YTHDF2 governs muscle size through a targeted modulation of proteostasis

Abstract

The regulation of proteostasis is fundamental for maintenance of muscle mass and function. Activation of the TGF-β pathway drives wasting and premature aging by favoring the proteasomal degradation of structural muscle proteins. Yet, how this critical post-translational mechanism is kept in check to preserve muscle health remains unclear. Here, we reveal the molecular link between the post-transcriptional regulation of m⁶A-modified mRNA and the modulation of SMAD-dependent TGF-β signaling. We show that the m⁶A-binding protein YTHDF2 is essential to determining postnatal muscle size. Indeed, muscle-specific genetic deletion of YTHDF2 impairs skeletal muscle growth and abrogates the response to hypertrophic stimuli. We report that YTHDF2 controls the mRNA stability of the ubiquitin ligase ASB2 with consequences on anti-growth gene program activation through SMAD3. Our study identifies a post-transcriptional to post-translational mechanism for the coordination of gene expression in muscle.

Fig. 1. YTHDF2 regulates postnatal skeletal muscle mass.

a Schematic of muscle-specific YTHDF2 knockout (Y2-KO) mouse generation, created in Adobe Illustrator. b qPCR analysis of Ythdf2 mRNA expression using the gastrocnemius of control (Ctrl) and Y2-KO mice. c Western blot analysis of YTHDF2 protein expression in Ctrl and Y2-KO gastrocnemius. d Quantification of YTHDF2 protein expression using total protein detection by Ponceau stain as a loading control. e Tibialis anterior weight, f quadriceps weight, and g gastrocnemius weight for 1- and 2-month-old Ctrl and Y2-KO mice, normalized to tibia length. h Body weight and i tibia length for Ctrl and Y2-KO mice at 2 months of age. j Representative hematoxylin/eosin (H&E) staining for Ctrl and Y2-KO tibialis anterior. k Quantification of m⁶A relative to total adenosine (m⁶A/A) as determined by ELISA in Ctrl and Y2-KO quadriceps. l Western blot analysis of YTHDF2 protein expression in Ctrl tibialis anterior compared to Y2-KO injected with AAV-GFP or AAV-YTHDF2 (AAV-Y2) at 2 months of age and analyzed 8 weeks following administration. m Tibialis anterior weight normalized to tibia length for the indicated groups. Biological animal replicates: n = 5 (Ctrl) and 4 (Y2-KO) in panel b; n = 3 (Ctrl) and 3 (Y2-KO) in panel c and d; n = 4 (Ctrl at 1 month), 5 (Y2-KO at 1 month), 12 (Ctrl at 2 months), and 8 (Y2-KO at 2 months) for panel e–g; n = 12 (Ctrl) and 8 (Y2-KO) in panel h, I; n = 5 (Ctrl) and 4 (Y2-KO) in panel k; n = 3 (Ctrl), 3 (Y2-KO + AAV-GFP), and 3 (Y2-KO + AAV-Y2) in panel l; n = 9 (Ctrl), 6 (Y2-KO + AAV-GFP), and 7 (Y2-KO + AAV-Y2) in panel m. Data were presented as the mean ± SEM with the individual biological samples shown. Significance was determined by two-tailed Student’s t-test for comparisons between Ctrl and Y2-KO mice, by two-way ANOVA with Tukey’s HSD multiple-comparison test for comparison of the means of Ctrl and Y2-KO mice at 1 and 2 months of age, or by one-way ANOVA for Ctrl, Y2-KO + AAV-GFP, and Y2-KO + AAV-Y2 mice: *p ≤ 0.05, **p ≤ 0.01, ****p ≤ 0.0001. Scale bar = 100 µm for panel j.

Fig. 2. Loss of YTHDF2 impairs myofiber growth irrespective of fiber-type distribution.

Representative wheat germ agglutinin (WGA, green) stained images for Ctrl and Y2-KO tibialis anterior at a 1 and b 2 months of age. Myofiber cross-sectional areas for c tibialis anterior and d quadriceps for Ctrl and Y2-KO mice at 1 and 2 months of age. e Fiber type distribution for Ctrl and Y2-KO tibialis anterior and f representative myosin type staining for type I (anti-BA-D5, red), type IIA (anti-SC-71, pink), type IIB (anti-BF-F3, pink), and all but IIX (anti-BF-35, pink) with counterstains anti-laminin (green) and DAPI (blue). Biological animal replicates n = 4 (Ctrl at 1 month), 4 (Y2-KO at 1 month), 7 (Ctrl at 2 months), and 6 (Y2-KO at 2 months) for panel c, d; n = 4 (Ctrl) and 4 (Y2-KO) for panel e. Data were presented as the mean ± SEM with the individual biological samples shown. Significance was determined by two-tailed Student’s t-test for comparisons between Ctrl and Y2-KO mice at 2 months, or by two-way ANOVA with Tukey’s HSD multiple-comparison test for comparison of the means of Ctrl and Y2-KO mice at 1 and 2 months of age. Welch’s correction was used for unequal variances: **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Scale bar = 125 µm for panel a, b. Scale bar = 100 µm for panel f.

Fig. 3. Deletion of YTHDF2 elicits disparity in muscle size and function by 8 months of age.

a Quadriceps weight, b tibialis anterior weight, c gastrocnemius weight, d plantaris weight, e soleus weight, f and heart weight for Ctrl and Y2-KO mice at 8 months of age, normalized to tibia length. g Tibia length for Ctrl and Y2-KO mice at 8 months of age. h Representative wheat germ agglutinin (WGA, green) stained images for Ctrl and Y2-KO tibialis anterior at 8 months of age. Myofiber cross-sectional areas for i quadriceps, j tibialis anterior, k gastrocnemius, l plantaris, and m soleus for Ctrl and Y2-KO mice at 8 months of age. n In vivo muscle twitch and tetanic torque measures for Ctrl and Y2-KO mice at 8 months of age. o Schematic for forced graded maximal exercise test, created in Adobe Illustrator, and p maximal running distance for Ctrl and Y2-KO mice at 8 months of age. Biological animal replicates: n = 13 (Ctrl) and 11 (Y2-KO) in panel a–g; n = 5 (Ctrl) and 5 (Y2-KO) in panel i–m; n = 10(Ctrl) and 15 (Y2-KO) in panel n; n = 12 (Ctrl) and 10 (Y2-KO) in panel p. Data were presented as the mean ± SEM with the individual biological samples shown. Significance was determined by a two-tailed Student’s t-test for comparisons between Ctrl and Y2-KO mice. Welch’s correction was used for unequal variances. Grubbs’ (ESD) tests were run using GraphPad Prism, and outliers were removed from analysis when applicable: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Scale bar = 125 µm for panel h.

Fig. 4. YTHDF2-null muscles favor catabolic signaling.

a Western blot analyses and quantification using total protein detection by Ponceau stain as a loading control for b phosphor-(p-)AKT, c AKT, d p-AKT/AKT ratio, e p-mTOR, f mTOR, g p-mTOR/AKT ratio, h LC3I, i LC3II, j LC3II/LC3I ratio, k p62, l Beclin-1, m monoubiquitin, and n polyubiquitin in Ctrl and Y2-KO quadriceps at 8 months of age. o Proteasomal activity as determined via colorimetric proteasome activity assay kit in Ctrl and Y2-KO quadriceps. Biological animal replicates: n = 4 (Ctrl) and 4 (Y2-KO) in panel a–n; n = 3(Ctrl) and 3 (Y2-KO) in panel o. Data were presented as the mean ± SEM with the individual biological samples shown. Significance was determined by a two-tailed Student’s t-test for comparisons between Ctrl and Y2-KO mice. Welch’s correction was used for unequal variances: *p ≤ 0.05, **p ≤ 0.01.

Fig. 5. YTHDF2 is dispensable for myogenic regeneration but essential for fiber growth following acute injury.

a Schematic for barium chloride (BaCl₂)-induced chemical injury to the tibialis anterior muscle, created in Adobe Illustrator. b Weight for Ctrl and Y2-KO tibialis anterior subjected to injection by PBS (uninjured) or BaCl₂ 4-, 7-, and 14 days post-injection at 2 months of age. c Representative wheat germ agglutinin (WGA, green) staining for Ctrl and Y2-KO tibialis anterior 0- (PBS), 4-, 7-, and 14 days post-injection with BaCl₂. d Representative hematoxylin/eosin (H&E) staining for Ctrl and Y2-KO tibialis anterior 0- (PBS), 4-, 7-, and 14 days post-injection with BaCl₂. e Representative myotube (Desmin, green) staining for differentiated myoblasts isolated from Ctrl and Y2-KO total hindlimb, with counterstain DAPI (blue). f Quantification of fusion index for primary myoblast differentiation experiment. Biological animal replicates: n = 7 (Ctrl Uninjured), 6 (Y2-KO Uninjured), 6 (Ctrl D4), 5 (Y2-KO D4), 5 (Ctrl D7), 5 (Y2-KO D7), 7 (Ctrl D14), and 6 (Y2-KO D14) for panel b. Primary myoblast culture replicates n = 3 (Ctrl) and 3 (Y2-KO) for panel e, f. Data were presented as the mean ± SEM with the individual biological samples shown. Significance was determined by two-way ANOVA with Tukey’s HSD multiple-comparison test for comparison of the means of Ctrl and Y2-KO mice across uninjured, and BaCl₂-injured D4, D7, and D14 groups: *p ≤ 0.05, **p ≤ 0.01, ****p ≤ 0.0001. Scale bar = 125 µm for panel c. Scale bar = 100 µm for panel d. Scale bar = 100 µm for panel e.

Fig. 6. YTHDF2 is essential to the hypertrophic response of skeletal muscle.

a Schematic of muscle overload-induced hypertrophy of the plantaris muscle, created in Adobe Illustrator. b Western blot analysis of YTHDF2 protein expression in the wild-type plantaris following muscle overload or sham operation. c Quantification of YTHDF2 protein expression using total protein detection by Ponceau stain as a loading control. d Plantaris weight for 2-month-old Ctrl and Y2-KO mice 14 days following sham or muscle overload surgery, normalized to tibia length. e Representative wheat germ agglutinin (WGA, green) staining and f myofiber cross-sectional areas for Ctrl and Y2-KO plantaris 14 days following sham or muscle overload surgery. g Venn diagram of LC-MS/MS-identified proteins upregulated in either Ctrl or Y2-KO plantaris muscles (fold change >2.0) following muscle overload. h Logarithmic fold change plot for peptides detected in both Ctrl and Y2-KO samples. i Venn diagram of transcripts detected with meRIP-Seq and YTHDF2-RIP-Seq. j Overlapping transcripts were filtered according to whether the corresponding protein was detected via LC-MS/MS (gray). Proteins with differential expression (fold change >1.3) have been highlighted. Biological animal replicates n = 3 (Sham) and 3 (Overload) for panel b, c; n = 4 (Ctrl Sham), 4 (Y2-KO Sham), 6 (Ctrl Overload), and 5 (Y2-KO Overload) for panel d; n = 4 (Ctrl Sham), 4 (Y2-KO Sham), 4 (Ctrl Overload), and 5 (Y2-KO Overload) for panel f; pooled plantaris samples (2 mice per sample) were used for LC/MS, where data reflect n = 2 (pooled Ctrl) and 2 (pooled Y2-KO) for panel g, h. Data are presented as the mean ± SEM with the individual biological samples shown. Significance was determined by two-tailed Student’s t-test for comparisons between WT sham and overload mice, or by two-way ANOVA with Tukey’s HSD multiple-comparison test for comparison of the means of Ctrl and Y2-KO mice following sham or overload surgery: ***p ≤ 0.001, ****p ≤ 0.0001. Scale bar = 125 µm for panel e.

Fig. 7. YTHDF2 modulates the expression of E3 ubiquitin ligase ASB2.

a Western blot analysis of ASB2 protein expression in Ctrl and Y2-KO quadriceps at baseline (2 months of age). b Quantification of YTHDF2 protein expression using total protein detection by Ponceau stain as a loading control. qPCR analysis of c total Asb2 mRNA and d cytosolic Asb2 mRNA using the quadriceps muscle of Ctrl and Y2-KO mice. qPCR analyses for Asb2 RNA enrichment in the wild-type quadriceps following RNA immunoprecipitation with e m⁶A and f YTHDF2, normalized to input RNA. g qPCR analysis for Asb2 following Actinomycin D treatment for the indicated times in H9C2 rat myoblasts transfected with an siRNA pool targeting YTHDF2 (si-Y2) or control non-targeting siRNA (si-Ctrl), normalized to respective t = 0 values. Best-fit values for half-life and decay rate were calculated using least squares regression analysis. h Schematic of muscle overload-induced hypertrophy of the plantaris muscle with siRNA intervention, created in Adobe Illustrator. i Plantaris weight for Ctrl (sham and overload) and siRNA-treated Y2-KO mice 14 days following muscle overload surgery, normalized to tibia length. Biological animal replicates: n = 4 (Ctrl) and 4 (Y2-KO) for panel a–f, and i. Biological cell replicates: n = 3 (si-Ctrl at 0, 0.5, 1, and 2 h) or 4 (si-Ctrl at 4 and 6 h), and 4 (si-Y2 at 0, 0.5, 1, 2, 4, and 6 h) for panel g. Data were presented as the mean ± SEM with the individual biological samples shown. Significance was determined by a two-tailed Student’s t-test for comparisons between Ctrl and Y2-KO mice, and between WT IgG and pulldown antibody (m⁶A/YTHDF2). Wilcoxon rank-sum test was used to compare means for nonparametric ASB2 protein quantification data. Grubbs’ (ESD) tests were run using GraphPad Prism and outliers were removed from analysis when applicable: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

Fig. 8. Altered ASB2 expression in the absence of YTHDF2 elicits dysregulated SMAD3 signaling.

a Western blot analysis of protein expression of SMAD proteins 1, 2, 3, 4, 5, 7, and 9 in Ctrl and Y2-KO quadriceps at 2 months of age. b Western blot analysis of phosphor-(p-)SMAD3 and FOXO3a protein expression in Ctrl and Y2-KO quadriceps at baseline and c, d respective protein expression using total protein detection by Ponceau stain as a loading control. qPCR analysis of e Murf1 and f Mafbx mRNA expression in Ctrl and Y2-KO gastrocnemius. Chromatin immunoprecipitation (ChIP) analysis of SMAD3 interaction with g MAfbx and h Murf1 promoters. i Western blot analysis of SMURF2 protein expression in Ctrl and Y2-KO quadriceps and j quantification of SMURF2 protein expression using total protein as a loading control. k Western blot analysis for SMURF2 and l quantification following cyclohexamide treatment for the indicated times in H9C2 rat myoblasts transfected with an siRNA pool targeting YTHDF2 (si-Y2) or control non-targeting siRNA (si-Ctrl), normalized to respective t = 0 values. Best-fit values for half-life and decay rate were calculated using least squares regression analysis. m Western blot analysis for SMURF2, YTHDF2, and ASB2 and n quantification following in H9C2 rat myoblasts transfected with an siRNA pool targeting YTHDF2 (si-Y2) and/or ASB2 (si-ASB2), or control non-targeting siRNA (si-Ctrl). o Schematic of YTHDF2-mediated muscle growth modulation, created in Adobe Illustrator. Biological animal replicates: n = 4 (Ctrl) and 4 (Y2-KO) for panels a–c, d, I, j; n = 5 (Ctrl) and 4 (Y2-KO) for panel e, f; n = 3 (Ctrl) and 3 (Y2-KO) for panel g, h, m, n. Pooled cell replicates were used for panels k and l. Data were presented as the mean ± SEM with the individual biological samples shown. Significance was determined by a two-tailed Student’s t-test for comparisons between Ctrl and Y2-KO mice (using respective control values for panel b data): *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.