CERS1-correlated pathways in human skeletal muscle associate with muscle function and myogenesis. (A) Overview over the de novo sphingolipid biosynthesis pathway. (B) CERS1 expression in young (22 ± 3.61yo) and old (74 ± 8.4yo) human skeletal muscle cells (n=3-5 biological replicates). (C) Gene set enrichment analysis of CERS1 mRNA correlated transcripts in the human skeletal muscle from the Genotype-Tissue Expression (GTEx) dataset (n=469). Normalized effect size (NES). (D) Spearman correlation between skeletal muscle expression of CERS1 and the first principal component (PC1) of the myogenesis pathway (left) from (C) and expression of key genes involved in myogenesis (right). (E-F) Gene expression of enzymes involved in the sphingolipid de novo synthesis pathway (E) and myogenic regulatory factors (F) upon horse serum induced differentiation of primary human skeletal muscle cells (n=2-3 biological replicates). The pharmacological inhibitor P053 or DMSO was used at 1uM in (E-F). Data: mean ± SEM. *P < 0.05, **P < 0.01, ****P<0.0001.

P053 administration inhibiting Cers1 in aged mice deteriorates skeletal muscle function and morphology. (A) Overview over the experimental pipeline that was used to administer P053 in aged C57BL/6J mice using intraperitoneal injections (I.P) three times a week for 6 months. (B) Skeletal muscle ceramide levels in aged mice treated with DMSO or P053 (n=7-8 per group). (C) Representative images of laminin-stained tibialis anterior of young control or mice injected with DMSO or P053 (n=5-6 per group). Scale bar, 50µM. (D) Quantification of minimal Ferret diameter distribution in skeletal muscle cross sections from young control or mice injected with DMSO or P053 (n=5-6 per group). (E) Automated detection of laminin/CD45+ stained muscle fibers from young control or aged mice injected with DMSO or P053 showing cross sectional area (top) and inflammation-adjacent signal intensities (bottom). (F) Representative brightfield images and quantification of Sirus red stained muscle sections from young and aged mice treated with DMSO or P053 (n=6 per group). (G) Representative brightfield images and quantification of muscle cross sections from young and aged mice treated with DMSO or P053 stained with hematoxylin/eosin (n=6 per group). (H-I) Phenotyping measurements of young control and aged mice injected with DMSO or P053 showing (H) lean body mass and (I) grip strength (n=7-14 per group). Data: mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Adeno-associated virus 9 (AAV9) mediated knockdown of Cers1 expression in aged skeletal muscle reduces skeletal muscle function and morphology. (A) Schematic showing a single gastrocnemius intramuscular injection of adeno-associated virus particles containing short hairpin RNA against Cers1 in aged C57BL/6J mice. (B) Skeletal muscle ceramide levels in aged mice intramuscularly injected with shRNA targeting Cers1 (n=6-8). (C) Representative images of laminin-stained aged mice muscle cross sections intramuscularly injected with shRNA targeting Cers1 (n=6 per group). Scale bar, 50µM. (D) Quantification of minimal Ferret diameter distribution in aged mice intramuscularly injected with shRNA targeting Cers1 (n=6 per group). (E) Automated detection of laminin/CD45+ stained skeletal muscle fibers from aged mice intramuscularly injected with shRNA-Cers1 showing cross sectional area (top) and inflammation-adjacent signal intensities (bottom). (F) Representative brightfield images and quantification of Sirus red stained skeletal muscle cross sections from mice intramuscularly injected with AAV9 particles containing scramble, or shRNA targeting Cers1 (n=6 per group). (G) Representative brightfield images and quantification of hematoxylin/eosin-stained muscle cross sections from mice intramuscularly injected with AAV9 particles containing scramble, or shRNA targeting Cers1 (n=6 per group). (H-I) Phenotyping measurements of aged mice intramuscularly injected with shRNA-Cers1 or a scramble shRNA with (H) gastrocnemius muscle weight and (I) grip strength (n=6-8 per group). Data: mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Inhibition of Cers1 in muscle cells impairs myogenic differentiation and C. elegans healthspan. (A) Primary human skeletal muscle cell ceramide levels infected with adenovirus containing silencing RNA targeting CERS1 (n=5-6 per group). (B) Representative confocal immunocytochemistry images of differentiating human primary muscle cells deficient of CERS1 (n=4 per group, left). (C) Gene expression profiling of differentiating human primary muscle cells deficient of CERS1 (n=4-5 per group, right). Scale bar, 50µM. (D) Travelled distance in transgenic RW1596 C. elegans (myo3p-GFP) treated with DMSO, 50uM P053 or 100uM P053 (n=40-57 per group). (E) Representative confocal muscle images of RW1596 C. elegans (myo3p-GFP) treated with DMSO or 100uM P053 (n=6 per group). (F) Travelled distance in transgenic RW1596 C. elegans (myo3p-GFP) treated with EV control, or RNAi against lagr-1 (n=46-56 per group). (G) Representative confocal muscle images of RW1596 C. elegans (myo3p-GFP) treated with control, or RNAi against lagr-1 (n=6 per group). Data: mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P<0.0001.

Common and rare genetic variants in CERS1 affect muscle function and health in humans. (A) Overall result of the Mendelian randomization analysis in the UK biobank cohort using skeletal muscle expression quantitative trait loci (cis-eQTL) of CERS1. (B) Scatter plot showing the effect of the three independent (R2<0.1, see Figure S9D) cis-eQTLs rs117558072, rs1122821, rs71332140 on overall health, right grip strength, and left grip strength in the UK biobank. The slope of the regression line depicts the estimated causal effect with the inverse-variance weighted Mendelian randomization method. (C) Lollipop plot depicting rare variants in the coding region of CERS1 and their effects on overall health, right grip strength, and left grip strength in the UK biobank. Colors indicate phenotypes, dot size indicates effect size, red dotted line indicates the suggested cut-off p<0.05. (D) Overview over the de novo sphingolipid biosynthesis pathway highlighting the hypothesis that CERS1 inhibition leads to a compensatory upregulation of CERS2, which might inhibit muscle function in aging. Data: mean ± SEM. P values are shown in the figure panels.

Cers1 regulates muscle contraction in human muscle and myogenesis in mouse myoblasts. (A) Spearman correlation between skeletal muscle expression of CERS1 and the first principal component (PC1) of the muscle contraction pathway (left) and expression of key genes involved in muscle contraction (right). (B) Gene expression of enzymes involved in the sphingolipid de novo synthesis pathway upon horse serum induced differentiation of mouse C2C12 myoblasts (n=3-4). (C-D) Ceramide levels during myogenic differentiation of mouse C2C12 myoblasts (n=3-4) (C) and human primary skeletal muscle cells (n=2-3) (D). Data: mean ± SEM. ****P<0.0001.

The Cers1 inhibitor P053 deteriorates myoblast maturation. (A) Gene expression of myogenic regulatory factors upon horse serum induced differentiation of mouse C2C12 myoblasts (n=3-4). (B-C) Gene expression of myosin light- and heavy chains upon horse serum induced differentiation of isolated human primary myoblasts (n=2-3 biological replicates) (B) and mouse C2C12 (n=3-4) muscle cells (C) upon horse serum induced myoblast differentiation. The pharmacological inhibitor P053 or DMSO was used at 1uM in (A-C). Data: mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P<0.0001.

The Cers1 inhibitor P053 exacerbates mouse skeletal muscle aging. (A) Skeletal muscle dihydroceramide levels in aged mice treated with DMSO or P053 (n=7-8 per group). (B) Average minimal Ferret diameter in muscle of young and aged mice treated with DMSO or P053 (n=5-6 per group). (C) Quantification of muscle cross-sectional area in young and aged mice treated with DMSO or P053 (n=5-6 per group). (D) Representative immunohistochemistry images of muscle cross sections from young and aged mice treated with DMSO or P053 stained with DAPI and CD45 (n=5-6 per group). Scale bar, 50µM. Data: mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

The Cers1 inhibitor P053 deteriorates mouse skeletal muscle aging. (A) Gene expression in young or aged muscle of mice treated with DMSO or P053 (n=5-6 per group). (B-C) Phenotyping of young control and aged mice injected with DMSO or P053 showing running time (B), and running distance (C) (n=7-14 per group). Data: mean ± SEM. *P < 0.05, **P < 0.01.

Genetic inhibition of Cers1 exacerbates mouse skeletal muscle aging. (A) Cers1 RNA expression in aged mouse muscle intramuscularly injected with scramble, or shRNA targeting Cers1 (n=6 per group). (B) Dihydroceramide levels in gastrocnemius muscle injected with AAV9 particles containing scramble, or shRNA targeting Cers1 (n=6-8 per group). (C) Average Ferret diameter in skeletal muscle of aged mice intramuscularly injected with scramble, or shRNA targeting Cers1 (n=6 per group). (D) Quantification of cross-sectional area in aged mice treated with AAV9-shRNA Cers1 (n=6 per group). (E) Representative immunohistochemistry images of muscle cross sections of CD45-stained muscle from mice intramuscularly injected with AAV9 particles containing scramble, or shRNA targeting Cers1 (n=6 per group). Data: mean ± SEM. *P < 0.05, **P < 0.01.

Genetic inhibition of Cers1 deteriorates mouse skeletal muscle aging. (A) Muscle gene expression in of mice intramuscularly injected with AAV9 particles containing scramble, or shRNA targeting Cers1 (n=6 per group). (B-C) Phenotyping measurements of aged mice intramuscularly injected with shRNA-Cers1 shows running time (B) and running distance (C) (n=6-8 per group). Data: mean ± SEM. *P < 0.05.

Lentivirus mediated silencing of Cers1 reduces ceramides and blunts myogenesis in mouse muscle cells. (A) Cers1 gene expression in C2C12 mouse muscle cells transduced with lentivirus containing scramble, or shRNA targeting Cers1 (n=5-6 biological replicates per group). (B) C2C12 myoblast ceramide levels upon lentivirus mediated silencing of Cers1 (n=3 per group). (C) Dihydroceramide levels in C2C12 cells transduced with lentivirus containing shRNA targeting Cers1 (n=3 per group). (D) Representative immunocytochemistry images of differentiating C2C12 muscle cells deficient of Cers1 (n=6 per group). Scale bar, 50µM. (E) Quantification of (D) showing myotube diameter (left), myotube area (middle) and % of multinucleated myotubes (right) in C2C12 mouse muscle cells transduced with lentivirus containing scramble, or shRNA targeting Cers1 (n=6 per group). (F) Gene expression profiling of differentiating C2C12 muscle cells deficient of Cers1 (n=8-12 per group). Data: mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P<0.0001.

Adenovirus mediated silencing of CERS1 reduces ceramides and blunts myogenesis in human muscle cells. (A) CERS1 gene expression in isolated primary human muscle cells transduced with adenovirus containing scramble, or shRNA targeting CERS1 (n=3 per group). (B) Dihydroceramide levels in differentiated primary human muscle cells transduced with adenovirus containing scramble, or shRNA targeting CERS1 (n=5-6 per group). (C-E) Quantification of myotube diameter (C), myotube area (D), and % of multinucleated myotubes (E) in differentiated primary human muscle cells transduced with adenovirus containing scramble, or shRNA targeting CERS1 (n=4 per group). Data: mean ± SEM. *P < 0.05, **P < 0.01.

CERS1 inhibition is associated with impaired health span. (A) Mean speed measured in transgenic RW1596 C. elegans (myo3p-GFP) treated with DMSO, 50uM P053 or 100uM P053 (n=40-57 per group). (B) Gene expression of the Cers1 orthologue lagr-1 in transgenic RW1596 C. elegans (myo3p-GFP) treated with empty vector control, or RNAi targeting lagr-1 (n=3 biological replicates per group). (C) Mean speed measured in transgenic RW1596 C. elegans (myo3p-GFP) treated with EV control, or RNAi against the Cers1 orthologue lagr-1 (n=46-56 per group). (D) Linkage disequilibrium of the common CERS1 expression quantitative trait loci (cis-eQTL) rs117558072, rs1122821, rs71332140 measured as r2 in all populations (https://ldlink.nci.nih.gov). (E) Violin plots showing the allelic effect of rs117558072 (left), rs1122821 (middle), rs71332140 (right) on CERS1 expression in human skeletal muscle in the Genotype-Tissue Expression (GTEx) dataset (n=706 human in total). Data: mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Cellpose training quality control.

List of mouse and human qPCR primers.

List of plasmids.

List of antibodies.