OPA1 deficiency promotes secretion of FGF21 from muscle that prevents obesity and insulin resistance

Abstract

Mitochondrial dynamics is a conserved process by which mitochondria undergo repeated cycles of fusion and fission, leading to exchange of mitochondrial genetic content, ions, metabolites, and proteins. Here, we examine the role of the mitochondrial fusion protein optic atrophy 1 (OPA1) in differentiated skeletal muscle by reducing OPA1 gene expression in an inducible manner. OPA1 deficiency in young mice results in non-lethal progressive mitochondrial dysfunction and loss of muscle mass. Mutant mice are resistant to age- and diet-induced weight gain and insulin resistance, by mechanisms that involve activation of ER stress and secretion of fibroblast growth factor 21 (FGF21) from skeletal muscle, resulting in increased metabolic rates and improved whole-body insulin sensitivity. OPA1-elicited mitochondrial dysfunction activates an integrated stress response that locally induces muscle atrophy, but via secretion of FGF21 acts distally to modulate whole-body metabolism.

Keywords: ER stress; FGF21; OPA1; mitochondrial dysfunction; skeletal muscle.

Figure 1. OPA1 deficiency disrupts mitochondrial cristae structure and progressively impairs mitochondrial respiratory capacity

1. Representative immunoblot of OPA1 in gastrocnemius muscle from 12‐week‐old WT and mOPA1 KO (KO) mice and densitometric analysis of OPA1 normalized by VDAC (n = 6).

2. Electron micrographs from soleus muscle of 12‐week‐old mice (arrows highlight mitochondria with abnormal cristae structure) (n = 4). Scale bar = 0.32 μm.

3. BN‐PAGE in gastrocnemius muscle of 12‐week‐old mice (n = 4). I‐V, mitochondrial OXPHOS complexes I‐V; SC1, super complex 1; SC2 super complex 2.

4. Citrate synthase activity in gastrocnemius muscle of 12‐week‐old mice (n = 6).

5. mtDNA copy number in gastrocnemius muscle of 12‐week‐old mice (n = 6).

6. mRNA expression of mitochondrial biogenesis genes (n = 4–6).

7. Maximally stimulated succinate‐supported mitochondrial respirations in soleus from 12‐ and 20‐week‐old mice (n = 4–7).

8. Succinate‐supported ATP synthesis rates in soleus from 12‐ and 20‐week‐old mice (n = 3–6).

9. Maximally stimulated mitochondrial respirations in white gastrocnemius muscle from 20‐week‐old mice (n = 3–4).

10. Maximally stimulated mitochondrial respirations in red gastrocnemius muscle from 20‐week‐old mice (n = 3–4).

Data information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test, using a significance level of P < 0.05. (*) Significantly different vs. WT mice. Source data are available online for this figure.

Figure EV1. Mitochondrial dynamics proteins and mitochondrial oxygen consumption

1. A, B

   Representative immunoblots and densitometric quantification in muscle homogenates from WT and mOPA1 KO mice. (A) OPA1 levels in soleus muscle normalized to GAPDH (n = 3). (B) Mfn1 and Mfn2 normalized to GAPDH (n = 4–5).

2. C

   Maximally stimulated succinate‐supported mitochondrial respirations in soleus from 40‐week‐old WT and mOPA1 KO (KO) mice (n = 5).

3. D

   Succinate‐supported ATP synthesis rates in soleus from 40‐week‐old WT and mOPA1 KO (KO) mice (n = 5).

tPData information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test, using a significance level of P < 0.05. (*) Significantly different vs. WT mice. Source data are available online for this figure.

Figure 2. Reduced OPA1 levels in muscle attenuates age‐induced weight gain and glucose intolerance and induces muscle atrophy

1. A

   Body weight over time (n = 8–13).

2. B

   Total fat mass over time (n = 8–13).

3. C

   Total lean mass over time (n = 8–13).

4. D

   Cross sections of gastrocnemius muscle from 20‐week‐old mice stained with wheat‐germ agglutinin (WGA). Scale bar = 20 μm (main images) and 40 μm (magnification inserts).

5. E

   Measurements of muscle fiber diameter in gastrocnemius muscle from 20‐week‐old mice (n = 3–5).

6. F

   Measurements of the number of myocyte/area (n = 3–5).

7. G, H

   (G) Western blot analysis and (H) densitometric quantification of MuRF1 and GAPDH in gastrocnemius muscle from 40‐week‐old mice. Densitometric quantification is relative to GAPDH (n = 5).

8. I

   mRNA expression of Atrogin‐1 in gastrocnemius muscle of 20‐week‐old mice (n = 5). Data are represented as fold change vs. WT mice.

9. J–L

   CLAMS data in 8‐ and 40‐week‐old WT and mOPA1 KO (KO) mice. (J) Daily food intake (n = 5–9). (K) Activity levels (n = 6–9). (L) Energy expenditure (n = 6–9).

10. M

    Glucose tolerance tests (n = 7–11).

11. N

    Area under the curve (AUC) for glucose tolerance tests (n = 7–11).

Data information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test, using a significance level of P < 0.05. (*) Significantly different vs. WT mice or 8‐week‐old mice from the same genotype. Source data are available online for this figure.

Figure EV2. Assessment of overall muscle health of mOPA1 KO mice

1. Cross sections of soleus muscle from 20‐week‐old mice stained with wheat‐germ agglutinin (WGA). Scale bar = 50 μm.

2. Measurements of muscle fiber diameter in soleus muscle from 20‐week‐old mice (n = 4).

3. Exhaustion test in 20‐ and 40‐week‐old WT and mOPA1 KO mice (n = 4–9).

4. Blood lactate after the exhaustion test (n = 4–9).

5. Measurements of vertical work performed by 20‐ and 40‐week‐old mice during the exhaustion test (n = 4–9).

6. Measurements of grip strength in 40‐week‐old mice (n = 5).

tPData information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test, using a significance level of P < 0.05. (*) Significantly different vs. WT mice. Source data are available online for this figure.

Figure 3. Muscle‐specific OPA1 deficiency prevents diet‐induced obesity and insulin resistance and increases circulating FGF21

1. A

   Weekly body weight during 12 weeks of control or HFD feeding (n = 5–9).

2. B

   Total fat mass after 12 weeks on control or HFD (n = 9–11).

3. C

   Total lean mass after 12 weeks on control or HFD (n = 9–11).

4. D–F

   CLAMS data in high‐fat‐fed WT and mOPA1 KO (KO) mice at the end of 12 weeks of feeding. (D) Daily food intake (n = 5–7). (E) Activity level (n = 5–7). (F) Energy expenditure (n = 6–7).

5. G

   Glucose tolerance tests in control and high‐fat‐fed WT and mOPA1 KO (KO) mice at the end of 12 weeks of feeding (n = 5–9).

6. H

   Area under the curve (AUC) for glucose tolerance tests (n = 5–9).

7. I

   Fasting insulin levels (n = 5–9).

8. J

   Insulin tolerance test in control and high‐fat‐fed WT and KO mice at the end of 12 weeks of feeding (n = 5–8).

9. K

   Area under the curve (AUC) for insulin tolerance tests (n = 5–8).

10. L–N

    FGF21 mRNA expression in 20‐week‐old mice at the end of 12 weeks of feeding. (L) Skeletal muscle (n = 3–5). (M) Liver (n = 4–5). (N) White adipose tissue (WAT) (n = 3–7).

11. O

    FGF21 plasma levels in control and high‐fat‐fed WT and KO mice at the end of 12 weeks of feeding (n = 3–6).

Data information: Data are expressed as means ± SEM. Significant differences were determined by ANOVA followed by Tukey multiple comparison test, using a significance level of P < 0.05 (*) vs. WT Cont (#) vs. WT HFD except panel (F) where significance was determined by Student's t‐test; (*) significantly different vs. WT HFD. Source data are available online for this figure.

Figure EV3. Metabolic parameters in mOPA1 KO mice

1. Triglycerides levels in mice fed either a control or a HFD for 12 weeks (n = 5–6).

2. Fasting triglyceride levels in the serum of mice fed either a control or a HFD for 12 weeks (n = 5–8).

3. IL‐15 and IL‐6 mRNA expression in gastrocnemius muscle of 20‐week‐old WT and mOPA1 KO mice (n = 4–5).

4. Representative immunoblot and densitometric quantification of OPA1 and FGF21 normalized to GAPDH in soleus muscle of 20‐week‐old mice (n = 3).

5. Representative immunoblot and densitometric quantification of PGC‐1α and UCP1 normalized to GAPDH in BAT of 20‐week‐old mice (n = 4).

6. Representative immunoblot and densitometric quantification of PGC‐1α and UCP1 normalized to GAPDH in sc‐WAT of 20‐week‐old mice (n = 3–4). Data are represented as fold change vs. WT mice.

tPPData information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test (D, F), using a significance level of P < 0.05 or by ANOVA followed by Tukey multiple comparison test (A, B), using a significance level of P < 0.05. (*) Significantly different vs. WT mice. (*) vs. WT Cont (#) vs. WT HFD. Source data are available online for this figure.

Figure 4. Reducing skeletal muscle OPA1 in mice with diet‐induced obesity and insulin resistance promotes weight loss and normalizes glucose tolerance in parallel with increased FGF21

1. Weekly body weight before and after tamoxifen injections (total of 20 weeks of control or HFD) (n = 6).

2. Final body weight after 20 weeks of control or HFD (n = 6).

3. Glucose tolerance tests in control and high‐fat‐fed WT and mOPA1 KO (KO) before tamoxifen injection (8 weeks of control or HFD) (n = 6–9).

4. Area under the curve (AUC) for glucose tolerance tests (n = 6–9).

5. Glucose tolerance tests in control and high‐fat‐fed WT and mOPA1 KO (KO) 12 weeks after tamoxifen injection (20 weeks total of control or HFD) (n = 5–8).

6. Area under the curve (AUC) for glucose tolerance tests (n = 5–8).

7. FGF21 plasma levels after 20 weeks of control or HFD (12 weeks after tamoxifen injections) (n = 6–8).

Data information: Data are expressed as means ± SEM. Significant differences were determined by ANOVA followed by Tukey multiple comparison test, using a significance level of P < 0.05. (*) vs. WT Cont (#) vs. WT HFD. Source data are available online for this figure.

Figure 5. Muscle‐derived FGF21 mediates the metabolic phenotype observed in OPA1‐deficient mice

1. A

   Representative immunoblot of FGF21 in gastrocnemius muscle from 12‐week‐old mice and densitometric analysis of FGF21 normalized by GAPDH (n = 4).

2. B

   FGF21 serum levels in 12‐week‐old mOPA1 KO (KO), DKO mice, and their respective WT controls (WT) (n = 5).

3. C–N

   Metabolic data in DKO mice and their respective WT controls after 10 weeks of control or HFD. (C) Final body weight (n = 5–9). (D) Total fat mass (n = 5–9). (E) Total lean mass (n = 5–9). (F) Daily food intake (n = 4). (G) Activity level (n = 4). (H) Energy expenditure (n = 4). (I) Glucose tolerance tests (n = 9–16). (J) Area under the curve (AUC) for the GTT (n = 9–16). (K) Fasting insulin levels (n = 5–11). (L) Insulin tolerance test (n = 7–10). (M) Area under the curve (AUC) for the ITT (n = 7–10). (N) FGF21 plasma levels (n = 5–10).

Data information: Data are expressed as means ± SEM. Significant differences were determined by ANOVA followed by Tukey multiple comparison test, using a significance level of P < 0.05 (*) vs. WT or WT Cont (#) vs. OPA1 KO. Source data are available online for this figure.

Figure EV4. Assessment of overall mitochondrial function and muscle health in mOPA1/FGF21 DKO mice

1. Representative immunoblots and densitometric quantification of OPA1 normalized to GAPDH in gastrocnemius muscle from 12‐week‐old WT and mOPA1/FGF21 KO mice (n = 3).

2. BN‐PAGE in gastrocnemius muscle of 12‐week‐old mice (n = 4).

3. mtDNA copy number in gastrocnemius muscle of 12‐week‐old mice (n = 3–4).

4. Maximally stimulated succinate‐supported mitochondrial respirations in soleus from 20‐week‐old mice (n = 4–6).

5. Succinate‐supported ATP synthesis rates in soleus from 20‐week‐old mice (n = 4–5).

6. Maximally stimulated succinate‐supported mitochondrial respirations in soleus from 40‐week‐old mice (n = 5).

7. Measurements of grip strength in 40‐week‐old mice (n = 3).

8. Body weight and body composition in 20‐week‐old WT and DKO mice (n = 5–9).

9. Cross sections of gastrocnemius muscle from 20‐week‐old mice stained with wheat‐germ agglutinin (WGA). Scale bar = 50 μm.

10. Measurements of muscle fiber diameter in gastrocnemius muscle from 20‐week‐old mice (n = 3).

tPData information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test, using a significance level of P < 0.05. (*) Significantly different vs. WT mice. Source data are available online for this figure.

Figure 6. ER stress activation is necessary for FGF21 induction in OPA1‐deficient muscle

1. Representative immunoblots in gastrocnemius muscle from 12‐week‐old WT or mOPA1 KO (KO) mice.

2. Densitometric analysis of immunoblots of 4‐hydroxynonenal (4‐HNE), nitrotyrosine, MnSOD (n = 5), and BiP (n = 3) normalized to GAPDH protein levels (n = 5) and of pAMPK/AMPK (n = 5) and peIF2α/eIF2α ratios (n = 5).

3. mRNA expression of ER stress genes in gastrocnemius muscle from 20‐week‐old WT or mOPA1 KO (KO) mice (data are expressed as fold change vs. WT mice, represented as the dashed line) (n = 3–5).

4. AKT phosphorylation at Ser 473 and S6 phosphorylation at Ser235/236, 30 min after an i.p. injection of either saline or insulin, normalized by total AKT and total S6 levels, respectively.

5. Densitometric analysis of pAKT/AKT and pS6/S6 immunoblots (n = 3).

6. Representative immunoblots in soleus muscle from 12‐week‐old WT or mOPA1 KO (KO) mice treated with TUDCA for 4 weeks.

7. Densitometric analysis of immunoblots of OPA1, FGF21, and BiP normalized to GAPDH protein levels (n = 4).

8. FGF21 serum levels in 12‐week‐old mOPA1 KO mice and their respective WT controls 4 weeks after TUDCA treatment (n = 3–4).

9. Body weight and body composition in 12‐week‐old mice 4 weeks after TUDCA treatment (n = 3–5).

10. Glucose tolerance test in 12‐week‐old mice 4 weeks after TUDCA treatment (n = 4–6).

Data information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test (B, C), using a significance level of P < 0.05; (*) significantly different vs. WT mice or by ANOVA followed by Tukey multiple comparison test (E, G, H), using a significance level of P < 0.05 (*) vs. WT, WT vehicle, or WT PBA; (#) vs. KO PBS. Source data are available online for this figure.

Figure 7. OPA1 deletion in primary myotubes confirms a role for OPA1‐induced mitochondrial dysfunction and ER stress on FGF21 secretion

1. Representative immunoblots in primary myotubes obtained from OPA1^fl/fl mice and infected with Ad‐GFP or Ad‐Cre. Cells were treated with either vehicle (PBS) or PBA for 3 days. Densitometric analysis of immunoblots of OPA1, CHOP, XBP‐1, and FGF21 normalized to GAPDH protein levels and of pAMPK/AMPK ratios (n = 6).

2. Secreted FGF21 levels measured in the media (n = 6).

3. Measurement of oxygen consumption rates (OCR) in primary myotubes 3 days after adenoviral infection (n = 8–10 technical replicates).

4. Measurements of extracellular acidification rates during a glycolysis stress test in primary myotubes 3 days after adenoviral infection (n = 8–10 technical replicates).

5. TMRM fluorescence over time after treatment with H₂O₂ (n = 3 technical replicates).

6. Representative pictures of primary myotubes stained with TMRM at times 0 (immediately after addition of H₂O₂) and 100 min after H₂O₂ treatment. Scale bar = 100 μm.

Data information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test (C–E), using a significance level of P < 0.05; (*) significantly different vs. WT cells or by ANOVA followed by Tukey multiple comparison test (A, B), using a significance level of P < 0.05 for the studies in which cells were treated with PBA. (*) vs. WT; (#) vs. KO. Source data are available online for this figure.

Figure EV5. Mitochondrial stress and ER stress in C2C12 myotubes

1. FGF21 protein levels measured by immunoblot, in cell lysates from C2C12 myotubes treated either with vehicle, oligomycin, or tunicamycin ± PBA for 8 h. Data are represented as fold change vs. vehicle normalized to GAPDH (n = 4–8).

2. FGF21 in the media secreted from C2C12 myotubes treated either with vehicle, oligomycin, or tunicamycin ± PBA for 8 h (n = 3).

3. mRNA expression of Fgf21, Atf4, and BiP in C2C12 myotubes treated with vehicle or oligomycin for 8 h (data are expressed as fold change vs. vehicle) (n = 4–6).

4. Western blot analysis and densitometric quantification of phosphorylated AMPK normalized to total AMPK in C2C12 cells treated with vehicle or AICAR for 12 h (n = 4).

5. FGF21 mRNA expression in C2C12 cells treated with vehicle or AICAR for 12 h (n = 3–4). Data are represented as fold change vs. vehicle.

tPPData information: Data are expressed as means ± SEM. Significant differences were determined by Student's t‐test (C, D), using a significance level of P < 0.05 or by ANOVA followed by Tukey multiple comparison test (A, B), using a significance level of P < 0.05. (*) Significantly different vs. vehicle. (*) and (#) vs. oligomycin or tunicamycin. Source data are available online for this figure.