Coupling of mitochondrial function and skeletal muscle fiber type by a miR-499/Fnip1/AMPK circuit

Abstract

Upon adaption of skeletal muscle to physiological and pathophysiological stimuli, muscle fiber type and mitochondrial function are coordinately regulated. Recent studies have identified pathways involved in control of contractile proteins of oxidative-type fibers. However, the mechanism for coupling of mitochondrial function to the muscle contractile machinery during fiber type transition remains unknown. Here, we show that the expression of the genes encoding type I myosins, Myh7/Myh7b and their intronic miR-208b/miR-499, parallels mitochondrial function during fiber type transitions. Using in vivo approaches in mice, we found that miR-499 drives a PGC-1α-dependent mitochondrial oxidative metabolism program to match shifts in slow-twitch muscle fiber composition. Mechanistically, miR-499 directly targets Fnip1, an AMP-activated protein kinase (AMPK)-interacting protein that negatively regulates AMPK, a known activator of PGC-1α. Inhibition of Fnip1 reactivated AMPK/PGC-1α signaling and mitochondrial function in myocytes. Restoration of the expression of miR-499 in the mdx mouse model of Duchenne muscular dystrophy (DMD) reduced the severity of DMD Thus, we have identified a miR-499/Fnip1/AMPK circuit that can serve as a mechanism to couple muscle fiber type and mitochondrial function.

Keywords: contractile fiber type; gene regulation; microRNA; mitochondrial function; muscle.

Figure 1. miR‐499 is dynamically regulated during muscle fiber type transition and affects mitochondrial function

1. RT‐qPCR analysis of Myh7b/miR‐499 and Myh7/miR‐208b levels during differentiation of myoblasts into mature myotubes (n = 3 independent experiments). *P = 0.0005 (Myh7b), *P < 0.0001 (miR‐499), *P = 0.0002 (Myh7), *P < 0.0001 (miR‐208b).

2. Oxygen consumption rates (OCR) in primary myoblasts and differentiated myotubes. Basal OCR was first measured, followed by administration of 10 mM sodium pyruvate, and 2 μM oligomycin (to inhibit ATP synthase), uncoupler FCCP (2 μM), or rotenone/antimycin (Rot/A; 1 μM) as indicted. n = 3 separate experiments done with 10 biological replicates. *P < 0.0001 (FCCP).

3. Mean expression levels (RT‐qPCR) in white vastus (WV) and soleus (Sol) muscle from 8‐week‐old male wild‐type mice (n = 5 mice per group). *P < 0.0001 (Myh7b), *P < 0.0001 (miR‐499), *P < 0.0001 (Myh7), *P < 0.0001 (miR‐208b).

4. Mitochondrial respiration rates were determined from mitochondria isolated from indicated muscle using pyruvate as substrate. ADP‐dependent respiration, oligomycin‐induced (oligo), uncoupler FCCP, and antimycin A (AA) are shown. n = 3 separate experiments done with 7–8 biological replicates. *P < 0.05 (ADP).

5. (Left) Schematic depicts the increments of speed over time. (Right) Respiratory exchange ratio (RER) during a graded exercise regimen as described in Materials and Methods (n = 5 mice per group). Notably, MCK‐miR‐499 mice were able to exercise at a higher speed before exhaustion. *P < 0.05.

6. Bars represent mean blood lactate levels for 3‐ and 15‐month‐old male MCK‐miR‐499 and NTG mice following a 25‐min run on a motorized treadmill. For 3‐month blood lactate, NTG, n = 9; MCK‐miR‐499, n = 10. For 15‐month blood lactate, NTG, n = 7; MCK‐miR‐499, n = 5. *P = 0.039 (3 months), *P = 0.018 (15 months).

7. Mitochondrial respiration rates were determined from the plantaris muscle of the indicated genotypes using pyruvate/malate as substrate. Pyruvate/malate (Py/M)‐stimulated, ADP‐dependent respiration, oligomycin‐induced (oligo), and the respiratory control ratio (RC) are shown. NTG, n = 6; MCK‐miR‐499, n = 4. *P = 0.015 (Py/M), *P = 0.017 (ADP), *P = 0.001 (Oligo).

Data information: All values represent the mean ± SEM. P‐value was determined using two‐tailed unpaired Student's t‐test.

Figure 2. MCK‐miR‐499 muscle is reprogrammed for increased capacity for mitochondrial oxidation

1. Gene ontology (GO) enrichment analysis of gene transcripts upregulated in MCK‐miR‐499 muscle identified a number of terms related to mitochondrial function and muscle contraction. Gene expression array data generated from the gastrocnemius muscle of 8‐week‐old male MCK‐miR‐499 mice were compared to littermate controls (NTG).

2. Expression of the nuclear‐encoded and mitochondrial‐encoded genes (RT‐qPCR) in the gastrocnemius muscle from the indicated genotypes (n = 5 mice per group). *P < 0.05.

3. Representative Western blot analysis performed with gastrocnemius muscle total protein extracts prepared from the indicated mice using myoglobin, cytochrome c, and β‐actin (control) antibodies (n = 8 mice per group).

4. Expression of genes (RT‐qPCR) involved in muscle glucose and fatty acid metabolism in the gastrocnemius muscle from the indicated genotypes (n = 5 mice per group). *P < 0.05.

5. LDH isoenzymes were separated by polyacrylamide gel electrophoresis using whole cell extracts from NTG heart (Ht, control) and gastrocnemius muscle from the indicated mice. A representative gel showing 3–4 independent mice per group is shown.

6. Quantification of LDH isoenzyme activity gel electrophoresis in (E). Values represent the mean % (± SEM) total LDH activity. *P < 0.05.

7. Cross section of the gastrocnemius muscle from 3‐month‐old male NTG and MCK‐miR‐499 mice stained for myosin I ATPase activity (top), SDH (middle), and α‐GPDH (bottom). Representative images are shown. Scale bar: 500 μm.

Data information: All values represent the mean ± SEM. P‐value was determined using two‐tailed unpaired Student's t‐test.Source data are available online for this figure.

Figure EV1. Increased PGC‐1α expression in MCK‐miR‐499 muscle

1. Upstream regulator analysis by Ingenuity Pathways Analysis (IPA) based on the gene expression array data generated from the gastrocnemius muscle of MCK‐miR‐499. The top five upstream regulators are shown with the predicted z‐scores.

2. Expression of genes encoding PGC‐1, ERR, and PPAR transcription factors (RT‐qPCR) in gastrocnemius muscle compared to NTG controls (n = 5 mice per group). *P < 0.05.

3. Representative Western blot analysis of PGC‐1α (Top) and myoglobin (Bottom) in white vastus (WV), gastrocnemius (GC), and soleus (Sol) muscle from the indicated genotypes (n = 4 mice per group).

4. Results of qPCR to determine mitochondrial DNA levels in WV muscle of the indicated genotypes using primers for NADH dehydrogenase (Nd1, mitochondria‐encoded) and lipoprotein lipase (Lpl, nuclear‐encoded). Nd1 levels were normalized to Lpl DNA content and expressed relative to NTG (= 1.0) muscle (n = 5 mice per group). P = 0.569 (NS, not significant).

Data information: All values represent the mean ± SEM. P‐value was determined using two‐tailed unpaired Student's t‐test.Source data are available online for this figure.

Figure 3. miR‐499 regulates muscle mitochondrial oxidative capacity through PGC‐1α

1. (Left) Representative Western blot analysis of PGC‐1α protein expression in the gastrocnemius muscle from the indicated genotypes, heart (Ht) lysate from WT mice and PGC‐1α mKO (MCK‐Cre) mice as positive and negative controls, respectively. (Right) Quantification of the PGC‐1α/tubulin signal ratios normalized (= 1.0) to the NTG control (n = 5 mice per group). *P = 0.00017.

2. Expression of the Ldhb, Ldha, Mb (myoglobin), and Cox5a genes (RT‐qPCR) in the gastrocnemius muscle from the indicated genotypes (n = 5 mice per group). *P < 0.001 (versus NTG), ^‡ P < 0.01 (versus 499Tg).

3. (Left) Representative Western blot analysis performed with gastrocnemius muscle total protein extracts prepared from the indicated mice using cytochrome c, myoglobin, and α‐tubulin (control) antibodies. (Right) Quantification of the Cyt c/tubulin signal ratios normalized (= 1.0) to the NTG control. NTG, n = 5; 499Tg, n = 6; PGC‐1α mKO, n = 5; 499Tg/PGC‐1α mKO, n = 5. *P < 0.01 (versus NTG), ^‡ P = 0.00018 (versus 499Tg).

4. Cross section of gastrocnemius muscle from the indicated mice stained for SDH activity (top) as well as MHC1 immunofluorescence (IF; bottom); representative images are shown. Scale bar: 500 μm.

5. Respiratory exchange ratio (RER) during a graded exercise regimen as described in Materials and Methods. Notably, no significant increase in exhaust speed was observed in 499Tg/PGC‐1α mKO mice compared to PGC‐1α mKO mice. PGC‐1α mKO, n = 5; 499Tg/PGC‐1α mKO, n = 8.

6. The bars represent the mean blood lactate levels from the indicated mice following a 25‐min run on a motorized treadmill. NTG, n = 9; 499Tg, n = 7; PGC‐1α mKO, n = 6; 499Tg/PGC‐1α mKO, n = 9. *P = 0.0009 (NTG versus 499Tg), P = 0.316 (NS, not significant).

Data information: All values represent the mean ± SEM. P‐value in (A and F) was determined using two‐tailed unpaired Student's t‐test; P‐value in (B and C) was determined using one‐way ANOVA coupled to a Fisher's least‐significant difference (LSD) post hoc test.Source data are available online for this figure.

Figure EV2. PGC‐1α‐independent regulation of the slow‐twitch muscle fiber program by miR‐499

1. Representative Western blot analysis of PGC‐1α protein expression in the gastrocnemius muscle from the indicated genotypes (n = 4 mice per group).

2. Morphology of dissected gastrocnemius/soleus (asterisk) and tibialis anterior (TA) muscle.

3. Expression of genes encoding slow‐twitch/fast‐twitch troponin and Sox6 (RT‐qPCR) in soleus (Sol), gastrocnemius (GC), and white vastus (WV) muscle from the indicated genotypes. For Sol, NTG, n = 10; 499Tg, n = 9; PGC‐1α mKO, n = 8; 499Tg/PGC‐1α mKO, n = 10. For GC and WV, n = 5 mice per group. *P < 0.05 (versus NTG), ^‡ P < 0.05 (versus 499Tg).

Data information: Values represent the mean (± SEM) and are shown as arbitrary units (AU) normalized (= 1.0) to the value of the NTG control. P‐value was determined using one‐way ANOVA coupled to a Fisher's least‐significant difference (LSD) post hoc test.Source data are available online for this figure.

Figure EV3. Identification of miR‐499 targets

1. Diagram shows the miR‐499 target identification. The TargetScan and MicroCosm programs were used to identify putative target mRNAs for miR‐499, and this list was cross‐matched for genes that were downregulated in MCK‐miR‐499 muscle (fold change < −1.2). The overlapping putative targets, together with those predicted targets that are known to be involved in the regulation of energy metabolism, were chosen for further UTR luciferase validation assay.

2. 3′UTR luciferase reporters containing the predicted binding site of miR‐499 were used in cotransfection studies in HEK293T cells in the presence or absence of plasmids expressing miR‐499 (n = 3 independent experiments). Sox6 3′UTR containing the binding site of miR‐499 was used as a positive control. *P < 0.01. All values represent the mean ± SEM and are shown as arbitrary units (AU) normalized to corresponding controls. P‐value was determined using two‐tailed unpaired Student's t‐test.

Figure 4. miR‐499 directly targets Fnip1

1. The schematic shows the putative conserved miR‐499 binding site within the 3′UTR of the Fnip1 gene.

2. Luciferase reporters containing the wild‐type Fnip1 3′UTR or Fnip1 3′UTR mutated in the predicted binding site of miR‐499 were used in cotransfection studies in HEK293T cells in the presence or absence of plasmids expressing miR‐499 (n = 3 independent experiments). *P < 0.0001 (Fnip1 3′UTR); *P = 0.023 (Mutated).

3. RT‐qPCR analysis of Fnip1 and Fnip2 mRNA levels in the gastrocnemius muscle of the indicated genotypes (n = 5 mice per group).

4. (Top) Fnip1 protein expression in the gastrocnemius muscle from the indicated mice. (Bottom) Quantification of the Fnip1/tubulin signal ratio (n = 4 mice per group). *P = 0.014.

Data information: All values represent the mean ± SEM. P‐value was determined using two‐tailed unpaired Student's t‐test.Source data are available online for this figure.

Figure 5. Inhibition of Fnip1 reactivated AMPK‐PGC‐1α signaling and mitochondrial function in myocytes

1. A

   (Left) Representative Western blot analysis performed on extracts of myotubes subjected to Fnip1 siRNA or control (Con) siRNA using phospho‐AMPKα (Thr172) and AMPKα antibodies. (Right) Quantification of the p‐AMPKα/AMPKα signal ratios (n = 3 independent experiments). *P = 0.037.

2. B, C

   Results of RT‐qPCR analysis on WT primary mouse myotubes after transfection with Fnip1 siRNAs or scrambled Con siRNA as indicated. For (C), 48 h post‐siRNA transfection, myotubes were treated for 24 h with DMSO or 10 μm compound C before harvest (n = 3 independent experiments). *P < 0.0001 (Fnip1 in B), *P = 0.0001 (PGC‐1α in B); *P < 0.0001 (versus Con siRNA in C), ^‡ P < 0.0001 (versus Fnip1 siRNA in C).

3. D

   (Top) Representative Western blot analysis performed on extracts of the gastrocnemius muscle isolated from NTG or MCK‐miR‐499 mice using phospho‐AMPKα (Thr172), AMPKα, phospho‐ACC (Ser79), or total ACC antibodies. (Bottom) Quantification of the p‐AMPKα/AMPKα and p‐ACC/ACC signal ratios normalized (= 1.0) to the NTG control (n = 8 mice per group). *P < 0.01.

4. E

   Oxygen consumption rates (OCR) in primary mouse myotubes transfected with Fnip1 siRNA or Con siRNA. n = 7 separate experiments done with 5 biological replicates. *P < 0.05 (Pyruvate), *P < 0.01 (FCCP).

5. F

   Results of RT‐qPCR analysis for WT primary mouse myotubes subjected to inhibition of both miR‐499 and miR‐208b (anti‐miRs) (n = 4 independent experiments). *P = 0.037 (Ppargc1a), *P = 0.0019 (Fnip1).

6. G

   Oxygen consumption rates (OCR) in primary mouse myotubes transfected with miR‐499/miR‐208b inhibitors alone and together with the presence of Fnip1 siRNA. n = 4 separate experiments done with 5 biological replicates. *P < 0.01 (anti‐miRs versus Control).

Data information: All values represent the mean ± SEM. P‐value in (A, B, D–F) was determined using two‐tailed unpaired Student's t‐test; P‐value in (C and G) was determined using one‐way ANOVA coupled to a Fisher's least‐significant difference (LSD) post hoc test.Source data are available online for this figure.

Figure 6. Restoring the expression of miR‐499 activates the slow‐oxidative muscle fiber program and ameliorates muscular dystrophy in mdx mice

1. Mean expression levels (RT‐qPCR) in gastrocnemius muscle of 8‐week‐old male WT and mdx mice (n = 5 mice per group). *P = 0.009 (Myh7b), *P = 0.0007 (miR‐499), *P = 0.0143 (Myh7), *P < 0.0011 (miR‐208b).

2. Mitochondrial respiration rates were determined from mitochondria isolated from the hindlimb muscle of the indicated mice using pyruvate as substrate. n = 3 separate experiments done with 7–8 biological replicates. *P < 0.01 (ADP).

3. (Left) Representative Western blot analysis of PGC‐1α (Top) and Fnip1 (Bottom) protein expression in the gastrocnemius muscle from the indicated genotypes with α‐tubulin as the loading control. (Right) Quantification of the PGC‐1α/tubulin and Fnip1/tubulin signal ratios normalized (= 1.0) to the WT control. WT, n = 8; mdx, n = 5; mdx/499Tg, n = 7. *P < 0.0001 (versus WT), ^‡ P < 0.0001 (versus mdx).

4. Representative Western blot analysis of myoglobin protein expression in the gastrocnemius muscle from the indicated genotypes with α‐tubulin as the loading control (n = 5 mice per group).

5. (Top) Representative MHC1 immunofluorescence (IF) in the soleus of the indicated genotypes. Scale bar: 500 μm. (Bottom) Representative H&E staining of the gastrocnemius muscle of the indicated genotypes. Scale bar: 100 μm.

6. Five‐week‐old male WT, mdx, and mdx/MCK‐miR‐499 mice were euthanized. and serum creatine kinase activity was determined. WT, n = 10; mdx, n = 10; mdx/499Tg, n = 14. *P < 0.0001 (versus WT), ^‡ P = 0.0003 (versus mdx).

7. The bars represent the mean running time and distance (± SEM) for 8‐week‐old male mice on a motorized treadmill. WT, n = 10; mdx, n = 13; mdx/499Tg, n = 8. *P = 0.01406 (Running time, versus WT), ^‡ P = 0.005756 (Running time, versus mdx). *P = 0.004558 (Running distance, versus WT), ^‡ P = 0.001668 (Running distance, versus mdx).

8. Respiratory exchange ratio (RER) during a graded exercise regimen on a motorized treadmill. mdx, n = 8; mdx/499Tg, n = 6. ^‡ P < 0.05.

Data information: All values represent the mean ± SEM. P‐value in (A, B and H) was determined using two‐tailed unpaired Student's t‐test; P‐value in (C, F and G) was determined using one‐way ANOVA coupled to a Fisher's least‐significant difference (LSD) post hoc test.Source data are available online for this figure.

Figure EV4. miR‐499 activation reverses the diminished oxidative muscle fiber program in mdx muscle

1. (Left) Representative Western blot analysis performed on extracts of the gastrocnemius muscle isolated from the indicated genotypes using phospho‐AMPKα (Thr172) and AMPKα antibodies. (Right) Quantification of the p‐AMPKα/AMPKα signal ratios normalized (= 1.0) to the NTG control. WT, n = 6; mdx, n = 6; mdx/499Tg, n = 5. *P < 0.0001 (versus WT), ^‡ P < 0.0001 (versus mdx).

2. Expression of the Ldhb and Ldha genes (RT‐qPCR) in muscle from the indicated genotypes (n = 5 mice per group). Ldhb: *P = 0.0006 (versus WT), ^‡ P < 0.0001 (versus mdx); Ldha: *P = 0.059 (versus WT), ^‡ P = 0.0052 (versus mdx).

3. A representative LDH isoenzyme activity gel is shown (n = 3 mice per group).

4. Mitochondrial respiration rates were determined from the extensor digital longus muscle of the indicated genotypes using pyruvate/malate as substrate. Pyruvate/malate (Py/M)‐stimulated, ADP‐dependent respiration, and oligomycin‐induced (oligo) are shown. mdx, n = 6; mdx/499Tg, n = 7. *P = 0.0474 (ADP).

5. Results of qPCR to determine mitochondrial DNA levels in WV muscle of the indicated genotypes using primers for NADH dehydrogenase (Nd1, mitochondria‐encoded) and lipoprotein lipase (Lpl, nuclear‐encoded). mdx, n = 5; mdx/499Tg, n = 6. P = 0.0989 (NS, not significant).

Data information: All values represent the mean ± SEM. P‐value in (A and B) was determined using one‐way ANOVA coupled to a Fisher's least‐significant difference (LSD) post hoc test; P‐value in (D and E) was determined using two‐tailed unpaired Student's t‐test.Source data are available online for this figure.

Figure EV5. Restoring the expression of miR‐499 reduces muscle damage in mdx mice

1. Bar histograms represent size distribution of the muscle fibers from the indicated genotypes. Relative fiber size was quantified using Image‐Pro Plus software (n = 5 mice per group).

2. Quantification of muscle fibers with centrally located nuclei in indicated genotypes. Values represent the mean % (± SEM) total muscle fibers from n = 5 mice per group and five images per muscle. *P = 0.0306 (versus mdx).

3. Representative images of Evans blue dye infiltration in damaged myofibers of the tibialis anterior (TA) muscle from indicated genotypes. Scale bar: 500 μm.

Data information: Values represent the mean ± SEM. P‐value was determined using two‐tailed unpaired Student's t‐test.

Figure 7. A model for the adaptive mitochondrial function during muscle fiber type transition

The schematic depicts a proposed model for the miR‐499/Fnip1/AMPK circuit that orchestrates mitochondrial function to match muscle contractile machinery during fiber type transition.