https://orcid.org/0000-0003-1825-8010
Abstract
The relationship between amino acids, B vitamins, and their metabolites with D3-creatine (D3Cr) dilution muscle mass, a more direct measure of skeletal muscle mass, has not been investigated. We aimed to assess associations of plasma metabolites with D3Cr muscle mass, as well as muscle strength and physical performance in older men from the Osteoporotic Fractures in Men cohort study.
Out of 1 425 men (84.2 ± 4.1 years), men with the lowest D3Cr muscle mass (n = 100), slowest walking speed (n = 100), lowest grip strength (n = 100), and a random sample (n = 200) serving as a comparison group to the low groups were included. Metabolites were analyzed using liquid chromatography–tandem mass spectrometry. Metabolite differences between the low groups and random sample and their relationships with the muscle outcomes adjusted for confounders and multiple comparisons were assessed using t-test/Mann–Whitney–Wilcoxon and partial correlations, respectively.
For D3Cr muscle mass, significant biomarkers (p < .001) with ≥10% fold difference and largest partial correlations were tryptophan (Trp; r = 0.31), kynurenine (Kyn)/Trp; r = −0.27), nicotinamide (Nam)/quinolinic acid (Quin; r = 0.21), and alpha-hydroxy-5-methyl-tetrahydrofolate (hm-THF; r = −0.25). For walking speed, hm-THF, Nam/Quin, and Quin had the largest significance and fold difference, whereas valine (r = 0.17), Trp (r = 0.17), HKyn/Xant (r = −0.20), neopterin (r = −0.17), 5-methyl-THF (r = −0.20), methylated folate (r = −0.21), and thiamine (r = −0.18) had the strongest correlations. Only hm-THF was correlated with grip strength (r = −0.21) and differed between the low group and the random sample.
Future interventions focusing on how the Trp metabolic pathway or hm-THF influences D3Cr muscle mass and physical performance declines in older adults are warranted.
Constituents of the diet and their metabolites have been linked to sarcopenia, a condition that is characterized by loss of muscle mass and reduced strength and physical performance. In particular, dietary proteins, B vitamins, and circulating amino acids (AAs) play a pivotal role in muscle physiology and are likely critical to maintaining muscle mass and function, with derangements in these nutrients being linked to sarcopenia and decrements in sarcopenic traits (1,2). For instance, lower levels of B vitamins and higher levels of homocysteine were reported in sarcopenic participants (3,4). In addition, participants in the Baltimore Longitudinal Study of Aging with low specific strength, defined as the ratio between muscle strength and mass, had significantly higher levels of plasma leucine, isoleucine, tryptophan, serotonin, and methionine (5). Interestingly, kynurenine pathway metabolites, which are products of tryptophan degradation, have also been linked to poor physical performance and muscle weakness (6,7).
Despite the growing evidence of a link between AA, B vitamins and their pathway metabolites, and the pathophysiology of sarcopenia, high heterogeneity exists among studies, especially those assessing the associations between AA levels and sarcopenia (2). Many of these inconsistencies stem from the methods used to assess sarcopenic components, specifically that of muscle mass (2). Most metabolomic studies of sarcopenia have measured lean mass (LM) or fat-free mass (FFM) rather than actual muscle mass. These surrogate measures do not fully depict functional muscle mass as they also include noncontractile components such as lipids and fibrotic tissue, both of which increase with aging. D3-creatine (D3Cr) dilution is a method to directly and accurately assess skeletal muscle mass through the measurement of the creatine pool size, and has been found to be a stronger predictor of functional decline in older adults when compared to LM assessed by dual-energy x-ray absorptiometry (DXA) (8).
Currently, there are no studies that have assessed the association of nutrient metabolites related to D3Cr muscle mass. Hence, the primary aim of this investigation was to assess the association between plasma levels of AAs, B vitamins, and their pathway metabolites with D3Cr muscle mass in older men from the prospective Osteoporotic Fractures in Men (MrOS) cohort study. Moreover, we sought to investigate the association of the aforementioned nutrient metabolites with the other components of sarcopenia (ie, muscle strength and physical performance) in order to determine if the patterns of association differ depending on the muscle characteristic assessed.
Materials and Methods
MrOS Cohort
From 2000 to 2002, 5 994 ambulatory community-dwelling men aged ≥65 years without bilateral hip replacements were enrolled in MrOS, a multicenter cohort study of aging and osteoporosis (9,10). The study was approved by the Institutional Review Board at each participating center and all participants provided written informed consent. In 2014–2016, 2 786 survivors were contacted to participate in a follow-up clinic visit (Year 14 visit) (8). Participants self-reported their smoking status and a physician’s diagnosis of chronic kidney disease or renal failure. The Block Dietary Systems Food Frequency Questionnaire was used to assess protein intake and multivitamin use (11). Age at Year 14 visit was calculated from birthdate self-reported at baseline. Physical activity was self-reported using the Physical Activity Scale for the Elderly (PASE) (12). Height was measured by wall-mounted stadiometers and weight by balance beam or digital scales; body mass index (BMI) was calculated as weight (kg)/height (m2).
Muscle Mass, Strength, and Physical Performance
Skeletal muscle mass was estimated using D3Cr dilution methods described previously (8). To account for variations in total muscle mass by body size, we analyzed D3Cr muscle mass divided by body mass (kg). Grip strength (kg) was assessed by analyzing the maximal value from 2 tests of each hand using Jamar handheld dynamometers (13). Walking speed (m/s) was determined by timing the completion of a 6-m course at the participant’s usual pace and was used as an indicator of physical performance (14).
Nutrient Metabolites
Fasting serum was collected at the Year 14 clinic visit. The following analytes were measured at the Nestlé Research nutrition analytics mass spectrometry lab: alanine, arginine, β-alanine, monomethylarginine, asymmetric dimethylarginine (ADMA), symmetric dimethylarginine (SDMA), asparagine, aspartic acid, citrulline, ethanolamine, glutamic acid, glutamine, glycine, histidine, 1-methylhistidine, 3-methylhistidine, a-aminobutyric acid (AABA), β-aminoisobutyric acid, γ-aminobutyric acid, isoleucine, leucine, lysine, methionine, ornithine, phenylalanine, proline, hydroxyproline, sarcosine, serine, taurine, threonine, tryptophan, tyrosine, and valine. Measurements were performed using a method adapted from Guiraud et al. (15). Briefly, this was achieved by adding isotope-labeled AAs as internal standards and measuring the analytes using ultra-high-pressure liquid chromatography–tandem mass spectrometry (UPLC-MS/MS).
All other nutrient analytes were measured using mass spectrometry coupled to relevant chromatographic separation by Bevital (Bergen, Norway; https://bevital.no/) using 2 019 panels of analytes. Hydro-soluble B vitamins and kynurenine pathway metabolites were measured by LC-MS/MS (Bevital panel D: anthranilic acid, cotinine, cystathionine, flavin mononucleotide, 3-hydroxyanthranilic acid, 3-hydroxykynurenine [HKr], kynurenic acid, kynurenine, N1-methylnicotinamide, nicotinamide, nicotinic acid, neopterin, 4-pyridoxic acid, pyridoxine, picolinic acid, pyridoxal, pyridoxal 5-phosphate, quinolinic acid, riboflavin, thiamine, thiamine monophosphate, trans-3’-hydroxycotinine, trigonelline, tryptophan, and xanthurenic acid). Measurements were performed by mixing serum samples with labeled internal standards and resolving the analytes on a C8 LC column by a gradient-type mobile phase, and detected using electrospray ionization tandem mass spectrometry as described in Midttun et al. (16). The kynurenine-to-tryptophan ratio (Kyn/Trp), an indicator of indoleamine-pyrrole 2,3-dioxygenase (IDO) activity, the major enzyme that catalyzes O2-dependent oxidation of L-tryptophan to N-formylkynurenine, which is the rate-limiting step in the kynurenine pathway, was calculated by dividing the concentration of kynurenine (in nmol/L) by the concentration of tryptophan (in µmol/L) (17). We calculated 2 ratio-based functional markers, based on the principle that when vitamin deficiency affects the catalytic capacity of an enzyme, the response is an increase in upstream metabolites and a parallel decrease in downstream metabolites. Hydroxykynurenine-to-xanthurenic acid ratio (HKyn/Xant) was calculated by dividing the concentration of 3-hydroxykynurenine by the concentration of xanthurenic acid (in nmol/L) (18). We also calculated the HKr ratio, which includes hydroxykynurenine and the 4 kynurenines that are products of the pyridoxal 5’-phosphate (PLP)-dependent enzymes, kynurenine transaminase (KAT), and kynureninase (KYNU). HKr was calculated as the ratio of hydroxykynurenine to the sum of kynurenic acid, anthranilic acid, xanthurenic acid, and 3-hydroxyanthranilic acid. HKr shows a steep inverse association with PLP at concentrations less than 20 nmol/L, serving as a marker for vitamin B6 deficiency (18). However, a notable increase in HKr in the interval from 40 to 20 nmol/L PLP also supports the concept of marginal vitamin deficiency in this interval. Compared to the index that only included XA as the downstream metabolite (HKyn/Xant), HKr demonstrated better specificity, sensitivity, and reproducibility across cohorts for 11 043 men and women (19). The ratio of nicotinamide to quinolinic acid (Nam/Quin) was calculated by dividing the concentration of nicotinamide by the concentration of quinolinic acid (in nmol/L).
Folate species and catabolites were measured by LC-MS/MS (Bevital panel E: acetamidobenzoylglutamate, 4-alpha-hydroxy-5-methyl-tetrahydrofolate [hm-THF], 5-formyl-THF, 5-methyl-THF [5m-THF], para-aminobenzoylglutamate). Because hm-THF is formed upon degradation of 5m-THF during storage, we analyzed the sum of hm-THF and 5m-THF as methylated folate (mF). Measurements were performed by adding 13C-labeled folate forms as internal standards and measuring the analytes using LC-MS/MS (20).
Homocysteine and 1-carbon metabolites were measured by GC-MS/MS (Bevital panel A from 2019: methylmalonic acid, total homocysteine, total cysteine, methionine, serine, and glycine). Measurements were performed by adding 2H-labeled metabolites as internal standards and measuring the analytes using gas chromatography–tandem mass spectrometry (GC-MS/MS) (21).
Study Sample
Of the 2 786 survivors invited to participate in the Year 14 Visit, 1 841 completed the visit with an additional 583 providing questionnaire data. Of these, 1 425 completed the D3Cr muscle mass measurement (Supplementary Figure 1). Of those men, we sampled participants from each of the following groups: 100 with the lowest muscle mass (D3Cr muscle mass/weight), 100 with the lowest walking speed, and 100 with the lowest grip strength. We also sampled 200 participants at random to serve as a comparison group to the groups above, and to complete analyses across the full range of values in the data set. Participants with stroke or Parkinson’s disease, those taking oral corticosteroids, and those on androgen deprivation therapy were not eligible for inclusion in analyses. Due to the overlap between the 3 low groups, there were a total of 409 men included in the analyses (Supplementary Figure 2).
Statistical Methods
Participant characteristics were summarized for the groups with the lowest muscle and performance characteristics and the random sample. Nutrient biomarkers 3-methylhistidine, cotinine, nicotinic acid, para-aminobenzoylglutamate, thiamine monophosphate, trans-3-hydroxycotinine, sarcosine, and pyridoxine were not included in the analysis due to over 8% with analyte concentrations below the level of detection. All nutrient biomarkers had skewed distributions and were log-transformed for all analyses with the exception of the following: AAs, AABA, ADMA, alanine, aspartic acid, glycine, isoleucine, leucine, lysine, methionine, phenylalanine, proline, serine, total cysteine, threonine, tryptophan, tyrosine, and valine; and the AA-related metabolite total homocysteine.
We compared mean and median values of the nutrient biomarkers between the lowest D3Cr muscle mass/wgt group and the random sample using t-tests or Mann–Whitney–Wilcoxon test, respectively. Means and standard deviations are reported. The log-fold difference between the low group and the random sample was plotted with the log p value, with greater values indicating greater nutrient biomarker presence in the low group. To account for multiple comparisons, we used the false discovery rate (FDR) correction of p values to determine q values as suggested by Benjamini and Hochberg (22). All significant differences are labeled.
We then assessed the partial correlation of each nutrient biomarker with the outcome of continuous D3Cr muscle mass using participants in the random sample and the low D3Cr muscle mass/wgt group, while adjusting for age, BMI, smoking status, multivitamin use, kidney disease, and physical activity. An FDR correction of p values was used to account for multiple comparisons. We repeated this analysis for the other 2 outcomes: walking speed and grip strength.
As a sensitivity analysis, we assessed if further adjustment for total dietary protein intake altered the partial correlations between the nutrient biomarkers and the muscle-related outcomes in a subsample of the cohort with available dietary data (n = 258).
Data were analyzed with SAS Software version 9.4 (SAS Institute Inc., Cary, NC).
Results
Participant Characteristics
The characteristics of the random sample and low muscle mass, low walking speed, and low grip strength groups are shown in Table 1. Participants with lower values for these metrics were generally older and had worse physical performance and lower D3Cr muscle mass, and were less physically active than those in the random sample.
Characteristics of Older Men in the Random Sample and Each of the Low Groups for Muscle Mass, Walking Speed, and Grip Strength
| Characteristics | Random Sample (N = 187) | Low D3Cr Muscle Mass/Weight (N = 92) | Low Walk Speed (N = 97) | Low Grip Strength (N = 93) |
|---|---|---|---|---|
| Age (yr) | 82.93 ± 3.32 | 86.02 ± 4.24 | 87.13 ± 4.76 | 87.48 ± 4.40 |
| Height (cm) | 172.30 ± 5.81 | 172.27 ± 6.30 | 169.95 ± 7.16 | 167.95 ± 6.99 |
| BMI (kg/m2) | 27.00 ± 3.21 | 30.45 ± 4.02 | 27.92 ± 4.10 | 26.49 ± 3.15 |
| DXA % fat | 27.24 ± 4.93 | 34.92 ± 5.28 | 29.25 ± 7.45 | 28.33 ± 6.22 |
| DXA ALM (kg) | 23.04 ± 2.76 | 22.55 ± 3.41 | 22.03 ± 3.41 | 20.60 ± 2.82 |
| ALM/height2 (kg/m2) | 7.76 ± 0.84 | 7.60 ± 1.01 | 7.61 ± 0.93 | 7.30 ± 0.82 |
| Muscle mass (kg) | 25.25 ± 3.78 | 19.73 ± 2.67 | 21.01 ± 3.63 | 21.04 ± 3.81 |
| % Muscle mass from creatine/weight | 0.32 ± 0.04 | 0.22 ± 0.01 | 0.26 ± 0.04 | 0.28 ± 0.04 |
| Grip strength (kg) | 37.31 ± 6.16 | 31.23 ± 6.89 | 29.19 ± 7.06 | 21.42 ± 2.88 |
| Completed 400-m walk (yes, %) | 174 (98.31) | 45 (78.95) | 29 (70.73) | 62 (91.18) |
| Walking speed (m/s) | 1.17 ± 0.19 | 0.83 ± 0.21 | 0.59 ± 0.12 | 0.86 ± 0.26 |
| Walking speed from 400-m walk (m/s) | 1.09 ± 0.18 | 0.84 ± 0.17 | 0.65 ± 0.22 | 0.87 ± 0.20 |
| Have FFQ data (yes, %) | 109 (58.29) | 64 (69.57) | 61 (62.89) | 56 (60.22) |
| Multivitamin use (yes, %) | 103 (55.08) | 56 (60.87) | 58 (59.79) | 61 (65.59) |
| Protein intake (g) | 60.37 ± 26.93 | 61.36 ± 27.94 | 61.35 ± 27.27 | 58.63 ± 23.41 |
| Renal disease (yes, %) | 6 (3.21) | 5 (5.43) | 8 (8.25) | 6 (6.45) |
| Smoking status: ever (n, %) | 87 (46.52) | 54 (58.70) | 51 (52.58) | 40 (43.01) |
| PASE Physical Activity Score | 130.70 ± 58.90 | 70.29 ± 48.69 | 66.70 ± 51.40 | 68.29 ± 46.74 |
| Characteristics | Random Sample (N = 187) | Low D3Cr Muscle Mass/Weight (N = 92) | Low Walk Speed (N = 97) | Low Grip Strength (N = 93) |
|---|---|---|---|---|
| Age (yr) | 82.93 ± 3.32 | 86.02 ± 4.24 | 87.13 ± 4.76 | 87.48 ± 4.40 |
| Height (cm) | 172.30 ± 5.81 | 172.27 ± 6.30 | 169.95 ± 7.16 | 167.95 ± 6.99 |
| BMI (kg/m2) | 27.00 ± 3.21 | 30.45 ± 4.02 | 27.92 ± 4.10 | 26.49 ± 3.15 |
| DXA % fat | 27.24 ± 4.93 | 34.92 ± 5.28 | 29.25 ± 7.45 | 28.33 ± 6.22 |
| DXA ALM (kg) | 23.04 ± 2.76 | 22.55 ± 3.41 | 22.03 ± 3.41 | 20.60 ± 2.82 |
| ALM/height2 (kg/m2) | 7.76 ± 0.84 | 7.60 ± 1.01 | 7.61 ± 0.93 | 7.30 ± 0.82 |
| Muscle mass (kg) | 25.25 ± 3.78 | 19.73 ± 2.67 | 21.01 ± 3.63 | 21.04 ± 3.81 |
| % Muscle mass from creatine/weight | 0.32 ± 0.04 | 0.22 ± 0.01 | 0.26 ± 0.04 | 0.28 ± 0.04 |
| Grip strength (kg) | 37.31 ± 6.16 | 31.23 ± 6.89 | 29.19 ± 7.06 | 21.42 ± 2.88 |
| Completed 400-m walk (yes, %) | 174 (98.31) | 45 (78.95) | 29 (70.73) | 62 (91.18) |
| Walking speed (m/s) | 1.17 ± 0.19 | 0.83 ± 0.21 | 0.59 ± 0.12 | 0.86 ± 0.26 |
| Walking speed from 400-m walk (m/s) | 1.09 ± 0.18 | 0.84 ± 0.17 | 0.65 ± 0.22 | 0.87 ± 0.20 |
| Have FFQ data (yes, %) | 109 (58.29) | 64 (69.57) | 61 (62.89) | 56 (60.22) |
| Multivitamin use (yes, %) | 103 (55.08) | 56 (60.87) | 58 (59.79) | 61 (65.59) |
| Protein intake (g) | 60.37 ± 26.93 | 61.36 ± 27.94 | 61.35 ± 27.27 | 58.63 ± 23.41 |
| Renal disease (yes, %) | 6 (3.21) | 5 (5.43) | 8 (8.25) | 6 (6.45) |
| Smoking status: ever (n, %) | 87 (46.52) | 54 (58.70) | 51 (52.58) | 40 (43.01) |
| PASE Physical Activity Score | 130.70 ± 58.90 | 70.29 ± 48.69 | 66.70 ± 51.40 | 68.29 ± 46.74 |
Notes: ALM = appendicular lean mass; D3Cr = D3-creatine; DXA = dual-energy x-ray absorptiometry; FFQ = Food Frequency Questionnaire; PASE = Physical Activity Scale for the Elderly. Bold values indicate a significance difference from the random sample (p < .05).
Characteristics of Older Men in the Random Sample and Each of the Low Groups for Muscle Mass, Walking Speed, and Grip Strength
| Characteristics | Random Sample (N = 187) | Low D3Cr Muscle Mass/Weight (N = 92) | Low Walk Speed (N = 97) | Low Grip Strength (N = 93) |
|---|---|---|---|---|
| Age (yr) | 82.93 ± 3.32 | 86.02 ± 4.24 | 87.13 ± 4.76 | 87.48 ± 4.40 |
| Height (cm) | 172.30 ± 5.81 | 172.27 ± 6.30 | 169.95 ± 7.16 | 167.95 ± 6.99 |
| BMI (kg/m2) | 27.00 ± 3.21 | 30.45 ± 4.02 | 27.92 ± 4.10 | 26.49 ± 3.15 |
| DXA % fat | 27.24 ± 4.93 | 34.92 ± 5.28 | 29.25 ± 7.45 | 28.33 ± 6.22 |
| DXA ALM (kg) | 23.04 ± 2.76 | 22.55 ± 3.41 | 22.03 ± 3.41 | 20.60 ± 2.82 |
| ALM/height2 (kg/m2) | 7.76 ± 0.84 | 7.60 ± 1.01 | 7.61 ± 0.93 | 7.30 ± 0.82 |
| Muscle mass (kg) | 25.25 ± 3.78 | 19.73 ± 2.67 | 21.01 ± 3.63 | 21.04 ± 3.81 |
| % Muscle mass from creatine/weight | 0.32 ± 0.04 | 0.22 ± 0.01 | 0.26 ± 0.04 | 0.28 ± 0.04 |
| Grip strength (kg) | 37.31 ± 6.16 | 31.23 ± 6.89 | 29.19 ± 7.06 | 21.42 ± 2.88 |
| Completed 400-m walk (yes, %) | 174 (98.31) | 45 (78.95) | 29 (70.73) | 62 (91.18) |
| Walking speed (m/s) | 1.17 ± 0.19 | 0.83 ± 0.21 | 0.59 ± 0.12 | 0.86 ± 0.26 |
| Walking speed from 400-m walk (m/s) | 1.09 ± 0.18 | 0.84 ± 0.17 | 0.65 ± 0.22 | 0.87 ± 0.20 |
| Have FFQ data (yes, %) | 109 (58.29) | 64 (69.57) | 61 (62.89) | 56 (60.22) |
| Multivitamin use (yes, %) | 103 (55.08) | 56 (60.87) | 58 (59.79) | 61 (65.59) |
| Protein intake (g) | 60.37 ± 26.93 | 61.36 ± 27.94 | 61.35 ± 27.27 | 58.63 ± 23.41 |
| Renal disease (yes, %) | 6 (3.21) | 5 (5.43) | 8 (8.25) | 6 (6.45) |
| Smoking status: ever (n, %) | 87 (46.52) | 54 (58.70) | 51 (52.58) | 40 (43.01) |
| PASE Physical Activity Score | 130.70 ± 58.90 | 70.29 ± 48.69 | 66.70 ± 51.40 | 68.29 ± 46.74 |
| Characteristics | Random Sample (N = 187) | Low D3Cr Muscle Mass/Weight (N = 92) | Low Walk Speed (N = 97) | Low Grip Strength (N = 93) |
|---|---|---|---|---|
| Age (yr) | 82.93 ± 3.32 | 86.02 ± 4.24 | 87.13 ± 4.76 | 87.48 ± 4.40 |
| Height (cm) | 172.30 ± 5.81 | 172.27 ± 6.30 | 169.95 ± 7.16 | 167.95 ± 6.99 |
| BMI (kg/m2) | 27.00 ± 3.21 | 30.45 ± 4.02 | 27.92 ± 4.10 | 26.49 ± 3.15 |
| DXA % fat | 27.24 ± 4.93 | 34.92 ± 5.28 | 29.25 ± 7.45 | 28.33 ± 6.22 |
| DXA ALM (kg) | 23.04 ± 2.76 | 22.55 ± 3.41 | 22.03 ± 3.41 | 20.60 ± 2.82 |
| ALM/height2 (kg/m2) | 7.76 ± 0.84 | 7.60 ± 1.01 | 7.61 ± 0.93 | 7.30 ± 0.82 |
| Muscle mass (kg) | 25.25 ± 3.78 | 19.73 ± 2.67 | 21.01 ± 3.63 | 21.04 ± 3.81 |
| % Muscle mass from creatine/weight | 0.32 ± 0.04 | 0.22 ± 0.01 | 0.26 ± 0.04 | 0.28 ± 0.04 |
| Grip strength (kg) | 37.31 ± 6.16 | 31.23 ± 6.89 | 29.19 ± 7.06 | 21.42 ± 2.88 |
| Completed 400-m walk (yes, %) | 174 (98.31) | 45 (78.95) | 29 (70.73) | 62 (91.18) |
| Walking speed (m/s) | 1.17 ± 0.19 | 0.83 ± 0.21 | 0.59 ± 0.12 | 0.86 ± 0.26 |
| Walking speed from 400-m walk (m/s) | 1.09 ± 0.18 | 0.84 ± 0.17 | 0.65 ± 0.22 | 0.87 ± 0.20 |
| Have FFQ data (yes, %) | 109 (58.29) | 64 (69.57) | 61 (62.89) | 56 (60.22) |
| Multivitamin use (yes, %) | 103 (55.08) | 56 (60.87) | 58 (59.79) | 61 (65.59) |
| Protein intake (g) | 60.37 ± 26.93 | 61.36 ± 27.94 | 61.35 ± 27.27 | 58.63 ± 23.41 |
| Renal disease (yes, %) | 6 (3.21) | 5 (5.43) | 8 (8.25) | 6 (6.45) |
| Smoking status: ever (n, %) | 87 (46.52) | 54 (58.70) | 51 (52.58) | 40 (43.01) |
| PASE Physical Activity Score | 130.70 ± 58.90 | 70.29 ± 48.69 | 66.70 ± 51.40 | 68.29 ± 46.74 |
Notes: ALM = appendicular lean mass; D3Cr = D3-creatine; DXA = dual-energy x-ray absorptiometry; FFQ = Food Frequency Questionnaire; PASE = Physical Activity Scale for the Elderly. Bold values indicate a significance difference from the random sample (p < .05).
Amino Acids
Compared to the random sample, several AAs were significantly different in men with the lowest D3Cr muscle mass/wgt and slowest walking speed, whereas no differences in AA concentrations were observed in men with the lowest strength (Supplementary Figure 3 and Supplementary Table 1).
After controlling for age, BMI, smoking status, multivitamin use, kidney disease, and physical activity, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, and valine were all positively correlated with D3Cr muscle mass (Supplementary Table 2). As illustrated by the volcano plot (Figure 1A), tryptophan, lysine, methionine, and the branch-chained AAs leucine and valine had the largest partial correlations and strongest relationships with D3Cr muscle mass. Similar to D3Cr muscle mass, higher concentrations of the AAs tryptophan and valine were significantly correlated with faster walking speed after adjustment for confounders (Figure 1B and Supplementary Table 3). None of the AAs were related to grip strength (Figure 1C and Supplementary Table 4) nor were there any differences in AA concentrations between men with the lowest strength compared to the random sample (Supplementary Figure 3C and Supplementary Table 1).
Volcano plots of partial correlations of nutrient biomarkers with (A) D3Cr muscle mass, (B) walking speed, and (C) grip strength adjusted for age, BMI, smoking status, multivitamin use, kidney disease, and physical activity. An FDR correction of p values was used to account for multiple comparisons. AABA = a-aminobutyric acid; BMI = body mass index; D3Cr = D3-creatine; FDR = false discovery rate; HKyn/Xant = hydroxykynurenine-to-xanthurenic acid ratio.
AA-Related Metabolites
The AA-related metabolites with ≥10%-fold difference and of greatest significance between men with the lowest D3Cr muscle mass/wgt and men with the slowest walking speed compared to the random sample were mainly those related to the tryptophan kynurenine pathway (see the illustrated pathway in Figure 2). No differences in the concentrations of AA-related metabolites between the lowest strength group and the random sample were observed (Supplementary Figure 3 and Supplementary Table 1).
Tryptophan kynurenine pathway. D3Cr = D3-creatine.
After controlling for confounders, metabolites of the tryptophan kynurenine pathway, as well as AABA and SDMA, were significantly correlated with D3Cr muscle mass, such that higher concentrations of 3-hydroxykynurenine, HKr ratio, HKyn/Xant, Kyn/Trp, quinolinic acid, and SDMA and lower concentrations of AABA, β-alanine, and Nam/Quin were indicative of lower D3Cr muscle mass (Figure 1A and Supplementary Table 2). The AA-related metabolites with the largest partial correlations and strongest relationships with D3Cr muscle mass were Kyn/Trp and Nam/Quin, AABA, and SDMA (Figure 1A). Slow walking speed was also related to lower levels of AABA and higher concentrations of the tryptophan kynurenine pathway metabolites HKyn/Xant and neopterin (Figure 1B and Supplementary Table 3). For grip strength, none of the AA-related metabolites were correlated after adjustment for confounders (Figure 1C and Supplementary Table 4).
B-Vitamin Metabolites
The B-vitamin metabolite hm-THF was consistently higher across men with low D3Cr muscle mass/wgt, slow walking speed, and low grip strength (Supplementary Figure 3 and Supplementary Table 1).
In the adjusted partial correlation analysis, low D3Cr muscle mass was associated with higher hm-THF and acetamidobenzoylglutamate; however, hm-THF had the larger partial correlation and strongest relationship (Figure 1A and Supplementary Table 2). 5m-THF, methylated folate, and thiamine were inversely correlated with walking speed with higher levels of these metabolites being associated with slower walking speed (Figure 1B and Supplementary Table 3). The only nutrient metabolite correlated with grip strength was the B-vitamin metabolite hm-THF, where higher levels were related to lower grip strength (Figure 1C and Supplementary Table 4).
When adjusting for total dietary protein intake, the relationships of phenylalanine, 3-hydroxykynurenine, β-alanine, HKr ratio, HKyn/Xant, Nam/Quin, quinolinic acid, and SDMA with D3Cr muscle mass were no longer significant (Supplementary Table 2), whereas for walking speed and grip strength, all previously significant nutrient biomarkers were no longer significant (Supplementary Tables 3 and 4). Despite changes in p values, the overall strength of the partial correlations of the nutrient biomarkers with the muscle-related outcomes did not substantially change.
A summary of the unique and overlapping AAs, AA-related metabolites, and B-vitamin metabolites that were significantly correlated with D3Cr muscle mass, walking speed, and grip strength after adjustment for confounders and multiple comparisons are displayed in Figure 3. A summary of the unique and overlapping nutrient metabolites that differed between men with the lowest D3Cr muscle mass/wgt, lowest walking speed, and lowest grip strength compared to the random sample can be found in Supplementary Figure 4.
Summary of the unique and overlapping relationships of amino acids, amino acid-related metabolites, and B-vitamin metabolites with D3Cr muscle mass, walking speed, and grip strength after adjusting for age, BMI, smoking status, multivitamin use, kidney disease, and physical activity, and applying an FDR correction of p values for multiple comparisons. AABA = a-aminobutyric acid; BMI = body mass index; D3Cr = D3-creatine; FDR = false discovery rate; HKr = 3-hydroxykynurenine; HKyn/Xant = hydroxykynurenine-to-xanthurenic acid ratio; hm-THF = 4-alpha-hydroxy-5-methyl-tetrahydrofolate; SDMA = symmetric dimethylarginine.
Discussion
This study assessed the association between plasma levels of AAs, B vitamins, and their pathway metabolites with D3Cr muscle mass along with other components comprising sarcopenia (ie, walking speed and grip strength) in older men. A number of nutrient biomarkers were identified, with the B-vitamin metabolite 4-alpha-hydroxy-5-methyl-THF, the AA tryptophan, and tryptophan-related metabolites within the kynurenine pathway, as well as the branched-chain AA leucine, displaying the strongest correlations with D3Cr muscle mass. Similar results were observed for walking speed, although the strength of the correlations was less than that of D3Cr muscle mass; 4-hm-THF was the only metabolite significantly correlated with grip strength.
Accumulating evidence has indicated a link between certain circulating AAs and sarcopenia. In particular, a meta-analysis of studies comparing AA profiles of individuals with and without sarcopenia found that the individuals with sarcopenia had significantly reduced levels of the branched-chain AA leucine and isoleucine and the aromatic amino acid tryptophan, although there was high heterogeneity in the metabolomic results across studies given the different definitions and measurement methods used to assess sarcopenia (2). Indeed, branched-chain AAs are primarily metabolized in muscle tissue, where they are involved in muscle protein synthesis, satellite cell activation, and proteolysis inhibition (1). Branched-chain AAs may further stimulate mitochondrial biogenesis by activating the mammalian target of rapamycin (mTOR), peroxisome proliferator‐activated receptor gamma coactivator 1‐alpha, endothelial nitric oxide synthase, Sirtuin 1, and AMP‐activated protein kinase, in addition to other targets (23). In particular, leucine has been shown to stimulate muscle protein synthesis through activation of the mTOR signaling pathway, reduce protein degradation, and increase LM (24). Hence, branched-chain AAs (leucine, isoleucine, and/or valine) have been consistently found to be positively correlated with DXA FFM index (24,25) and appendicular LM (ALM) (25–27) in older adults after adjustment for relevant confounders and multiple comparisons. Our results corroborate these previous findings while using a more appropriate measure of total body skeletal muscle mass (ie, D3Cr muscle mass), where we observed leucine to be one of the AAs with the strongest association with D3Cr muscle mass. Similar results were observed for walking speed, with the branched-chain AAs (leucine [p = .057] and valine) exhibiting positive correlations, although the strength of these relationships was not as strong as other AA and AA-related metabolites in our study, as well as in others (28). Likewise, although the branched-chain AAs leucine and isoleucine were reported to be positively related to grip strength in a cohort of older men (29), the majority of studies in older adults did not observe any relationship with the branched-chain AAs, similar to our findings (6,30,31).
The other nutrient metabolites with consistent associations in our study were tryptophan and its associated metabolites in the kynurenine pathway, with these biomarkers having the strongest correlations with D3Cr muscle mass and walking speed. Moreover, several of the kynurenine pathway metabolites were found to be altered in participants with low muscle mass and walking speeds when compared to the random sample. The kynurenine pathway begins with the degradation of tryptophan to kynurenine via IDO and tryptophan 2,3-dioxygenase (32). Levels of kynurenine tend to be elevated in older adults due to high levels of IDO-1 activity, which is induced by the chronic inflammatory condition associated with aging (ie, inflammaging) (32). In fact, the Kyn/Trp ratio has been found to be significantly correlated with inflammatory cytokines including IL-6, IFN-γ, and TNF-α (6). Once kynurenine is produced, it can be metabolized into kynurenic acid by kynurenine aminotransferases, or alternatively to anthranilic acid (by kynureninase) or 3-hydroxykynurenine (by kynurenine mono-oxygenase) (32). 3-Hydroxykynurenine is further metabolized to quinolinic acid through a series of reactions mediated by 3-hydroxyanthraniate 3,4-dioxygenase (see Figure 2 for pathway). Kynurenic acid has neuroprotective properties (33,34), whereas kynurenine, 3-hydroxykynurenine, and quinolinic acid are considered neurotoxic (6,34,35). Interestingly, we observed that the neurotoxic kynurenine metabolites were inversely associated with D3Cr muscle mass, and were consistently higher in older men with low walking speed. Given that the function of the major enzymes controlling metabolic flux through the kynurenine pathway (ie, kynurenine aminotransferases and kynureninase) is dependent on PLP (ie, the active form of vitamin B6) and plasma PLP levels tend to decline with age (36), a possible explanation for the increased neurotoxic metabolites observed in the older men is low PLP. Indeed, PLP was inversely correlated with 3-hydroxykynurenine (r = −0.24, p < .0001), quinolinic acid (r = −0.13, p = .004), and Kyn/Trp (r = −0.14, p = .002) after adjustment of age, BMI, smoking status, multivitamin use, kidney disease, and physical activity. Nevertheless, PLP did not differ between men in the low groups compared to the random sample, nor was it significantly correlated with D3Cr muscle mass after adjustment for confounders and multiple comparisons. Hence, there are likely other factors beyond PLP driving the inverse associations of the neurotoxic metabolites with muscle mass and function.
Many recent studies have described a link between certain kynurenine pathway metabolites with age-related degenerative diseases and frailty (6,7,32,37,38); however, to date, few studies have assessed the relationship between kynurenine and related metabolites with physical performance indices indicative of muscle function. In a sample of older adults, kynurenine levels were shown to be associated with grip strength, gait speed, and chair stand after adjusting for age, sex, and BMI (38). Similarly, Westbrook et al. (6), using a larger sample size, showed that there was a negative association between the ratio of Kyn/Trp with walking speed and a positive association of tryptophan with both walking speed and grip strength. In our study, we observed that older men with the lowest walking speeds had higher levels of kynurenine and Kyn/Trp and lower levels of tryptophan, consistent with the results of these previous studies. After adjustment, tryptophan remained positively correlated with walking speed, although the association with kynurenine did not reach significance. Unlike the previous studies, we did not observe a significant relationship between tryptophan or kynurenine and grip strength. It is possible that differences in the populations could explain variations in results as the MrOS cohort consisted of relatively healthy men (ie, 16.5% and 5.6% of the total sample were classified as having mild and severe sarcopenia, respectively, according to European working group definition (39)) who mostly identified as White, while the other study cohorts included both sexes (6,38), with ~80% of the men and women in Jang et al. being classified as prefrail or frail (38).
The relationship between kynurenine and related metabolites with muscle mass has been less clear. In mice, kynurenine was shown to induce muscle atrophy and lipid peroxidation affecting muscle mass and function (40), indicating that kynurenine may play a mechanistic role in the development of sarcopenia. On the other hand, using ALM as a proxy for muscle mass, Murphy et al. (26) did not identify kynurenine as a metabolite being associated with ALM in older male adults. In contrast, using a more precise measure of total body muscle mass (D3Cr), we found a positive correlation of tryptophan, and negative correlations of kynurenine and Kyn/Trp, with muscle mass, with both tryptophan and Kyn/Trp being among the most strongly correlated biomarkers. It is possible that the inherent inclusion of nonmuscle tissue in the ALM measurement used in the previous study (Murphy et al.) may have masked the relationship between kynurenine and muscle. For instance, in a mouse study (40), although the overall quadricep weight did not change in kynurenine-treated mice, there was a significant decrease in muscle fiber size, indicating that there was likely an increase in the noncontractile tissue within the muscle. Thus, kynurenine seemed to alter the contractile to noncontractile ratio of the quadricep muscle, yet not the overall muscle mass of the mice. Hence, the use of ALM, which is a measure of whole muscle weight, including both contractile and noncontractile tissue mass, may be obscuring the relationship between the contractile/functional muscle mass and kynurenine.
Beyond kynurenine, we also observed several other metabolites in the kynurenine pathway to be related to low D3Cr muscle mass and walking speed. Of these were quinolinic acid and NAM/QA. Quinolinic acid is crucial for the endogenous production of nicotinamide and nicotinamide adenine dinucleotide (NAD+), a key electron carrier necessary for optimal mitochondrial energetics. As the terminal reaction of the pathway, quinolinate phosphoribosyl transferase converts quinolinic acid to nicotinamide. Preclinical studies have shown that supplementation with nicotinamide riboside (an exogenous NAD+ precursor) is capable of improving muscle function (41,42). Moreover, in the skeletal muscle of older adults with sarcopenia, decreased NAD+ levels have been found (43), whereas increased levels have been observed in older individuals who exercise regularly (44). Intriguingly, we observed that the NAM/QA ratio was one of the metabolites most positively related to D3Cr muscle mass and was also one of the metabolites with the greatest fold difference when comparing men with the lowest D3Cr muscle mass and walking speed with the random sample. These results support the idea that altered mitochondrial energetics and potentially, impairments in de novo NAD+ synthesis through the conversion of quinolinate to nicotinamide likely have a causal link to the loss of muscle mass and performance in aging (45).
As previously mentioned, quinolinic acid is considered to be neurotoxic as it is involved in the production of free radicals, oxidative stress, as well as excitotoxicity, all of which can have direct and indirect deleterious effects on muscle (34). The kynurenine aminotransferase 4 (otherwise known as GOT2), which converts kynurenine to kynurenic acid, the latter being neuroprotective, is located within the mitochondria of skeletal muscle and is under the control of the mitochondria biogenesis regulator PGC-1α (33). Considering its relatively high mass and associated mitochondrial content, it is plausible that skeletal muscle is a major tissue responsible for kynurenine detoxification to kynurenic acid and thus is a major determinant of the divergence of kynurenine away from other toxic metabolites such as 3-hydroxykynurenine and quinolinic acid. In support of this hypothesis, we observed that quinolinic acid and 3-hydroxykynurenine were both increased in men with low D3Cr muscle mass and walking speed compared to the random group. Moreover, after adjusting for important covariates, we observed that these metabolites were inversely related to D3Cr muscle mass. Along the same lines, Al Saedi et al. (7) reported that the ratio of kynurenic acid to quinolinic acid, which is representative of the balance between neuroprotective and neurotoxic metabolites, was positively related to gait speed as well as grip strength after adjustment for age and sex.
An interesting finding of our study was the consistent association between the folate metabolism marker hm-THF and low levels of D3Cr muscle mass, walking speed, and grip strength. There is scant literature about the biological activity of hm-THF, aside from suggestions that it is biologically inactive in rat (46) and in humans (47). It appears that hm-THF is an oxidation product of calcium L-methylfolate (L-5-MTHF-Ca), which is used as an alternative to folic acid to supplement food. We speculate that hm-THF may be a marker of poor diet quality. Since 1998, the Food and Drug Administration has required manufacturers to fortify cereal grain products (bread, rice, pasta, and cereal) labeled as enriched with folic acid. Nevertheless, when we adjusted our models by protein intake, which is highly correlated with diet quality (Supplementary Table 2), hm-THF remained inversely associated with D3Cr (FDR p value ≤.001) and borderline associated with walking speed (FDR p value = .24).
Our study has several strengths including the use of a well-characterized study population and state-of-the art methods to simultaneously assess multiple nutritional biomarkers. However, a few limitations must be considered. First, our population was limited to older men who were primarily White. Thus, our findings may not be generalizable to other groups including women, other racial and ethnic groups, and younger adults. Second, our results are cross-sectional, so the temporal relation of these biomarkers to muscle mass, strength, and performance measures cannot be established. Third, although we corrected for multiple comparisons, it is possible that these results were observed by chance, even though similar findings have been reported by others (6,28,30,31,38). Nevertheless, these findings should be confirmed in other data sets. Finally, although suggestive, our results cannot establish a causal relationship between the compounds we measured and muscle mass or performance.
Conclusion
In summary, we found that a number of nutrient biomarkers, particularly some branched-chain AAs, tryptophan, metabolites along the kynurenine pathway, as well as a marker of folate metabolism, were associated with D3Cr muscle mass and walking speed. No biomarkers were associated with grip strength with the exception of the folate metabolite, hm-THF. The complex and dynamic nature of sarcopenia makes the quest for identifying nutrient biomarkers challenging. Nevertheless, our findings strengthen the growing body of evidence indicating a relationship between specific AAs and kynurenine metabolites with sarcopenia. Future research should focus on intervention studies investigating whether perturbations to the kynurenine pathway and AA metabolism influence muscle loss and declines in mobility in older adults.
Funding
The Osteoporotic Fractures in Men Study is supported by National Institutes of Health funding. The following institutes provide support: the National Institute on Aging, the National Institute of Arthritis and Musculoskeletal and Skin Diseases, the National Center for Advancing Translational Sciences, and NIH Roadmap for Medical Research under the following grant numbers: U01 AG027810, U01 AG042124, U01 AG042139, U01 AG042140, U01 AG042143, U01 AG042145, U01 AG042168, U01 AR066160, R01 AG066671, and UL1 TR002369.
Conflict of Interest
None.
Acknowledgments
We thank Denis Breuillé for discussion of results and Julie Deuquet for project management.
Author Contributions
M.H.R.: formal analysis, writing; E.J.: formal analysis, visualization, writing; E.M.: data curation, methodology, writing; N.P.: formal analysis, writing; L.L.: formal analysis, project administration, writing; R.T.H.: conceptualization, writing; Y.G.: data curation, methodology; J.C.: data curation, methodology; T.E.R.: investigation, writing; L.F.: conceptualization, writing; J.N.F.: conceptualization, writing; E.S.O.: investigation, project administration, writing; P.M.C.: conceptualization, funding acquisition, project administration, supervision, writing.
