Metformin treatment results in distinctive skeletal muscle mitochondrial remodeling in rats with different intrinsic aerobic capacities

Abstract

The rationale for the use of metformin as a treatment to slow aging was largely based on data collected from metabolically unhealthy individuals. For healthspan extension metformin will also be used in periods of good health. To understand the potential context specificity of metformin treatment on skeletal muscle, we used a rat model (high-capacity runner/low-capacity runner [HCR/LCR]) with a divide in intrinsic aerobic capacity. Outcomes of metformin treatment differed based on baseline intrinsic mitochondrial function, oxidative capacity of the muscle (gastroc vs soleus), and the mitochondrial population (intermyofibrillar vs. subsarcolemmal). Metformin caused lower ADP-stimulated respiration in LCRs, with less of a change in HCRs. However, a washout of metformin resulted in an unexpected doubling of respiratory capacity in HCRs. These improvements in respiratory capacity were accompanied by mitochondrial remodeling that included increases in protein synthesis and changes in morphology. Our findings raise questions about whether the positive findings of metformin treatment are broadly applicable.

Keywords: deuterium oxide; geroscience; healthspan; protein synthesis; proteomics.

FIGURE 1

(a) Study design. Age‐matched, 18‐month‐old high‐capacity runner (HCR) and low‐capacity runner (LCR) male rats were randomly assigned to one of three groups: a control group (CON), an acute metformin‐treated group (MET), a 48‐h washout metformin group (MET‐WO). The MET and MET‐WO groups received 100 mg/kg/day of metformin in drinking water for the first 7 days of treatment. After day 7, the metformin dose was increased to 200 mg/kg/day. The MET rats received metformin until the day of sacrifice, whereas the MET‐WO rats were switched back to tap water 48 h prior to euthanasia. The CON rats received regular tap water. All rats underwent deuterium oxide (D₂O) stable isotope labeling during the last week of the intervention. (b–d) Body composition measures from CON (black), MET (red), and MET‐WO (blue). (e–i) Skeletal muscle masses normalized to tibial length. (j) Total body water percentage. (k) Gastroc glycogen content. Data were analyzed by two‐way ANOVA (group × treatment) and are represented as mean ± SEM from 6 to 7 rats per group.

FIGURE 2

Determination of the nutrient sensing marker AMPK, (a) total, b, phosphorylated (Thr172), and c, phospho:total ratio in the gastroc. Data were analyzed by two‐way ANOVA (group × treatment) and are represented as mean ± SEM from 6 to 7 rats per group.

FIGURE 3

Determination of the fractional synthesis rates in (a–c), mitochondrial (Mito), (d–f), myofibrillar (Myo) fractions, and (g–i) cytosolic (Cyto) of the gastrocnemius (gastroc), soleus, and tibialis anterior (TA) muscles. Data were analyzed by two‐way ANOVA (group × treatment) and are represented as mean ± SEM from 4 to 7 rats per group.

FIGURE 4

We performed a mitochondrial targeted quantitative proteomics in rat gastroc and soleus skeletal muscles. The size of the dot indicates the relative abundance of the protein in the (a) gastroc and (b) solues. Visualization of the synthesis rates (K[1/day]) of the mitochondrial electron transport system complexes (CI, CII, CIII, CIV, CV) in the (c) gastroc and (d) soleus. Blue indicates a faster synthesis rate, whereas red indicates a slower synthesis rate. Gray proteins indicate that the protein was not detected in our analyses. The complexes were generated using complex IDs (CI: 7AK5, CII: 1ZOY, CIII: 7TZ6, CIV: 7COH, and CV: 8H9V) from the RCSB Protein Data Bank and rendered using ChimeraX. Data were analyzed by two‐way ANOVA (group × treatment) and are represented as mean from 5 to 7 rats per group. Within each strain $ indicates a significant difference from control group (CON), † indicates a significant difference from metformin‐treated (MET) group, and ‡ indicates significant differences from CON and MET.

FIGURE 5

(a–e) Mitochondrial dynamics markers. The average cross‐sectional area (CSA) of the (f) total mitochondria, (g) subsarcolemmal population, and (h) intermyofibrillar population. (i) The total subsarcolemmal CSA normalized to total mitochondrial CSA. (j) The total intermyofibrillar CSA normalized to total mitochondrial CSA. (k) The total mitochondrial CSA normalized to total fiber CSA. (l) The total subsarcolemmal CSA normalized to total fiber CSA. (m) The total intermyofibrillar CSA normalized to total fiber CSA. (n) Representative images showing the tom20tdtomato channel. Data were analyzed by two‐way ANOVA (group × treatment) and are represented as mean ± SEM from 3 to 7 fibers taken from a single image from 6 to 7 rats per group.

FIGURE 6

(a) Maximal mitochondrial respiration (V _max). Mitochondrial respiration in (b) HCRs and (c) LCRs. The apparent Km of ADP was calculated using Michaelis–Menten kinetics. Data were analyzed by two‐way ANOVA (a and d) or by one‐way ANOVA for each substrate addition (b and c) within each strain. Data are represented as mean ± SEM from 4 to 7 rats per group. HCR, high‐capacity runner; LCR, low‐capacity runner.