Targeting necroptosis in muscle fibers ameliorates inflammatory myopathies

Abstract

Muscle cell death in polymyositis is induced by CD8⁺ cytotoxic T lymphocytes. We hypothesized that the injured muscle fibers release pro-inflammatory molecules, which would further accelerate CD8⁺ cytotoxic T lymphocytes-induced muscle injury, and inhibition of the cell death of muscle fibers could be a novel therapeutic strategy to suppress both muscle injury and inflammation in polymyositis. Here, we show that the pattern of cell death of muscle fibers in polymyositis is FAS ligand-dependent necroptosis, while that of satellite cells and myoblasts is perforin 1/granzyme B-dependent apoptosis, using human muscle biopsy specimens of polymyositis patients and models of polymyositis in vitro and in vivo. Inhibition of necroptosis suppresses not only CD8⁺ cytotoxic T lymphocytes-induced cell death of myotubes but also the release of inflammatory molecules including HMGB1. Treatment with a necroptosis inhibitor or anti-HMGB1 antibodies ameliorates myositis-induced muscle weakness as well as muscle cell death and inflammation in the muscles. Thus, targeting necroptosis in muscle cells is a promising strategy for treating polymyositis providing an alternative to current therapies directed at leukocytes.

Fig. 1. Expression of necroptosis-associated proteins in dying muscle fibers in PM.

a–c Representative images of muscle specimens of PM patients (n = 9). Scale bar indicates 20 μm. a Immunofluorescence staining against PAX7 (green) and the TUNEL staining (red). Nuclei were counterstained with DAPI (blue). Arrowheads indicate TUNEL positive PAX7 positive satellite cells. b Hematoxylin & Eosin (HE) and immunofluorescence staining against RIPK1, RIPK3, MLKL, phosphorylated MLKL at S358 (phospho-S358 MLKL), CASP8 (57 kDa), FAS, and CFLAR (green). The arrows indicate the dying muscle fibers, which showed reduced eosin staining in the cytoplasm. c Immunofluorescence staining against PAX7 (green) and active 18 kDa CASP8 subunit (red). Nuclei were counterstained with DAPI (blue). Arrowheads indicate active 18 kDa CASP8 subunit positive PAX7 positive satellite cells.

Fig. 2. Involvement of PRF1, GZMB, and FAS/FASLG in CTL-induced cell death of myoblasts and myotubes.

a The viability of H2K^bOVA-myoblasts co-cultured with OT-I CTLs (n = 80), Prf1^−/−OT-I CTLs (n = 59), or Gzmb^−/− OT-I CTLs (n = 61). The viability was visualized by PI staining and assessed with time-lapse imaging. Data are presented as mean and SD. Two-way analysis of variance (ANOVA) test, followed by Dunnett’s test. *p < 0.05, **p < 0.01. b The viability of H2K^bOVA-myotubes co-cultured with OT-I CTLs (n = 30) or Prf1^−/− OT-I CTLs (n = 26). c The viability of H2K^bOVA-myotubes co-cultured with OT-I CTLs (n = 42) or Gzmb^−/− OT-I CTLs (n = 40). d The viability of H2K^bOVA-myoblasts co-cultured with OT-I CTLs in the presence of FAS-Fc chimera protein (1 μM: n = 42, 10 μM: n = 50) or the control Fc protein (n = 48). Data are presented as mean and SD. e The viability of H2K^bOVA-myotube co-cultured with OT-I CTLs in the presence of FAS-Fc chimera protein (0.1 μM: n = 57, 1 μM: n = 50, 10 μM: n = 44) or the control Fc protein (0.1 μM: n = 36, 1 μM: n = 33, 10 μM: n = 34). Representative data of three independent experiments are shown. Log-rank test, followed by Holm–Sidak multiple comparisons. **p < 0.01.

Fig. 3. Different pattern of cell death between myoblasts and myotubes in the co-culture with CTLs.

a, b The representative confocal z-stack images of time-lapse analysis of H2K^bOVA-myoblasts (a) or H2K^bOVA-myotubes (b) co-cultured with OT-I CTLs in the presence of Annexin V (green) and PI (red). Nuclei of the cells were stained with Hoechst 33342 (blue). Images were taken at the indicated times from the starting of the co-culture. The cells surrounded by dotted line indicate dying myoblast and myotube. Scale bars indicate 20 μm. Representative data of three independent experiments are shown. c, d The TUNEL staining (red) of H2K^bOVA-myoblasts (c) or H2K^bOVA-myotubes (d) co-cultured with OT-I CTLs for 4 or 16 h, respectively. The myoblasts and myotubes were pre-labelled with CellTracker (green). Confocal z-stack images are shown. Scale bars indicate 20 μm. Representative data of two independent experiments are shown. e The immunofluorescence staining of FAS, CFLAR, RIPK1, RIPK3, MLKL, and phosphorylated MLKL at S345 (phospho-S345 MLKL; green) in H2K^bOVA-myotubes co-cultured with OT-I CTLs for 8 h. Nuclei were counterstained with DAPI (blue). Scale bar indicates 10 μm. Representative data of three independent experiments are shown.

Fig. 4. Effect of apoptosis or necroptosis inhibition on CTL-induced cell death in myotubes and myoblasts.

a, b The viability of z-VAD-fmk- or Nec1s-pretreated H2K^bOVA-myotubes in the co-culture with OT-I CTLs. z-VAD-fmk 0 μM: n = 30, 20 μM: n = 30, 100 μM: n = 30. Nec1s 0 μM: n = 56, 10 μM: n = 55, 50 μM: n = 46, 100 μM: n = 50. Log-rank test, followed by Holm–Sidak multiple comparisons. **p < 0.01. c, d The viability of z-VAD-fmk- or Nec1s-pretreated H2K^bOVA-myoblasts in the co-culture with OT-I CTLs. z-VAD-fmk 0 μM: n = 50, 20 μM: n = 48, 100 μM: n = 59. Nec1s 0 μM: n = 33, 100 μM: n = 30. Data are presented as mean and SD. Two-way ANOVA test, followed by Dunnett’s test. *p < 0.05, **p < 0.01. e The viability of H2K^bOVA-myotubes transfected with scrambled siRNA (siControl: n = 66), siRNA specific for Ripk3 (siRIPK3 #1: n = 67, siRIPK3 #2: n = 63), or without transfection (siRNA (−): n = 61) in the co-culture with OT-I CTLs. Log-rank test, followed by Holm–Sidak multiple comparisons. **p < 0.01. a–e The data represent three independent experiments. f The protein expression of RIPK3, MYOG, FAS, and GAPDH evaluated with western blotting in H2K^bOVA-myotubes transfected with siControl or siRIPK3. The data represent two independent experiments. g–i The levels of HMGB1 (g), IL-1α (h), and IL-6 (i) in the co-culture of Nec1s-pretreated or untreated H2K^bOVA-myotubes and OT-I CTLs for 20 h. IL-1β was below detectable level in the culture supernatants of all conditions. The levels in the mono-culture of OT-I CTLs are subtracted from those in the co-culture. Data are presented as mean and SD of triplicate (g) or quadruplicate (h, i) experiments. Student’s t-test. *p < 0.05, **p < 0.01.

Fig. 5. The involvement of necroptosis in the muscle inflammation and muscle weakness in CIM.

a HE, immunofluorescence staining against PAX7 (green), and the TUNEL staining (red). Nuclei were counterstained with DAPI (blue). The arrowhead indicates PAX7 positive TUNEL positive satellite cell. b HE and immunofluorescence staining (green) for the expression of RIPK1, RIPK3, MLKL, phosphorylated-S345 (phospho-S345) MLKL, CFLAR, and FAS in the muscle specimens of CIM mice. a, b The arrows indicate the dying muscle fibers. Scale bars indicate 20 μm. Each histological examination was performed three times and the representative images are shown. c, d The histological scores (c) of the severity of myositis and the area of necrotic muscle fibers (d) in wild-type (WT, n = 13) or Ripk3^−/− (n = 12) mice on day 14 of CIM. e, f The histological scores (e) and the area of necrotic muscle fibers (f) in WT (n = 12) or MLKL^−/− (n = 13) mice on day 14 of CIM. c–f Data are presented as median ± interquartile range. Mann–Whitney U test. **p < 0.01. g The grip strength of CIM mice treated with Nec1s or the vehicle (n = 12). The treatment was started immediately after the immunization in the prophylactic group (Prophylactic Tx, n = 13), or from day 7 in the therapeutic group (Therapeutic Tx, n = 14). BW; Body weight of the mice. Data are presented as mean ± SD. Two-way ANOVA test, followed by Dunnett’s multiple comparison test. **p < 0.01. h, i The histological scores (h) and the area of necrotic muscle fibers (i) on day 21 of CIM. Vehicle; n = 12, Therapeutic Tx; n = 14, Prophylactic Tx; n = 13. Data are presented as median ± interquartile range. Kruskal–Wallis test, followed by Dunn’s test. **p < 0.01. j Proportion of TUNEL⁺ cells out of PAX7⁺ cells in the muscles of CIM on day 21. Vehicle; n = 3, Therapeutic Tx; n = 3, Prophylactic Tx; n = 3. Data are presented as mean ± SD. One-way ANOVA test, followed by Dunnett’s test. *p < 0.05. c–j Data represent two independent experiments.

Fig. 6. The involvement of HMGB1 in the muscle inflammation and muscle weakness in CIM.

a The serum levels of HMGB1 in CIM mice on day 21. The data are presented as mean ± SD. One-way ANOVA test, followed by Dunnett’s test. **p < 0.01. b HE and immunofluorescence staining of the muscle specimens of CIM for the expression of HMGB1. The arrows indicate the dying muscle fibers. Scale bars indicate 20 μm. c The histological scores on day 14 in CIM mice treated with 1.0 mg/kg (n = 10), 3.0 mg/kg (n = 9), or 10 mg/kg (n = 9) of anti-HMGB1 antibodies (anti-HMGB1) or 10 mg/kg of the control antibodies (Control: n = 10). The data are presented as median ± interquartile. Kruskal–Wallis test, followed by Dunn’s test. *p < 0.05. d The areas of necrotic muscle fibers on day 14 of CIM. Data are presented as median ± interquartile range. Kruskal–Wallis test, followed by Dunn’s test. *p < 0.05. e The grip strength of CIM mice treated with anti-HMGB1 antibodies or the control antibodies. Data are presented as mean ± SD. Two-way ANOVA test, followed by Dunnett’s test. **p < 0.01. c–e Data represent two independent experiments.