Fig 2: AAV6-mediated overexpression of Musclin in skeletal muscle attenuates left ventricular dysfunction and myocardial fibrosis during long-term pressure overload.a Scheme depicting the experimental time line. b mRNA level of Ostn (Musclin) 9 weeks after TAC and intramuscular application of AAV6 control (Co) or AAV6 Musclin (Mu) vector (n = 12/group), *p = 0.036 (c) Immunoblot for Musclin 9 weeks after TAC and intramuscular application of AAV6 Co or AAV6 Mu. GAPDH, Actin and Tubulin are loading controls. The size of the proteins is indicated in kDa. d Musclin plasma levels in AAV6 Co (n = 21) or AAV6 Mu (n = 22) treated mice, **p = 0.0029. e Heart weight/tibia length (HW/TL) ratio in AAV6 Co (sham n = 5, TAC n = 7) or AAV6 Mu (sham n = 5, TAC n = 6) treated mice 9 weeks after sham or TAC surgery, ***p = 0.0004, *p = 0.0316. f–h Echocardiographic ejection fraction 3 weeks (all sham n = 5/group, TAC AAV6 Co n = 17, TAC AAV6 Mu n = 15), ****p < 0.0001, **p = 00015 and *p = 0.0293 (f), and 6 weeks after surgery, ****p < 0.0001 in AAV6 Control and AAV6 Musclin groups for sham vs. TAC surgery, **p = 0.0021 (g), and 9 weeks after surgery, ****p < 0.0001, ***p = 0.0006 and **p = 0.0051 (h). Cardiac systolic contractility (dp/dt max, i), *p = 0.0122 and diastolic relaxation (dp/dt min, j), *p = 0.0164 determined by left ventricular catheterization in the indicated mice (all sham n = 5/group, TAC AAV6 Co n = 9, AAV6 Mu n = 8) 9 weeks after sham or TAC surgery. Representative Sirius red-stained heart sections (k) and quantified myocardial fibrotic area (l) and of mice treated as shown (all sham n = 5/group, all TAC n = 8/group, 9 weeks after surgery), **p = 0.0093 and *p = 0.0206. Scale bar: 500 µm. m–p qPCR analysis of the indicated fibrosis genes 9 weeks after sham or TAC surgery (sham AAV6 Co n = 5, sham AAV6 Mu n = 4, TAC AAV6 Co n = 7, TAC AAV6 Mu n = 8), **p = 0.0064 (m), **p = 0.0081 (n), *p = 0.0248 (o), *p = 0.0369 (p). q Mature CNP plasma levels in AAV6 Co or AAV6 Mu treated mice 3 weeks after TAC surgery (n = 4/group), *p = 0.0462. Data are shown as mean ± standard error of the mean (SEM). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by two-tailed Student’s t test (d, q) or by one-way ANOVA followed by the Holms–Sidak’s multiple comparisons test (other bar graphs). Source data are provided as a source data file.

Fig 3: Musclin increases cGMP and cAMP levels and cardiomyocyte contractility through NPR-B-dependent inhibition of PDE3.Cardiomyocytes were isolated from either cGES-DE5 transgenic (for cGMP measurements) or from Epac2-camps (for cAMP measurements) transgenic mice. a, b Fluorescence resonance energy transfer (FRET)-based measurements of cGMP in single cultured cardiomyocytes (n = 7 cells). The stimulation with CNP and Musclin was conducted as indicated. Representative traces (a) and a quantitative analysis (b) are shown, *p = 0.0342. c, d FRET-based measurements of cAMP in single cardiomyocytes treated as indicated (n = 13 cells). The cells were stimulated with Isoproterenol (Iso), CNP and subsequently Musclin as indicated. The graph illustrates the FRET-responses to CNP in % of the maximal Iso effect. Representative traces (c) and a quantitative analysis (d) are shown, **p = 0.0019. e–h ELISA-based measurements of cGMP and cAMP in cultured cardiomyocytes under the indicated conditions from wild-typ (WT) (e, g), **p = 0.0012 for vehicle-treated cells vs. stimulated with CNP (100 nM), ****p < 0.0001 for cells stimulated with Musclin vs. with CNP (100 nM) and Musclin, **p = 0.0013 for cells treated with CNP (100 nM) vs. cells stimulated with Musclin and CNP (100 nM) (e), *p = 0.034 (g) and global Npr2 knockout mice (f, h) (for cGMP measurement n = 4/condition and for cAMP n = 3/condition). i, j FRET-based measurements of cAMP in single WT cardiomyocytes treated as indicated. Representative traces (i) and a quantitative analysis (j) are shown (n = 6 cells without Musclin treatment and n = 8 cells with Musclin). k Representatives traces of sarcomere length and l quantification of cell shortening (from traces as shown in k) in isolated wild-type mouse cardiomyocytes treated as indicated (n = 9 cardiomyocytes per group, *p = 0.017). Cilostamide was used as PDE3 inhibitor. Data in bar graphs are shown as mean ± standard error of the mean (SEM). *p < 0.05, **p < 0.01, ****p < 0.0001 determined by two-tailed Student’s t test (b, d, j), one-way ANOVA followed by the Holms–Sidak’s multiple comparisons test (e, f, l) or by Kruskal–Wallis with Dunn’s multiple comparisons test (g, h). n.s. denotes “not significant”. Source data are provided as a source data file.

Fig 4: Musclin inhibits the activation of cardiac fibroblasts through NPR-B dependent activation of protein kinase G (PKG).a Measurement of cardiac fibroblast proliferation by BrDU incorporation ELISA with addition of CNP and/or Musclin (n = 5/condition, **p = 0.0039 for vehicle vs. CNP, **p = 0.0019 for vehicle vs. musclin and ***p = 0.0002). b Assessment of cardiac fibroblast migration through detection of recovery of a scratch wound after 4 h during stimulation as indicated (n = 6/condition, **p = 0.0023, ***p = 0.0003 for vehicle-treated cells vs. stimulated with Musclin, and ***p = 0.0005 vs. cells stimulated with CNP and Musclin). c–e Measurement of cardiac fibroblast proliferation by BrDU incorporation ELISA under the indicated conditions (c–e) (n = 6/condition, except siNPR2 treatment n = 7); *p = 0.0103 for siControl cells, vehicle treated vs. stimulated with CNP, **p = 0.0016 vs. stimulated with Musclin and ***p = 0.0001 vs. stimulated with CNP and Musclin, *p = 0.0103 for siNPR1 cells, vehicle treated vs. stimulated with CNP, ****p < 0.0001 vs. stimulated with Musclin and vs. stimulated with CNP and Musclin (c); *p = 0.0243, **p = 0.0055 and ***p = 0.0003 (d); **p = 0.0021 for siControl cells, vehicle treated vs. stimulated with CNP, ***p = 0.0002 vs. stimulated with Musclin, and ****p < 0.0001 vs. treated with both CNP and Musclin, ***p = 0.0002 for siNPR3 cells, vehicle treated vs. stimulated with CNP, ****p < 0.0001 vs. treated with Musclin and vs. cells treated with both CNP and Musclin (e). Assessment of cardiac fibroblast migration by scratch assay in cardiac fibroblasts (n = 6/condition) treated with siRNA, CNP and Musclin as indicated (f–h), *p = 0.036 for siControl cells, vehicle treated vs. treated with CNP, ***p < 0.0001 vs. Musclin and vs. stimulated with CNP and Musclin, *p = 0.046 for siNPR1 cells, vehicle treated vs. stimulated with CNP, **p = 0.0013 vs. Musclin and ****p < 0.0001 vs. cells treated with both CNP and Musclin (f); *p = 0.0148 for siControl cells, vehicle treated vs. treated with CNP, **p = 0.0013, and *p = 0.0148 between siControl and siNPR2 cells treated with both CNP and Musclin (g); *p = 0.0162 for siControl cells, vehicle treated vs. cells stimulated with CNP, **p = 0.0042 vs. cells treated with Musclin and ***p = 0.0004 vs. cells stimulated with both CNP and Musclin, *p = 0.0162 for siNPR3 cells, vehicle treated vs. stimulated with CNP or Musclin and ***p = 0.0002 vs. stimulated with Musclin and CNP, ***p = 0.0005 for siControl cells treated with Musclin und CNP vs. siNPR3 cells stimulated with Musclin and CNP, ***p = 0.0008 for vehicle-treated siControl cells vs. vehicle-treated siNPR3 cells (h). i, j Assessment of cardiac fibroblast proliferation and migration after stimulation with Musclin, CNP or the PKG inhibitor DT3 as indicated (for proliferation n = 7/condition, **p = 0.0018 for vehicle vs. Musclin treated cells and *p = 0.016 for cells stimulated with Musclin vs. cells treated with Musclin and DT3 (i), and for migration n = 4/condition, **p = 0.0043 for vehicle-treated cells vs. stimulated with Musclin, ***p = 0.0005 for cells treated with Musclin vs. treated with DT3 (j)). k Immunoblot from cardiac fibroblasts’ protein lysate for the indicated proteins after stimulation as indicated. GAPDH was used as loading control. The size of the proteins is indicated in kDa. Data in bar graphs are shown as mean ± standard error of the mean (SEM). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 determined by one-way ANOVA followed by the Holms–Sidak’s multiple comparisons test. Mu stands for Musclin. Source data are provided as a source data file.