Muscle metabolic resilience and enhanced exercise adaptation by Esr1-induced remodeling of mitochondrial cristae-nucleoid architecture in males

Abstract

Reduced estrogen action is associated with obesity and insulin resistance. However, the cell and tissue-specific actions of estradiol in maintaining metabolic health remain inadequately understood, especially in men. We observed that skeletal muscle ESR1/Esr1 (encodes estrogen receptor α [ERα]) is positively correlated with insulin sensitivity and metabolic health in humans and mice. Because skeletal muscle is a primary tissue involved in oxidative metabolism and insulin sensitivity, we generated muscle-selective Esr1 loss- and gain-of-expression mouse models. We determined that Esr1 links mitochondrial DNA replication and cristae-nucleoid architecture with metabolic function and insulin action in the skeletal muscle of male mice. Overexpression of human ERα in muscle protected male mice from diet-induced disruption of metabolic health and enhanced mitochondrial adaptation to exercise training intervention. Our findings indicate that muscle expression of Esr1 is critical for the maintenance of mitochondrial function and metabolic health in males and that tissue-selective activation of ERα can be leveraged to combat metabolic-related diseases in both sexes.

Keywords: estrogen action; exercise adaptation; insulin sensitivity; mitochondrial cristae architecture; mitochondrial function; mtDNA replication; oxidative metabolism.

Graphical abstract

Figure 1

Muscle ERα controls insulin sensitivity and oxidative metabolism (A) ESR1 in muscle from men shows a high number of gene-gene correlations and limited overlap with other metabolic tissues outside of adipose (GTEx n = 210 males, age 20–70 years). (B and C) Muscle Esr1 is inversely correlated with the HOMA-IR insulin resistance index and adiposity (% fat) in a 16-strain panel of inbred mice (n = 4 mice per strain for individual closed circles). (D and E) Muscle Esr1 expression is reduced in male high-fat diet-fed and Lep^Ob mice vs. lean controls (n = 6/group). (F and G) Esr1 mRNA and protein in muscle from control f/f and MERKO male mice (representative immunoblot, n = 6/genotype). (H–K) (H) Fasting plasma insulin in MERKO vs. control f/f (n = 17 mice/genotype). (I) Glucose tolerance test in control f/f and MERKO male mice (n = 6–8/group). Skeletal muscle insulin sensitivity assessed by (J) hyperinsulinemic-euglycemic clamp (IS-GDR, insulin-stimulated glucose disposal rate; n = 6–8/genotype) and (K) ex vivo soleus muscle glucose uptake (n = 5–6 mice/genotype) in control f/f and MERKO male mice. (L) Representative immunoblots of insulin-stimulated p-Akt in basal 6-h-fasted quadriceps muscle (n = 6/genotype). (M) GLUT4 protein levels in basal 6-h-fasted quadriceps muscle (n = 5–6/genotype). Densitometric analyses expressed in arbitrary units (AU). All values expressed as means ± SEM, ∗p < 0.05 between genotype comparison, # within genotype between condition comparison. Significance detected by Student’s t test and repeated measures ANOVA where appropriate.

Figure 2

Muscle-specific ERα deletion reduces basal energy expenditure promoting muscle lipid accumulation (A) Lipidomic analyses of muscle from control f/f and MERKO mice, n = 6/genotype. (B) Quantitative reverse-transcription PCR (RT-PCR) analysis of quadriceps muscle transcripts reflecting lipid metabolism (n = 6/genotype). (C–G) (C and D) Metabolic caging studies using indirect calorimetry to determine (C) VO₂, (D) energy expenditure, (E) respiratory exchange ratio (RER), (F) food consumption, and (G) ambulatory activity (n = 6 per genotype). All values are expressed as means ± SEM detected by Student’s t test and ANCOVA where appropriate. ∗p < 0.05 between genotype difference.

Figure 3

ERα deletion alters mitochondrial morphology and respiration (A) Real-time respirometry was performed on frozen quadriceps muscle homogenates from control f/f and MERKO mice (n = 6 mice/genotype). (B and C) Representative immunoblots and corresponding densitometry of representative subunits of the mitochondrial electron transport complexes in muscle from control f/f and MERKO (n = 6/genotype). (D and E) Representative immunoblot and densitometry of muscle PGC1α from control f/f and MERKO mice (n = 6/genotype). (F–H) (F and G) Primary muscle cells were stained with Mitosox reflecting superoxide and (H) TMRM reflecting mitochondrial membrane potential and analyzed by flow cytometry (n = 3 studies performed in duplicate). (I) Transmission electron microscopy images showing elongated and hyperfused mitochondria in MERKO (right) vs. control f/f (left). Scale bars: 2 μm. All values are expressed as means ± SEM detected by Student’s t test and repeated measures ANOVA where appropriate. ∗p < 0.05 between genotype difference.

Figure 4

Muscle ERα deletion alters mitochondrial fission-fusion signaling and mtDNA replication (A–E) Representative immunoblots and corresponding densitometry of mitochondrial fission-fusion proteins and cristae junction-related proteins from control f/f vs. MERKO mouse quadriceps muscle (n = 6/genotype). (F) RNA sequencing of gastrocnemius muscle from the 100-strain UCLA Hybrid Mouse Diversity Panel (HMDP) with Venn diagram showing the gene overlap between Esr1, Dnm1L, and Polg1 (n = 4 mice/strain). (G) Enrichment score analysis of muscle RNA sequencing from HMDP mice shows that Esr1, Polg1, and Dnm1L associate with transcripts involved in protein transport and mitochondria. (H and I) (H) Polg1 and (I) Polrmt expression are markedly reduced in muscle from NC-fed MERKO vs. control f/f mice (n = 6/genotype). (J) Reduced PRPP in MERKO vs. control f/f detected by metabolomics analysis (n = 5–6/genotype). (K) Reduced Brdu incorporation into newly synthesized mtDNA from C2C12 myotubes with Esr1-KD vs. scrambled control (Scr) (n = 3 in triplicate). (L and M) (L) Mitochondrial DNA copy number in muscle is identical between MERKO and control f/f mice (n = 6/genotype). (M) Confocal microscopy images show enlarged mtDNA-containing nucleoids (stained with Picogreen) in cells lacking Esr1, along with reticular-networked mitochondria (stained with Mitotracker red). All values are expressed as means ± SEM detected by Student’s t test. ∗p < 0.05 between genotype difference.

Figure 5

Muscle-specific Esr1 overexpression increases mitochondrial function and metabolic health (A and B) Polg1 transcript and PolG protein expression (representative immunoblot and densitometry) are markedly increased in quadriceps muscle from mERα^OE vs. control f/f (n = 6 mice/genotype). (C and D) (C) Muscle mitochondrial DNA copy number and (D) expression of mitochondrial-encoded transcripts are increased in mERα^OE vs. control f/f (n = 6 mice/genotype). (E) Transmission electron microscopy images of muscle from mERα^OE (right) vs. control f/f (left) showing that Esr1 overexpression promotes a more spherical (lower) electron-dense organelle with increased cristae volume (n = 4 mice/genotype). Scale bars: 2 μm for top 4 panels and 200 nm for lower 2 panels of higher magnification images. (F) Heatmap showing top 40 differentially expressed mitochondrial proteins between mERα^OE and control f/f (n = 5 per genotype; quadriceps muscle). (G–J) (G) Hindlimb muscles from mERα^OE show a deeper red color compared with control f/f (n = 6 mice/genotype). Metabolic caging studies show that mERα^OE mice have an (H and I) increased VO₂ and (J) body temperature compared with control f/f (n = 6–7 mice/genotype). (K–M) Muscle-specific Esr1 overexpression protects male mice from HFD-induced (K) glucose intolerance (AUC, area under the curve) and (L and M) muscle insulin resistance determined by ex vivo muscle 2-deoxyglucose uptake and hyperinsulinemic-euglycemic clamps, respectively (n = 6 mice/genotype). All values are expressed as means ± SEM detected by Student’s t test. ∗p < 0.05 between genotype difference.

Figure 6

Muscle ERα overexpression reorganizes chromatin structure and enhances exercise-induced adaptation of mitochondria (A and B) Polg1 locus showing peak locations relative to the transition start site (TSS) determined by ATAC sequencing. (C) E₂-stimulated binding of ERα to the Polg1 proximal promoter (Pgr; positive control) (n = 3 in triplicate). (D and E) Motif enrichment for top 200 and bottom 200 regions of open chromatin as defined by ATAC sequencing in skeletal muscle of mERα^OE mice compared to control f/f. (F and G) (F) Acute exercise by treadmill running to maximum speed and (G) muscular endurance assessed by duration of dynamic hanging (latency to fall test), MERKO vs. control f/f (n = 5–7 mice/genotype). (H–J) (H) Total running volume over 30 days of volitional activity was similar between the genotypes, MERKO vs. control f/f (n = 6 mice/genotype). However, the well-described exercise training-induced increase in (I) Pgc1a transcript and (J) mitochondrial CN observed for control f/f (open bars, relative to sedentary) was blunted in MERKO (black bars) and not significantly different from SED (n = 5/genotype). (K) Complex I and II respiration rates were reduced in frozen muscle homogenates (quadriceps) from MERKO vs. control f/f following training as assessed by Seahorse analysis (n = 6/genotype). (L) Average daily volitional running in control f/f (open bars) vs. mERα^OE (closed bars) (n = 5–6/genotype). (M and N) Adipose tissue mass and fasting insulin concentrations were reduced in mERα^OE compared with control f/f mice following the 30-day running intervention (n = 6/genotype). (O and P) (O) Mitochondrial DNA CN and (P) transcript expression in skeletal muscle of control f/f (open bars) vs. mERα^OE (closed bars) following 30 days of volitional wheel running (n = 6/genotype; SED, sedentary; TRN, exercise trained). All values are expressed as means ± SEM detected by Student’s t test. ∗p < 0.05 between genotype difference, # within genotype, between conditions comparison.