Edited* by Bruce M. Spiegelman, Dana-Farber Cancer Institute/Harvard Medical School, Boston, MA 02115, and approved January 28, 2011 (received for review December 30, 2010)
A causal role for mitochondrial DNA (mtDNA) mutagenesis in mammalian aging is supported by recent studies demonstrating that the mtDNA mutator mouse, harboring a defect in the proofreading-exonuclease activity of mitochondrial polymerase gamma, exhibits accelerated aging phenotypes characteristic of human aging, systemic mitochondrial dysfunction, multisystem pathology, and reduced lifespan. Epidemiologic studies in humans have demonstrated that endurance training reduces the risk of chronic diseases and extends life expectancy. Whether endurance exercise can attenuate the cumulative systemic decline observed in aging remains elusive. Here we show that 5 mo of endurance exercise induced systemic mitochondrial biogenesis, prevented mtDNA depletion and mutations, increased mitochondrial oxidative capacity and respiratory chain assembly, restored mitochondrial morphology, and blunted pathological levels of apoptosis in multiple tissues of mtDNA mutator mice. These adaptations conferred complete phenotypic protection, reduced multisystem pathology, and prevented premature mortality in these mice. The systemic mitochondrial rejuvenation through endurance exercise promises to be an effective therapeutic approach to mitigating mitochondrial dysfunction in aging and related comorbidities.
The mitochondrial theory of aging postulates that the lifelong accumulation of somatic mitochondrial DNA (mtDNA) mutations leads to mitochondrial abnormalities resulting in a progressive decline in tissue function (1, 2). Mitochondrial abnormalities and mtDNA mutagenesis are well-established intrinsic instigators that drive multisystem degeneration, stress intolerance, and energy deficits during aging in humans (3), monkeys (4), and rodents (5). Reduced mitochondrial quality and content in multiple tissues is also implicated in several aging-associated conditions, including cancer, obesity, cardiovascular diseases, hypertension, type 2 diabetes, osteoporosis, and dementia, as well as in the pathogenesis of neurometabolic syndromes, psychiatric disorders, end-stage renal disease, and mitochondrial cytopathies (6–10). Current treatment strategies for conditions associated with mitochondrial dysfunction address the secondary symptoms but not the deficiency itself (11). One possible approach to mitigating the primary deficiency is to boost the residual mitochondrial oxidative capacity by increasing functional mitochondrial mass in the affected tissues.
The epidemic emergence of modern chronic diseases largely stems from the adoption of a sedentary lifestyle and excess energy intake (12). There is incontrovertible evidence from epidemiologic studies that endurance exercise extends life expectancy and reduces the risk of chronic diseases (7–10, 13–21). Endurance exercise is the most potent physiological inducer of mitochondrial biogenesis in skeletal muscle (12) and also has profound effects on metabolism in various other tissues, including heart, brain, adipose tissue, and liver (22, 23). These adaptations result in improved healthspan, reduced risk of morbidity and mortality, and enhanced quality of life (12, 24). In this work, we used the mtDNA mutator mouse (designated the PolG mouse), a model of progeroid aging that exhibits elevated mtDNA point mutations and systemic mitochondrial dysfunction and phenocopies human aging (25, 26), to investigate whether endurance exercise can effectively counteract the entrenched multisystem degeneration and mitochondrial dysfunction to mitigate premature aging in these mice.
As early as 6 mo of age, sedentary PolG mice (PolG-SED) displayed symptoms of accelerated aging, as described previously (25, 26), including alopecia, graying hair, weight loss, poor body condition, and impaired mobility (Fig. S1A and Movie S1). At 8 mo of age, PolG mice that had undergone 5 mo of forced endurance exercise (PolG-END; 15 m/min for 45 min, 3 times/wk) lacked visible features of the accelerated aging phenotype (alopecia and graying hair) and were visually indistinguishable from age-matched WT littermates (Fig. S1A and Movie S1). Endurance exercise also attenuated the decline in body weight and body condition in PolG mice (Fig. S1 D and E). In addition, the PolG-END mice exhibited similar levels of physical activity and motor performance as the WT ice (Movies S2 and S3), indicating improved muscle function. When subjected to a progressive exhaustive exercise test, PolG-END mice had significantly greater functional endurance capacity compared with PolG-SED mice (Fig. 1A and Movie S4). Remarkably, the endurance capacity of PolG-END mice surpassed that of WT mice in trials 2–4 (Fig. 1A). We also observed a significant and complete prevention of early mortality in PolG-END mice (P < 0.01; Fig. S1 B and C). Given the large effect size, these findings are highly significant despite the relatively small number of animals per group; however, future studies with larger groups of animals subjected to lifelong endurance exercise are needed to define the full extent of this life-prolonging effect.
Endurance exercise mitigates the exercise intolerance and systemic pathology observed in PolG-SED mice. (A) The time to exhaustion (in min) of PolG-END mice in monthly endurance stress test trials was significantly greater compared with both PolG-SED and WT mice (n = 10/group). (B and C) Weights of skeletal muscle (quadriceps femoris and gastrocnemius) (B) and brain (C) from WT, PolG-SED, and PolG-END mice at 8 mo of age (n = 18/group). (D) Trichrome-stained cross-sections of heart from WT, PolG-SED, and PolG-END mice (n = 6/group). Representative images of heart from each group are displayed. (Scale bar: 50 mm.) (E) Weight of abdominal (ABDO) and retroperitoneal (RETRO) fat pads from WT, PolG-SED, and PolG-END mice (n = 10/group). (F) H&E-stained sections of dorsal skin from WT, PolG-SED, and PolG-END mice (n = 3/group). Open arrowheads indicate dermal skin; closed arrowheads, subcutaneous fat; diamond arrowheads, subcutaneous muscle. (G) Weights of gonads (ovaries and testes) from WT, PolG-SED, and PolG-END mice (n = 10/group). (H) Hemoglobin values in from WT, PolG-SED, and PolG-END mice (n = 10/group). *P < 0.05, **P < 0.01, and ***P < 0.001, PolG-SED versus both WT and PolG-END; †P < 0.05, ‡P < 0.01, PolG-END versus WT. Error bars represent SEM.
Aging is characterized by a loss of muscle mass (sarcopenia) and brain atrophy (9, 10, 27, 28). Consistent with the causal role of mtDNA mutations in multiorgan pathology in PolG mice, we found significant reductions in skeletal muscle (25, 26, 29) and brain mass in PolG-SED mice compared with WT mice (Fig. 1 B and C and Fig. S2A). PolG-SED mice also displayed significant cardiac hypertrophy (Fig. 1D and Fig. S2 B and C) and heart pathology (Fig. S2D) compared with WT mice. Endurance exercise mitigated age-associated sarcopenia, brain atrophy, and cardiomyopathy in PolG-END mice (Fig. 1 B–D and Fig. S2 A–D).
The PolG-SED mice had decreased body fat (lipodystrophy), consistent with their overall cachectic appearance and reduced weight of abdominal and retroperitoneal fat pads (Fig. 1E). Endurance exercise induced significant gains in the mass of abdominal and retroperitoneal fat pads, thereby attenuating lipodystrophy in PolG mice (Fig. 1E). In addition, histological examination of dorsal skin and subcutis showed that exercise suppressed the depletion of subcutaneous adipose tissue and thinning of the dermis in PolG mice (Fig. 1F). PolG-END mice also maintained a normal melanin content comparable to that of WT mice (Fig. S2E). Profoundly reduced fertility has been reported in PolG mice (26), congruent with reduced sperm count and decreased female fecundity with advancing age in these mice (26). We noted gonadal atrophy in 8-mo-old PolG-SED mice compared with WT mice (Fig. 1G). Endurance exercise attenuated the gonadal atrophy present in PolG mice of both sexes (Fig. 1G).
Anemia is a frequent clinical problem seen in elderly humans (30) and patients with acute leukemias (31). Consistent with previous reports (26, 31), we found signs of macrocytic anemia with abnormal erythroid maturation and megaloblastic changes, characterized by significantly higher mean corpuscular volume and lower hemoglobin, erythrocyte and leukocyte concentrations in peripheral blood of PolG-SED mice compared with WT mice (Fig. 1H and Fig. S3 A–C). We also noted splenic enlargement in PolG-SED mice, indicative of stress erythropoiesis (Fig. S3D). Endurance exercise prevented the development of anemia in the PolG mice, and hemoglobin, mean corpuscular volume, erythrocyte, and leukocyte levels and spleen size were indistinguishable between WT and PolG-END mice (Fig. 1H and Fig. S3 A–D).
Development of the progeroid phenotype in PolG mice is causally associated with reduced mtDNA copy number and accumulation of mtDNA point mutations (25, 26). We found a significant depletion in full-length mtDNA content (skeletal muscle, heart, and liver) and an increase in mtDNA point mutations (skeletal muscle) in PolG-SED mice compared with WT mice (Fig. 2 A and B and Fig. S3 E–G). Endurance exercise completely rescued mtDNA depletion in skeletal muscle, heart, and liver of the PolG mice (Fig. 2 A and B and Fig. S3 E and F). A causative role for mtDNA point mutations via impaired assembly of respiratory chain complexes in driving the premature aging in PolG mice has been reported (32). This phenomenon is exemplified by mitochondrial cytochrome c oxidase (COX) complex assembly, wherein the mtDNA-encoded subunits are highly conserved and form the catalytic core of the complex. The assembly of the mtDNA-encoded subunits of COX complex provides a platform for the subsequent incorporation of the nuclear subunits (33). Amino acid substitutions in the mtDNA-encoded subunits of the COX complex, a consequence of mtDNA point mutations, are deleterious to the function and stability of this complex. Thus, such amino acid substitutions are strongly selected against in the germ line (34). Endurance exercise reduced the frequency of point mutations in the PolG mice, resulting in a concomitant increase in mitochondrial COX complex assembly (Fig. 2 B and C and Fig. S3G). Our results indicate that endurance exercise-mediated normalization of the systemic degenerative pathology in PolG mice is closely associated with maintenance of high levels of mtDNA copy number and a reduction in mtDNA point mutation load. Clearly, exercise in PolG mice prevents mtDNA mutations from reaching a critical threshold above which pathology manifests, and it represents a viable presymptomatic therapy for patients carrying polymerase gamma mutations known to cause pathology (35).
Endurance exercise rescues mtDNA depletion, mitigates mtDNA random point mutations, and enhances COX assembly in PolG mice. (A) Quantification of mtDNA levels relative to diploid nuclear genome in skeletal muscle (soleus) from WT, PolG-SED, and PolG-END mice at 8 mo of age (n = 10/group). (B) Frequency of mtDNA point mutations in 3,325, 2,473, and 3,842 reads of mtDNA sequences from multiple mtDNA populations, yielding 1.46 × 106, 1.25 × 106, and 9.21 × 105 bp of mtDNA sequences from the PolG-SED, PolG-END, and WT mice, respectively. (C) COX assembly in skeletal muscle (quadriceps femoris) from WT, PolG-SED, and PolG-END mice (n = 4–6/group) using 2D Blue-Native PAGE. Blots were probed with COX subunits I, IV, Vb, and VIc. *P < 0.05, **P < 0.01, and ***P < 0.001, PolG-SED versus both WT and PolG-END; †P < 0.05, PolG-END versus WT. Error bars represent SEM.
Endurance exercise is known to induce metabolic adaptations via activation of the transcriptional coactivator peroxisome proliferator-activated receptor γ coactivator-1 α (PGC-1α), the master regulator of mitochondrial metabolism and biogenesis that has been touted as a potential therapeutic target for aging-associated diseases (36). Studies in primary cells from patients with mitochondrial disorders and in skeletal muscle of COX10 conditional knockout mice have indicated that inducing mitochondrial biogenesis via ectopic expression of PGC-1α (37) or endurance training (38) has beneficial effects on the mitochondrial pathology of genetic origin. Whether PGC-1α is the central regulator of mitochondrial biogenesis in other tissues remains unknown (36, 39). Interestingly, mild overexpression of PGC-1α in skeletal muscle alone is known to be protective against sarcopenia, to attenuate inactivity-induced fiber atrophy, to ameliorate Duchenne muscular dystrophy and Huntington's pathology, to reduce systemic chronic inflammation, and to maintain systemic glucose and insulin homeostasis in aged mice (40–44). We found lower basal PGC-1α mRNA expression and nuclear abundance in skeletal muscle of PolG-SED mice compared with WT mice (Fig. 3A). The reduced nuclear PGC-1α content in PolG-SED mice was associated with no change in the nuclear abundance of receptor-interacting protein (RIP) 140, a negative regulator of PGC-1α (45). We also detected significant decreases in the mRNA and protein content of mitochondrial transcription factor A (Tfam), a mediator of mtDNA integrity and transcription (46), in concert with lower levels of other downstream PGC-1α targets (Fig. 3A and Fig. S4 A–C). This is consistent with a recent study reporting profound down-regulation of gene sets associated with mitochondrial metabolism in PolG mice (29). Endurance exercise abrogated the nuclear accumulation of RIP140, while increasing the content of nuclear PGC-1α and Tfam, as well as mRNA expression of several downstream targets in skeletal muscle of PolG mice, thus shifting the cellular dynamics toward mitochondrial biogenesis (Fig. 3A and Fig. S4 A–C). PGC-1α mRNA expression remained unchanged in other organs (Fig. S4C), further supporting the association between maintenance of muscle-specific PGC-1α expression and exercise-mediated multisystem rejuvenation of PolG mice. Because PGC-1α lacks functional DNA- and ligand-binding domains, pharmacologic interventions that directly promote PGC-1α function without affecting the activity of its upstream metabolic regulators remain elusive. Endurance exercise is the only practical way to “selectively” modulate PGC-1α function within a therapeutically beneficial window, thereby circumventing the unknown and unwanted side effects of the drugs and of nonspecific activation of PGC-1α.
Endurance exercise induces systemic mitochondrial biogenesis, enhances systemic COX activity, and mitigates dysregulated systemic apoptosis in PolG mice. (A) Gene expression and protein content of nuclear PGC-1α and Tfam in skeletal muscle (tibialis anterior) of PolG-SED and PolG-END mice versus WT mice (n = 10/group) at 8 mo of age. (B) Representative Western blot of PGC-1α–mediated downstream proteins—(i) complex V ATP synthase subunit α, (ii) complex III subunit core 2, (iii) complex IV subunit I, (iv) complex II subunit 30 kDa, and (v) complex I subunit NADH-ubiquinone oxidoreductase 1β subcomplex 8—in skeletal muscle (quadriceps femoris), lung, and heart extracts from WT, PolG-SED (PS), and PolG-END (PE) mice (n = 8/group). (vi) Actin was used as a housekeeping loading control. (C) COX activity in skeletal muscle (quadriceps femoris), heart, and ovary of PolG-SED and PolG-END mice versus WT mice (n = 10/group). (D) Nuclear DNA fragmentation (apoptotic index) in the cytosolic fractions prepared from skeletal muscle (quadriceps femoris), heart, and liver from WT, PolG-SED, and PolG-END mice (n = 4–6/group). *P < 0.05, **P < 0.01, and ***P < 0.001, PolG-SED versus both WT and PolG-END; †P < 0.05, PolG-END versus WT. Error bars represent SEM.
Enhanced mitochondrial biogenesis in response to endurance exercise is supported by an increase in mitochondrial electron transport chain (ETC) subunits and COX activity in skeletal muscle of PolG-END mice compared with PolG-SED mice (Fig. 3 B and C). Electron microscopy studies of skeletal muscle and heart also demonstrated increased mitochondrial abundance in PolG-END mice versus PolG-SED mice (Fig. 4 A–C and Figs. S5 A–F and S6D). Electron microscopy revealed an accumulation of swollen, pleomorphic, oversized mitochondria in skeletal muscle and heart of PolG-SED mice (Fig. 4 B and D–G and Figs. S5 C and D and S6C). The abnormal mitochondria in the skeletal muscle of PolG-SED mice exhibited cristae fragmentation (Fig. 4D), vacuolization (Fig. 4E), disrupted membranes (Fig. 4F), and large myelin-like structures (Fig. 4G). Similar alterations in mitochondrial morphology have been documented to occur with age and in humans with mitochondrial myopathy (47, 48). Endurance exercise abrogated these morphological irregularities in mitochondrial morphology in PolG-END mice (Fig. 4 A, C, H and I and Figs. S5 A, B, E, and F and S6 C and D).
Endurance exercise restores mitochondrial abundance and morphology in PolG mice. (A–C) Electron micrographs of myofibers (quadriceps femoris) from WT (A), PolG-SED (B), and PolG-END (C) mice (n = 6/group) at 8 mo of age. (Scale bar: 1 μm.) (D–I) Myofibers of PolG-SED mice (D–G) are populated with enlarged, abnormally shaped mitochondria containing vacuoles, fragmented cristae, disrupted external membranes, and large myelin-like figures compared with mitochondria observed in WT (H) and PolG-END (I) mice. (Scale bar: 100 nm.)
Although numerous epidemiologic studies have clearly shown that exercise reduces morbidity and mortality (19, 21), data describing the systemic effects of endurance exercise are scarce. Given our findings of dramatic suppression of the accelerated aging phenotype and rescue of multisystem degenerate pathology, we sought to determine whether endurance exercise systemically promoted mitochondrial biogenesis in PolG mice. We found a significant reduction in the protein levels of mitochondrial ETC subunits and mitochondrial COX activity in heart, liver, brain, pancreas, and gonadal tissue in PolG-SED mice compared with WT mice (Fig. 3 B and C and Fig. S6 A and B). This decline in mitochondrial ETC subunit content and COX activity was prevented with endurance exercise in all tissues studied, indicating multisystem mitochondrial restoration (Fig. 3 B and C and Fig. S6 A and B).
Sustained mitochondrial dysfunction leads to activation of the caspase cascade culminating in DNA fragmentation, a hallmark of apoptosis that has been observed in aged tissues, acute leukemia, and neurometabolic disorders (29, 49, 50). The PolG-SED mice displayed increased DNA fragmentation in skeletal muscle, heart, liver, spleen, intestine, kidney, and gonads (Fig. 3D and Fig. S6E), suggesting that dysregulated systemic apoptosis is integral to driving multisystem degenerative pathologies (25, 29). Indeed, suppression of apoptosis successfully prevented cardiomyopathy in the heart-specific mitochondrial mutator mouse model (51). A recently proposed, intriguing mechanism postulates that somatic mtDNA mutation-driven generation of misfolded mitochondrial proteins may bind to proapoptotic proteins and activate systemic apoptosis (52). This might explain the mitochondrial ETC complex assembly defect and the disproportionate increase in apoptosis seen in the PolG-SED mice. Endurance exercise systemically abrogated this apoptosis index in the PolG mice, indicating that induction of prosurvival mechanism(s) is crucial for multisystem maintenance (Fig. 3D and Fig. S6E).
Although a plethora of previous studies found strong correlations among mtDNA mutations, mosaic respiratory chain dysfunction, and mammalian aging (2, 53–56), PolG mice provided the first direct cause-and-effect evidence that mtDNA mutagenesis and mitochondrial dysfunction results in progeroid aging phenotypes and associated multisystem pathologies (25, 26). Here we report that an increased burden of somatic mtDNA point mutations in PolG mice results in profound declines in mitochondrial biogenesis and systemic oxidative metabolism, reductions in mtDNA copy number, defects in the assembly of ETC functional complexes, accumulation of degenerate mitochondria, and a pathological increase in systemic apoptosis (25, 26, 32). Strikingly, 5 mo of endurance exercise promoted systemic mitochondrial biogenesis and increased multiorgan oxidative capacity, contributing to the complete phenotypic protection of the PolG mice. Whether the central mechanism driving mammalian aging and associated pathologies is mtDNA mutagenesis and depletion, enhanced systemic apoptosis, or some other form of mitochondrial dysfunction remains unknown (25, 26). Clearly, the therapeutic effects of endurance exercise are unprecedented and multifactorial in nature. The obvious question is how exercise can alleviate the mutational load despite the continued presence of defective mitochondrial polymerase γ. We hypothesize that endurance exercise-mediated regulation of muscle-specific PGC-1α may impose selective mitochondrial biogenesis of healthy mitochondria via modulation of mitochondrial dynamics (fusion and fission) and targeted autophagy of mitochondria carrying pathological levels of mutated mtDNA. This, together with the induction of secondary polymerase γ-independent mtDNA repair pathways with exercise, may maintain a pool of bioenergetically functional mitochondria. In addition, we speculate that endurance exercise modulates the release of systemic factors (e.g., chemokines, cytokines, metabolites) that may promote organ cross-talk, resulting in systemic mitochondrial biogenesis and multisystem rejuvenation. These adaptations may mitigate systemic mitochondrial dysfunction and accelerated cell death by diluting the pathological effects of mtDNA point mutations incurred systemically in PolG mice.
Our data clearly support endurance exercise as a medicine and a lifestyle approach to improving systemic mitochondrial function, which is critical for reducing morbidity and mortality across the lifespan (7, 19, 21). Our findings also have substantial implications for exercise therapy in young asymptomatic or paucisymptomatic patients harboring known pathogenic mutations in mtDNA regulatory proteins, such as POLG1, twinkle helicase, and others. Understanding the multiple molecular cues that lead to endurance exercise-mediated systemic mitochondrial rejuvenation in the mtDNA mutator mouse could also lead to the development of novel nutritional, pharmacologic, and exercise-based therapeutic interventions designed to ameliorate the structural and functional mitochondrial alterations associated with aging and metabolic diseases.
See SI Materials and Methods for details regarding animal breeding, exercise protocol, anthropometric measurements, endurance stress testing, and survival analyses. Molecular analyses including electron microscopy, melanin assays, mRNA expression, mtDNA copy number and point mutation analyses, subcellular fractionation, 2D BN-PAGE, immublotting, COX activity assays, apoptosis cell death detection ELISA, and statistical analyses are described in SI Materials and Methods. mtDNA mutator mice used in this study were described in ref. 25.
This work was supported by the Canadian Institutes of Health Research (Grant MOP97805), a kind donation from Mr. Warren Lammert and family (to M.A.T.), and Nathan Shock Award P30AG025708 (to S.M.). A.S. is funded by a Canadian Institutes of Health Research Institute of Aging doctoral research scholarship and a Canada PRESTIGE fellowship. D.I.O. and J.P.L. are funded by National Sciences and Engineering Research Council scholarships. G.C.K. and T.A.P. were awarded US Patent 7,126,040 for the PolgD257A mouse.
Author contributions: A.S. and M.A.T. designed research; A.S., J.M.B., D.I.O., J.P.L., B.P.H., M.A., J.E.T., S.M., and N.J.M. performed research; J.E.T., S.M., G.C.K., T.A.P., and M.A.T. contributed new reagents/analytic tools; A.S., J.M.B., and B.P.H. analyzed data; and A.S. wrote the paper.
The authors declare no conflict of interest.
↵*This Direct Submission article had a prearranged editor.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1019581108/-/DCSupplemental.
