Early-stage Alzheimer’s disease: are skeletal muscle and exercise the key?

Abstract

Alzheimer’s disease (AD) is the most common form of dementia affecting approximately 6.5 million people in the United States alone. The development of AD progresses over a span of years to possibly decades before resulting in cognitive impairment and clinically diagnosed AD. The time leading up to a clinical diagnosis is known as the preclinical phase, a time in which recent literature has noted a more severe loss of body mass and more specifically lean muscle mass and strength prior to diagnosis. Mitochondria dysfunction in neurons is also closely associated with AD, and mitochondrial dysfunction has been seen to occur in skeletal muscle with mild cognitive impairment prior to AD manifestation. Evidence from animal models of AD suggests a close link among skeletal muscle mass, mitochondria function, and cognition. Exercise is a powerful stimulus for improving mitochondria function and muscle health, and its benefits to cognition have been suggested as a possible therapeutic strategy for AD. However, evidence for beneficial effects of exercise in AD-afflicted populations and animal models has produced conflicting results. In this mini-review, we discuss these findings and highlight potential avenues for further investigation that may lead to the implementation of exercise as a therapeutic intervention to delay or prevent the development of AD.

SKELETAL MUSCLE HEALTH AS AN EARLY INDICATOR OF AD

Alzheimer’s disease (AD) is the prevailing form of dementia accounting for over 60% of all dementia cases with nearly 6.5 million adults diagnosed in the United States, which is projected to triple by the year 2050 (1, 2). AD develops along a continuum, beginning with a long preclinical phase, in which little to no cognitive abnormalities are present, before manifesting as mild cognitive impairment (MCI), then eventually, dementia (3, 4). Although substantial research efforts have been conducted to understand the pathology and risk factors for AD, no effective treatments have been established (5). The preclinical phase of AD precedes development of cognitive decline by several years or possibly decades (6) and, therefore, may provide an early window for interventional treatment. A better understanding of the phenotypic characteristics of preclinical AD and their mechanistic underpinnings could lead to innovation of new strategies to delay or prevent development of AD.

AD is an age-associated disease of the brain, but evidence of accelerated weight loss in the years before AD diagnosis compared with those that did not develop AD (7) suggests indicators of AD may manifest in peripheral systems during the preclinical phase. This notion substantiated by evidence of weight loss in individuals before AD diagnosis is due to loss of lean mass and that declines in lean mass correlate with reductions in brain volume (8). Although loss of skeletal muscle mass with age is a common characteristic of aging, known as sarcopenia, other studies have since shown loss of lean mass in patients later diagnosed with AD is more severe than in those that develop sarcopenia but remain cognitively intact (9–11). In addition, abnormal motor functions, such as loss of muscle strength (i.e., dynapenia) and altered gait, are all associated with an increased risk for developing AD (12–15). Furthermore, genetic predisposition to early development of AD due to possession of an allele of apolipoprotein, APOEε4 (16), also coincides with more pronounced muscle dysfunction with age (17, 18). Alternatively, possession of APOEε2 allele, which is associated with milder manifestation of AD, has been shown to be associated with attenuated declines in muscle function (19). Although the pathogenesis of AD is multifactorial and can be independent of a genetic predisposition, declines in skeletal muscle health are associated with the preclinical phase of AD and, therefore, may play an underappreciated role in AD etiology.

It is not entirely clear whether skeletal muscle abnormalities are merely reflective of the early AD pathology or whether skeletal muscle could be an interventional target during the preclinical phase of AD to slow progression of the disease. However, recent evidence from mouse models of AD suggests that manipulation of muscle may indeed have benefits on the development of AD and cognitive health. Inducing muscle atrophy through casting of hindlimbs in 5xFAD mice, a mouse model of AD, accelerates cognitive decline (20), suggesting muscle loss is tied to the rate of cognitive decline in AD. APP/PS1 mice, another mouse model of AD, have increased expression of myostatin, a myokine that promotes expression of atrophy-related genes in skeletal muscle (21). Interestingly, shRNA-mediated knockdown of myostatin in APP/PS1 mice increased grip strength and muscle mass, as well as improved cognition (21). These results in mouse models of AD suggest a link between muscle and cognitive decline with AD and support the notion that interventions that promote muscle health may also mitigate cognitive impairments associated with the development of AD.

Poor Mitochondrial Quality Is a Shared Phenotype between AD Neuropathology and Muscle

It has been postulated that the decline in skeletal muscle mass and function during the early stages of AD suggests the underlying mechanisms between AD neuropathology and muscle may be shared (8, 22). Mitochondria meet a large portion of the energetic demand of skeletal muscle and neurons by oxidizing nutrient substrates to generate ATP. Impaired mitochondrial respiration in synaptic mitochondria has been observed in both AD-diagnosed humans and mouse models of AD (23). In skeletal muscle, the capacity for mitochondria to produce ATP declines with age (24, 25) and evidence in mice, primates, and humans suggests that mitochondrial dysfunction parallels sarcopenia (24, 26) or may even precede it (27, 28), as well as the associated mobility impairments (29). Gait speed, a common measure for assessing functional decline and risk for AD (15), is limited by skeletal muscle bioenergetics primarily through mitochondrial respiration (29, 30). Therefore, impaired mitochondrial function associated with AD may also be present outside of neurons and affect peripheral tissues, such as skeletal muscle. Indeed, Morris et al. (31) recently found impaired mitochondrial respiration in skeletal muscle of humans with mild cognitive impairment (MCI), a clinical precursor to AD, compared with age-matched cognitively intact subjects. Similar derangements in mitochondrial function before overt cognitive decline have been seen in some AD mouse models as well (32, 33). Furthermore, dysfunctional mitochondria may produce increased reactive oxygen species (ROS) and can disrupt the oxidative balance, leading to greater oxidative stress. Increased oxidative stress in neurons has been seen to further drive AD pathology (34) and increased mitochondrial ROS production may further contribute to muscle loss due to oxidation of contractile proteins (35). Thus, mitochondrial dysfunction appears to be a shared characteristic of both neurons and skeletal muscle, which is present early in the development of AD.

Mitochondrial function is maintained by coordination between quality control processes that synthesize new mitochondria (biogenesis), remodel the reticulum (fission/fusion), and degrade damaged or dysfunctional regions of the reticulum (mitophagy; 36). Mutations in mitochondrial DNA are observed in the brain early in the development of AD pathology (37), which could lead to widespread mitochondrial dysfunction as the reticulum turns over. Accumulation in dysfunctional mitochondria would presumably necessitate mitophagy to ensure mitochondrial health. However, in human hippocampus from patients with AD, mitophagy events, as evidenced by the colocalization of lysosomal marker LAMP2 with mitochondrial protein TOM20, have been shown to be decreased by as much as 50% (38). In addition, expression of mitophagy-related proteins PINK1, BCL2L13, and BNlP3L and unphosphorylated autophagy-activating proteins TBK1 and ULK1 was all decreased in the hippocampus of subjects with AD, despite the increased activation of upstream AMPK (38), which phosphorylates Ulk1 to promote mitophagy (39). Pharmacologically inducing mitophagy using two mitophagy-inducing agents, urolithin A (UA) and actinonin (AC), restores cognition and improves AD-like amyloid beta (Aβ) plaque-forming pathologies in Caenorhabditis elegans and mouse models of AD (40), suggesting upregulation of mitophagy as having a potential therapeutic benefit for AD-associated mitochondrial dysfunction. Whether mitophagy, and/or other mitochondrial quality control processes, may underly mitochondrial dysfunction in skeletal muscle during early-stage AD is unknown.

Impaired mitochondrial quality in neurons may also be partly caused by one of the hallmarks of AD; accumulation of amyloid beta (Aβ) senile plaques (41). Aβ plaque load has been seen to progressively increase with age in neuronal mitochondrial pools in AD rodent models and was found to be 70% higher in postmortem AD human hippocampal mitochondria. In brain mitochondria isolated from rats, exposure to micromolar concentrations of Aβ resulted in impaired respiration (42). However, formation of intracellular Aβ and its association with decline in mitochondrial function in vivo are not fully understood. One possible explanation as to how Aβ forms within the cell and disrupts mitochondria function is through formation along the endoplasmic reticulum (ER) and mitochondrial-associated membranes (MAMs), which form junctions with mitochondria to allow intracellular communication, phospholipid transport, and calcium homeostasis between the mitochondria and ER. Within the MAM, lipid rafts, which are specialized detergent-insoluble domains, containing γ-secretases and amyloid precursor protein (APP) make up part of the junction and it is thought that amyloidogenic splicing of APP within the MAM may be a source for plaque formation in a cell leading to mitochondria impairment and the unfolded protein response (UPR) leading to neuronal loss through apoptosis as damage accrues (43). These amyloidogenic incidents could further implicate Aβ-induced mitochondrial function in the preclinical phase of AD; however, this is an area of the preclinical phase that needs to be further explored.

It is unclear as to how skeletal muscle mitochondria function may correspond to or affect neuronal mitochondrial function and what role Aβ may have in skeletal muscle function in early AD. In the context of inclusion body myositis, an age-related muscle disease characterized by progressive weakness, inflammation, and amyloid characteristics, Aβ accumulation in muscle leads to disrupted mitochondria function, decreased mitochondrial membrane potential, abnormal mitochondria morphology, and reduced respiration (44, 45). There is a paucity of evidence demonstrating effects of Aβ on skeletal muscle mitochondria or whether it contributes to the preclinical progression of AD. Yet skeletal muscle mitochondrial dysfunction is already present with MCI (31), which may allude to the early accumulation of Aβ in the brain resulting in peripheral dysfunction. The numerous roles that mitochondria have in skeletal muscle and neuronal health suggest that interventions that promote mitochondrial health across all bodily systems, such as exercise, may be beneficial for those at risk for developing AD.

CAN EXERCISE WARD OFF AD?

Mitochondrial health, particularly of skeletal muscle, can be maintained into advancing age through regular exercise (46). Exercise promotes mitophagy in skeletal muscle (39), which may be beneficial for AD (40). Similarly, exercise also promotes mitochondrial respiration, biogenesis, and remodeling in neurons, as well as hippocampal neurogenesis and increased blood flow to the brain (47). In combination with benefits of regular exercise on cognition (48), maintained mitochondrial health in both neurons and skeletal muscle suggests exercise may be a therapeutic strategy for AD. Additional pathologies, such as mid-life obesity, mid-life hypertension, and diabetes mellitus, are all risk factors for AD that are modifiable through the use of exercise (49). Indeed, physical activity has been identified as a modifiable risk factor for AD (49) and self-report data suggest engaging in regular exercise may reduce the risk of developing AD (50). Exercise, therefore, may be effective in preventing or delaying AD onset during the preclinical phase by improving mitochondrial function in both muscle and neurons.

Support in the literature, however, for exercise as an effective interventional strategy for AD is less clear. Data from two multidecade-long studies suggest that although aerobic exercise is beneficial for preventing incidence of some forms of dementia, there is no preventative effect of exercise on AD (51, 52). Recently, Yu et al. (53) demonstrated that in adults diagnosed with AD, 6 mo of cycling resulted in attenuated declines in memory and executive functions, however, there were still significant declines in attention-processing speed and language. After 12 mo, those in the exercise group had continued declines in all previous measures (53), suggesting that although there may be initial cognitive benefits with exercise, these are not maintained long-term. Elsewhere, aerobic exercise has been shown to improve functional independence in performance of activities of daily living (ADL) in individuals with early AD, but no benefits in cognition were noted (54). However, secondary analysis revealed that in the intervention group, those who improved in their cardiorespiratory fitness levels did improve in memory performance (54), which may allude to a disease stage-dependent response to exercise in patients with AD. In sum, it appears that aerobic exercise may have limited benefit to prevent the progression of AD, or a more personalized approach in exercise prescription that accounts for disease stage is required. However, the underlying mechanism(s) for the disconnect between aerobic exercise and AD is unclear.

Evidence from studies using a voluntary wheel running model of aerobic exercise in AD mouse models suggests that volume of exercise may play a role in the therapeutic efficacy of aerobic exercise in AD. Improved cognition, increased hippocampal neurogenesis, and reduced Aβ plaque load were reported in 5xFAD mice when access to voluntary running wheels was restricted to 3 h a day (55). In contrast, studies using 5xFAD mice, where access to running wheels was unlimited, reported no benefits on cognition, Aβ plaque burden, or neuroinflammatory markers (51, 56). The 5xFAD model develops AD-like pathology at a relatively young age (4–6 mo of age). Importantly, mice in the abovementioned studies all began wheel running between 2 and 3 mo of age, which continued for either 4 mo [Choi et al. (55)] or 6 mo [Hansson et al. (51) and Svensson et al. (56)], and thus, well into the age when AD-like pathology is present. These findings suggest a tipping point in exercise volume where a possible maladaptive response to exercise occurs, thus determining whether development of AD pathology is mitigated or not as a result of exercise. However, neither Choi et al. nor Hansson et al. reported running volume during wheel exposure and although Svensson et al. reported running volume, no nontransgenic mice were used for comparison. Therefore, future studies should consider the response to exercise itself in the context of neuropathological changes that occur with AD. Furthermore, it is unclear what relationship the divergent response to exercise access, and presumably running volume, has to the underlying mitochondrial dysfunction in both skeletal muscle and brain that coincides or may even precede AD pathology (31, 42). Whether mitochondrial health (or lack thereof) in muscle and/or neurons dictates the adaptive versus maladaptive response to aerobic exercise before the development of the AD phenotype is an important area for future investigation.

Alternatively, resistance exercise promotes skeletal muscle mass and may even improve skeletal muscle mitochondria function (58). Resistance exercise has been explored in aging human (>65 yr old) populations to aid in activities of daily living and fall risk mitigation (59). In cognitively intact individuals, resistance exercise has cognitive benefits similar to aerobic exercise (60). Altogether, resistance exercise may be effective in slowing the progression of AD. Recently, in subjects with MCI, 6-mo resistance exercise improved cognition and offered some protection from brain atrophy and Aβ plaque burden (61). In the 3xTG mouse model of AD, 4 wk of weighted ladder climbing, meant to mimic resistance exercise-like training, preserved cognitive function, increased hippocampal synaptic proteins, and reduced Aβ plaque load in both the frontal cortex and the hippocampus (62). Similarly, weighted ladder climbing decreased amyloid plaques and improved cognition in the APP/PS1 transgenic AD mouse model (63). As findings in AD mouse models have suggested that manipulation of muscle mass can dictate the rate of cognitive decline (20, 21), preserving muscle mass through resistance exercise during the preclinical phase of AD may be a viable means to slow progression of the pathology, though the molecular events leading to beneficial outcomes in AD remain poorly understood. Investigation into resistance exercise and its potential benefits on early muscle declines in AD and related neuropathology should be expanded upon.

CONCLUSIONS

Although AD is largely understood as a disease of the brain, it is becoming increasingly clear that the disease has implications for peripheral systems, such as skeletal muscle, during the early stages of the pathology. Accumulating evidence for impaired mitochondrial health as a shared characteristic of the early stages of AD in skeletal muscle and neurons suggests intervention strategies that target mitochondria and muscle health (such as exercise) have therapeutic potential. We have discussed findings that, interestingly, suggest that the therapeutic benefit of aerobic exercise on AD is not conclusive. Understanding potential mechanisms that underlie apparent volume thresholds for exercise efficacy may move toward effective implementation of exercise in those at risk of developing AD. However, our discussion of these findings highlights the challenges associated with differing study designs and comparing outcomes from single and mixed modalities of exercise, as well as the wide variety of animal models used to recapitulate AD pathology. A holistic, integrated approach that considers mitochondrial health in peripheral systems, like skeletal muscle, in context with the known roles of mitochondria in AD neuropathology could break new ground in this challenging and detrimental disease.

GRANTS

Dr. Drake is funded by National Institutes of Health National Institute on Aging R00AG057825 and start-up funds from the Virginia Tech College of Agriculture and Life Sciences.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.H.B. and J.C.D. conceived and designed research; M.H.B. and J.C.D. drafted manuscript; M.H.B. and J.C.D. edited and revised manuscript; M.H.B. and J.C.D. approved final version of manuscript.