To the best of our knowledge, this was the first population-based study that investigated the relationship between the serum metabolome and the longitudinal reduction rates in hand grip, knee extension, and leg muscle strength over 10-year follow-up period in a large sample size of older adults that were randomly selected from a general population. The reduction rate of the hand grip and leg muscle strength in the current study were comparable with previous studies [18,19,20], but the reduction rate of the knee extension was lower in the current study than the previous report [21], which might be due to the difference in study populations, follow-up times, study designs, and measurement methods.

We reported that elevated serum concentrations of dimethylarginines, especially ADMA, were significantly associated with the longitudinal reduction rate in the hand grip and knee extension strength. Interestingly, the elevated ADMA and total dimethylarginine blood levels were also associated with functional impairments including run errands and shopping, vacuuming, and bathing over the 10-year follow-up period. We also found that elevated uric acid concentration was significantly associated with the decline rate in leg muscle strength over a 10-year follow-up period. The increased uric acid level was also associated with the longitudinal complication of putting on socks, and climbing up five steps over 10-year follow-up period.

However, there is still much more in common between a leg muscle strength and a knee extension, than between hand grip and knee extension, it is interesting that both hand grip and knee extension had negative correlation with ADMA and total dimethylarginine blood levels, while leg muscle strength decline had a significant association with the elevated blood level of uric acid. This might be due to the hand grip and knee extension common functionalities and mechanisms. Wrist and finger flexion are mostly initiated by the muscles in the anterior and posterior compartments of the forearm (extrinsic muscles), and only the thin tendons of these muscles are found directly in the hand. The flexor tendons of the forearm anterior compartments run in the anterior of the hand through the palms to the tips of the fingers to facilitate flexing of the wrist and fingers leading to wrist flexion and hand grip force production [22]. Moreover, the extensor tendons of the forearm posterior compartments used for wrist extension and hand grip relaxation run through the back of the hand to the figures [23]. While the extrinsic muscles of the hand are responsible for stronger movements of the wrist and hand, the intrinsic muscles of the hand have no direct effect on wrist action but can contribute to grip force via the extensor mechanism [24]. The intrinsic muscles produce finer, more controlled movements and play important roles in rotating fingers toward the palm to maintain and improve hand grip [25]. Thus, hand grip strength is relatively specific for the muscles in the anterior compartment of the forearm that are involved in finger/wrist flexion. Similarly, while the quadriceps femoris in the anterior compartment of the thigh are activated to extend the knee in the knee extension strength test, the hamstrings in the posterior compartment of the thigh are predominantly involved to flex the knee [26]. Also, while the hand grip and knee extension tests are mostly used to assess the upper and lower body’s muscle strength and power, leg muscle strength test is implemented to evaluate the body balance and risk of fall in older adults, because balance consists of multiple body systems including the ability to align different body segments and to generate multi-joint movements to effectively control body position and movement [27]. Morover, As neural decrements present earlier than loss of strength, this would likely affect more complex movements more drastically than measures of specific, relatively isolated muscle groups performed in a stable setting. Indeed, the change in leg strength was more pronounced than grip strength.

Data on ADMA and muscle strength are sparse in the literature. In the cross-sectional study of 550 individuals [28], high serum level of ADMA was associated with lower hand grip, quadriceps strengths, and slower gait speed [28]. Cancer patients [29] were found to have higher levels of ADMA in the skeletal muscle compared with healthy controls, suggesting that increased levels of ADMA may contribute to impaired muscle protein synthesis in cancer cachexia. In the longitudinal setting, our data documented that the elevated ADMA level was associated with the reduction of muscle strength over time, especially hand grip strength and knee extension. Further studies to investigate the causal relationship between ADMA and muscle strength reduction is warranted. The increased blood concentration of the total dimethylarginine was also associated with the strength reduction in the hand grip and knee extension over the follow-up period. However, the effect size was similar to that of ADMA, suggesting that the association was most likely driven by ADMA rather than symmetric dimethylarginine (SDMA).

Dimethylarginines are products of degraded methylated proteins. Two enzymes—protein arginine methyltransferase type I and II (PRMT-I, PRMT-II), are involved in the methylation of arginine residues within proteins or polypeptides with the methyl groups derived from S-adenosylmethionine [30]. PRMT-I catalyzes the formation of NG-monomethyl-l-arginine (LNMMA) and NG,NG-dimethyl-l-arginine (ADMA) while PRMT-II methylate proteins to release NG,N’G-dimethyl-l-arginine (SDMA) and LNMMA. Free dimethylarginines are released into the cytoplasm during proteolytic breakdown of proteins and can be detected in blood, and eliminated from the body by renal excretion [31]. ADMA, but not SDMA, is metabolized via hydrolytic degradation to citrulline and dimethylamine by the dimethylarginine dimethylaminohydrolase-1 (DDAH-1) and -2 (DDAH-2) enzymes. Thus, the increased ADMA levels could be due to increased PRMT-I activity, reduced elimination by the kidney, decreased DDAH-1 and 2 enzymtic activities, or a combination. However, our GWAS analysis did not find any association between ADMA and these genes including PRMT-I and DDAH-1 and 2, suggesting that the increased ADMA level may not be genetic. Instead, we found that SNP rs1125718 on chromosome 8 was associated with ADMA concentration at GWAS significance level. This SNP is located in a gene desert and has not been reported to be associated with any disease or traits yet. However, several genes are located in the nearby region including NDRG1, WISP1, ST3GAL1, and ZFAT. Among them, WISP1 gene is interesting because a study showed that WISP1 as fibro-adipogenic progenitor (FAP)-derived matricellular signal is lost during aging. WISP1 is required for efficient muscle regeneration, and it controls the expansion and asymmetric commitment of tissue-resident muscle stem cells (MuSCs) through Akt signaling [32]. Also, previous studies showed that nitric oxide (NO) level positively correlated with WISP1 gene expression, and elevated levels of NO increased the WISP1 mRNA and protein expression levels through a beta-catenin signaling [33]. Interestingly, ADMA is known as an endogenous competitive inhibitor of NO synthase [34]. Our GWAS analysis showed that the second most associated SNP with ADMA was rs816296 which is located in the intron 1 of the NOS1 gene. Thus, we hypothesize that possible age-related muscle protein breakdown may lead to an increased release of ADMA which in turn inhibits NO production. The reduced NO synthesis may result in lower expression of WISP1 which leads to the matricellular signals in the skeletal muscle stem cell niche being disturbed [32, 35]. Hence, the MuSC number, activity, adhesion, migration, proliferation, self-renewal, and differentiation in skeletal muscle regeneration could be considerably deteriorated leading to the reduction of muscle strength [36,37,38].

Uric acid is an enzymatic waste endproduct from the degradation of purine nucleosides that is renally excreted. Uric acid plays both protective and damaging roles in the skeletal muscles [39], most likely due to its strong antioxidant properties at low levels and pro-inflammatory effect at high levels [35]. It has been proposed that oxidative stress might contribute to muscle weakness and wasting. Uric acid at a low level might stabilize the excessive production of free radicals that causes muscle protein damage leading to muscle strength reduction [40]. However, at high levels, uric acid stimulates the pro-inflammatory pathway and elevates the production of pro-inflammatory cytokines including interleukin-1 (IL-1), IL-6, and the tumor necrosis factor (TNF), which have an impact on muscle mass and function in aged muscles [41, 42]. While we did not find a significant cross-sectional association between uric acid and leg muscle strength (p=0.56) at the 2.6-year follow-up phase, we found that there was a positive association between uric acid concentrations and leg muscle strength at the baseline time point. This is consistent with previous studies [39, 40]. We also documented a negative association between uric acid levels and longitudinal leg muscle strength, consistent with previous studies [41, 42]. Thus, our findings suggested the importance of maintaining optimal levels of uric acid in the blood for muscle strength [40].

The strength of the current study was its longitudinal nature which allowed us to detect significant metabolite associations for muscle strength changes overtime within an individual. This can not be achieved in a cross-sectional analysis. Indeed, when we analyzed the data cross-setionally for the 2.6-year follow-up point, the significance for the identified metabolites became weaker or even non-significant. The current study also underscored the importance of the longer follow-up time with multiple time point measurements as it could minimize the effect of fluctuating variability on the measurements and provide more accurate estimate of changes over time. However, there are a number of limitations in this study. The present study used a commercially available metabolomics assay kit that offers limited coverage of metabolome. Thus, we might miss some metabolites that may contribute to the longitudinal reduction of muscle strength. Since metabolomics profiling was performed at only 2.6-year follow-up point, we cannot make any inference regarding the association between the changes in metabolite profiles over time and the muscle strength decline over time. Further studies with multiple time point metabolomic profilings are needed. Loss to follow-up might have influenced our results, especially for leg muscle strength as we had 6–8% of missing values at phase 3 and phase 4 follow-up points. Indeed, those lost to follow-up had a lower leg muscle strength measurement at baseline than those included in the analysis (data not shown). However, there was no difference in uric acid levels between those included and excluded in the final analysis, suggesting that loss to follow-up was unlikely to bias the observed association. We cannot rule out the potential confounding effect of gout on the association between uric acid and leg muscle strength as we did not have data on gout in our cohort. However, gout mostly affects big toes and associated with reduced muscle strength of the ankle and foot, not leg muscle strength, suggesting the observed association was less likely to be biased. Finally, our results may not be generalized to populations that have different area-specific socioeconomic indexes and health provisions than that in Tasmania, Australia.

In conclusion, our data demonstrates that baseline elevated serum concentrations of ADMA and uric acid were associated with age-dependent muscle strength reduction. Confirmation of these findings would establish new insights into the pathogenesis of age-related muscle strength decline and uncover novel targets for developing strategies to prevent muscle strength loss over time.