Scavenging mitochondrial hydrogen peroxide by peroxiredoxin 3 overexpression attenuates contractile dysfunction and muscle atrophy in a murine model of accelerated sarcopenia

Abstract

Age-related muscle atrophy and weakness, or sarcopenia, are significant contributors to compromised health and quality of life in the elderly. While the mechanisms driving this pathology are not fully defined, reactive oxygen species, neuromuscular junction (NMJ) disruption, and loss of innervation are important risk factors. The goal of this study is to determine the impact of mitochondrial hydrogen peroxide on neurogenic atrophy and contractile dysfunction. Mice with muscle-specific overexpression of the mitochondrial H₂ O₂ scavenger peroxiredoxin3 (mPRDX3) were crossed to Sod1KO mice, an established mouse model of sarcopenia, to determine whether reduced mitochondrial H₂ O₂ can prevent or delay the redox-dependent sarcopenia. Basal rates of H₂ O₂ generation were elevated in isolated muscle mitochondria from Sod1KO, but normalized by mPRDX3 overexpression. The mPRDX3 overexpression prevented the declines in maximum mitochondrial oxygen consumption rate and calcium retention capacity in Sod1KO. Muscle atrophy in Sod1KO was mitigated by ~20% by mPRDX3 overexpression, which was associated with an increase in myofiber cross-sectional area. With direct muscle stimulation, maximum isometric specific force was reduced by ~20% in Sod1KO mice, and mPRDX3 overexpression preserved specific force at wild-type levels. The force deficit with nerve stimulation was exacerbated in Sod1KO compared to direct muscle stimulation, suggesting NMJ disruption in Sod1KO. Notably, this defect was not resolved by overexpression of mPRDX3. Our findings demonstrate that muscle-specific PRDX3 overexpression reduces mitochondrial H₂ O₂ generation, improves mitochondrial function, and mitigates loss of muscle quantity and quality, despite persisting NMJ impairment in a murine model of redox-dependent sarcopenia.

Keywords: aging; hydrogen peroxide; mitochondria; peroxiredoxin3; sarcopenia.

FIGURE 1

Cre‐Lox approach increases hPRDX3 expression in skeletal muscle mitochondria without upregulation of other antioxidant enzymes. (a) A schematic of peroxiredoxin3 human transgene construct (PRDX3) demonstrates a flanked STOP codon by LoxP sites. The PRDX3^fl/lf mice were bred to mice containing Cre recombinase driven by Human Skeletal Actin (HSA)‐Cre promotor to induce constitutive expression of PRDX3 in skeletal muscle. (b) The mRNA expression of human PRDX3. n = 6. Data were analyzed using ordinary two‐way ANOVA with Tukey post hoc tests. (c) Representative immunoblot images showing PRDX3 expression of mitochondrial and cytosolic fractions. Human PRDX3 antibody was used for the assay. (d) Representative immunoblots demonstrating human PRDX3 and mouse Prdx3 protein expression in WT, Sod1KO, WT‐PRDX3tg, and Sod1KO‐PRDX3tg from whole muscle homogenate. The antibody detects both human PRDX3 and mouse Prdx3 proteins. (e) Protein expression of key antioxidant enzymes using a targeted proteomics approach in muscle homogenates. n = 4. Student t‐tests were used comparing means between WT and mPRDX3 groups. Values are shown mean ± SEM. *p < 0.05. Prdx, peroxiredoxin; m Prdx3, mouse Prdx3; PRDX3, human PRDX3; txn, thioredoxin; txnrd, thioredoxin reductase; Sod, superoxide dismutase; Cat, catalase; Gpx, glutathione peroxidase; GSR, glutathione reductase

FIGURE 2

Impaired mitochondrial function in Sod1KO mice is prevented in the Sod1KO/mPRDX3Tg mice. (a) Rate of hydrogen peroxide generation in basal state without substrates or inhibitors (i.e. State 1). Significant main effects by Sod1 and mPRDX3. n = 4–7. (b) Oxygen consumption rate (OCR) with sequential addition of substrates for complex I and II using isolated skeletal muscle mitochondria. n = 4–7. (c) Calcium retention capacity (CRC) using isolated mitochondria from skeletal muscle. Significant main effects by Sod1 and mPRDX3, and interaction effect. n = 5–7. (d) Representative calcium tracings during CRC assay. Individual spikes represent calcium chloride, which was sequentially injected every 1 min until the opening of permeability transition pore (PTP) opening (inset). Values are shown as mean ± SEM. Data were analyzed using ordinary two‐way ANOVA with Tukey post hoc tests. *p < 0.05. GM, glutamate (10 mM) and Malate (2 mM); ADP (2 mM); Suc, succinate (10 mM); Ca, calcium (2 μM)

FIGURE 3

Contractile properties of skeletal muscle are impaired in Sod1KO mice, but protected by PRDX3 overexpression. (a) Maximum isometric force (mN). Significant main effect by Sod1 and interaction effect. n = 6–8. (b) Maximum isometric specific force, force per estimated cross‐sectional area (N/cm²). Significant main effect by Sod1 and PRDX3, and interaction effect. n = 6–8. (c) Peak twitch tension (N/cm²). n = 6–8. (d) Time to reach peak twitch (ms). n = 6–8. (e) Time to reach one‐half relaxation after twitch contraction (ms). Significant main effect by Sod1 and PRDX3, and interaction effect. n = 6–8. (f) SERCA activity plotted in response to increasing calcium concentration in gastrocnemius homogenates. n = 4. (g) Maximum SERCA ATPase activity. Significant main effect by Sod1 and PRDX3, and interaction effect. n = 4. Values are shown mean ± SEM. Data were analyzed using ordinary two‐way ANOVA with Tukey post hoc tests. *p < 0.05. SERCA, sarco/endoplasmic reticulum calcium ATPase

FIGURE 4

Sod1KO mice exhibit a significant reduction in muscle mass, which are partially protected by PRDX3 overexpression. (a) Body weights were significantly decreased in the Sod1KO and Sod1KO/mPRDX3Tg mice. Main effect by Sod1KO. n = 8–12. (b) Relative weights of gastrocnemius and quadriceps (mg muscle mass normalized by g body mass) were significantly reduced in the Sod1KO mice, but Sod1KO/mPRDX3Tg mice are protected. n = 8–13. (c) The number of fibers in gastrocnemius did not differ by Sod1 or mPRDX3. (d) Mean fiber cross‐sectional area was decreased in Sod1KO, but partially protected by mPRDX3 overexpression. Data were analyzed using ordinary two‐way ANOVA with Tukey post hoc tests. (e) Representative images of gastrocnemius cross‐sections stained by H & E. Arrows indicates centrally localized nuclei in myofibers. (f) Relative frequency distribution of fiber cross‐sectional area comparing WT and WT‐mPRDX3tg, and Sod1KO and Sod1KO/mPRDX3Tg. Frequencies are for WT and Sod1KO are shown separately for clarity. n = 4–5. (g) Fibers with central nuclei are shown in percent of total (%).4–5 sections were analyzed per mouse. 4–5 mice were analyzed per group. Values are shown mean ± SEM. Data were analyzed using ordinary two‐way ANOVA with Tukey post hoc tests. *p < 0.05. Gastroc, gastrocnemius; Quad, quadriceps

FIGURE 5

Increased expression of MuRF1 in Sod1KO is normalized by PRDx3 overexpression. (a) Immunoblot images showing MuRF1 and Atrogin1 expression. (b) Quantified immunoblot results of proteins n = 7–8. Values are shown mean ± SEM. Data were analyzed using ordinary two‐way ANOVA with Tukey post hoc tests. *p < 0.05

FIGURE 6

PRDX3 overexpression does not improve neuromuscular disruption in Sod1KO. (a) Maximum isometric absolute force (N) and specific force (N/cm²) either direct stimulation on muscle or through the sciatic nerve. * Significant difference compared with muscle‐stimulated force of WT. # Significant difference compared with nerve stimulation force of WT. & Significant difference compared with muscle‐stimulated force of Sod1KO. n = 6–8. (b) Representative neuromuscular junction (NMJ) immunofluorescence images from gastrocnemius muscle. Acetylcholine receptors (AchR) are pseudo‐colored in green and stained with Alexa‐488 conjugated α‐bungarotoxin. n = 3–5. (c) Quantification of NMJ area. Significant main effect by Sod1. (d) Percent of fragmented NMJs. n = 3–5. (e) mRNA levels of genes elevated in response to denervation and aging. Significant main effects by Sod1. n = 6. Values are shown mean ± SEM. Data were analyzed using ordinary two‐way ANOVA with Tukey post hoc tests. p < 0.05. NMJ, neuromuscular junction