Metabolic changes linked to diet slowing muscle loss in fast-aging mice with DNA repair problems

Ercc1^∆/− mice were obtained by crossing Ercc1^∆/+ with Ercc1^+/− mice, which have been previously described.32Xpg^−/− mice have been generated and characterized previously33 by crossing Xpg^+/− with Xpg^+/− mice. Wild‐type (WT) F1 littermates (+/+ genotypes from Ercc1^∆/− and Xpg^−/− breedings) were used as controls. Only male mice were used in this paper to minimize the influence of gender on DR study. All of the animals used were in full accordance with European legislation, and detailed in the supporting information.

Mice were housed in individual ventilated cages under specific pathogen‐free conditions (20–22°C, 12–12 h light–dark cycle with light phase adjusted to between 24:00 and 12:00 h) and provided food and water ad libitum (AL). Because the Ercc1^∆/− and Xpg^−/− mice were smaller, food was administered within the cages, and water bottles with long nozzles were used from around 2 weeks of age. Mice were weighed, visually inspected weekly, and scored blindly for gross morphological and motor abnormalities weekly. All efforts were made to ameliorate the suffering of the animals. Animals were bred and maintained on AIN93G synthetic pellets (Research Diet Services B. V.; gross energy content 4.9 kcal/g dry mass, digestible energy 3.97 kcal/g). Animals were divided randomly over all groups to prevent selection bias. Dietary restriction (DR) was applied as published before.17 On average, Ercc1^∆/− and Xpg^−/− mice ate 2.3 g food per day. DR was initiated gradually starting from 7 weeks of age with 10% food reduction (2.1 g/day), when animals reached almost‐maximum bodyweight and development was completed, and was increased weekly by 10%, until it reached 30% DR (1.6 g/day) from 9 weeks of age until 16 weeks. WT mice ate on average 3.0 g food per day, resulting in 2.1 g/day for 30% DR. Food was given to the animals just before the start of the dark (active) period, Zeitgeber Time (ZT) 12:00, to minimize disturbance of the biological clock. All mice were sacrificed and muscle samples were collected within a time frame of 1–3.5 h after feeding in their active period, between ZT 13:00 and 15:30, to avoid nutritional or circadian fluctuations of metabolites due to differences in the time after feeding and the biological clock.

Motor coordination performance was assessed by measuring the average time spent on an (2–40 rpm) accelerating rotarod (Ugo Basile) and detailed in Supplementary Information.

Mice were killed by CO₂ asphyxiation and a large piece of quadriceps femoris muscle (Quad) was dissected and rapidly frozen in liquid nitrogen, stored in −80°C, then lyophilized for 24 h, and weighed before analysis. Blood glucose was determined with a Freestyle mini blood glucose meter just before mice collection.

The chemicals, solvents, internal standards (ISTDs) used for muscle tissue extraction as well as the muscle tissue extraction method were described in He et al.34 and detailed in the supporting information.

A portion of extracted mouse muscle tissue for a pre‐experiment was used as quality control (QC) samples. A QC sample was injected once each six to eight samples to evaluate and correct for changes in sensitivity of the instruments.

The lipid metabolites were analysed by a validated ultra‐performance liquid chromatography tandem mass spectrometry (UPLC‐MS/MS) method.35 Each sample was measured with two complementary reverse phase methods using mobile phases with different pH, method details in the supporting information.

The energy‐production‐related metabolites were analysed by a hydrophilic interaction liquid chromatography (HILIC) mass spectrometry platform,36 the method details are described in the supporting information.

For each metabolite analysed by UPLC‐MS(/MS), the response ratio was obtained by Equation (), normalized by the muscle tissue dry weight, and corrected by the QC samples: 1

For metabolites that can be measured by multiple platforms, that is, some fatty acids can be determined with both low and high pH lipid method, the results with smaller QC RSD were utilized (Table). S1

The fold change of metabolites regulated by DR was calculated by the Equation (): 2

RStudio (Version 1.4.1106) and R (Version 4.0.5) were used for data statistical analysis, and figure plotting.

We previously have shown that interventions like DR can yield significant benefits to health and lifespan in progeroid DNA‐repair‐deficient mice.17, 30, 37 Here, we first investigated the effects of DR (starting at the age of 7 weeks) on sarcopenia by monitoring mouse body weight, muscle mass, grip strength, and accelerating rotarod performance in 16‐week‐old Ercc1^∆/− mice and 14‐week‐old Xpg^−/− mice, with corresponding WT littermates. These ages correspond to approximately 75% of their maximum lifespan and at which both mouse models display progressive age‐related pathologies across many organs, including severe muscle wasting.17, 30 Under AL feeding conditions, body weight (P < 0.0001, Figure 1A) and muscle mass (P < 0.05, Figure 1B) of Ercc1^∆/− and Xpg^−/− mice were substantially lower compared to WT littermates, and all were significantly (P < 0.05) decreased by DR, ranging from 11% to 44% for body weight (Figure 1A) and from 13% to 20% for muscle mass (Figure 1B), consistent with previous data.22Ercc1^∆/− and Xpg^−/− mice showed a significantly (P < 0.001) reduced muscle strength (Figure 1C,D), which might be the result of their muscle mass waste (Figure 1B).30 However, DR did not affect their fore limbs and all limbs grip strength even with significantly decreased muscle mass (Figure 1C,D). This indicates improved muscle quality, which is defined as the amount of strength per unit of muscle mass and is a critical biomarker of muscle health.38 These results are consistent with the recent observation in human muscle.20 One of the detrimental effects of sarcopenia in the elderly is loss of motor coordination.39 To further investigate the effects of DR on muscle quality, four trials of accelerating rotarod performance were conducted. In line with their accelerated aged phenotype, AL‐fed Ercc1^∆/− and Xpg^−/− mice showed a significantly (P < 0.05) reduced motor coordination compared to WT mice (Figure 1E), consistent with the adverse effect of sarcopenia on motor coordination, and fully in line with previous observations.30, 39 DR strongly (P < 0.05) improved the motor coordination and learning (Figure 1E). In particular, Ercc1^∆/− and Xpg^−/− mice were able to equally control motor coordination as normal WT mice in the fourth trial, further (partially) demonstrating the improved muscle quality by DR. Ercc1^∆/− DR mice were poor in rotarod performance at the first trial but systematically improved. These mice have a longer lifespan and seem to profit more from DR than the Xpg^−/− mutants,17 which are already more advanced in overall aging compared to Ercc1^∆/− at the start of DR. This indicates that the improved neurological performance in Ercc1^∆/− mice also noted before,17 which may also partially contribute to the improved motor coordination and learning performance.

The significantly decreased grip strength, motor coordination and learning performance in AL progeroid mice compared to AL‐WT mice (Figure 1) may be related to altered muscle function and quality. Hence, muscle specimens of the quadriceps femoris muscle (Quad) were profiled using recently developed metabolomics methods for the detection of signalling lipids and polar metabolites34 and the identified molecules related to muscle function and quality, that is, oxidative stress, pro‐ and anti‐inflammatory markers, and energy status, analysed by Partial Least Squares Discriminant Analysis (PLS‐DA). Separation between WT and Ercc1^∆/− or Xpg^−/− mice was observed, especially for the anti‐inflammatory metabolites (Figure 2A), indicating large alterations of accelerated aging in these metabolites.

Three indicators of oxidative stress, anti‐inflammation, and energy status were used to profile the effects of accelerated aging. The ratio of oxidized glutathione (GSSG) to reduced glutathione (GSH) is an important marker of oxidative stress.40 The trend for increased GSSG/GSH ratios in Ercc1^∆/− (and Xpg^−/−) animals compared to WT mice, which did not reach statistical significance (Figure 2B) might be related to the slightly improved oxidation defence, suggested by the increase of β‐hydroxybutyrate in Xpg^−/− mice (Figure 2C), an important indicator of anti‐aging effects [S1]. The ratio of ATP to ADP in Ercc1^∆/− and Xpg^−/− mice (Figure 2D) was significantly (P < 0.05) elevated comparing with the WT mice.

To identify the fundamental molecular changes of DR on inflammation and oxidative stress in sarcopenia, lipidomics analysis was performed on muscle of Ercc1^∆/−, Xpg^−/− and WT mice by measuring pro‐inflammatory and oxidative‐stress‐related metabolites, and lipid peroxidation products, that is, ω6 polyunsaturated fatty acids (PUFAs), prostaglandin (PG) series‐2, thromboxanes (TX) series‐2, hydroxyoctadecadienoic acids (HODEs, 12(S)‐hydroxyheptadecatrienoic acid (HHTrE), and hydroxyeicosatetraenoic acids (HETEs).8, 11 The resulting metabolic profiles of PLS‐DA showed clear differences between AL and DR in Ercc1^∆/− and Xpg^−/− mice (Figure 3A), indicating that DR exerts large effects on pro‐inflammatory and oxidative‐stress‐related metabolites in the progeroid mice muscle. This is consistent with the very strong anti‐aging effect of DR in the progeroid repair mutants.17 In view of the high impact of DR in the Ercc1^∆/− and Xpg^−/− mice, we decided to investigate the effect of DR in more detail. Significantly (P < 0.01) decreased GSSG/GSH in Ercc1^∆/− and Xpg^−/− mice demonstrated reduced oxidative stress by DR in progeroid mice (Figure 3B). The heatmap profiles of the metabolites (Table S3) related to oxidative stress and pro‐inflammation clearly show an overall reduction by DR in both Ercc1^∆/− and Xpg^−/− mice (Figure 3C). To better visualize the DR effects, the fold changes (Equation 2) of significantly (P < 0.05) regulated metabolites in the schematic pathways related to pro‐inflammatory and oxidative stress were profiled (Figure 3D). The linoleic acid (LA) and arachidonic acid (AA) pathways and metabolites (Figure 3E,F) related to pro‐inflammation and oxidative stress, that is, dihomo‐γ‐linolenic acid (DGLA), AA, 12(S)‐HHTrE,11 9‐HODE,8 11‐HETE,8, 10 PGE₂, PGD₂, PGF_2α,8 and TXB₂,11 and muscle growth (PGF_2α) and regeneration (PGE₂) stimulators [S2–S4] were all significantly (P < 0.05) and consistently downregulated by DR in both models.

The altered concentrations of oxidative stress markers and pro‐inflammation signalling lipids associated with DR may be not only due to the reduced levels of reactive oxygen species (ROS) or reactive nitrogen species (RNS), but also caused by the modulation of anti‐inflammatory cytokines [S5], which presented protective antioxidant defences in Ercc1^∆/− and Xpg^−/− mice [S6], such as eicosapentaenoic acid (EPA), PG series‐1 and series‐3, TX series‐1, epoxyeicosatrienoic acids (EETs), dihydroxy‐eicosatetraenoic acids (DiHETEs), and dihydroxy octadecadienoic acids (DiHODEs) [S7–S11]. PLS‐DA profiles of these anti‐inflammatory metabolites clearly segregated AL and DR groups in Ercc1^∆/− and Xpg^−/− mice (Figure 4A), demonstrating the more potent robustness in regulation of anti‐inflammatory mediators by DR in the muscle of progeroid mice. β‐Hydroxybutyrate is reported as an important marker of the anti‐aging effects of DR and fasting [S1]. Significantly (P < 0.001) increased β‐hydroxybutyrate indicated improved anti‐inflammation by DR in progeroid Ercc1^∆/− mice (Figure 4B). The heatmap profile of analysed anti‐inflammation metabolites (Table S4) in Ercc1^∆/− and Xpg^−/− muscle showed an obvious downregulation of ω3 PUFAs (i.e., α‐Linolenic acid (ALA), EPA, docosapentaenoic acid (DPA), and docosahexaenoic acid (DHA)), and several clearly upregulated anti‐inflammatory mediators (i.e., 14,15‐DiHETE, TXB₁, PGF₁, and 12,13‐DiHODE) by DR in the progeroid sarcopenic mice (Figure 4C). Using a cut‐off P‐value of <0.05 for the fold change (Equation 2) of metabolites (Figure 4D), we clearly observed significantly down‐regulated (i.e., ALA, EPA, DPA, DHA, and PGE₃), and up‐regulated mediators (i.e., PGF₁, 8,9‐EET, 12,13‐DiHODE, 14,15‐DiHETE, and GSH) by DR in the anti‐inflammation‐related pathways in Ercc1^∆/− and Xpg^−/− mice (Figure 4E,F).

Energy production in muscle tissue is highly related to grip strength and muscle function [S12]. To determine the DR effects on muscle energy generation in sarcopenia, the related metabolites of glycolysis, tricarboxylic acid (TCA) cycle, gluconeogenesis, and saturated fatty acids (SFAs) lipolysis, were all analysed. A clear segregation between AL and DR was observed in Ercc1^∆/− and Xpg^−/− mice (Figure 5A), indicating high modulation of energy‐production‐related metabolites by DR in the muscle of progeroid mice. Significantly (P < 0.05) increased ATP/ADP ratio demonstrated improved energy status by DR in Ercc1^∆/− and Xpg^−/− mice (Figure 5B). The influence of fasting on the ATP/ADP ratio and energy status was minimized by sacrificing mice and collecting muscle samples 1–3.5 h after feeding, when mice had just finished or were finishing their food intake. Phosphocreatine is considered as the “energy pool” in muscle cells, and is preferentially consumed by generating creatine in case of an insufficient energy supply [S13]. A heatmap profile of all analysed energy‐production and storage‐related metabolites (Table S5) showed that most of the analytes were highly upregulated by DR (Figure 5C). To profile the DR effects, the fold change (Equation 2) of significantly (P < 0.05) modulated metabolites in the main energy generation pathways, that is, glycolysis and TCA, and metabolites related to pentose phosphate pathway (PPP), gluconeogenesis, SFAs lipolysis, and phosphocreatine were shown in their pathways in Figure 5D. The significantly increased fold change of key metabolites in glycolysis and TCA cycle indicated upregulation of energy production and storage pathways by DR in Ercc1^∆/− and Xpg^−/− mice.

Phosphocreatine is generated from creatine by receiving a high‐energy phosphate group split from ATP [S13], therefore higher ratio of phosphocreatine to creatine represents a higher energy status of muscle tissue. The significantly (P < 0.05) increased ratio of phosphocreatine to creatine further revealed that the energy status of muscle was notably improved by DR in Ercc1^∆/− and Xpg^−/− mice (Figure 5E). Glycogen, primarily in liver and muscle cells, is an important source to maintain energy balance. Fasting and DR can stimulate the transformation of liver and muscle glycogen to glucose and glucose‐1‐phosphate (glucose‐1‐P), respectively, to feed into glycolysis for energy production. The significantly (P < 0.001) increased glucose‐1‐P in muscle tissue demonstrated the high consumption of muscle glycogen during DR (Figure 5E). Muscle consumes nearly 80% of the body's glucose content [S14, S15]. The significantly (P < 0.01) upregulated glucose levels in muscle indicated that glucose is transported and absorbed from blood to muscle tissue because of DR (Figure 5E). This is consistent with the decreased blood glucose content by DR in Figure 5E and Vermeij et al.17 Notably (P < 0.05) increased glycose‐6‐phosphate (glycose‐6‐P), fructose‐6‐phosphate (fructose‐6‐P), and pyruvate by DR revealed significantly upregulated glycolysis by DR in progeroid mice (Figure 5E). Similarly, the notable (P < 0.05) increased citrate, succinyl‐CoA, and malate indicated a significant upregulation of TCA by DR in progeroid mice (Figure 5E).

Alanine was reported to be able to enhance gluconeogenesis in starvation [S16]. Significantly (P < 0.01) increased alanine also revealed upregulated gluconeogenesis by DR in muscle of progeroid mice (Figure 5F). SFAs are another potential energy source during DR and three important SFAs, that is, myristic acid, palmitic acid, and stearic acid, were determined in Ercc1^∆/− and Xpg^−/− mouse muscle specimens. The results showed that all these SFAs were unaffected by DR (Figure 5F), indicating that DR did not induce lipolysis of SFAs in progeroid mice. Hence, the increased energy should originate from glycolysis, TCA, and gluconeogenesis from glycogen.

The significant (P < 0.05) increase of 6‐phosphogluconate by DR may be due to the elevated level of its precursor glucose‐6‐P (Figure 5F). Ribose 5‐phosphate (ribose‐5‐P) is the source for the synthesis of nucleotides. The unaffected ribose‐5‐P and ribulose‐5‐phosphate (ribulose 5‐P) levels by DR (Figure 5F) indicated that these precursors of nucleotides synthesis were not accumulating, consistent with the general suppression of growth and enhanced salvage pathways by DR.