γ-Aminobutyric acids (GABA) and serum GABA/AABA (G/A) ratio as potential biomarkers of physical performance and aging

Declining physical performance with age and disease is an important indicator of declining health. Biomarkers that identify declining physical performance would be useful in predicting treatment outcomes and identifying potential therapeutics. γ-aminobutyric acid (GABA), a muscle autocrine factor, is a potent inhibitor of muscle function and works as a muscle relaxant. L-α-aminobutyric acid (L-AABA) is a biomarker for malnutrition, liver damage, and depression. We sought to determine if GABA and L-AABA may be useful for predicting physical performance. Serum levels of GABA and L-AABA were quantified in 120 individuals divided by age, sex, and physical capacity into low, average, and high performer groups. Analyses explored correlations between serum levels and physical performance. Both GABA and the ratio of GABA/AABA (G/A), but not AABA, were highly positively associated with age (Pearson correlations r = 0.35, p = 0.0001 for GABA, r = 0.31, p = 0.0007 for G/A, n = 120). GABA showed negative associations in the whole cohort with physical performance [fast gait speed, 6 min walk test (6MWT), PROMIS score, and SF36PFS raw score] and with subtotal and femoral neck bone mineral density. L-AABA was positively associated with usual gait speed, 6MWT, total SPPB score, and SF36PFS raw score in the total cohort of 120 human subjects, also with 6MWT and SF36PFS raw score in the 60 male subjects, but no associations were observed in the 60 females. As both GABA and L-AABA appear to be indicative of physical performance, but in opposite directions, we examined the G/A ratio. Unlike GABA, the G/A ratio showed a more distinct association with mobility tests such as total SPPB score, usual and fast gait speed, 6MWT, and SF36PFS raw score in the males, regardless of age and metabolic status. Serum G/A ratio could be potentially linked to physical performance in the male population. Our findings strongly suggest that GABA, L-AABA, and the G/A ratio in human serum may be useful markers for both age and physical function. These new biomarkers may significantly enhance the goal of identifying universal biomarkers to accurately predict physical performance and the beneficial effects of exercise training for older adults.


Introduction
In the United States with the increase of life expectancy, the population with an age above 65 will be 20 -25 % of the total population by 2050 (1).Aging is believed to be a combination of body deterioration and disease (2).The gradual process of bodily deterioration, inevitable changes in cellular structure and function, that takes place throughout the life is known as primary aging.Secondary aging is the process resulting from disease, lifestyle and environmental in uences, and is often preventable through lifestyle choice or modern medicine (2).Aging has deleterious effects on muscle and bone, thus reducing physical performance (3) and increasing frailty, risk of fall and fractures, leading to chronic disability to the elderly population (4).The human body consists of more than 500 muscles, regulated by the nervous systems to interact with the skeleton (5).Their synergistic actions allow standing and moving -physical activity, and leads energy consumption (6), and this physical activity, exercise, is well-established as a countermeasure against secondary aging (7).Exercise has many bene cial effects on the cardiorespiratory system and muscle mass/function, preventing age-associated insulin resistance and diminished mitochondrial capacity in skeletal muscle (8).But the mechanisms behind this bene cial effect on skeletomuscular health are still unclear.Therefore, there is an unmet clinical need for biomarkers to detect the effects of physical activity on muscle performance in order to identify new strategies to achieve healthy aging.
Biomarkers have been postulated as essential variables to measure the effects of physical activity on the human body.Generally, these biomarkers are related with performance (muscle status, endocrine response, and oxygen transport), health (nutritional and hydration status, allergies), and recovery (in ammation, injury status and risk, muscle damage) (9).Factors with endocrine, autocrine, and paracrine action secreted during physical activity, including myokines and metabolites, can serve as potential biomarkers for muscle physical performance.Although some published studies have found associations between physical activity, performance, and biomarkers in cohorts with different age, race, and physiological or pathological conditions (10)(11)(12)(13)(14), there is still a need for additional universal biomarkers, useful for generalized populations.For example, a published study performed in a crosssectional population-based sample reported that physical activity and performance are associated with lower levels of in ammatory biomarkers (C-reactive protein, interleukin-6, and brinogen) in plasma from the elderly population (aged 65 years or older) (14).However, only the 400-meter walking test was used in this study as a performance-based measurement, and the levels of physical activity were classi ed based on self-reported information which limits the validity and strength of this study.Aminobutyric acids are nonproteinogenic amino acids which may be released during physical activity and impact physical performance.They comprise three types of isomers: α-aminobutyric acid (2aminobutyric acid, AABA), β-aminobutyric acid (3-aminobutanoic acid, BABA), β-aminoisobutyric acid (3aminoisobutyric acid, BAIBA), and γ-aminobutyric acid (4-aminobutyric acid, GABA).Except for GABA, each isomeric aminobutyric acid has two mirror-image enantiomers with the same molecular structures but signi cantly different biological functions.Our research laboratory has developed a fast and sensitive method for complete separation and accurate measurement of the aminobutyric acids in biological samples using liquid chromatography with tandem mass spectrometry (LC-MS/MS) (15).AABA has been postulated to have antioxidant activity and found to be a biomarker for alcohol liver injury, sepsis, malnutrition, and depression (16)(17)(18)(19).L-AABA was further reported as the only identi ed enantiomer in human and mouse biological uid samples including serum, plasma, and cerebrospinal uid, and showed signi cant association with expression levels of UPB1, a gene encoding βureidopropionase belonging to the CN hydrolase family and signi cant association with BMD (15).GABA is present in bacteria where it is functionally involved in spore germination, and it also confers resistance to acidic pH to several bacteria (20) in fungi where it works as a source of carbon and nitrogen (21) and in plants where it plays hey roles in growth and development (22).In humans, it is a major inhibitory neurotransmitter in the central nervous system that regulates internal neuronal communication but also acts as a muscle relaxant during sleep and in setting overall muscle tone (23).It is also an antioxidant and anti-in ammatory amino acid (24).Our previous studies have shown that GABA has potential for early prevention of post-menopausal osteoporosis (15).
In this study, serum concentrations of GABA, L-AABA, and their ratio were quanti ed from 120 healthy non-Hispanic human subjects (aged between 20-85 years) with completed information of physical characteristics and performance.We identi ed associations between serum levels and physical performance assessments, and further investigated the potential of these aminobutyric acids to be used as biomarkers of physical performance and age.These markers can also potentially serve as the cornerstone to investigate new therapies for aging-related musculoskeletal diseases.

Recruitment of human subjects
Serum samples and data were obtained from 120 individuals who had visited the Musculoskeletal Function, Imaging, and Tissue Resource Core (FIT Core) of the Indiana Center for Musculoskeletal Health's Clinical Research Center (Indianapolis, Indiana) between 3/2018 and 4/2019.The FIT Core serves to provide: 1) standardized performance of physical function tests and patient reported outcomes related to physical function, 2) imaging outcomes for body composition and bone health, and 3) the collection and banking of biological samples within the Indiana Biobank.
Participants are recruited to the FIT Core by investigators seeking outcomes related to musculoskeletal health for their research subjects, as well as via self-referral from the local community.The Core has Institutional Review Board approval from Indiana University to test all-comers who provide written informed consent, irrespective of current or previous health status.
The FIT Core collected samples and data from 1,518 individuals between 3/2018 and 4/2019.To be included in the current analyses, individuals needed to be 20-85 years of age, self-reported white and non-Hispanic, and without a self-reported major chronic disease.Individuals within each sex and each 15 year age group (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)35-50, 50-65, and 65+ yrs) were ranked for their performance on the FIT Core's hand grip strength test and test of the number of chair stands completed in 30 seconds.The 5 individuals within each sex and age range with the lowest, average, and highest composite rank were selected and grouped as low (LP), average (AP), and high (HP) performers, respectively (Table 1).The detailed characteristics of the study subjects shown in Supplementary Table 1.

Physical function
The FIT Core assessed dominant hand grip strength (Jamar Plus+ digital hand dynamometer; Sammons Preston, Bolingbrook, IL), the number of chair stands completed in 30 seconds, and the time taken to complete 5 chair stands, as we have previously described (25).In addition to raw values, grip strength and repeat chair stand outcomes were converted to age-and sex-matched z scores relative to reference data obtained in the FIT Core (25).Time to walk 4-m from a stationary start at normal speed (usual gait speed) and as quickly as possible without running (fast gait speed) were measured with a stopwatch and converted to speed (m/s), as we have previously reported (26).
Results from the repeat chair stand, usual gait speed, and a static balance test (ability balance for 10 seconds with feet side-by-side, semi-tandem, and tandem) were used to calculate the Short Physical Performance Battery (SPPB) score out of 12 (27).Distance walked in 6 minutes was measured according to the American Thoracic Society (28).The physical function (PF) domain of the NIH Patient Reported Outcomes Measurement Information System (PROMIS) computerized adaptive test (CAT) (PROMIS-CAT-PF) (version 1.2) and the physical functioning subscale of the Short Form 36 (SF-36 PF) were used to assess self-reported functional health.

Body composition and bone health
Appendicular skeletal muscle mass relative to height (ASM/height 2 ; kg/m 2 ) and whole-body aBMD, fat mass, percent were assessed by whole-body dual-energy x-ray absorptiometry (DXA) (Norland Elite; Norland at Swissray, Fort Atkinson, WI).Regional DXA using the same scanner assessed hip and spine aBMD.
Quanti cation of isomeric aminobutyric acids in human serum samples was followed the LC-MS/MS method published by our laboratory (15).Brie y, mobile phases are methanol (A) and water containing 0.005% formic acid and 2.5 mM ammonium formate (B).The MS instrument was operated and optimized under positive electrospray (+ESI) and multiple reaction monitoring modes (MRM).The m/z transitions (precursor to product ions) and their tuning voltages were selected from published paper (15) and further optimized based on the best MRM responses from instrumental method optimization software.All analyses and data processing were completed on Shimadzu LabSolutions V5.91 software (Shimadzu Scienti c Instruments, Inc., Columbia, MD).

Sample Preparation for LC-MS/MS analysis
Ten microliter human serum samples and equal volume of IS mixture solution (1.2 µM, 0.1% formic acid in methanol, v/v) were added to 35 µL 0.1% (v/v) formic acid in methanol, followed by 20 min-shaking at room temperature and another 15 min-centrifugation at 15,000 ×g, 4°C to precipitate the proteins.The supernatant was directly transferred to autosampler vial and 45 μL of each sample was injected for LC-MS/MS analysis.
The samples of standard calibration curves were prepared by spiking the pure standards in surrogate matrix 5% (w/v) BSA in PBS (pH7.4).The samples for ten-point calibration curves were prepared by diluting the working solution to 0.02-10.24µM for GABA, and 0.08-81.92µM for AABA.Then ten microliters of each standard sample were taken and treated following the same preparation procedures of serum samples for LC-MS/MS analysis.

Statistical analysis
Data were summarized as mean ± SD.Comparisons among groupss were performed using Student's ttest and one-way ANOVA with Tukey's post-Hoc test (α=0.05).Association analysis was performed using both Pearson (r) correlations and Spearman (ρ) correlations, while scatter plots were used to decide the appropriate correlations for interpretations.To control for the effects of age and BMI, partial correlations were further calculated.SAS 9.4 (SAS Institute, Cary, NC, USA) was used for statistical analysis.Twosided p-values <0.05 were considered as signi cant.

Quanti cation of GABA and AABA in human serum and correlations with age
Quantitation of the aminobutyric acids GABA and L-AABA from all 120 human samples gave concentrations of 0.146 ± 0.035 µM, and 20.6 ± 7.3 µM, respectively (Figure 1).Concentrations of GABA and L-AABA, and the ratio of GABA to L-AABA levels (G/A ratio) are summarized in Table 2. Next serum levels were analyzed comparing age, gender, and physical parameters.Serum GABA levels signi cantly increased with age in the overall cohort (p=0.0008,n=120) and HP group (p=0.0025,n=40).The average serum levels of GABA in each age group were 0.131 µM (20-35 yrs), 0.142 µM (35-50 yrs), 0.145 µM (50-65 yrs), and 0.166 µM (65 yrs and up), respectively.These results indicate a substantial increase in the serum levels of GABA at age 65, which is signi cantly higher than that in the younger populations aged 20-35 yrs group (p=0.00038) and 35-50 yrs group (p=0.032)(Figure 2A).Similar results were also obtained when calculating G/A ratio in serum.Average values of serum G/A ratio in different age groups were 0.0069 for 20-35 yrs, 0.0074 for 35-50 yrs, 0.0075 for 50-65 yrs, and 0.0097 for 65 yrs and up, respectively, and signi cant differences were also observed between the most senior group (age 65+) and other younger age groups (p<0.01, Figure 2B).No signi cant gender-speci c differences (p < 0.05) were observed (Table 3).
Next both Pearson and Spearman correlation tests were applied to investigate the associations of GABA and AABA serum levels with age in all 120 and on men and women populations (n=60) separately.GABA showed a signi cant and positive association with age in the overall cohort (r = 0.349, ρ = 0.317, both p < 0.001).Furthermore, this association is stronger in women with increased correlation coe cients, but weaker in men with reduced correlation coe cients (Figure 3).No signi cant association was obtained between serum L-AABA levels with age.The serum G/A ratio, exhibited similar positive associations with age as GABA (Figure 3).A signi cant and positive correlation was observed between serum G/A ratio and age in the overall cohort (r = 0.305, p=0.00072; ρ = 0.273, p = 0.0025) and in men (r = 0.369, p=0.0037; ρ = 0.312, p = 0.016), but it becomes much weaker in women (r = 0.231, p=0.076; ρ = 0.251, p = 0.053).

Signi cant correlations between aminobutyric acids with physical parameters and physical performance
All associations of GABA and AABA with p-values < 0.1 for physical performance and parameters are summarized in Table 3. Whereas GABA has positive association with age, it showed negative associations with physical performance (fast gait speed, 6MWT, PROMIS score, and SF36 PFS raw score) in whole cohort.This suggests that elevated physical performance might link with reduced serum GABA levels.GABA levels in serum also negatively correlated with two BMD values (subtotal, and femoral neck) in 120 human subjects.Furthermore, remarkable gender-related differences were observed when investigating the associations between GABA levels and physical characteristics and performance as GABA was only correlated negatively with femoral neck BMD in female subjects.
As GABA signi cantly associated with age (Pearson r=0.3486, p=0.0001;Spearman ρ=0.3168, p=0.0004Table 3, Supplementary Figure 1), and it is well known that both age and BMI are highly associated with physical performance, BMD values, and lean/fat mass in both genders (Supplementary Table 2), we further assessed these correlations after controlling the effect of age and/or BMI by partial correlation tests.The results summarized in Table 4, Supplementary Figure 2 indicating that associations between serum GABA levels and physical performance scales, BMDs, and lean/fat mass values diminished in partial correlation tests in all physical capacity groups.Only two performance scales, best grip strength in the male HP group (Pearson and Spearman) and SF36 PFS raw score in male AP group (Pearson), showed negative associations with GABA levels after controlling the effects of age and BMI (Table 4, Supplementary Figure 2).
3.3 Signi cant positive correlations between AABA with physical parameters and physical performance, but not with age.
L-AABA also exhibited associations with physical performance, but unlike GABA, all associations of L-AABA are positive.L-AABA is positively associated with usual gait speed, 6MWT, total SPPB score, and SF36 PFS raw score in 120 human subjects, also with 6MWT and SF36 PFS raw score in the 60 male subjects, but no correlations were observed in the 60 female participants (Table 3, Supplementary Figure 1).Interestingly, L-AABA showed more noteworthy gender-dependent differences in the individuals with high (HP) and low physical performance (LP) based on the results of physical performance assessment.Results summarized in Table 5 show that L-AABA positively associated with usual gait speed, fast gait speed, 6MWT, total SPPB score, and SF36 PFS raw score in 80 subjects from HP and LP groups, and stronger positive associations were found in males from both young (20-50 years old) and the aged (50-85 years old) populations.For example, associations (Pearson correlation) with 6MWT in all participants (HP and LP) is 0.236 (p = 0.038, n = 80) for both genders and 0.387 (p = 0.015, n = 40) for men.In the young participants (HP and LP), associations with 6MWT are 0.290 (p = 0.069, n = 40) for both genders and 0.439 (p = 0.053, n = 20) for men.In addition, almost no correlations between L-AABA and physical performance scales were observed in females in the HP group.All these results suggest that the serum levels of both GABA and L-AABA might be indicative of physical performance, but in opposite directions, which together could make them extremely useful biomarkers.

Correlations between serum G/A ratio with physical parameters and performance
As the associations of serum levels of GABA and L-AABA with human physical capacity were found to be opposite, we proposed that the ratio of GABA and L-AABA levels in serum could potentially serve as a more precise predictor of physical performance.Thus, we explored the relationships of serum G/A ratio with physical parameters in males and females separately, by using both Pearson and Spearman correlation tests (Table 6, Figure 4).Serum G/A ratio in males was found to be negatively associated with 6 physical performance scales, including total SPPB score (r = -0.401,p = 0.002), usual gait speed (r = -0.334,p = 0.009), fast gait speed (r = -0.344,p = 0.007), 6MWT (r = -0.399,p = 0.002), and SF36 PFS raw score (r = -0.515,p < 0.0001).Additionally, serum G/A ratios negatively correlated with femoral neck BMD and total lean mass, and positively correlated with total fat percent in men (Table 6, Figure 4).However, no associations were observed for serum G/A ratio with any physical performance and physical parameters but age and BMI in females.These results indicate that serum G/A ratio could be potentially linked to physical performance in the male population.Then we further performed partial Pearson and partial Spearman tests to evaluate the correlations of serum G/A ratio with physical parameters after controlling the effect of age and/or BMI.Pearson correlation results revealed that 5 out of 6 statistically signi cant associations without the covariate adjustments (ordinary) still have signi cantly negative associations after controlling the effect of age and age + BMI (partial) in the male population (Table 6, Figure 4).All associations with BMD and lean/fat mass diminished greatly in partial correlation tests.This suggests that serum G/A ratios might negatively associate with physical performance in men, no matter the age and or metabolic status.A positive Spearman and Pearson association of G/A ratio with the total fat percentage (ρ 0.56, p 0.01, and r 0.616, p 0.004) and negative with the total lean mass (ρ -0.52, p=0.02 and r -0.518, p=0.019) was found in the subgroup of HP males, and still present after controlling the effect of age and age + BMI (partial).This group also had a negative Pearson correlation with the grip strength, PROMIS score and SF36 PFS raw score.The LP male group similarly showed a negative Pearson correlation with the physical parameters of fast gait speed, SF36 PFS raw score and total SPPB score.The HP female group had a negative Pearson and Spearman correlation with the usual gait speed and the LP female group with the PROMIS score.

Discussion
Aging muscle undergoes several morphological changes along with declining regenerative capacity, which affects muscle strength and physical performance (8).As the inability to modulate corticospinal excitability has been linked to declined motor performance with advancing age (29), probing the roles GABA, a main inhibitory neurotransmitter in brain, in the age related musculoskeletal disorders is signi cant and can lead to new therapeutic venues for these diseases, including sarcopenia.GABA is a highly conserved molecule from bacteria to humans with essential roles in all kingdoms (20)(21)(22).It has a dual action on neurotransmission in the mammalian central nervous system via two distinct classes of GABA receptors, ionotropic GABA-A and metabotropic GABA-B receptors, which differ in their pharmacological, electrophysiological, and biochemical properties (30).GABA possesses an excitatory effect mediated by GABA-A receptors, while activation of presynaptical GABA-B receptors impedes opening of voltage-activated Ca 2+ channels and reduces neurotransmitter release, thus GABA can inhibit its own release through a negative feedback loop via GABA-B receptors present on GABAergic axons (31).This dual action of GABA could be dose-dependent and inhibitory at low concentrations and excitatory to some GABA receptors at higher concentrations (32).Currently most studies on aging -related GABA changes focus on altered brain GABA levels (33,34), while the associations between circulating GABA levels, age, and aging-related physical performance are seldom reported.Age related CNS GABA disorders are noteworth and include: Parkinson's disease, Alzheirmer's disease, anxiety, depression, schiszophrenia, autism spectrum disorder, etc (35).Another very large group of diseases that nds its cause in defects in GABA metabolism are spasticity disorders.It is interesting that these diseases possess a very large peripheral muscle component to their pathophysiology.Since GABA is an isomer of BAIBA and we have reported that ex vivo contracting EDL and soleus muscles release BAIBA in a myokine fashion (36) , and C2C12 muscle cells release both BAIBA and GABA (37), and GABA has been considered to have limited ability to cross well the blood brain barrier, in humans it is feasible that to assume that GABA measured in serum derives mostly from skeletal muscles.
In the 120 serum samples collected from healthy humans covering both genders, from young to the elderly, and with different physical capacities, we demonstrated that GABA levels in the circulation are signi cantly associated with age.It increases consistently with aging in both men and women, however a positive association with physical activity was also observed but only in the female HP and female LP groups.The G/A ratio could be another good biomarker for aging with the same positive, but stronger correlations than GABA.Interestingly, when considering physical performance, a signi cant serum G/A ratio was only obtained in the male HP and male LP groups.This gender-related difference between GABA and the G/A ratio may come from L-AABA.Even though L-AABA levels did not show any associations with the age in the whole cohort, it was found to be negatively correlated with age in HP and LP males.Taking both blood GABA levels and G/A ratios into consideration may signi cantly increase the strength of studies to discover more universal aging biomarkers in the general population.
Skeletal muscle physical performance is regulated by factors associated with the nervous, muscular, and skeletal systems.It has been well studied that muscle function is signi cantly related to the measures of BMD (femoral neck BMD, spine BMD, hip BMD, etc.) and body composition (fat percentage, lean mass, etc.) (38)(39)(40).Aging-related muscle function loss is often accompanied by an increased fat mass and percentage along with reduced lean mass and BMD.Our group has recently reported that two genes coding for the GABA receptors GABBR2 and GLRA1 and the gene coding for glutamate decarboxylase GAD1, showed an association with total body BMD, and that GABBR2 is associated with bone fractures (15).These genes are involved in both bone and muscle function and likely further in uence overall musculoskeletal function and body metabolism (15).In the present study, femoral neck BMD was found to be negatively correlated with GABA in the whole cohort and female group, and with G/A ratio in male group.When physical performance was taken into consideration, a strong negative association was observed between GABA and spine BMD in female HP, while the G/A ratio negatively associated with femoral neck BMD in male AP and LP groups.Moreover, serum GABA levels and G/A ratios also exhibited generally negative associations with lean mass and positive associations with fat mass/percentage in the different physical performance groups.
GABA levels were also negatively associated with various physical performance scales in our study, except for the 5min-chair-stand.But this positive association with the chair-stand further con rmed that serum GABA concentrations increased when the physical performance declines.This negative correlation between serum GABA levels and physical performance might be age-related, as the results of further partial association tests indicated that these correlations between GABA and performance scales diminished when controlling for the effects from age.All these ndings suggest that aging is a key factor affecting GABA levels in circulation.
Even though no direct signi cant positive correlations were observed between GABA and physical performance assessments, GABA levels in the body are known to be affected by physical activity.Exercise can stimulate PGC1α-mediated GABA secretion from muscle (41).GABA concentrations in systemic circulation, gastrocnemius and quadriceps were increased in exercising mice and humans, suggesting that GABA may function as an exercise/PGC1α-mediated myokine (41).Exercise also increases GABA levels in brain (42).Our previous studies in 136 young to middle-aged healthy Caucasian women (aged 21-41 years) reported that serum GABA showed a positive association (ρ=0.31,p=0.0055) with physical activity (times per week) in lean women (BMI 18.5-24.9kg m -2 ), but no association in obese women with BMI ≥ 25.0 kg m -2 (15).This might be caused by a differential metabolomic pro le of GABA in muscle and adipose tissues, indicating that metabolic homeostasis (anabolic/catabolic balance, hormone/protein/amino acid de ciencies, etc.) during aging may affect GABA levels in the circulation.In the present study, regardless of the in uence of age, serum GABA concentrations in the HP group were higher than that in AP and LP groups in the whole cohort.As the level of performers is classi ed according to performance scales on grip strength and repeat chair stand test, this result further supports a widely accepted concept that the higher physical performance is always linked to more dynamic physical activities.Exercise can increase myo ber size, whole muscle mass, muscle quality, and attenuates age-related decreases in muscle strength (8).Elevated GABA level in serum induced by physical activity may help enhance muscle hypertrophy, then assist muscle growth and maintain the physical performance during aging, and GABA also plays key roles in maintaining optimal muscle tonicity (32).
Unlike GABA levels, the associations between G/A ratio with physical performance assessments appear to be more independent from age.Particularly in the 60 male participants, 5 out of 6 negative associations with various physical performance tests of mobility (total SPPB score, usual gait speed, fast gait speed, 6 min walk test, and SF36 PFS raw score) were still signi cant after controlling for the effect of age using partial Pearson correlation analysis.SF36 PFS raw scores also showed medium to strong correlations (r-values ranged from -0.4 to -0.6) with serum GABA levels in all three male HP, AP and LP groups.The short Form-36 Physical Functioning Scale (SF36 PFS) questionnaire is a set of patient selfreporting measures to indicate health status.The SF-36 PFS is composed of 10 items encompassing a hierarchical range of di culties on everyday activities, including vigorous activities, moderate activities, lifting/carrying, climbing stairs, bending/kneeling/stooping, and waling, etc., and it has been used as a reliable and valid "stand-alone" instrument in research describing activity limitations in the elderly living in the community, and patients with different minor pathophysiological conditions and chronic diseases (43)(44)(45).These results strongly suggest that G/A ratio could be a potential circulating biomarker to indicate change in physical performance, particularly mobility performance in men.
One of major limitations of published studies on the biomarkers for physical activity/performance is its cross-sectional nature.The direction of the causal pathways between physical activity and potential biological markers remains unclear.It is di cult to know if the altered muscle status (mass, strength, power, fatigue, etc.) is responsible for or affects the levels of biological factors in circulation, or vice versa.However, the signi cant associations with age in our study directly indicate that the change in serum GABA levels and G/A ratio in humans are caused by the aging progress.

Conclusion
Together, these observations suggest that serum aminobutyric acids, especially GABA and G/A ratio, are closely associated with aging-related physical performance.As both intrinsic (in ammation, apoptosis, mitochondria, calcium metabolism, etc.) and extrinsic (endocrine, nutritional status, immobility, etc.) factors contribute to progressively defective myogenesis and muscle dysfunction during aging, considering multiple serum biomarkers may more accurately predict the correlations among aging, physical performance, and the bene cial effects from exercise.These studies could also open up new therapeutical venues for muscle disorders such as stiff man syndrome and sarcopenia.wrote the manuscript.MB, LB, and SW provided critical advice, discussed the work, and edited the manuscript.All authors revised and nalized the manuscript before submission.
Tables Table 1.Characteristics on females and males by performance (low, average, high).virtually indicate the correlations between aminobutyric acids (GABA, L-AABA, and ratio) and age in whole cohort (both gender), or male and female separately.

Figures
Figures

Figure 1 Comparison
Figure 1

Figure legend :
Figure legend: Scatter plots with con dence ellipses (red line, 95% con dence level) were applied to

Table 2 .
Concentrations of aminobutyric acids GABA and L-AABA, and ratios of GABA/L-AABA (G/A ratio) in serum from different age or gender populations.

Table 3 .
Correlations of aminobutyric acids GABA and L-AABA with physical parameters and performance in 120 white non-Hispanic individuals (aged 20-85 years).Pearson correlations (r) and Spearman Rank correlations (ρ) with p-value < 0.1 are shown in the Table.