Rapamycin doses sufficient to extend lifespan do not compromise muscle mitochondrial content or endurance

Introduction

Aging is the most important risk factor for morbidity and mortality in Western society today. Due to the parallel rise in the risk for many different conditions, individuals often present with multiple comorbidities, and there is a limit to the benefit that can be obtained through therapies for any individual disease. For example, it has been estimated that a complete cure for cancer would extend the average human lifespan by about 3 years [1]. On the other hand, reducing calorie intake by ~40% while maintaining adequate nutrition slows the progression of most age-related changes simultaneously and extends life by 30-50% in rodents [2, 3]. Unfortunately, dietary restriction (DR) has major limitations as an approach to improve human health and longevity. First, it is likely that many would be unwilling or unable to maintain the requisite lifestyle [4]. Second, the regimen must be started early in life to obtain the maximal benefit [5-7]. Finally, studies in primates have yielded conflicting results. While there is general agreement that DR improves health and decreases age-related diseases, only one of the two ongoing studies has demonstrated an effect on overall survival [6, 8]. Identifying new, more generally applicable ways to target the aging process is an important goal for gerontology, and a promising approach to the prevention and treatment of age-related diseases.

Rapamycin, an inhibitor of the mammalian/mechanistic target of rapamycin (mTOR), presents a tantalizing possibility for a longevity drug [9]. It is the only compound that has extended both mean and maximum lifespan in both genders of mice by the rigorous standards of the National Institute on Aging's Intervention Testing Program [10, 11], and has been shown to slow the progression of multiple age-related phenotypes in mice [12-16]. Rapamycin works even when treatment is delayed until 20 months of age (equivalent to ~60 years for a human), and would not require any dietary modification. Because rapamycin has been used clinically as an immunosuppressant and chemotherapeutic, there is an extensive body of literature documenting its tolerability and side effects [17]. Rapamycin increases the risk of developing diabetes [18-20], increases cardiovascular risk factors [17, 21], causes hair, skin, and nail problems [21, 22], and has complex effects on the immune system [22, 23]. Although it has been suggested that the diabetes-like condition induced by rapamycin might be benevolent, resembling starvation-induced diabetes [24], the complete spectrum of side effects is likely to mask any anti-aging effects in humans, and to have a detrimental effect on lifespan overall. Thus, it is unlikely that rapamycin in its current form would have a beneficial effect in healthy humans, and it remains uncertain whether mTOR signaling could ever be targeted without the development of side effects.

There are two major protein complexes that contain mTOR, mTORC1 and mTORC2 [25]. Although rapamycin has been considered a specific inhibitor of mTORC1, chronic exposure to the drug can also disrupt mTORC2 in some cell lines [26] and in vivo [27]. We have previously demonstrated that rapamycin-induced insulin resistance is caused mainly by the “off-target” disruption of mTORC2, and that more specific targeting of mTORC1 using a genetic strategy can extend life without interfering with glucose metabolism [27]. This raises the hope that more specific pharmacological targeting of mTORC1 will be possible, and could replicate the beneficial aspects of rapamycin treatment with fewer negative consequences.

While it remains to be tested whether mTORC1 inhibition per se accounts for many of the detrimental effects of rapamycin, it is clear that this complex mediates the drug's effects on mitochondria in mammalian cells. Rapamycin decreases the expression of mitochondrial mRNAs in cultured muscle cells [28, 29] and suppresses oxygen consumption [28, 30, 31]. Decreased mitochondrial respiration is observed even in short-term experiments, suggesting that the effects of rapamycin are mediated in part by a post-translational mechanism. These effects are replicated by loss of mTORC1 function, but not by loss of mTORC2 function [28, 30]. Moreover, mTORC1 binds to the promoters of affected mitochondrial transcripts [29], providing further evidence that mTORC1, and not mTORC2, mediates the mitochondrial effects of rapamycin. These findings raise the possibility that rapamycin-treated mice might become frail and prone to bioenergetic failure, despite having increased longevity. Such effects in the face of mTORC1 inhibition might be considered a trade-off that could compromise survival in the wild, and possibly in humans, but would lead to increased longevity in the protected setting of a mouse colony. Therefore, we tested whether defects in mitochondrial biogenesis and function are apparent in the skeletal muscles of rapamycin-treated mice.

Results

Rapamycin treatment (2 mg/kg daily by intraperitoneal injection) decreased the mRNA expression of genes involved in mitochondrial biogenesis, including mitochondrial transcription factor A (TFAM), nuclear respiratory factor 1 (NRF1), and estrogen-related receptor α (ERRα), as well as genes involved in oxidative phosphorylation, including cytochrome c oxidase subunit 5B (COX5b), ATP synthase subunit O (ATP5O), and cytochrome c in gastrocnemius and soleus muscles, but not in the liver (Figures 1 and S1). These changes were most prominent in the highly oxidative soleus muscle, consistent with the findings of Cunningham et al. [29] and Blattler et al. [32].

Despite clear changes in message levels, we found that the expression of mitochondrial proteins involved in oxidative phosphorylation was unchanged by rapamycin treatment. We employed a series of monoclonal antibodies that detect representative subunits of each oxidative phosphorylation complex. This approach is predicted to give a reliable indication of overall complex assembly, since the subunits targeted by the monoclonal antibodies are labile when not properly incorporated into their respective oxidative phosphorylation complexes. No consistent changes in mitochondrial protein expression were observed in either the gastrocnemius or soleus muscles (Figure 2), or in the liver (Figure S2). Therefore, expression of mitochondrial proteins in the skeletal muscles of C57BL/6 mice was not affected by two weeks of intraperitoneal injection of rapamycin at a dose sufficient to cause metabolic dysfunction and to extend life.

Given the range of dosing strategies that have been used for rapamycin [9], as well as the long half-life of some mitochondrial proteins [33], we chose to examine the expression of mitochondrial proteins under the specific conditions that have been shown to increase longevity. Accordingly, these experiments employed HET3 mice (offspring of a cross between Balb/cByJ × C57BL/6J F1 mothers and C3H/HeJ × DBA/2J F1 fathers), which are expected to be free from any recessive defects present in the parental lines, display a long lifespan, and are known to respond to DR [34, 35]. Young or old HET3 mice were fed a diet containing encapsulated rapamycin at 14 ppm for 2 or 5 months and compared to age-matched controls fed the same diet lacking rapamycin. The effects of this regimen on glucose homeostasis were described previously [36]. As was the case in C57BL/6 mice injected with rapamycin, mitochondrial protein expression was unchanged in HET3 mice fed the rapamycin-containing diet for up to five months (Figure 3). Therefore, we were not able to detect any consistent change in mitochondrial protein expression using the rapamycin treatment regimen that extends mean and maximum lifespan in mice.

Our previous studies in cultured myoblasts suggested that rapamycin can impact mitochondrial function even in the absence of changes in protein expression [28], consistent with findings in two previous reports that employed other cell types [30, 31]. To determine whether mitochondrial performance might be impaired in rapamycin-treated animals, we studied treadmill endurance in a second cohort of animals. Rapamycin-treated mice were able to run the same distance as untreated littermates (Figure 4), suggesting that there was no overt deficit in skeletal muscle mitochondrial function.

To directly test mitochondrial function after rapamycin treatment in vivo, oxygen consumption was measured in isolated, permeabilized soleus muscle fibers by high resolution respirometry using an Oroboros Oxygraph 2K. Under all conditions tested, respiration was equivalent in soleus muscles isolated from rapamycin-treated or control mice (Figure 5). Thus, we were not able to detect any consequence of rapamycin treatment on mitochondrial function in vivo. This result contrasts somewhat with the findings of Cunningham et al., who were able to detect a decrease in respiration from soleus muscle homogenates following rapamycin treatment, albeit using a slightly higher dose in a different strain [29]. In order to test whether the permeabilization regimen might have masked any differences in oxygen consumption in our experiments, we also studied cultured myoblasts before and after the addition of digitonin. The decrease in oxygen consumption in intact cells that had been pre-treated with rapamycin remained readily apparent following the addition of digitonin and exogenous substrates (Figure 6), indicating that the permeabilization step would not have masked any differences in oxygen consumption between treated and control soleus muscles.

Discussion

Rapamycin impairs mitochondrial biogenesis and function in cultured muscle cells, which suggests a potential complication of targeting mTOR signaling as a means of promoting longevity in healthy humans. Interventions that promote longevity, including dietary restriction and genetic manipulations, often lead to increased mitochondrial biogenesis, and this has been hypothesized to contribute to the effects on lifespan [37-41]. Moreover, a growing body of evidence supports the concept that mitochondrial “reserve” capacity declines with age and is a major determinant of the ability to overcome stresses [42]. Thus, the inhibition of mitochondrial function by rapamycin despite extending longevity appears somewhat paradoxical. Our results suggest that rapamycin does not decrease mitochondrial protein expression or impair function in skeletal muscle in vivo. Although this conclusion resolves the paradox, it is also necessary to note that our findings imply that rapamycin extends life in the absence of any obvious increase in mitochondrial biogenesis, calling into question the concept that tissue mitochondrial content might be limiting for lifespan in mice.

We found that rapamycin clearly decreased the mRNA expression of mitochondrial genes in skeletal muscle, consistent with its effects in cell culture [29], and the observations of Blattler et al. [32]. The ultimate effect on protein levels might be predicted to be even more dramatic, since inhibition of mTORC1 can also have a negative influence on translation [43]. Therefore, it was somewhat surprising to see that mitochondrial protein expression and endurance were not significantly affected. Although these results could be reconciled by reduced turnover of mitochondria in the presence of rapamycin, we do not favor this explanation because inhibition of mTORC1, if anything, should increase degradation of cellular components via autophagy [44]. Indeed, during the preparation of this manuscript, Drake et al. [45] have provided direct evidence that mitochondrial protein turnover in skeletal muscle is unaffected by rapamycin treatment. Another possibility is that decreases in the mRNA abundance for mitochondrial genes are compensated by increased translation. Nuclear-encoded mitochondrial transcripts display distinguishing structural characteristics in the untranslated regions that can contribute to preferential translation under specific conditions, such as dietary restriction [46, 47], and translation of mtDNA-encoded transcripts is subject to many factors that remain incompletely understood [48]. A third possibility is that the assembly of oxidative phosphorylation complexes is better coordinated in the presence of rapamycin, which might prevent the degradation of partial or misfolded complexes [49]. The mechanism by which mitochondrial abundance is maintained in the presence of fewer transcripts will be an interesting area for further investigation.

It is known that total loss of mTORC1 in skeletal muscle, due to deletion of the essential subunit Raptor, or the catalytic subunit mTOR, causes profound mitochondrial dysfunction [50, 51]. Mitochondria are also affected in mice lacking YY1, the transcription factor that recruits mTORC1 to many mitochondrial genes [32]. The fact that rapamycin-treated mice are able to compensate for the reduction in mTORC1 activity suggests a number of possibilities. First, it may be that only a small fraction of normal mTORC1 activity is required to maintain mitochondrial function, and that higher or continuously delivered doses of rapamycin, or more potent mTOR inhibitors, will indeed cause overt mitochondrial dysfunction. In support of this, Cunningham et al detected a reduction in the respiratory capacity of soleus muscle (but not gastrocnemius muscle) using only a slightly higher dose than the present study (2.5 mg/kg in 6-week old Balb/c mice, [29]). Alternatively, mTORC1 activity may be more important for establishing a healthy pool of mitochondria during development than in adult mice, or the rapamycin-resistant functions of mTORC1 [52] may be sufficient to preserve mitochondria. Interestingly, mTORC1 activity is not an absolute requirement for healthy mitochondria, as mitochondrial function can be restored in mice lacking mTOR through the genetic or pharmacological stimulation of mitochondrial biogenesis pathways [53].

In these experiments, we have focused on skeletal muscle because the impairment of mitochondrial biogenesis by rapamycin has been previously demonstrated in myotubes [29], which is consistent with our own findings, and because mTORC1 inactivation has been shown to cause mitochondrial dysfunction in muscle in vivo [50, 51]. Although we also observed no significant effects of rapamycin on the expression of mitochondrial components in the liver, we cannot exclude the possibility that rapamycin may impair mitochondrial function in other tissues or cell types. For instance, neurons have very high energetic demands and are reliant on mitochondrial function [54]. Although the effects of rapamycin on neuronal mitochondria have not been measured directly, it is interesting to note that the drug has therapeutic effects in several neurodegenerative disease models [55], which might alleviate concerns that it could be inhibiting bioenergetics. Nevertheless, a comprehen-sive assessment of rapamycin's effects on mitochondria throughout the body is still lacking, and the possibility of impaired mitochondrial function in a specific tissue should be considered carefully in relation to any effects of rapamycin in vivo.

Developing therapeutic approaches to target the underlying aging process has the potential to delay or prevent multiple age-related diseases. Inhibition of mTOR signaling with rapamycin has provided some of the strongest evidence to date that this might be possible in mammals. Rapamycin itself is unlikely to have a net benefit in healthy humans due to its side effects, but it remains unclear whether more specific targeting of mTORC1 would alleviate enough of these concerns to constitute a viable strategy. Our present results argue that inhibiting mTORC1 sufficiently to prolong life does not lead to depletion of mitochondrial proteins in skeletal muscle or impair exercise performance in mice. However, many other detrimental side effects of rapamycin have been observed, and some of these may also be mTORC1-dependent. Given the potential benefit of slowing the changes that lead to age-related diseases in humans, understanding and overcoming these side effects should be a high priority for future studies.

Methods

Supplementary Materials

Supplemental Figures