1. Introduction
Osteoporosis and sarcopenia are the most important musculoskeletal diseases of the elderly [
1]. Both develop with aging and are risks for loss of activity of daily living (ADL) and quality of life (QOL), as well as mortality, in the elderly [
2,
3]. As the both the number and rate in the elderly increase in various populations, preventing these diseases before they develop is mandatory. To do this, factors predictive of future disease are needed. However, both osteoporosis and sarcopenia are multifactorial and not attributable to a single factor, making prediction difficult.
Osteoporosis is characterized by reduced bone strength and increased risk for fragility fractures, and morbidity rates reportedly increase with age [
4]. Known risks for osteoporosis include aging, menopause, smoking, excessive alcohol consumption, steroid use, genetic factors, and inflammatory or metabolic disease. Metabolites associated with low bone mass have been identified [
5,
6], but no longitudinal predictors have been reported. The fracture risks assessment tool (FRAX
®) can predict fragility fractures based on responses to questionnaires relevant to age, sex, weight, height, previous fracture, parent fractured hip, current smoking and alcohol habits, glucocorticoid use, rheumatoid arthritis, and secondary osteoporosis; however, no biological markers are used in this assessment [
7].
Sarcopenia is defined as a progressive loss of muscle mass with reduced muscle strength and/or function with age [
8,
9]. Known risks for sarcopenia include aging, immobilization, and cachexia [
10]. The European Working Group on Sarcopenia in Older People (EWGSOP) developed a practical clinical definition and diagnostic criteria for sarcopenia in 2010 [
8]. After that, the Asian Working Group for Sarcopenia (AWGS) provided appropriate diagnostic cutoff values for Asian populations [
9]. Osteoporosis and sarcopenia frequently simultaneously develop in the elderly [
11], suggesting that common predictors in addition to aging may exist for these diseases.
The present study was performed using Osteoarthritis/Osteoporosis Against Disability (ROAD) cohorts established in 2005. The ROAD study is a national, prospective study of musculoskeletal diseases that consists of population-based cohorts from several communities in Japan. Details of the cohort profiles have been reported elsewhere [
4,
12]. Briefly, between 2005 and 2007, a baseline database including the clinical and genetic information of 3040 residents (1061 men and 1979 women with a mean age of 70.3 years (standard deviation (SD): 11.0); 71.0 (10.7) years for men and 69.9 (11.2) years for women) was created. The study revealed that sarcopenia prevalence increases by 60 years of age and that four years after sarcopenia development, subjects become at risk for osteoporosis [
11].
Metabolites are the end products of various metabolic pathways, and their levels overall reflect individuals’ metabolomic status [
13,
14]. Metabolomic analysis can detect changes or differences in metabolite levels and is thus useful to diagnose diseases such as cancer and to detect low bone mass or bone status in human and animals [
15,
16,
17,
18]. Since osteoporosis and sarcopenia are metabolic diseases that disrupt bone and muscle homeostasis, metabolomic analyses could predict diseases before development.
Here, we undertook the metabolomic analysis of serum samples collected at the second follow-up of ROAD and found that levels of the amino acid glycine (Gly) were significantly higher in subjects not identified as having osteoporosis at that second survey but who had newly developed osteoporosis during the four-year follow-up, compared with osteoporosis-negative subjects at the second and third surveys, after adjustment for age, sex, and BMI. Similarly, levels of taurine were significantly lower in subjects newly defined as exhibiting sarcopenia at the third survey after adjustment for age, sex, and BMI. Thus, Gly and taurine could serve as predictors of the future development of osteoporosis and sarcopenia, respectively.
3. Discussion
Osteoporosis and sarcopenia are age-related diseases that are increasing in the number of cases and prevalence in developed countries. Both worsen ADL and QOL in the elderly. These conditions may be correlated and influence each other. In support of this idea, we found that among patients with newly developed osteoporosis, 15.2% also had newly developed sarcopenia. In contrast, 5.8% of subjects enrolled in the study who did not develop osteoporosis newly developed sarcopenia (
Table 1). By contrast, 17.2% of patients with newly developed sarcopenia also newly developed osteoporosis, while 6.7% of subjects who did not newly develop sarcopenia newly developed osteoporosis (
Table 4). Moreover, we found that elevated serum Gly levels predicted new osteoporosis development within four years but not the development of sarcopenia, while decreased serum taurine levels were associated with the new development of sarcopenia but not osteoporosis in a four-year period. Gly levels were higher in subjects with newly developed sarcopenia compared with those who did not, but differences were not statistically significant (non-SP vs. new SP: 494.3 ± 118.0 vs. 499.8 ± 138.6, respectively;
p > 0.05). Moreover, taurine levels were significantly higher in subjects that newly developed osteoporosis relative to those who did not (non-OP vs. new OP: 252.0 ± 80.0 vs. 283.1 ± 126.1, respectively;
p = 0.036). However, that difference also was not statistically significant after adjustment for age and BMI. Thus, although osteoporosis and sarcopenia are closely related diseases, methods to predict them likely differ overall.
Five subjects developed both osteoporosis and sarcopenia (the “both” group), but their Gly levels were comparable to those seen in individuals who developed osteoporosis but not sarcopenia (new OP alone) (both vs. new OP alone: 484.3 ± 93.7 vs. 561.0 ± 145.4, respectively;
p > 0.05). Likewise, taurine levels were similar between the “both” group and subjects who developed sarcopenia but not osteoporosis (new SP alone) (both vs. new SP alone: 242.5 ± 49.7 vs. 221.3 ± 47.0, respectively;
p > 0.05). In contrast, levels of the amino acid citrulline were significantly higher in the “both” group relative to new SP alone (57.8 ± 22.7 vs. 41.2 ± 10.3, respectively;
p < 0.05). Citrulline supplementation reportedly promotes increased skeletal muscle mass [
19], but at present, the mechanisms underlying increased citrulline levels in subjects who develop both osteoporosis and sarcopenia remain unclear.
Bone and muscle interact [
20,
21,
22,
23], and myokines produced from muscle regulate bone homeostasis [
24]. For example, irisin, a myokine secreted from muscle with exercise, controls bone metabolism [
25,
26]. Extracellular matrix vesicles produced in and secreted by muscles reportedly regulate bone homeostasis [
27]. Insulin-like growth factor 1 (IGF1), which is produced by osteoblasts and stored in extracellular matrix proteins in bones, is a bone-remodeling factor [
28] that is released upon osteoclastic bone resorption and stimulates subsequent osteoblastic bone-forming activities. IGF1 also stimulates anabolic pathways and inhibits catabolic pathways in muscles [
29]. Previously, we demonstrated that reduced IGF1 levels promoted decreased muscle mass in adult mice [
30]. Indeed, here we showed lower serum IGF1 levels in the new- versus non-SP cases (
Table 5). Transforming growth factor beta 1 (TGFβ1) is also a bone-remodeling factor released from the bone matrix [
31]. TGFβ1 is activated by osteoclastic activity and stimulates bone formation by osteoblasts [
31]. On the other hand, TGFβ1 activity promotes muscle atrophy [
32]. Thus, the activities of bone and muscle are tightly linked, and the disruption of that interaction could promote the development of both osteoporosis and sarcopenia. Indeed, some subjects assessed here developed osteoporosis and sarcopenia concomitantly, although we showed that factors predictive of these conditions differ.
Several studies combining cohort and metabolomic analysis have been reported. For example, gut microbiome α-diversity is reportedly explained by a subset of 40 plasma metabolites in two independent human cohort blood samples [
33]. Metabolite profiles are also reported to predict future diabetes [
34]. Low plasma lysophosphatidylcholine (LPC) 18:2 is reportedly a predictor of decline in gait speed in older adults [
35]. Here, although small sample size was a limitation of our study, we combined comprehensive metabolome analysis with a longitudinal cohort study to show for the first time that Gly and taurine levels in sera are predictors of osteoporosis and sarcopenia, respectively.
Glycine (Gly) is a non-essential amino acid, and increased glycine levels are reportedly seen in male idiopathic osteoporosis patients [
36]. Moreover, in humans, supplementation with six amino acids including glycine is associated with a higher BMD in the spine and forearm [
37]. We found that Gly levels were significantly and negatively correlated with bone mineral density (BMD) at either the lumbar spine L2–4 levels or the femoral neck (
r = −0.1818 (
p < 0.05) or −0.2134 (
p < 0.05), respectively) at the second survey, supporting the idea that increased Gly levels are associated with decreased BMD. Gly is the most abundant amino acid residue in collagen [
38], and collagen degradation products such as collagen cross-linked N-telopeptide (NTx) are frequently detected at high levels in the blood and urine of osteoporosis patients. Taurine is a non-essential amino acid synthesized from methionine and cysteine by an cysteine sulfinic acid decarboxylase (CSD). Older adults with sarcopenia reportedly exhibit higher levels of several amino acids in serum, including taurine [
39]. Sarcopenia is known to be associated with undernutrition [
40], and nutritional supplementation with taurine reportedly counteracts the development and progression of sarcopenia in human subjects [
41]. We found that taurine levels were positively correlated with walk speed, one of the criteria used to assess sarcopenia, but that association was not statistically significant (
r = 0.0378 and
p > 0.05). The mechanisms underlying the changes in the serum levels of these amino acid levels before osteoporosis or sarcopenia development remain unclear; however, we conclude that the monitoring of these levels in individuals could be predictive of musculoskeletal disorders.