Somatic mutagenesis in satellite cells associates with human skeletal muscle aging

Human aging is associated with a decline in skeletal muscle (SkM) function and a reduction in the number and activity of satellite cells (SCs), the resident stem cells. To study the connection between SC aging and muscle impairment, we analyze the whole genome of single SC clones of the leg muscle vastus lateralis from healthy individuals of different ages (21–78 years). We find an accumulation rate of 13 somatic mutations per genome per year, consistent with proliferation of SCs in the healthy adult muscle. SkM-expressed genes are protected from mutations, but aging results in an increase in mutations in exons and promoters, targeting genes involved in SC activity and muscle function. In agreement with SC mutations affecting the whole tissue, we detect a missense mutation in a SC propagating to the muscle. Our results suggest somatic mutagenesis in SCs as a driving force in the age-related decline of SkM function.

Threshold used to define the expressed enhancers, promoters and genes obtained from the "Myoblast differentiation to myotube" dataset from the FANTOM5 project http://fantom.gsc.riken.jp/data/. The expression data were available at 8 different time points during the myoblast to myotube differentiation (CNhs13847-14585), as follows: day 00 (basal); day 01, day 02, day 03, day 04, day 06, day 08, day 10, and day 12 following exposure to the differentiation stimuli. The expression levels before (day 0) and after the induction of the differentiation were separately analyzed. (a) The size of the genomic region considered "expressed enhancers" on day 0 (basal) or at all time points is shown as a function of the enhancer expression levels. In this study, only enhancers expressed with score ≥1 were included. The threshold is shown as a vertical bar. (b) For the promoters, the size of the expressed regions is shown as a function of the expression levels. The threshold was set as ≥30. (c) For the exons, the gene names were derived from the expressed regions that were included in the promoter list, and the number of genes as a function of the expression levels is shown.
Representative examples of 2D plots from the rare event detection ddPCR analysis of variant c.7825C>T in gene HSPG2 in satellite cell clone CES6 P2703_127 DNA (a), CES6 blood DNA (b), CES6 skeletal muscle DNA (c), control skeletal muscle DNA (d), CES6 skeletal muscle cDNA (e), control skeletal muscle cDNA (f), CES6 skeletal muscle noRT control (g) and control satellite cell population cDNA (h). Channel 1 fluorescence (mutant probe labeled with FAM) is plotted against channel 2 fluorescence (wild-type probe labeled with HEX). Blue dots represent droplets containing only mutant alleles, orange dots represent droplets containing both wild-type and mutant alleles, green dots represent droplets containing only wild-type alleles, while gray dots represent empty droplets. SCC: satellite cell clone, SkM: skeletal muscle, control SkM: skeletal muscle biopsy from unrelated individual.
Supplementary Note 1: List of tested genes associated with muscle diseases.  TCAP, TIA1, TK2, TMEM43, TMEM5, TNNI2, TNNT1, TNNT3, TNPO3, TOR1A, TOR1AIP1, TPM2,  TPM3, TRAPPC11, TRAPPC11, TRIM32, TRIM54, TRIM63, TTN, TTR, TUBB3, VCP, YARS2 The list of genes responsible for muscle diseases was adopted from the gene table of monogenic neuromuscular disorders (nuclear genome), which is updated yearly and available at www.musclegenetable.fr. Only disease groups associated with myocyte dysfunction were selected, while genes associated with neuron or cardiomyocyte function were excluded. Genes are shown in alphabetic order. The characterization of the satellite cell purity in the CD56 + sorted populations obtained from every biopsy used in the study was performed either in parallel with the original cell sorting (freshly isolated cells, left part of the table) or using the unsorted cells left by the sorting and expanded in vitro for 4 weeks (expanded populations, right part of the table). The expression of MyoD was measured by qPCR. The transcript levels of Pax7 were measured by qPCR in the sorted CD56 + cells plated as a population and expanded for 1 week. Values are shown as a relative expression, using the CES3 expanded population as a reference (see Supplementary Fig. 1b). The % of Pax7 and CD45 (bonemarrow derived) positive cells was assessed by FACS. The % of TE7 (fibroblast) cells was assessed by immunofluorescence on cytospun cells.
Supplementary Satellite cell clones (SCCs) sequenced in the study were derived from 7 donors (first column) of different ages. The myogenic origin of 21/29 SCCs was tested with either a differentiation assay (myotubes) or qPCR for the expression of the myogenic genes Pax7, MyoD and myogenin. All tested clones resulted positive. Coverage indicates the percentage of autosomes covered by at least 15 reads in each sample. SNVs: single nucleotide variants. Indels: insertion deletions. SNVs and indels were normalized on coverage (last columns).
Supplementary To validate somatic variants found in the study, 2 SCCs (P2703_113 and P2703_116) were subjected to a second round of whole genome sequencing (validation set), with independent library preparation. Clone P2703_113 was sequenced twice with independent library preparations. Clone P2703_116 was split into 2 wells during cell culture (1000 cell-stage) and resulted in 2 independently grown clones (P2703_116 and P2703_119) derived from the same ancestor cell. Validation was also tested in unrelated SCCs (P2703_111/120) to test the unspecific background signal of the method when comparing 2 random samples. Variants were called with our pipeline in the discovery set and then tested in the validation set. Variants were considered validated when present with a minimum coverage of 15x and a minimum of 3 reads supporting the alternative allele.
Supplementary Variants with predicted high impact on the encoded protein were selected across all clones. 11 variants suitable for Agena genotyping (sequenom maldi-tof technology) were tested in all samples included in the study. Variants were considered validated when they were found in the corresponding SCC and in none of the other samples, including the corresponding blood DNA. Six additional variants were selected and tested with ddPCR rare event detection assays in SCC, blood and muscle of the donor where they were discovered (see also Supplementary Tables 11 and 12).
Supplementary Number of indels/SCC according to different deletion and insertion classes. Specific gene sets were tested for the occurrence of somatic mutations in young and old SCCs. Lists of differentially expressed genes after training were obtained from the published studies indicated in the first column.

Supplementary
Supplementary Specific gene sets were tested for the occurrence of somatic mutations in young and old SCCs. Lists of the genes involved in the pathway that modulates the response to training were obtained from curated gene sets from GO, KEGG, Biocarta and Reactome. Curated databases were accessed through the GSEA (Molecular Signature Database, Broad Institute). "Insulin," "AMPK" and "ribosome" pathways are not specific to the skeletal muscle. Conversely, all curated gene sets included in "hypertrophy" include the words "skeletal muscle" in their description, and all the sets in "activation and recruitment of resident stem cells" include the word "muscle".
Supplementary  The average number of mutations in different gene-related regions (exons, introns, 5 kb upstream and 5 kb downstream) in old SCCs (N=16) is shown. List of tested genes is shown in supplementary note 1. All genes present in supplementary note 1 and not mentioned in here did not show any somatic variant in old SCCs. Genes on chromosome X (ALG13, DMD, EMD, FHL1, LAMP2, MTM1, PGK1, PHKA1,  VMA21) could not be addressed, as only autosomic variants were analyzed in the study (see methods). Supplementary