Muscle-specific Keap1 deletion enhances force production but does not prevent inactivity-induced muscle atrophy in mice

Abstract

Immobilization-associated muscle atrophy and weakness appear to be driven in part by oxidative stress. Nuclear Factor Erythroid 2-Related Factor 2 (NRF2) is a critical redox rheostat that regulates oxidative stress responses, and its deletion is known to accelerate muscle atrophy and weakness during aging (sarcopenia) or denervation. Conversely, pharmacologic activation of NRF2 extends mouse lifespan and attenuates sarcopenia. Similarly, deletion of Kelch-like ECH-associated Protein 1 (Keap1), a negative regulator of NRF2, enhances exercise capacity. The purpose of this study was to determine whether muscle-specific Keap1 deletion is sufficient to prevent muscle atrophy and weakness in mice following 7 days of hindlimb unloading (HU). To test this hypothesis, control (Ctrl) and tamoxifen-inducible, muscle-specific Keap1 knockout (mKO) mice were subjected to either normal housing (Sham) or HU for 7 days. Activation of NRF2 in muscle was confirmed by increased mRNA of NRF2 targets thioredoxin 1 (Txn1) and NAD(P)H quinone dehydrogenase 1 (NQO1) in mKO mice. Keap1 deletion had an effect to increase force-generating capacity at baseline. However, muscle masses, cross-sectional area, and ex vivo force were not different between mKO and Ctrl HU mice. In addition, muscle 4-hydroxynonenal-modified proteins and protein carbonyls were unaffected by Keap1 deletion. These data suggest that NRF2 activation improves muscle force production during ambulatory conditions but is not sufficient to prevent muscle atrophy or weakness following 7 days of HU.

Keywords: NRF2; carbonyl stress; oxidative stress; redox; sarcopenia.

FIGURE 1

Validation of skeletal muscle‐specific knockout of Keap1. (A) Schematic of Keap1 inducible knockout allele. (B) Genotyping for loxP (top gel) and Cre (bottom gel) performed on ear snips of mice. (C) Timeline of the experimental design. Ctrl and mKO male mice were injected with tamoxifen for 5 days, followed by a 2‐week washout before undergoing either 7 days of hindlimb unloading (HU) or normal housing (Sham) before terminal measures were assessed. (D) Genotyping for successful Cre recombination following tamoxifen injection. Only successfully recombined alleles are amplified to produce a 288 bp fragment (Arrow) in gastrocnemius muscle from mKO mice (top) but not in liver (bottom) or in loxP Ctrl mice. (E) Western blot of Keap1 in mouse gastrocnemius from Ctrl and mKO mice. (F, G) RT‐PCR for NRF2 target genes Txn1 and Nqo1 in gastrocnemius muscles from Ctrl and mKO mice. Data are Mean ± SEM and analyzed via two‐way ANOVA with Bonferroni post hoc tests. Statistical significance was set to p < .05. For all panels, Sham Ctrl N = 5, Sham mKO N = 5, HU Ctrl N = 8, and HU mKO N = 9. Schematics in panels (A) and (C) were made by BioRender.

FIGURE 2

Skeletal muscle‐specific knockout of Keap1 does not prevent muscle atrophy following 7 days of hindlimb unloading. (A) Post‐intervention body mass was lower following HU and in Sham mKO mice. (B) Body composition via NMR reveals lower lean mass in Sham mKO mice and in both Ctrl and mKO mice following HU. (C–G) The average of left and right hindlimb muscle masses taken from mice at the time of dissection was lower following HU but not altered by mKO. Data are Mean ± SEM and analyzed via two‐way ANOVA with Bonferroni post hoc tests. Statistical significance was set to p < .05. For all panels, Sham Ctrl N = 5, Sham mKO N = 5, HU Ctrl N = 8, and HU mKO N = 9.

FIGURE 3

Skeletal muscle‐specific knockout of Keap1 decreases fast‐twitch fiber CSA in EDL but does not alter fiber type in SOL following HU. (A) Representative image of EDL cross section and immunofluorescence targeting Type IIa fibers (green). (B) Frequency of EDL fiber cross‐sectional area in Sham and HU (C) was not different between genotypes. (D) Fast, oxidative 2a fibers represent a greater proportion of EDL muscles in HU mKO mice compared to HU Control mice. (E) Representative image of SOL cross section and immunofluorescence targeting Type I (red) and IIa (green) fibers. (F) Frequency of SOL fiber cross‐sectional area in Sham and HU (G) was not different between genotypes. (H) Fiber type proportion of SOL muscles in Sham and HU mice was not altered by Keap1 mKO. Data are Mean ± SEM and analyzed via two‐way ANOVA with Bonferroni post hoc tests. Statistical significance was set to p < .05. *Indicates post hoc significance. For panels (B–D), Sham Ctrl N = 3, Sham mKO N = 3, HU Ctrl N = 7, and HU mKO N = 8. For panels (F–H), Sham Ctrl N = 4, Sham mKO N = 4, HU Ctrl N = 9, and HU mKO N = 7.

FIGURE 4

Skeletal muscle‐specific knockout of Keap1 enhances ex vivo force but does not prevent muscle weakness following 7 days of hindlimb unloading. (A) Force frequency curves quantifying absolute ex vivo force in soleus muscles were higher in Sham mKO mice but were no different between Ctrl and mKO mice following HU. Specific force was calculated by normalizing ex vivo force production to muscle cross‐sectional area. (B) Soleus ex vivo specific force was higher in Sham mice but was not affected by Keap1 knockout. Ex vivo absolute (C) and specific force (D) were highest in mKO sham mice. Data are Mean ± SEM and analyzed via two‐way ANOVA with Bonferroni post hoc tests. Statistical significance was set to p < .05. Post hoc significance is indicated as follows: a—Sham Ctrl versus Sham mKO, b—Sham Ctrl versus HU Ctrl, c—Sham Ctrl versus HU mKO, d—Sham mKO versus HU Ctrl, e—Sham mKO versus HU mKO, and f—HU Ctrl versus HU mKO. For all panels, Sham Ctrl N = 5, Sham mKO N = 5, HU Ctrl N = 8, and HU mKO N = 9.

FIGURE 5

Skeletal muscle‐specific knockout of Keap1 does not affect autophagic flux or carbonyl stress. (A) Representative western blot for LC3I, LC3II, p62, and 4HNE. (B) Quantification of 4HNE western blots revealed no effect of mKO or 7 days of HU on 4HNE‐modified proteins. (C) Global protein carbonylation was also not altered by mKO or 7 days of HU. Both LC3I (D) and LC3II (E) protein abundance was not affected by mKO but was increased following 7 days of HU. However, the proportion of LC3II:LC3I (F) was unaffected by mKO nor 7 days of HU. (G) p62 was also unaffected by mKO and 7 days of HU. (H) The expression of the stress response gene Atf4 was unaltered by mKO or HU. Data are Mean ± SEM and analyzed via two‐way ANOVA with Bonferroni post hoc tests. Statistical significance was set to p < .05. For all panels except C, Sham Ctrl N = 5, Sham mKO N = 5, HU Ctrl N = 8, and HU mKO N = 9. For panel (C), Sham Ctrl N = 4 due to lack of sample.