Journal Logo

Research Article: Observational Study

The minor C-allele of the rs2014355 variant in ACADS gene is associated with exercise-induced increase in HDL cholesterol levels in Taiwanese adults

Yang, Tzi-Peng PhDa; Shih, Fen-Fen MSb; Hsu, Ming-Yi PhDb,c; Tsai, Meng-Hsiun PhDd; Nfor, Oswald Ndi DVM, PhDe; Chen, Pei-Hsing MSe; Ho, Chien-Chang PhDf,g; Lin, Chuan-Chao MD, PhDh,i; Liaw, Yung-Po PhDe,j,∗

Editor(s): Zhao., Yingjie

Author Information
doi: 10.1097/MD.0000000000023838
  • Open


1 Introduction

Short-chain acyl-coenzyme A dehydrogenase is a homotetrameric flavoenzyme that catalyzes the initial step of the mitochondrial fatty acid beta-oxidation pathway.[1] It cuts long-chain free fatty acids (FFA) into short-chains (C3/C4) when they are inside the mitochondria.[1] This pathway plays an important role in energy metabolism especially during a physiological response to tissue energy depletion during periods of fasting, illnesses, and increased muscular activity. Mutations of the ACADS gene are associated with deficiency of the short-chain acyl-coenzyme A dehydrogenase protein (SCADD) which is marked by an increased amount of fatty acids, and the presence of increased butyrylcarnitine (C4) in blood plasma, and increased ethylmalonic acid (EMA) concentrations in urine.[2,8,17]

It is well known that HDL-C has a strong correlation with the risk of coronary atherosclerosis.[3,11,14] There is much evidence to show that exercise and diet are important in improving serum HDL-C levels, hence minimizing the risk of cardiovascular disease.[5,10,12] It has also been reported that regular endurance exercise training may increase HDL-C levels.[13] ACADS is one of the quantitative effectors of HDL-C.[15] An experiment on ACADS gene deletion in mice (Balb/cBy) showed higher levels of HDL regardless of gender or diet.[15] The mechanism underlying the relationship between HDL and ACADS is still unclear. Another research in ACADS-deficient mice also showed this gene protects mice against diet-induced obesity and insulin resistance.[2] Triglycerides, free fatty acids, cholesterol, glucose, and hepatic lipid were found to be lower in Balb/cBy mice.[17]

The ACADS gene is approximately 13 kb in length and has 10 exons at chromosome 12 (12q24.31).[18] Several single nucleotide polymorphisms (SNPs) of this gene have been reported.[8,18] In a genome-wide association study (GWAS), an association was found between ACADS rs2014355 SNP and the ratio between the short-chain acylcarnitines C3 and C4.[9] Mirkov et al also found that ACADS SNPs affect serum metabolomics traits: mRNA of ACADS in liver tissue of people with rs2014355 TC/CC than in those with TT genotype.[16] The minor C-allele of rs2014355 in the ACADS gene affects insulin release following oral glucose load.[4] The above data imply that ACADS rs2014355 genotypes will affect glucose and lipid metabolism.

However, studies that explore the effects of exercise and ACADS on HDL-C are limited. The purpose of this study was to investigate the association of exercise and ACADS rs2014355 genotypes with HDL-C.

2 Material and methods

The entire data used in this study were retrieved from the TWB database (2008–2015). Enrollment in the Taiwan biobank is restricted to Taiwanese adults between the ages of 30 and 70 with no personal history of cancer. In our study, data were available for 9684 participants (that is, 5515 physically inactive and 4169 active participants). Baseline characteristics included exercise, rs2014355 (TT, TC+CC genotypes), sex (women/men), age, smoking habits (never/former/current), drinking habits (never/former/current), BMI (normal/underweight/overweight/obese), vegetarian (no/yes), coffee-drinking habits (no/yes), and disease (coronary heart disease, hyperlipidemia, diabetes, and hypertension) status (yes/no). We obtained exercise data through TWB questionnaires and defined exercise as participating in any exercise activity for over 30 minutes per session (at least 3 times per week) for the last 3 months. Other lifestyle variables used in our analysis have been previously described.[6] Disease conditions including hyperlipidemia and coronary heart disease were determined using self-answered TWB questionnaires. Diabetes was determined by self-answered questionnaires and confirmed by a fasting blood glucose of over 126 mm Hg or HbA1c of over 6.5. Hypertension was also determined through questionnaires and confirmed by SBP and DBP of over 140 mm Hg and 90 mm Hg, respectively. All eligible TWB participants signed informed consent prior to data collection. The Institutional Review Board of Chung Shan Medical University Hospital approved this study (CS2–16114 and CS1-20009).

2.1 Statistical analysis

We performed statistical analyses using SAS version 9.4 (SAS Institute Inc., Cary, North Carolina, USA) and Plink 1.9. The rs2014355 SNP passed quality control (i.e., HWE P value ≥.001, minor allele frequency = 0.12, and call rate = 0.999). HDL-C levels across different exercise groups were compared using the Student t-test. Differences between categorical variables were tested using the chi-square test. Backward stepwise linear regression was used to assess the associations of rs2014355 genotypes and exercise with HDL-C levels.

3 Results

Overall, 9,684 participants were enrolled in this study (Table 1). There were 5515 physically active (exercise group) and 4169 inactive individuals (no exercise group). HDL-C levels were significantly different between the exercise and no exercise group (P < .0001). The mean HDL-C concentration was 52.63 mg/dl for no exercise and 54.68 mg/dl for the exercise group. Of the participants in the no-exercise group, 76.63% were those with the rs2014355 TT genotype while 23.37% were those with the TC+CC genotype. Among those in the exercise group, 78% were those with the TT genotype while 22% were those with the TC+CC genotype. There was no significant difference in HDL-C concentrations of those with TT+CC and those with the TT genotype (P = .1024) (Table 2). However, HDL-C levels were significantly higher among individuals in the exercise, compared to no exercise group (β=1.79856; P < .0001). There was a significant interaction between ACADS rs2014355 and exercise on HDL-C concentration (P = .0412). Regression analysis showed a significant association between TC+CC genotype and HDL-C only in the exercise group (i.e., β = 1.09785; P value = .0146). The β-value was −0.03754 (P = .9154) in the no-exercise group (Table 3). When no exercise and TT genotype was used as the reference group (Table 4), the β value was −0.02881, (P = .9372), for the TC+CC/no exercise group, 1.53846 (P < .0001) for the TT/exercise group, and 2.66237 (P < .0001) for the TC+CC/exercise group (Table 4).

Table 1 - Basic characteristics of the study participants.
No exercise Exercise
Variables n = 5515 n = 4169 P value
HDL-C 52.6272 ± 0.1759 54.6781 ± 0.2111 <.0001
rs2014355 .1097
 TT 4226 (76.63) 3252 (78.00)
 TC+CC 1289 (23.37) 917 (22.00)
Sex .1105
 Women 2937 (53.25) 2152 (51.62)
 Men 2578 (46.75) 2017 (48.38)
Age 46.8711 ± 0.1394 54.6006 ± 0.1579 <.0001
Cigarette smoking habit <.0001
 No 4167 (75.56) 3260 (78.20)
 Former 599 (10.86) 617 (14.80)
 Current 749 (13.58) 292 (7.00)
Alcohol drinking habit <.0001
 No 4905 (88.94) 3698 (88.70)
 Former 149 (2.70) 176 (4.22)
 Current 461 (8.36) 295 (7.08)
BMI <.0001
 Normal 2486 (45.08) 2072 (49.70)
 Underweight 180 (3.26) 74 (1.78)
 Overweight 1660 (30.10) 1278 (30.65)
 Obese 1189 (21.56) 745 (17.87)
Vegetarian .1686
 No 5230 (94.83) 3979 (95.44)
 Yes 285 (5.17) 190 (4.56)
Coffee drinking habit .0901
 No 3529 (63.99) 2737 (65.65)
 Yes 1986 (36.01) 1432 (34.35)
Coronary heart disease .0024
 No 5464 (99.08) 4102 (98.39)
 Yes 51 (0.92) 67 (1.61)
Hyperlipidemia <.0001
 No 5209 (94.45) 3853 (92.42)
 Yes 306 (5.55) 316 (7.58)
Diabetes <.0001
 No 4993 (90.53) 3642 (87.36)
 Yes 522 (9.47) 527 (12.64)
Hypertension <.0001
 No 4529 (82.12) 3088 (74.07)
 Yes 986 (17.88) 1081 (25.93)
Continuous variables are presented as mean ± standard error and categorical variables as n (%).

Table 2 - Multiple linear regression showing the association of ACADS rs2014355 and exercise with HDL-C levels.
Variables β P value
rs2014355 (ref: TT)
 TC+CC 0.45517 .1024
Exercise (ref: no)
 Yes 1.79856 <.0001
Sex (ref: Women)
 Men −7.68782 <.0001
Age 0.04148 .0007
Cigarette smoking habit (ref: No)
 Former −0.73999 .0600
 Current −3.29346 <.0001
Alcohol drinking habit (ref: No)
 Former −0.73943 .2713
 Current 3.34271 <.0001
BMI (ref: Normal)
 Underweight 7.43222 <.0001
 Overweight −5.14607 <.0001
 Obese −8.12695 <.0001
Vegetarian (ref: No)
 Yes −6.18723 <.0001
Coffee drinking habit (ref: No)
 Yes 1.05275 <.0001
Coronary heart disease (ref: No)
 Yes −2.22465 .0398
Hyperlipidemia (ref: No)
 Yes −1.67701 .0007
Diabetes (ref: No)
 Yes −3.92351 <.0001
Hypertension (ref: No)
 Yes −0.95193 .0022

Table 3 - Association between HDL-C and rs2014355 based on exercise habits.
No exercise Exercise
Variables β P value β P value
rs2014355 (ref: TT)
 TC+CC −0.03754 .9154 1.09785 .0146
Sex (ref: Women)
 Men −7.87164 <.0001 −7.39905 <.0001
Age 0.05524 .0004 0.01946 .3193
Cigarette smoking habit (ref: No)
 Former −0.50632 .3390 −1.06173 .0736
 Current −3.26175 <.0001 −3.25918 <.0001
Alcohol drinking habit (ref: No)
 Former −0.60034 .5295 −0.92207 <.0001
 Current 3.48232 <.0001 3.08553 <.0001
BMI (ref: Normal)
 Underweight 6.80027 <.0001 8.95338 <.0001
 Overweight −5.57660 <.0001 −4.63021 <.0001
 Obese −8.15255 <.0001 −8.24442 <.0001
Vegetarian (ref: No)
 Yes −5.79936 <.0001 −6.79596 <.0001
Coffee drinking habit (ref: No)
 Yes 1.12684 .0003 0.94829 .0165
Coronary heart disease (ref: No)
 Yes −3.17466 .0455 −1.50048 .3187
Hyperlipidemia (ref: No)
 Yes −2.05037 .0025 −1.32929 .0682
Diabetes (ref: No)
 Yes −3.73565 <.0001 −4.13118 <.0001
Hypertension (ref: No)
 Yes −0.68059 .1116 −1.15985 .0109
Interaction (rs2014355∗exercise) P value = .0412

Table 4 - Association of HDL-C based on exercise habits and rs2014355 genotypes.
Variables β P value
Exercise, rs2014355 (ref: No exercise, TT)
 No exercise, TC+CC −0.02881 .9372
 Exercise, TT 1.53846 <.0001
 Exercise, TC+CC 2.66237 <.0001
Sex (ref:Women)
 Men −7.68390 <.0001
Age 0.04146 .0007
Cigarette smoking habit (ref: No)
 Former −0.74611 .0579
 Current −3.29365 <.0001
Alcohol drinking habit (ref: No)
 Former −0.75567 .2609
 Current 3.35266 <.0001
BMI (ref: Normal)
 Underweight 7.42941 <.0001
 Overweight −5.15605 <.0001
 Obese −8.12641 <.0001
Vegetarian (ref: No)
 Yes −6.19380 <.0001
Coffee drinking habit (ref: No)
 Yes 1.05097 <.0001
Coronary heart disease (ref: No)
 Yes −2.21828 .0404
Hyperlipidemia (ref: No)
 Yes −1.70297 .0006
Diabetes (ref: No)
 Yes −3.92081 <.0001
Hypertension (ref: No)
 Yes −0.93990 .0025

4 Discussion

To our knowledge, this is the first study to examine the relationship between ACADS rs2014355 genotypes and HDL based on exercise habits. We found that the association between HDL-C and exercise differed with respect to the rs2014355 genotypes. Compared with the TT genotype, the TC+CC genotype together with exercise was more effective in improving HDL-C levels (that is, there was a 2.7 mg/dl increase in HDL-C levels of those with the TC+CC genotype compared to just 1.6 mg/dl among those with the TT genotype). In the absence of exercise, HDL-C concentrations were low and there were no differences regardless of the genotype.

A genome-wide GWAS carried out in Germany in 2010 pointed out that the metabolic capacity of the ACADS rs2014355 C allele is stronger than that of the TT genotype.[9] This is consistent with our study. Genes mRNA levels and their regulations/expressions are different in different tissues. Mirkov et al reported that the ACADS mRNA levels in liver tissues were higher among TC+CC (rs2014355) than in TT individuals.[16] ACADS is highly expressed in the small intestine, colon, and liver, but there is no data to show its expression in white blood cells.[7] In Taiwan biobank, DNA was extracted from white blood cells.

So far, it is well known that exercise can improve the performance of ACADS, HDL-C, and insulin sensitivity. Our results suggest that this phenomenon could be more pronounced in those with the minor C allele. Besides, studies by Hornbak et al in 2011 showed that the minor C-allele of rs2014355 was associated with reduced glucose-stimulated insulin released during an oral glucose tolerance test (OGTT) and may be mediated through an impaired β-oxidation of fatty acids.[4] This study also indicated that levels of HDL-C under OGTT did not differ between people with TT or TC+CC genotypes. In contrast, the studies cited above did not include exercise in their study design, as we have done. Of note, exercise plays a major role in the regulation of ACADS.

Our study was limited in that information on the intensity and frequency of exercise was not available in the Taiwan Biobank dataset. Further investigations in this area are recommended. HDL-C levels might be influenced by exercise type and intensity. This study adds more knowledge on exercise-related HDL-C elevation to improve cardiovascular disease.

5 Conclusions

The association between HDL-C and exercise differed based on the ACADS rs2014355 genotypes. Compared to the TT genotype, the TC+CC genotype in combination with exercise was associated with higher levels of HDL-C. These results suggest that people with the rs2014355 TT genotype may improve their HDL-C levels by increasing their physical activity. This might in turn help to reduce the risk of cardiovascular disease.

Author contributions

Conceptualization: Tzi-Peng Yang, Fen-Fen Shih, Ming-Yi Hsu, Meng-Hsiun Tsai, Oswald Ndi Nfor, Pei-Hsing Chen, Chien-Chang Ho, Chuan-Chao Lin, Yung-Po Liaw.

Data curation: Oswald Ndi Nfor, Pei-Hsing Chen, Chien-Chang Ho.

Formal analysis: Pei-Hsing Chen.

Methodology: Tzi-Peng Yang, Fen-Fen Shih, Ming-Yi Hsu, Meng-Hsiun Tsai, Oswald Ndi Nfor, Pei-Hsing Chen, Chien-Chang Ho, Chuan-Chao Lin, Yung-Po Liaw.

Supervision: Chuan-Chao Lin, Yung-Po Liaw.

Validation: Tzi-Peng Yang, Fen-Fen Shih, Ming-Yi Hsu, Meng-Hsiun Tsai, Yung-Po Liaw.

Writing – original draft: Tzi-Peng Yang.

Writing – review & editing: Fen-Fen Shih, Ming-Yi Hsu, Meng-Hsiun Tsai, Oswald Ndi Nfor, Pei-Hsing Chen, Chien-Chang Ho, Chuan-Chao Lin, Yung-Po Liaw.


[1]. Modre-Osprian R, Osprian I, Tilg B, et al. Dynamic simulations on the mitochondrial fatty acid beta-oxidation network. BMC Syst Biol 2009;3:2.
[2]. Wolfe L, Jethva R, Oglesbee D, et al. Short-Chain Acyl-CoA Dehydrogenase Deficiency. Seattle: University of Washington; 1993.
[3]. Siebel AL, Heywood SE, Kingwell BA. HDL and glucose metabolism: current evidence and therapeutic potential. Frontiers Pharmacol 2015;6:258.
[4]. Hornbak M, Banasik K, Justesen JM, et al. The minor C-allele of rs2014355 in ACADS is associated with reduced insulin release following an oral glucose load. BMC Med Genetics 2011;12:4.
[5]. Stefanick ML, Mackey S, Sheehan M, et al. Effects of diet and exercise in men and postmenopausal women with low levels of HDL cholesterol and high levels of LDL cholesterol. New Eng J Med 1998;339:12–20.
[6]. Tantoh DM, Lee K-J, Nfor ON, et al. Methylation at cg05575921 of a smoking-related gene (AHRR) in non-smoking Taiwanese adults residing in areas with different PM 2.5 concentrations. Clin Epigenetics 2019;11:69.
[7]. Wei M, Gibbons LW, Mitchell TL, et al. Alcohol intake and incidence of type 2 diabetes in men. Diabetes Care 2000;23:18–22.
[8]. Tein I, Elpeleg O, Ben-Zeev B, et al. Short-chain acyl-CoA dehydrogenase gene mutation (c.319C>T) presents with clinical heterogeneity and is candidate founder mutation in individuals of Ashkenazi Jewish origin. Molecular Genetics & Metabolism 2008;93:179–89.
[9]. Illig T, Gieger C, Zhai G, et al. A genomewide perspective of genetic variation in human metabolism. Nature Genetics 2010;42:127–41.
[10]. Sarzynski MA, Ruiz-Ramie JJ, Barber JL, et al. Effects of increasing exercise intensity and dose on multiple measures of HDL (High-Density Lipoprotein) function. Arteriosclerosis, Thrombosis Vascular Biol 2018;38:943–52.
[11]. Kronenberg F. HDL in CKD-The Devil Is in the Detail. J Am Society Nephrol 2018;29:1356–71.
[12]. Spate-Douglas T, Keyser RE. Exercise intensity: its effect on the high-density lipoprotein profile. Arch Physical Med Rehabilit 1999;80:691–5.
[13]. Couillard C, Despres JP, Lamarche B, et al. Effects of endurance exercise training on plasma HDL cholesterol levels depend on levels of triglycerides: evidence from men of the health, risk factors, exercise training and genetics (HERITAGE) family study arteriosclerosis. Thrombosis Vascular Biol 2001;21:1226–32.
[14]. Marz W, Kleber ME, Scharnagl H, et al. HDL cholesterol: reappraisal of its clinical relevance. Clin Res Cardiol 2017;106:663–75.
[15]. Su Z, Leduc MS, Korstanje R, et al. Untangling HDL quantitative trait loci on mouse chromosome 5 and identifying Scarb1 and Acads as the underlying genes. J Lipid Res 2010;51:2706–13.
[16]. Mirkov S, Myers JL, Ramirez J, et al. SNPs affecting serum metabolomic traits may regulate gene transcription and lipid accumulation in the liver. Metabolism: Clinical & Experimental 2012;61:1523–7.
[17]. Chen Y, Chen J, Zhang C, et al. Deficiency in the short-chain acyl-CoA dehydrogenase protects mice against diet-induced obesity and insulin resistance. FASEB J 2019;33:13722–33.
[18]. Corydon MJ, Andresen BS, Bross P, et al. Structural organization of the human short-chain acyl-CoA dehydrogenase gene. Mammalian Genome 1997;8:922–6.

ACADS gene; exercise; HDL-C; SNP

Copyright © 2021 the Author(s). Published by Wolters Kluwer Health, Inc.