Musculoskeletal Deficits and Cognitive Impairment: Epidemiological Evidence and Biological Mechanisms

Sarcopenia is generally defined as age-related progressive loss of muscle mass and function, although the working definition varies among different groups [29, 30]. Thus, several operational definitions of sarcopenia are in use, including those provided by the European Working Group on Sarcopenia in Older People [31], the United States Foundation for the National Institutes of Health (FNIH) [32], the Asian Working Group for Sarcopenia (AWGS) [33], and the Sarcopenia Definition and Outcomes Consortium [34]. Therefore, the prevalence of sarcopenia reported in the literature varies substantially within and across geographic areas, due to differences in the criteria and cutoff points applied [35, 36]. Researchers conducting the Geelong Osteoporosis Study (GOS) [37] described the consequences of applying different criteria and cutoff points (international and population-specific) in a homogeneous sample for prevalence estimates for the Australian population [38, 39]. The main finding was that, across all the definitions, the prevalence of sarcopenia increased with increasing age; however, the varied criteria and cutoff points resulted in inconsistent case ascertainment. The GOS explored the relationship between several muscle parameters and overall cognitive function, specific cognitive domains, and cognitive impairment among men [40]. The findings suggested that muscle parameters, especially muscle function [41••], muscle quality [42•], and muscle density [43••], are associated with certain cognitive domains (including working memory, attention, and information processing speed) independent of age, physical activity levels, education, and lifestyle factors. However, these associations were not detected for all the cognitive domains tested. Similar results on the association between dynapenia (age-associated muscle strength loss not caused by neurologic or muscular diseases) and low cognition were reported in female GOS participants [44•].

In summary, inconsistencies in studies associating muscle health with cognition could be due to differences in criteria used to define sarcopenia, sarcopenic obesity, and cognitive deficits. Evidence for mechanisms linking muscle health with certain brain functions also remain unclear.

Sarcopenic obesity is a condition characterized by concurrent high adiposity levels and low muscle mass and function with advanced age [45]. The association between sarcopenic obesity and cognition was assessed in 353 community-recruited USA participants aged 40 years and over [46]. Body composition was measured by bioelectrical impedance analysis. Obesity was determined by body mass index (BMI) and fat percentage. Global cognition was assessed using the Montreal Cognitive Assessment. Specific cognitive domains in verbal fluency and mental flexibility were assessed using the animal naming test (participants are asked to name as many animals as they could within one minute). Visual search speed, scanning, and processing speed were assessed using Trail Making A. The authors recommended that sarcopenia and sarcopenic obesity should be regarded as clinical indicators of cognitive impairment, listing potential mechanisms that explain sarcopenic obesity and cognitive deficits association, including low-grade chronic inflammation, oxidative stress, and insulin resistance; however, no biomarker data was included in the report.

The association between sarcopenic obesity and cognitive impairment was examined in 948 community-based Chinese participants aged 60 years and over (51% female) [47]. Body composition was measured by bioelectrical impedance analysis and cognition by MMSE. Sarcopenia was defined using the AWGS criteria, and obesity was determined by body fat percentage (fat mass/weight). Six percent of participants were identified as sarcopenic obese. This study reported an independent association between sarcopenic obesity and cognitive impairment and proposed that the mechanisms underlying the link between muscle, fat, and cognition were inflammation, insulin resistance, and decreased growth hormone secretion, but these hypotheses were not tested.

The association between sarcopenic obesity and cognitive performance was also examined in 1235 Singaporean patients (~48% female) aged 45 years and over with type 2 diabetes; all were attending diabetes care [48]. Body composition was assessed using bioelectrical impedance analysis, and the Repeatable Battery for the Assessment of Neuropsychological Status and MMSE were applied for assessing cognition. This study identified an association between sarcopenic obesity and poor cognitive performance, particularly in the domains of memory and language. The authors acknowledged that the mechanisms underlying these dual deficits in brain and body were not fully understood; however, it is likely that in these patients, abnormal levels of insulin affect amyloid β metabolism, which controls neuronal function.

Further, in a longitudinal, population-based study in the USA, the association between sarcopenic obesity and incident cognitive function was determined in 5822 participants (~56% female) aged 65 years old and over, without cognitive impairment at baseline [49]. They adopted the FNIH definition of sarcopenia, and defined obesity by BMI. Cognition was assessed using AD-8 score or immediate/delayed recall, orientation, clock-draw test, or date/person recall. At baseline, 12.9% of subjects were identified as having sarcopenic obesity; 21.2% were identified with cognitive impairment at follow-up. Sarcopenia obesity and sarcopenia alone were significantly associated with higher risk of cognitive impairment. On the other hand, a study including 542 participants 21–90 years recruited from the Chinese community produced contrasting findings [50]. The researchers obtained body composition data using dual-energy x-ray absorptiometry (DXA) and defined obesity as the upper two quintiles of fat mass index. They employed a sarcopenia definition based on the 2019 AWGS criteria. Cognitive impairment was determined using the Repeatable Battery for the Assessment of Neuropsychological Status. This battery included 12 tests to assess immediate memory, visuospatial/constructional abilities, language, attention, and delayed memory. Sarcopenia and sarcopenic obesity were not associated with cognitive impairment, but obesity alone and muscle function (grip strength or gait speed) were. The authors speculated that insulin resistance underpins the association.

During contraction, skeletal muscle releases molecular factors that may affect cognitive function, such as BDNF, a neurotrophin required in adults for the maintenance of synaptic connections and adaptive neuronal plasticity, regulating cognitive processes such as learning and memory [79]. A study showed that after long-term voluntary exercise, adult male mice exhibited a lactate-dependent increase in hippocampal BDNF [80]. Interestingly, lactate (a metabolite released from muscle during exercise) was responsible for improvement in both learning and memory in these mice, and the induction of the hippocampal BDNF expression was found to be dependent on the sirtuin 1 (SIRT1)/peroxisome proliferator–activated receptor-γ coactivator 1-α (PGC1α)/fibronectin type III domain-containing 5 (FNDC5) pathway [80]. Furthermore, SIRT1 knockout male mice were found to be more anxious than wild-type mice and suffering cognitive impairment characterized by reduced learning abilities and memory [81]. In a rat model of AD, direct intervention in the hippocampus with the 42 amino acid form of amyloid β (Aβ1–42) resulted in cognitive impairment by suppressing PGC1α/FNDC5/BDNF signaling [82]. However, the cognitive impairment was partially reversed by moderate physical activity, revealing a recovered PGC1α/ FNDC5/BDNF pathway [82].

In addition, BDNF is released in response to muscle contraction [83], and percutaneous electrical stimulation of the hindlimb muscles in a rat model of spinal cord injury significantly increased BDNF levels in both the anterior tibialis and the vertebral column [84]. Importantly, deletion of BDNF in skeletal muscle in mice resulted in a fatigue-resistant muscle phenotype, migrating from fast to slow muscle fibers in glycolytic muscles tibialis anterior and extensor digitorum longus [85]. In contrast, BDNF overexpression increased the glycolytic and fast fiber phenotype of the muscles [85]. This is consistent with clinical findings, since BDNF levels in skeletal muscles induced by controlled physical activity were found to be correlated positively with muscle phenotypic changes favoring type II muscle fibers (fast and glycolytic) [86]. Moreover, serum BDNF was increased in sedentary subjects 1 h after training, but this was not found in trained young and adult patients, suggesting the relevance of physical conditioning when assessing the effect of training on BDNF induction [86]. In addition, the decrease of BDNF after training was correlated with improvement in cognitive processes such as visuospatial and verbal skills (measured using a before-and-after Addenbrooke’s Cognitive Examination-Revised test) [86]. Consistent results have been obtained in experimental models, involving young and aged rats, suggesting a role of the PGC1α/ FNDC5/BDNF pathway in the protection of cognition from a musculoskeletal health approach [87].

BDNF expression can also be affected by altered mastication. In growing mice receiving a soft diet, learning and memory processes were impaired, compared to mice eating standard chow, by the apparent decrease in masticatory function [88]. In addition, BDNF expression in the hippocampus of mice receiving soft chow was decreased compared to those fed with standard chow, whereas no changes on the BDNF receptor were found in either group [88]. Moreover, the reduced expression of BDNF was consistent with a decrease on the synapses, leading to degraded neuronal structure and therefore neural function [88]. These findings are consistent with those of a recent systematic review of animal studies that identified a relationship between altered mastication and cognitive impairment characterized by decreased expression of BDNF in the hippocampus, decreased synapses, low performance in behavioral evaluations, and diminished memory and spatial location [89••]. Interestingly, male rats fed with standard chow exhibited higher expression of BDNF hippocampus than those fed with either soft or hard chow [90]. This result poses the question of whether reduced and increased masticatory functions are risk factors for cognitive impairment, suggesting a focus on clinical conditions ranging from loss of teeth (and therefore decreased masticatory function) to masticatory muscle parafunction. Short-term exposure of young adult male mice to a soft diet results in dysregulated expression neurodegenerative condition–related genes such as TREM2, DAP12, APOE, and CD33 in the microglia, suggesting that soft diet has an immunomodulatory role as a risk factor for cognitive impairment [91]. Also, mastication on one side only has been shown to affect BDNF gene expression in the hippocampus, with cognitive impairment evaluated using the Morris Water Maze test in young adult male mice [92•]. In addition, using the MWM test, it was determined that reduced physical activity and reduced masticatory function in adult and aged mice affected their memory and learning skills, but these were restored when normal mastication was enabled [93]. The reduction of the branches in the astrocytes of the group with reduced physical activity and masticatory function [93] is intriguing, suggesting the need for more research into the role of the mastication as a neuroprotective musculoskeletal activity. For instance, in humans, masticatory function has been evaluated as a neuroprotective activity based on its clinical correlation with increased brain blood perfusion [27].

Irisin is a myokine released in response to physical activity, downstream of PGC1α/FNDC5 pathway activation, after FNDC5 cleavage [94, 95]. Irisin stimulates BDNF expression in the hippocampus [96], and is believed to mediate the effect of physical activity on BDNF expression [94, 95]. Continuous physical training increases BDNF and Irisin serum levels, with benefits for cognitive performance, measured as the working memory (part of short memory that is a cognitive ability that can hold the information in mind for executing cognitive function tasks [97]) in adults aged 50 to 70 years [98]. In male mice, injection of Irisin to the hippocampus after physical restraint improved the cognitive response to memory tasks. However, female mice did not benefit from Irisin administration, suggesting a sex-dependent effect [99]. Additionally, in FNDC5 knockout mice, absence of Irisin diminished cognitive skills (spatial and learning memory) after voluntary physical activity when compared with wild-type animals [100]. Interestingly, the same study showed that systemically administered Irisin was able to cross the blood–brain barrier and partially rescue cognitive impairment in two AD mouse models [100]. The mechanism of Irisin action remains to be fully understood. However, the use of AD mouse models has revealed a potential role of Irisin in neuroinflammation control [100] and cognition improvement after physical activity, with increased levels of FDNC5, BDNF, and IL-6 [101]. These findings can be compared and contrasted with clinical data about the correlation between Irisin levels in serum [102] or cerebrospinal fluid [103] and neurodegenerative/inflammatory biomarkers in conditions that affect cognition [102, 103].

Osteokines are molecules released by osteoblasts and osteocytes. Among them, the osteoblast-derived protein osteocalcin or bone γ-carboxyglutamic acid (Gla) protein has been proposed to impact cognition [76, 104, 105]. Compared to baseline measurements, serum levels of osteocalcin increase after intense controlled physical activity to a similar extent in women and men [106]. A correlation study in young men exposed to reduced physical activity followed by a single session of high intensity training found that serum levels of both BDNF and undercarboxylated osteocalcin—the hormonally active form of the protein—were increased [107]. Irisin serum levels were also elevated after the intervention [107]. However, whether there is a molecular link among these molecules remains to be determined.