% (diluted in PBS, Fisher scientific, BP9701). Incubation with anti-MYOG (BD Pharmingen, 556358) and anti-DES (Abcam, ab32362) primary antibodies diluted in BSA 4% was performed overnight. After washing several times with PBS the samples, they were incubated with streptavidin dylight (Vector laboratories, SA-5549) for 20 min at 37°C. Cells were washed thoroughly with PBS, and cells were briefly incubated in Hoechst 33342 (Sigma–Aldrich, 14533) for nuclei visualization. Slides were mounted using Fluoromount G (Invitrogen, 15586276).

Acute cytokine assay

In order to assess the effect of acute treatment with cytokines on proliferation, cells were grown as follows: MuSCs were passaged and seeded at 3000 cells/cm² in growth medium. The next day, medium was replaced with fresh one with cytokines. TNF-α was added at 5 ng/ml (Invivogen, rcyc-htnfa) while IFN-β at 1000 U/ml (Peprotech, 300–02BC). Fifty per cent of the medium was changed the next day with fresh growth medium with cytokines. At day 2, cells were either processed for proliferation assay as described above or for RNA-extraction followed by reverse transcription-quantitative PCR (RT-qPCR), or for chromatin immunoprecipitation (ChIP)-qPCR (see below).

In order to assess the effect of acute treatment with cytokines on differentiation, cells were grown as follows: MuSCs were passaged and seeded at 1500 cells/cm² in growth medium. The next day, medium was replaced with fresh one with cytokines. TNF-α was added at 5 ng/ml (Invivogen, rcyc-htnfa) while IFN-β at 1000 U/ml (Peprotech, 300-02BC). Fifty per cent of the medium was changed the next 2 days with fresh growth medium with cytokines. At day 3, cells were either processed for differentiation assay as described above or for RNA-extraction followed by RT-qPCR (see below).

Mid-term cytokine assay

In order to assess the effect of mid-term treatment with cytokines on proliferation, cells were grown as follows: MuSCs were passaged and seeded at 3000 cells/cm² in growth medium. The next day, medium was replaced with fresh one with cytokines. TNF-α was added at 5 ng/ml (Invivogen, rcyc-htnfa) while IFN-β at 1000 U/ml (Peprotech, 300-02BC). Fifty per cent of the medium was changed every day with fresh growth medium with cytokines. Cells were passaged upon reaching 80% confluency as detailed above. Cells were grown for 5 days before passaging them for either proliferation or differentiation assay as described in the acute cytokine assay protocol (see above).

H3.3 HA lentivirus constructs generation, transduction, and H3.3 overexpression in DM-MuSCs

SNAP-H3.3-HA and empty control vector plasmids generation, packaging into lentivirus and transduction was done as previously described [16–18]. Shortly, SNAP-H3.3-HA plasmid was obtained by polymerase chain reaction (PCR) using pBABE-H3.3-SNAP-HA3 as template (kind gift by Dr Lars Jansen) and cloned into puromycin-resistant pLVX-TetOne plasmid with EcoRI restriction enzyme. pLVX plasmid control empty vector or encoding H3.3-SNAP-HAx3 were co-transfected with PsPAX.2 and psMD2.G plasmids by the calcium phosphate method into HEK 293T cells to package lentiviral particles. After 48 h, supernatant containing replication-incompetent retroviruses was filtered and applied for 24 h on the target DM-MuSCs seeded at 3000 cells/cm² in growth medium containing polybrene 8 μg/ml (Sigma). Twenty-four hours later, stable transfectants were selected with puromycin (1 μg/ml, Invivogen) until cells reached 50% confluency, then passaged and processed as described above for the proliferation assay. One-hundred nanograms per milliliter of doxycycline was supplemented or not into the growth medium 24 h and 48 h after seeding the cells.

Knockdown of H3.3 in HC-MuSCs with siRNAs

H3.3 knockdown was performed with small interfering RNA (siRNA) as previously described [18]. HC-MuSCs were seeded either in at a density of either 3000 cells/cm² in growth medium for proliferation assay or 1500 cells/cm² in growth medium for differentiation assay onto either labtek for EdU- and immuno-stainings or onto petri-dishes for RT-qPCR. Twenty-four hours after seeding the cells, either control siLuc (EUROGENTEC, FR-CL000-005) at 40 nM or human siRNA targeting H3F3A and H3F3B (Key resources Supplementary Table S7 in Appendix), were added to HC-MuSCs at 20 nM each using Lipofectamine RNAiMax reagent (Invitrogen, 13778-075) and Opti-MEM medium (Gibco, 31-985-070). Twenty-four hours after transfection, medium was removed and changed to either growth medium for proliferation assay or differentiation medium for differentiation assay. Fifty per cent of the medium was changed every day. Forty-eight hours after transfection for proliferation or 72 h after differentiation, either RNA were harvested for reverse-transcription followed by quantitative PCR, or proliferation and differentiation were assessed as described above in respectively the proliferation assay or the differentiation assay.

Microscopy, imaging, and quantification

Images were acquired with the Axio Observer Z1 inverted wide-field epifluorescence microscope (Zeiss) and a CoolSnap HQ2 camera from Photometrics. Identical settings and contrast were applied for all images of the same experiment to allow data comparison. For each sample, 5–10 pictures were recorded at 10× magnification. Raw images were treated with Fiji software. Proliferation index was defined as the number of EdU-positive or MKI67-positive nuclei over the total number of nuclei. Differentiation index was defined as the number of Myogenin-positive nuclei over the total number of nuclei. For differentiation at high density, fusion index was calculated as the number of nuclei present in myotubes (>2 nuclei in the same desmin-positive cell) over the total number of nuclei and differentiation index was defined as the number of Myogenin-positive single nuclei added to nuclei present in myotubes over the total number of nuclei.

Reverse transcription

At the end of the proliferation or differentiation assay, cells were detached by trypsin–EDTA incubation, washed with PBS and pelleted at 400 g for 5 min. Pellets were snap-frozen and stored at −80°C before RNA-extraction. TRIzol reagent protocol (Invitrogen, 15596026) was used to isolate total RNAs, resuspended in ddH2O according to the manufacturer instructions. Contaminant DNA was removed by digestion with DNAse for 15 min at room temperature (Promega, M610A) and reaction was stopped with STOP buffer (Promega, M199A) incubation for 2 min at room temperature. We used 500 ng of RNA for RT. RNAs were annealed with Random Primers (Invitrogen, 58875) and RT was performed with the RevertAid H Minus Reverse Transcriptase (Thermo Scientific, FEREP0451) according to the manufacturer instructions. Complementary DNAs were stored at −20°C before qPCR analysis.

ChIP

At the end of the proliferation assay, cells were crosslinked directly in the culture dishes as described previously [18]. After the PBS washes, cell pellets were snap-frozen in liquid nitrogen and stored at −80°C before immunoprecipitation. Cells were de-frozen on ice and chromatin was prepared following the TruChIP protocol from Covaris, as described in [18]. We used the Covaris M220 Focused-ultrasonicator to shear through chromatin (7 min at 140 W, Duty off 10%, Burst cycles 200). After shearing, ChIP was performed. We used 20 μl of protein A magnetic dynabeads (Invitrogen, 10001D) for immunoprecipitation with 2 μg of anti-H3.3 antibody (Diagenode, C15210011). DNA was purified with phenol-chloroform (Sigma–Aldrich, 77617). Samples were centrifuged at 20 000 g at 4°C for 5 min. Liquid phase was harvested, then 1 μg/ml of glycogen, 200 mM of NaCl, and 2 volumes of cold ethanol 100% was added to each sample. DNA was precipitated for 2 h at −20°C, then centrifuged for 10 min at 4°C and 20000g. Supernatant was discarded. Pellets were washed with ethanol 70% and centrifuged for 5 min at 4°C and 20 000 g. DNA pellets were dried then resuspended in ddH20 and stored at −20°C before qPCR analysis.

Quantitative PCR

qPCRs were performed using the ONE SYBR qPCR Master Mix (Ozyme, OZYA008). Primers used for qPCR are described in the Key Resources Supplementary Table S7 in Appendix.

Western blot analysis

At the end of the proliferation or differentiation assay, total cellular extracts were obtained by directly lysing the cells in 2 X Laemmli sample buffer [125 mM Tris–Hcl, pH 6.8, 20% glycerol, 4% sodium dodecyl sulfate (SDS), bromophenol blue] containing 100 mM dithiothreitol (DTT). Lysate was subjected to 15% acrylamide gel electrophoresis, and transfer was then run overnight on nitrocellulose membrane (Thermo Scientific, 88018). Blocking was done by incubating membranes 1 h at room temperature with 5% BSA (diluted in PBS). Antibodies were incubated overnight at 4°C or 1 h at room temperature. Anti‐rabbit antibody coupled to HRP was then incubated for 2 h at room temperature, and the membrane was revealed with either ECL (Thermo Scientific, 32109) or ECL femto (Thermo Scientific, 34096).

RNA-seq

At the end of the proliferation or differentiation assay, cells were detached by trypsin–EDTA incubation, washed with PBS, and pelleted at 400 g for 5 min. Pellets were snap-frozen and stored at −80°C before RNA-extraction. QIAGEN Rneasy mini kit was used to extract and purify RNA according to manufacturer instructions (Qiagen, 74104). Purified RNAs were stored at −80° before library preparation and sequencing. NEBNext^® Ultra™ II Directional RNA Library kit (New England Biolabs) was used to prepare an RNA-seq library. Paired-end sequencing was performed on an Element AVITI™ system (HELIXIO, Clermont-Ferrand, France).

RNA-seq analysis

RNA-seq paired-reads were aligned and quantified using salmon (version 1.9.0) on gencode v41 for index generation. Resulting counts were processed and analyzed on R (version 4.4.1) using Deseq2 package (version 1.44.0). DEG were selected with a P-value under .05 and a log₂-fold change either above 1 or under −1. Heatmap and lineplots were produced using counts normalized by variance stabilizing transformation. Gene set enrichment analysis was performed using GSEA (version 4.3.2) and using log₂-fold change values as ranking. Over-representation analysis was performed with enrichplot (version 1.24.0), org.Hs.eg.db (3.19.1), and clusterProfiler (4.12.0) packages. Transcription factor (TF) motif enrichment analysis was performed on DEG using the findMotifs.pl function of HOMER (version 3.12).

CUT&RUN

At the end of the proliferation or differentiation assay, 250 000 cells were detached by trypsin–EDTA incubation, washed with growth medium and pelleted at 600 g for 3 min. Samples were then processed with active motif CUT&RUN kit (Active motif, 53180) following manufacturer instructions with some modifications inspired from [19]. Nuclei were first isolated by resuspending pellets in nuclei isolation buffer and incubating them for 10 min on ice. Nuclei were then washed twice in wash buffer and bound to activated Concavalin A beads for 10 min at room temperature. Following that, nuclei were incubated with antibodies overnight at 4°C with 1 μg of anti-H3.3 antibody (Active motif, 91191). After antibody incubation, unbound antibodies were washed twice with permeabilization buffer and pAG-MNase was added. Samples were incubated for 1 h at 4°C. Cells were washed twice with permeabilization buffer then cooled to 0°C and incubated with ice-cold 100 mM CaCl2 for 2 h at 4°C. MNase digestion was terminated with the addition of STOP buffer. RNA and protein digestion was performed by incubating samples overnight at 37°C with 0.1% of SDS and 0.3 mg/ml of proteinase K. DNA was purified by phenol-chloroform as detailed in the ChIP protocol. DNA was stored at −20°C before library preparation. The libraries were paired-end sequenced on a DNBSEQ Technology platform (BGI Genomics Co., Shenzhen, China).

CUT&RUN analysis

After removing adaptor sequences and low-quality reads by cutadapt (version 4.6), CUT&RUN paired-end reads were aligned to the human genome (GRCh38) using Bowtie 2 (version 2.3.5.1 by Ben Langmead) with a default setting. Samtools (version 1.10) was used to generate bam files and quality controls were done using deepTools. Normalization factors were calculated using CUT&RUN greenlist following the protocol of [20]. We use the deepTools “bamCoverage” with option --scaleFactor to generate scaled bigwig tracks. Tracks were visualized with IGV (version 2.12.3). Mean bigwig files were generated using WigToBigWig (version 2.9). Peak calling was done using SEACR [21] with the following parameters: normalization (non), peak calling (stringent), threshold (0.01). Peaks with <300 bp of distance were merged using bedtools. Profile plots were generated using deepTools functions computeMatrix and plotProfile. We downloaded from the ENCODE portal [22] (https://www.encodeproject.org/) the bigwig files with the following identifiers: ENCFF635GGC (H2AZ), ENCFF446DQZ (H3K27ac), ENCFF232NCT (H3K4me1), ENCFF885ESG (H3K4me3); and the bed files with the following identifiers: ENCFF446DQZ (H3K27ac), ENCFF232NCT (H3K4me1). R (version 4.4.1) with package DiffBind (version 3.14.0) were used to perform differential binding analysis. Peak annotation was performed with CHIPseeker package (version 1.40.0) with a TSS range of a −500, 500 bp window. Over-representation analysis was performed as described above for the RNA-seq.

Statistical analyses and figures

Histograms and statistical analyses were performed using R. Each replicate corresponds to MuSC derived from a different HC or DM patient. To perform Student’s t-test or Anova, we verified normal distribution of samples using Shapiro test and variance equality with Fisher test. Wilcoxon or Mann–Whitney u-test was applied in absence of normality for the sample distribution. P-values are depicted on graphs as follows: *<.05; **<.01; ***<.001; ****<.0001. Biorender.com was used to generate figures, schemes, and models.

Antibodies

All the primary and secondary antibodies used in this study, together with the species, the references and the dilutions for immunofluorescence and western blotting, are summarized in the Key Resources Supplementary Table S7 in Appendix.

Results

DM-MuSCs exhibit widespread suppression of the myogenic program and an enhanced inflammatory transcriptomic signature

We previously showed that DM-MuSCs present intrinsic myogenesis defects [8]. As a follow up, an RNA-seq analysis was conducted on those cells to compare with that of HC-MuSCs (Supplementary Fig. S1A and B). Patient cells were age-matched, and no significative differences was observed in the percentage of CD56-positive cells between HC-MuSCs and DM-MuSCs, indicating no loss of myogenic identity in DM-MuSCs (Supplementary Fig. S1C and D). Proliferation refers to cells cultured in growth medium (Prolif), early differentiation occurs 1 day after transitioning to differentiation medium (DiffD1), and late differentiation occurs 3 days after the shift to differentiation medium (DiffD3). Differential gene expression analysis of DM- versus HC-MuSCs was performed in each culture condition. GSEA of the differentially expressed genes (DEGs) show that in all conditions, DM-MuSCs overexpressed genes involved in inflammatory processes including TNF-⍺, Interleukin (IL)-6, or IFN signaling (Fig. 1A-left and Supplementary Table S1). They also exhibited a significant downregulation of genes involved in myogenesis, consistently identified as the most significantly downregulated term in each GSEA analysis (Fig. 1A-right and Supplementary Table S1). DM-MuSCs in Prolif and DiffD1 conditions also showed a downregulation of the genes associated with proliferation and cell-cycle progression such as G2M checkpoints and E2F targets (Fig. 1A-right and Supplementary Table S1). This is coherent with DM-MuSCs having a lower proliferation rate than HC-MuSCs [8]. In contrast, they showed a higher expression of those genes at DiffD3 (Fig. 1A-left and Supplementary Table S1), suggesting a delay in cell cycle exit upon differentiation. DEGs between DM and HC-MuSCs showed little overlap between the three conditions along the myogenesis process. Twenty-five genes (out of 749) showed common upregulation, including pro-inflammatory genes (HLA-DRB1, IFI44) and pro-fibrotic genes (FN1) (Fig. 1B). Expression of only 21 genes was commonly decreased in all states including genes involved in development (HOTAIR, HOX) (Fig. 1B). The myogenesis process is known to have a drastic impact on both transcriptome and epigenome [23–25]. Principal component analysis revealed that the primary transcriptomic variations across samples was driven by the differentiation process rather than by the origin from DM or HC muscle (Supplementary Fig. S2A). To go further, we looked at the kinetics of expression of DEGs between DM- and HC-MuSCs along the differentiation process. Genes which expression was upregulated in DM-MuSCs at Prolif or DiffD3 stages remained highly expressed as compared with HC-MuSCs during the whole myogenic process (Fig. 1C-left and Supplementary Fig. S2B). Conversely, genes that expression was downregulated in DM-MuSCs remained weakly expressed as compared with HC-MuSCs during all the myogenic process (Fig. 1C-right and Supplementary Fig. S2B). Overall, although the DM-MuSC transcriptome followed the myogenic program similarly to that of HC-MuSCs, there was a significant global expression shift between DM and HC-MuSCs, suggesting that DM-MuSCs have a slower dynamic and a reduced capacity to timely remodel their transcriptome during myogenesis.

Figure 1.

Transcriptomic comparison of HC- and DM-MuSCs. (A) Molecular Signatures Database Hallmark (MSigDB Hallmarks) GSEA results on the differential expression analysis between DM- and HC-MuSCs in the three conditions, ranked by enrichment scores. Left panels show Hallmark terms enriched in DM-MuSCs, right panels show Hallmark terms under-represented in DM-MuSCs. (B) Overlap of the DEGs in Prolif, DiffD1, and DiffD3 conditions between DM- and HC-MuSCs. The lists of the common upregulated (left panel, in red) and downregulated (right panel, in blue) DEGs in DM-MuSCs are given. (C) Mean expression in DM- and HC-MuSCs at the three time points (Prolif, Diff1, Diff3) of the DEGs that were identified by comparing DM- and HC-MuSC gene expression at Prolif [green circles in panel (B)] and at DiffD3 [red circles in (B)] with upregulated DEGs presented in the left panel (red) and downregulated DEGs presented in the right panel (blue)s.

DM-MuSCs show an overall reduced transcriptomic plasticity

The impact of chronic inflammatory state on the transcriptomic dynamics of DM-MuSCs in proliferation and differentiation was then evaluated. As standard control, DEG analyses in HC-MuSCs along the three steps of myogenesis was first performed (Supplementary Fig. S3A–D and Supplementary Table S2). Genes associated with chemotaxis and hypoxia were over-represented in Prolif as compared with DiffD1 (Supplementary Fig. S3A). As expected, genes related to muscle development and differentiation were more enriched at DiffD1 as compared with Prolif (Supplementary Fig. S3B). In the subsequent step of myogenesis, i.e. DiffD1 versus DiffD3, terms associated with DNA replication and mitotic division were over-represented in DiffD1 as compared with DiffD3 (Supplementary Fig. S3C). Inversely, and as expected, an enrichment was observed for genes associated with muscle function, development and differentiation, at DiffD3 as compared with DiffD1 (Supplementary Fig. S3D). These results are coherent with what is expected for the in vitro myogenesis process. We next compared in HC- and DM-MuSCs the average expression of genes in the four DEG sets described above (Supplementary Fig. S3E–H and Supplementary Table S2). Genes that were more expressed in HC-MuSCs in Prolif versus DiffD1 showed a decreased expression along the differentiation process in both HC- and DM-MuSCs, though they were more expressed in the latter. Those genes were notably related to chemotaxis (for instance CXCL2) (Supplementary Fig. S3E). On the other hand, DEGs upregulated in DiffD1 versus Prolif showed a sustained increased expression during the entire differentiation process and were related to muscle cell differentiation and development like MYMX (Supplementary Fig. S3F). The expression of DEGs upregulated in DiffD1 versus DiffD3 in HC-MuSCs decreased exponentially along differentiation, as exemplified by genes involved in cell-cycle progression such as CCNB1 (Supplementary Fig. S3G). Importantly, DM-MuSCs fail to properly repress this DEG set (Supplementary Fig. S3G). This suggests a reduced capacity of DM-MuSCs to exit the cell-cycle during differentiation. Finally, DEG upregulated between DiffD3 versus DiffD1 (Supplementary Fig. S3H) showed an enrichment in genes also related to muscle differentiation such as MYH3 (Supplementary Fig. S3H). Interestingly, DEG sets upregulated in DiffD1 versus Prolif and DiffD3 versus DiffD1 maintained a lower expression level in DM-MuSCs versus HC-MuSCs (Supplementary Fig. S3F and H), suggesting a delay for DM-MuSCs in entering differentiation. Overall, these results mark a reduced transcriptomic plasticity of DM- as compared with HC-MuSCs throughout differentiation, with a delay in exiting the cycle cycle and entering into terminal myogenic differentiation.

DM-MuSCs show a reduced transcriptomic plasticity on genes related to myogenesis

A specific focus was then made on genes related to myogenesis. As expected, the expression of most of the genes associated with myogenesis and muscle functions, including myosin heavy chains coding genes, RYR1, MYMX, and MYMK, showed a strong increase in the DiffD3 condition in HC-MuSCs (Fig. 2A and B). Although their expression kinetics followed a similar increase in DM-MuSCs, their initial expression level was lower, and they did not reach those observed in HC-MuSCs (Fig. 2A and B). The initial low expression was confirmed by RT-qPCR of five myogenic genes in proliferating cells and showed that all but MYOD1 had a lower expression in DM- as compared with HC-MuSCs (Supplementary Fig. S4A). The deficit in myogenic differentiation of DM-MuSCs was confirmed by the lower expression of MYOG mRNAs in DM- versus HC MuSCs (Fig. 2C) and a lower number of DM-MuSCs expressing MYOG protein as compared with HC-MuSCs (Fig. 2D) in the DiffD3 condition.

Figure 2.

DM-MuSCs show reduced plasticity in response to differentiation stimulus. (A) Heatmap of the expression of MRFs and their targets in DM- and HC-MuSCs in Prolif, DiffD1 and DiffD3 conditions. (B) Mean expression of MRFs and their targets in the three conditions by DM- (red) and HC- (green) MuSCs. (C) Expression of MYOG from the RNA-seq, blue crossbars indicate median for each cell type in each condition. P-value significance between HC- and DM-MuSC is indicated on the graph. (D) Immunofluorescence for MYOG (red) in HC- and DM-MuSCs, from four different patients each. Cells are labeled for desmin (green) and nuclei with hoechst (blue) and quantification of the percentage of MYOG+ nuclei in each sample. P-value significance between HC- and DM-MuSC is indicated on the graph. (E) Heatmap of the expression of genes coding for cell-cycle regulators in DM- and HC-MuSCs in Prolif, DiffD1, and DiffD3 conditions. (F) Mean expression per condition of all genes coding for cell-cycle regulators between DM- (red) and HC- (green) MuSCs. (G) Expression of MKI67 from the RNA-seq data. Blue crossbars indicate median for each cell type in each condition. P-value significance between HC- and DM-MuSCs is indicated on the graph. Immunofluorescence for MKI67 (red) in HC- and DM-MuSCs (nuclei blue) and quantification as percentage of positive cells in Prolif (H) and Diff3 (I) conditions.

Further investigation of DEG by RT-qPCR confirmed the decreased expression of MYMX in Prolif and DiffD3 DM-MuSC, while proliferation-associated genes CCNB1 and MKI67 were significantly downregulated in Prolif DM-MuSC as compared with HC-MuSCs, but not in DiffD3 (Supplementary Fig. S4B). This blunted plasticity of DM-MuSCs was confirmed in the enrichment maps generated from the over-represented terms in HC and DM-MuSCs in the DiffD3 condition. As expected, a strong enrichment of terms associated with muscle development and contraction was observed in HC-MuSCs (Supplementary Fig. S4C), whereas DM-MuSCs remain abnormally enriched in terms associated with proliferation and cell cycle (Supplementary Fig. S4D). Hence, cell cycle regulators, such as cyclins and cyclin-dependent kinases, were more strongly expressed in HC-MuSCs as compared with DM-MuSCs during proliferation (Fig. 2E Prolif, and Supplementary S4D). They sustained a quicker repression in HC-MuSCs as compared with DM-MuSCs as the differentiation process progressed to DiffD1 then DiffD3 (Fig. 2F and Supplementary Table S3). For instance, the proliferation marker MKI67 showed a marked tendency toward higher expression in HC-MuSCs as compared with DM-MuSCs in Prolif, a pattern that appeared reversed in the DiffD3 condition (Fig. 2G). This was confirmed at the protein level, MKI67+ cells were more abundant in HC- as compared with DM-proliferating MuSCs (Fig. 2H), while conversely, DM-MuSCs retained a higher number of MKI67+ cells than HC-MuSCs in the DiffD3 condition (Fig. 2I). Similar results were obtained using EdU (Supplementary Fig. S4E and F). Taken together, these results indicate that DM-MuSCs show impaired proliferation and further fail to both suppress proliferation-associated genes during differentiation and upregulate MRFs. They thus exhibit an intrinsic impairment in rewiring their myogenic program.

DM-MuSCs show an intrinsic activation of pro-inflammatory signaling pathways

Pro-inflammatory stimulation alters MuSC differentiation [26]. The above transcriptomic data showed that DM-MuSCs overexpressed genes involved in several pro-inflammatory pathways (Fig. 1A). The expression of 84 genes including receptors, ligands and targets of IFN type I and II, IL-1, IL-6, and TNF-⍺ pathways was investigated. Overall, these genes were highly expressed in DM- as compared with HC-MuSCs at all stages of the myogenesis process, with their expression decreasing along the course of myogenesis (Fig. 3A and B, and Supplementary Fig. S5A). Notably, DM-MuSCs expressed higher levels of chemokines and major histocompatibility complex coding genes than HC-MuSCs (Supplementary Fig. S5B). This tendency was confirmed on a subset of genes by RT-qPCR (Supplementary Fig. S5C). Strikingly, the expression levels of the two ISGs MX1 and ISG15, remained comparable between DM-MuSCs and HC-MuSCs in Prolif (Fig. 3C and Supplementary Fig. S5C), even though they are known to be over-expressed in DM skeletal muscle [27–29]. DM is characterized by a strong IFN-I signature, present both systemically in leukocytes and locally in skeletal muscle, with a correlation to disease severity [5, 28, 30, 31]. However, when the expression of 48 ISGs [17] was analyzed, only a few of them showed a significant overexpression in DM-MuSCs (labeled in red in Fig. 3D and Supplementary Table S4). These results suggest that overall pro-inflammatory pathways are maintained active in DM-MuSCs. The strong IFN-I signature described in DM is not observed in isolated MuSCs.

Figure 3.

DM-MuSCs show an intrinsic activation of pro-inflammatory pathways. (A) Heatmap of the expression of all the genes involved in type I, II and III IFN signaling (blue), IL1 signaling (red), IL6 signaling (purple), and TNF-α (orange) family signaling between DM- and HC-MuSCs in Prolif, DiffD1, and DiffD3 conditions. (B) Mean expression of all the genes involved in type I, II, and III IFN signaling, IL1 signaling, IL6 signaling, and TNF-⍺ family signaling between DM- (red) and HC-MuSCs (green) in Prolif, DiffD1, and DiffD3 conditions. (C) Expression of MX1 and ISG15 from the RNA-seq data. Blue crossbars indicate median for each cell type in each condition. P-value significance between HC- and DM-MuSC is indicated on the graph. (D) Heatmap of the expression of 48 ISGs from [17] between DM- and HC-MuSCs in Prolif, DiffD1, and DiffD3 conditions. Gene names are colored in red and blue if they are significantly up- and downregulated between DM- and HC-MuSCs, respectively.

To further investigate IFN signaling, MuSCs were treated with IFN-β to analyze if they could reactivate a memory IFN-I response [32]. IFN-β treatment had no effect on proliferation and commitment into differentiation of both DM- and HC-MuSCs (Supplementary Fig. S6A). Accordingly, IFN-β treatment of HC-MuSCs did not impact on their capacity to express myogenic genes (but induced the expression of its targets MX1 and OAS1) both in Prolif and DiffD3 conditions (Supplementary Fig. S6B and C). Previous studies showed that IFN-β treatment represses high density-cultured HC-MuSC fusion [33, 34]. We successfully replicated these findings (Supplementary Fig. S6D). Taken together, these data indicate that IFN signaling alone has a limited impact on the early stages of myogenesis in normal MuSCs in culture, suggesting that other signaling pathways may be responsible for the impaired plasticity observed in MuSCs from DM patients.

Activation of pro-inflammatory pathways induces loss of myogenesis

As detailed above, DM-MuSCs exhibited a much higher expression of genes associated with various inflammatory pathways as compared with HC-MuSCs (Fig. 3A and Supplementary Fig. S5A). Common TF signature on the promoters of the DEGs revealed that in all three conditions (Prolif, DiffD1, and DiffD3) genes that were underexpressed in DM-MuSCs were targets of myogenic TFs (e.g. MEF2, MYF5, MYOD, MYOG) (Fig. 4A and Supplementary Table S5). In contrast, overexpressed genes were enriched in TF motifs of NFkB-p65-Rel and the AP-1 family (e.g. JUNB, BATF, FOS, FOSL2, FRA1, FRA2), which are canonical targets of the TNF-⍺ signaling pathway. Additionally, an increase of the cell cycle homology region (CHR) motifs was specifically found in DM-MuSCs in the DiffD3 condition (Fig. 4A and Supplementary Table S5). CHR motifs are linked to the transcriptional regulation of cell cycle genes, consistent with the elevated expression of proliferation-associated genes in DM-MuSCs (Figs 1A and 2E, and Supplementary Fig. S3C and G). These findings suggest that an intrinsic activation of the canonical TNF-⍺ signaling pathway is at work in DM-MuSCs during the differentiation process. We therefore analyzed molecular and cellular parameters of HC-MuSCs treated with TNF-α in comparison with IFN-β. Similarly to IFN-β, no effect of TNF-α, in combination or not with IFN-β was observed on HC-MuSC proliferation (Fig. 4B). Interestingly, unlike IFN-β, TNF-α strongly repressed MuSC differentiation, as monitored by MYOG+ cells, and without synergizing with IFN-β (Fig. 4C). Accordingly, TNF-α treatment strongly upregulated the expression of its targets (TNFAIP3 and IL6) while also increasing that of IFN (MX1 and OAS1) in proliferating (Fig. 4D) and differentiating (Fig. 4E) conditions. While the stimulation of ISGs by TNF-α may not be entirely unexpected, it is not a commonly reported finding in the literature. We thus re-analyzed data from a previous study investigating the impact of TNF-α treatment on skeletal myogenesis using murine C2C12 myoblasts [35], and confirmed the TNF-α-dependent increase in the expression of representative ISGs (Supplementary Fig. S7). Moreover, TNF-α, unlike IFN-I, significantly repressed the expression of MRFs PAX7 and MYOD1 in proliferating HC-MuSCs (Fig. 4D) and repressed the expression of MYOG and MYH3 in differentiating cells (Fig. 4E). To minimize potential biases from acute stimulation, HC-MuSCs were treated for 8/9 days with TNF-α or IFN-β (Supplementary Fig. S8A). The results were similar to those using the acute treatment, showing no impact on proliferation (Supplementary Fig. S8B) and a decrease in differentiation (Supplementary Fig. S8C). The transcriptomic expression of MRFs, IFN, and TNF-⍺ targets mirrored that observed in the short treatment experiment (Supplementary Fig. S8D and E). Altogether, these results show that IFN-I has only a minimal direct impact on myogenesis while TNF-α stimulation significantly impairs the differentiation of HC-MuSCs. TNF-α also induces the expression of ISGs, thereby mimicking the dysfunctions and altered gene expression observed in DM-MuSCs. This suggests that DM-MuSCs may have a cell-intrinsic activation of TNF-⍺ signaling, which both sustains inflammation and disrupts differentiation.

Figure 4.

NFkB stimulation impairs myogenesis. (A) TF motif analysis of the promoters of DEGs between DM- and HC-MuSCs in Prolif (left panel), DiffD1 (middle panel), and DiffD3 (right panel). In y-axis, the name of the enriched TF motif signatures, with in parenthesis the corresponding DNA-binding domain family. Motifs enriched in promoters of DEGs up- and downregulated in DM-MuSCs appear in red and green, respectively. (B) Staining for EdU (green) in HC-MuSCs from three different patients in Prolif condition with or without addition of IFN-β and/or TNF-α (nuclei blue), and quantification as percentage of positive cells. (C) Immunofluorescence for MYOG (green) and DES (red) in HC-MuSCs from three different patients in DiffD3 condition with or without addition of IFN-β and/or TNF-α (nuclei blue), quantification as percentage of positive cells. (D) Expression of myogenic genes, IFN- and TNF-stimulated genes of HC-MuSCs from four different patients treated with or without IFN-β and/or TNF-α in the Prolif condition (RT-qPCR). P-value significance between HC- and DM-MuSC is indicated on the graph. (E) Expression of myogenic genes, IFN- and TNF-stimulated genes of HC-MuSCs from four different patients treated with or without IFN-β or TNF-α in the DiffD3 condition. P-value significance between HC- and DM-MuSC is indicated on the graph.

DM-MuSCs show global histone H3.3 reduction and loss of transcriptomic plasticity

Our data show that DM-MuSCs exhibit severe myogenesis defects and maintain pro-inflammatory pathways active, even when removed from their inflammatory tissue environment and long-term cultured. We thus hypothesized that DM-MuSCs might sustain epigenetic dysregulations, acting as a long-term memory that contributes to their intrinsic myogenesis defects. Strikingly, H3.3 coding genes (H3F3A and H3F3B) and several of its chaperones were among the significantly repressed DEGs in DM-MuSCs in the Prolif condition (Fig. 5A and Supplementary S9A) as well as for H3F3A in the DiffD3 condition (Fig. 5B). Dynamics of the histone variant H3.3 is impacted by inflammatory stresses such as senescence [15], viral infection, and IFN stimulation [16, 17]. A decrease of H3.3 was confirmed at the protein level using western blot in the Prolif condition (Supplementary Fig. S9B). To determine whether loss of H3.3 directly contributes to the pathology or represents a secondary consequence, we performed rescue experiments in DM-MuSCs using a doxycycline-inducible H3.3-HA overexpression system (Supplementary Fig. S9C). Despite increased H3.3 expression in more than two-thirds of transfected cells, no significant increase in proliferation was observed at the population level (Supplementary Fig. S9D). However, when stratifying cells based on their H3.3-HA expression, H3.3-overexpressing cells showed a higher proportion of EdU+ cells as compared with H3.3-HA-negative cells (Fig. 5C). Conversely, siRNA-mediated knockdown of H3.3 in HC-MuSCs had no impact on proliferation and differentiation (Supplementary Fig. S9E and F). At the transcriptomic level, the expression of H3F3A and H3F3B was decreased, along with alterations in myogenic regulatory genes. Notably, PAX7 expression was strongly reduced under proliferative conditions, whereas MYOD1 and MYOG expression show a modest increase in differentiation condition (Supplementary Fig. S9G and H). Taken together, these results suggest that reduced H3.3 levels alone may not be enough to explain the pathological phenotype observed in DM-MuSCs.

Figure 5.

DM-MuSCs show a global loss of H3.3, which impairs myogenesis. Heatmaps of the expression of H3.3 coding genes, H3.3 chaperones and PML and SP100 in HC- and DM-MuSCs in the Prolif condition (A) and DiffD3 condition (B). Genes significantly down- or upregulated between HC- and DM-MuSCs are labeled with red and green asterisks, respectively (the number of asterisks refers to the P-value). (C) Quantification of the percentage of EdU positive cells between DM-MuSC HA negative and HA positive, after transfection with Dox inducible H3.3-HA vector. Each cell line was transfected with a doxycycline-inducible H3.3-HA overexpressing vector. After puromycin selection for 48 h, cells were seeded in proliferation medium with or without 100 ng/ml of doxycycline. Each dot is one DM patient derived MuSC (n = 4). (D) Volcano plots of the differential binding analysis of H3.3 between DM- and HC-MuSCs in the Prolif condition (upper panel) or in the DiffD3 condition (lower panel). Peaks with a significant differential binding of H3.3 appear in blue. (E) ChIP-qPCR for H3.3 of HC-MuSCs from three different patients treated with or without TNF-α in the Prolif condition. Results in log₂-FC of the input percentage of H3.3 in the transcription end sites (TES) in TNF-α stimulated HC-MuSCs normalized to untreated HC-MuSCs. (F) Expression of H3F3A and H3F3B of HC-MuSCs from three different patients cultured with or without TNF-α in the Prolif condition (RT-qPCR).

DM-MuSCs also showed a decrease in the expression of essential components of the H3.3 chaperone HIRA complex UBN1 in the Prolif condition (Fig. 5A) and ASF1A in both Prolif and DiffD3 conditions (Fig. 5A and B). Hence, the genome-wide deposition of H3.3 in HC- and DM-MuSCs was assessed by CUT&RUN in Prolif and DiffD3 conditions. Differential binding analysis of H3.3 between HC- and DM-MuSCs revealed a general reduction of H3.3 binding at most consensus peaks in DM- as compared with HC-MuSCs in both Prolif and DiffD3 conditions (Fig. 5D). These results fit with a global loss of H3.3 across the chromatin, as expected from the reduced expression of histone H3.3 and H3.3 chaperones coding genes in DM-MuSCs. Interestingly, treatment of proliferating HC-MuSCs with TNF-α tended to reduce H3.3 deposition at the transcription end sites (TES) of both myogenic regulators and inflammatory genes (Fig. 5E), yet it did not repress the expression of the H3.3-encoding genes, H3F3A and H3F3B (Fig. 5F). These findings suggest that DM-MuSCs have a deficiency in H3.3 deposition on chromatin, resulting from both reduced expression of H3.3 and its chaperones, as well as pro-inflammatory cytokine-driven reduction of H3.3 deposition.

We next investigated whether the loss of H3.3 was uniform throughout the chromatin. We looked at the deposition of H3.3 on MRFs, ISGs and inflammation-related genes in HC- as compared with DM-MuSCs. In the Prolif condition, all gene sets had decreased H3.3 binding, while in the DiffD3 condition, H3.3 signal was highly decreased on MRFs, unchanged on ISGs, and moderately decreased on inflammation-related genes in DM-MuSCs as compared with HC-MuSCs (Fig. 6A). Analysis of the CUT&RUN results for the distribution of H3.3 in DM-MuSCs and HC-MuSCs on the DEGs (as described in Fig. 1 and Supplementary Fig. S2) showed that both up- and downregulated genes in DM-MuSCs incorporated significantly less H3.3 than HC-MuSCs in the Prolif condition (Supplementary Fig. S10A). At DiffD3, the differential incorporation of H3.3 in downregulated genes in DM-MuSCs is even more exacerbated (Supplementary Fig. S10B). To assess the functional impact of altered H3.3 deposition, we integrated CUT&RUN data with RNA-seq under proliferative conditions. Overall, changes in gene expression and H3.3 chromatin binding between DM- and HC-MuSCs showed only a weak correlation (Fig. 6B). However, genes with both reduced H3.3 enrichment and decreased expression were significantly enriched for myogenesis-related signatures (Fig. 6C). These results suggest that although DM-MuSCs exhibit a global decrease in H3.3 incorporation across the genome, this deficiency is particularly pronounced at MRFs, and becomes even more marked during differentiation.

Figure 6.

DM-MuSCs loss of H3.3 is focalized on myogenic regulators and enhancers. (A) Intensity plot of the mean H3.3 signal on myogenic regulatory genes (top panel), IFN type I stimulated genes (middle panel), or inflammation-related genes (bottom panel) in DM- (light red) and HC- (light green) MuSCs in Prolif condition, and in DM- (dark red) and HC- (dark green) MuSCs in DiffD3 condition. Differential mean H3.3 signal within the −5/+5 kb of the TSS and TES of the selected genes was evaluated between HC- and DM-MuSCs, with each gene of the selected panel counted as a replicate. P-value significance between HC- and DM-MuSC is indicated on the graph. (B) Integration of the differential binding of H3.3 and differential expression analyses between DM- and HC-MuSCs in the Prolif condition. In x-axis log₂-FC of the differential expression of genes. In y-axis log₂-FC of the differential binding of H3.3 in peaks located in genes. (C) MSigDB Hallmarks over-representation analysis of genes bound with H3.3 and showing a decreased H3.3 deposition in DM-MuSCs. Left, middle, and right panel show genes that with increased, unchanged or decreased expression in DM-MuSCs as compared with HC-MuSCs in the Prolif condition. (D) Intensity plot on differentially bounded peaks with decreased (left panel) or increased H3.3 binding (right panel) in DM-MuSCs. Mean H3.3 signal in DM (red) and HC- (green) MuSCs in Prolif condition. Mean signal of H2AZ in yellow, H3K27ac in purple, H3K4me1 in dark blue, and H3K4me3 in light blue in HC-MuSCs (datasets from ENCODE [22], https://www.encodeproject.org/).

It is well established that post-translational modifications of histones can influence their effect on the chromatin environment, with H3K4me3 primarily found at active promoters, H3K4me1 at enhancers, and H3K27ac at both active promoters and enhancers [36]. To further explore the correlation between H3.3 loss in DM-MuSCs and the dysregulation of specific genomic regions, we compared existing ENCODE ChIP-seq datasets for histone modifications in Prolif HC-MuSCs [22] (https://www.encodeproject.org/) with our results on the genome-wide deposition of H3.3 in DM-MuSCs. Interestingly, H3.3 peaks with decrease signal in DM-MuSCs were more enriched in active chromatin marks H3K27ac, H3K4me1, H3K4me3, and histone variant H2AZ than the H3.3 peaks with increase signal in DM-MuSCs (Fig. 6D). More specifically, ~16.4% of the peaks with decreased H3.3 level in DM-MuSCs were shared with peaks of H3K4me1 and H3K27ac (Supplementary Fig. S10C). This would predict that H3.3 is predominantly lost at active enhancers in DM-MuSCs, potentially explaining their reduced capacity to remodel the transcriptome in response to differentiation cues. In conclusion, loss of H3.3 in both MRFs and in regulatory elements could be at the roots of the trancriptomic dysregulation we observe in DM-MuSCs, and lead to their inability to properly undergo myogenesis.

Discussion

In this study, major transcriptional defects are observed in DM-MuSCs, with a global upregulation of pro-inflammatory genes and repression of myogenic-associated genes. DM-MuSCs exhibit reduced transcriptional plasticity, failing to properly exit the cell cycle and to activate the myogenic transcriptional program. Although they retain their myogenic identity, they fail to activate key MRFs. DM-MuSCs show an intrinsic activation of canonical TNF-⍺ signaling, with the upregulation of targets of NFκB and AP-1 family of TFs. Strikingly, this intrinsic inflammation is not marked by intracellular elevated IFN-I signaling, which corroborates with the fact that IFN-I only weakly impacts MuSC dynamics in vitro (only the last fusion step). TNF-α stimulation of HC-MuSCs partly mimics DM-MuSCs by directly blunting their differentiation. Additionally, DM-MuSCs exhibit reduced expression of H3.3 and its chaperones, which correlates with lower H3.3 levels on the chromatin. Along with the intrinsic effects of TNF-α, this suggests that defects in H3.3 dynamics could further impair the activity of myogenic enhancers, directly contributing to the dysfunction of MuSCs in DM. However, it remains unclear whether the failure of DM-MuSCs to upregulate MRFs is a consequence of reduced H3.3 deposition or whether the impaired upregulation of MRFs itself leads to decreased H3.3 deposition.

DM-MuSCs present an intrinsic elevation of pro-inflammatory signaling. However, while DM is characterized by a high IFN-I signature, DM-MuSCs have a moderate intrinsic IFN-I signaling activity. Given the lack of effect of IFN-I on the first steps of in vitro myogenesis observed in this study, this suggests that IFN signaling may not directly influence proliferating or early differentiating MuSCs. Instead, IFN-I stimulation represses cell fusion. Other studies reported a negative effect of IFN-I on MuSC fusion and myofiber size [33, 34] while another publication showed that on a 3D myobundle model, IFN-I stimulation leads to reduced muscle contraction strength and increased fatigue [37]. Therefore, elevated IFN-I levels may have a greater impact on myofibers than on MuSCs. Another potential implication of elevated IFN-I could be on the skeletal muscle microenvironment. Indeed, another characteristic of DM is the infiltration of immune cells in the perifascicular area, mainly monocytes and neutrophils [38–41]. Immune cells are believed to be the main source of IFN-I and other pro-inflammatory cytokines found in DM [38, 42]. However, the mechanism driving their activation and infiltration into the muscle remains poorly understood. Interestingly, DM-MuSCs upregulate the expression of chemoattractants, such as CXCL chemokines. There is a growing number of publications reporting beneficial effect of JAK inhibitors on DM patients [6, 34, 43, 44]. JAK inhibitors suppress IFN-I signaling, potentially protecting myofibers, while also reducing activation of the immune system [34].

In spite of the reported high IFN-I activity in DM patients, IFN-I signaling pathway appears weakly active in DM-MuSCs in vitro. Instead, they show intrinsic activation of the canonical TNF-α signaling pathway. TNF-α superfamily ligands and downstream effectors are well known for their capacity to block differentiation of MuSCs [45–50]. In addition, elevated TNF-α signaling has been described in other context with chronic inflammation, such as sarcopenia [51, 52], aging [53, 54], and cancer cachexia [55]. Overexpression of TNF-α in mice leads to early onset of rheumatoïd arthritis, loss of muscle mass, and interstitial lung disease [56], the latter being also common among MDA5 + DM patients [57]. Nonetheless, our results, together with the existing literature, suggest that strong of TNF-α stimulation alone is insufficient to fully replicate the physiological defects of DM-MuSCs, especially since it does not impact on MuSC proliferation [50, 58]. While it has been demonstrated that low concentration of TNF repress proliferation, it also increases myogenic differentiation [45]. This opens at least two possibilities: either an alternative mechanism is responsible for persistently repressing the proliferation of DM-MuSCs, or long-term, chronic exposure to TNF-α could have broader effects as compared with short-term stimulation. A recent publication further complexified the role of TNF-α in the regulation of myogenesis [59]. The authors performed knockdown of TNF-α in MuSCs and observed that it increases both proliferation and differentiation. Thus, it seems that autocrine TNF-α signaling in MuSCs inhibits both proliferation and differentiation, which is highly reminiscent of the phenotype of DM-MuSCs. The differences between endogenous and exogenous TNF-α stimulation in MuSCs remain to be explored, particularly given the very low expression of the TNFA gene in both DM-MuSCs and HC-MuSCs.

The upregulation of inflammatory signaling pathways can partly explain DM-MuSC myogenesis defects. However, the mechanism sustaining their intrinsic inflammation remains elusive. Indeed, how can inflammation be sustained without the presence of the immune system and its proinflammatory signals? Several recent publications show that following acute inflammation, AP1 family of TFs are retained on some enhancers, associated with retention of euchromatin marks H3K27ac and H3K4me1 and highly open chromatin [60, 61]. These loci and the binding of AP1 TF allow an inflammatory memory response following a second stimulation. One of these publications showed that the inflammatory exposure of pancreatic cells leads to improper cell fate decision and maintenance of an inflammatory specific population of de-differentiated cells [60]. This is reminiscent of our results that show failure of DM-MuSCs to properly engage in differentiation and their maintenance of an inflammatory phenotype. The authors also showed that this correlates with the retention of H3K4me1 and H3K27ac on inflammation memory loci, with an associated AP1 signature. In our study, we observed an elevated expression of targets of AP1 TF in DM-MuSCs no matter the condition, even though there is a poor conservation of the dysregulated genes from one condition to the other. This could be explained by AP1 targeted enhancers being constantly active in DM-MuSCs, but shifting their target genes during the differentiation process as gene regulatory networks are remodeled. Chronic exposure to inflammation of DM-MuSCs could have led to a pathological memory response, which would maintain an inflammatory specific population of DM-MuSCs with elevated activation of AP1 and NFkB signaling pathways. Inflammatory MuSCs could propagate inflammation to healthy cells or at least prevent their proper differentiation by the repression of differentiation induced by their secretion of TNF-α or other proinflammatory cytokines.

In addition to intrinsic TNF-α activity, DM-MuSCs present blunted response to myogenesis and fail to remodel their transcriptome to properly differentiate. Interestingly, there is a big overlap between MuSC specific enhancers and regions with reduced H3.3 in DM-MuSCs. Deposition of H3.3, notably by HIRA, has been associated with the regulation of cell fate and cell identity [9–12]. Knock-out of either HIRA or H3.3 in mouse embryonic stem cells reduces the binding of de novo expressed TFs to their target loci, preventing proper differentiation. This highlights the crucial role of H3.3 deposition for cell fate [62]. Moreover, H3.3 is also enriched in poised enhancers marked by H3K4me1 but no other active marks [36]. It was shown in mice that knock-out of HIRA leads to the reduced incorporation of H3.3 in the chromatin and consequently the reduction of active mark H3K27ac at ATAC peaks loci. This ultimately impedes myogenesis by MuSCs [12]. Finally, retinoic acid stimulation followed by time-ChIP revealed that H3.3 is already present on retinoic acid activated enhancers before the onset of the upregulation of retinoic acid induced genes [63]. Thus, H3.3 deposition would bookmark poised enhancers and maintain their chromatin open to be rapidly activated upon intrinsic or extrinsic stimulation. In DM-MuSCs, loss of H3.3 could lead to the erosion of the euchromatin on myogenic-specific enhancers, reducing their capacity to be rapidly activated upon differentiation. This could explain the loss of transcriptional plasticity observed in DM-MuSCs. Our study reveals for the first time a potential link between chronic inflammation and loss of H3.3, associated with an intrinsic activation of TNF-α signaling. This opens new avenues to investigate how inflammation can become chronic and pathological in myopathies.

Supplementary data

Supplementary data are available at NAR Molecular Medicine online.

Conflict of interest

None declared.

Funding

INMG-PGNM laboratory is funded by grants from the Centre National de la Recherche Scientifique (CNRS), Institut National de la Santé et de la Recherche Médicale (Inserm), University Claude Bernard Lyon 1, and l'Association Française contre les Myopathies (AFM)-téléthon. This work was supported by the Agence Nationale de la Recherche (ANR) through grants ANR-21-CE17-0018 IFN-Epi-IM to P.L., B.C., O.B., AFM-téléthon (MyoNeurALP alliance to P.L. and B.C.),ANR-10-LABX-61 LabEX DEVweCAN and ANR-20-COV9-0004-01 to P.L., ligue contre le cancer to A.C. P.L. team is member of the Groupements de Recherche, “Dynamique des interactions entre chromatines virale et cellulaire” (DYNAVIR, GDR2194), and “Architecture et Dynamique du Noyau & des Génomes” (ADN&G) funded by the CNRS. P.L. is a CNRS Research Director.

Data availability

The RNA-seq and CUT&RUN datasets have been deposited in the Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nig.gov/geo/) under the accession numbers GSE298208 and GSE298210. ChIP-seq datasets of histone PTM in HC-MuSC were retrieved from ENCODE (https://www.encodeproject.org/).