Abstract
Aim
Several genetic loci related to lean mass have been identified in healthy individuals by genome‐wide association studies; however, the contribution of these loci to body composition in type 2 diabetes remains to be investigated. Here, we aimed to clarify the genetic determinants of body composition in individuals with type 2 diabetes.
Methods
A total of 176 Japanese outpatients (70 women and 106 men) with type 2 diabetes were studied using a cross‐sectional design. Body composition was measured using bioimpedance analysis with a commercially available device (InBody770). Single‐nucleotide polymorphisms in IRS1 (rs2943656), HSD17B11 (rs9991501), VCAN (rs2287926), ADAMTSL3 (rs4842924) and FTO (rs9936385) were evaluated by genotyping. The contributions of single‐nucleotide polymorphisms to body composition were examined, considering known clinical determinants.
Results
Sex, body composition and age were identified as clinical predictors. IRS1 rs2934656 was identified as an independent predictor of skeletal muscle mass (β = 0.11, P = 0.026), and ADAMTSL3 rs4842924 was an independent predictor of body fat mass (β = 0.15, P = 0.0095) and appendicular lean mass (β = −0.13, P = 0.017).
Conclusions
The findings clarified the contribution of genetic factors – IRS1 and ADAMTSL3 – to interindividual variation in body composition, independent of clinical factors, in type 2 diabetes patients. These data will contribute to the establishment of effective methods for the prediction, prevention, and intervention of sarcopenia and frailty in diabetes patients. Geriatr Gerontol Int 2021; 21: 932–938.
Keywords: ADAMTSL3 , aging, body compositions, IRS1 , type 2 diabetes
Introduction
Lean body mass, consisting primarily of skeletal muscle and internal organs, is a component of the body composition calculated by subtracting fat and bone mass from the total body mass. Body composition is closely associated with glucose metabolism, because skeletal muscle is a major site for insulin‐stimulated glucose uptake, 1 and adipose tissue contributes to insulin resistance. 2 , 3 Progressive changes in lean and fat mass are observed with aging in healthy individuals and in patients with diabetes, and often cause unfavorable metabolic changes in the elderly. 4 In particular, the age‐related loss of skeletal muscle mass (SMM) and strength could lead to functional impairment, physical disability, and even mortality, referred to as sarcopenia and frailty. 5 We have recently shown that the diabetic state, as well as hyperglycemia, are correlated with reductions in SMM and strength, indicating a close relationship between glucose metabolism and lean mass. 6 Additionally, sex is a well‐known factor affecting body composition. Accordingly, age, sex and diabetic state are important clinical determinants of lean mass.
With respect to genetic factors for lean mass, family and twin studies have estimated a heritability of 0.52–0.60, 7 , 8 suggesting that this trait is highly heritable in healthy adults. A large meta‐analysis of genome‐wide association studies supported the contribution of genetic factors to lean mass in healthy adults. 9 However, the contributions of these loci to indices of body composition have not been fully elucidated in individuals with type 2 diabetes. Therefore, in the present study, we evaluated the contribution of genetic factors to body composition in type 2 diabetes, considering established clinical determinants.
Methods
Participants
A total of 176 consecutive Japanese ambulatory outpatients (70 women and 106 men) with type 2 diabetes were recruited from November 2016 to January 2018 at Kindai University Hospital, located in Japan, and studied using a cross‐sectional design (Appendix S1. Supplementary methods). The clinical characteristics of participants are described in Table 1. The present study was approved by the appropriate ethics committee, and written informed consent was obtained from all participants.
Table 1.
Clinical characteristics of participants with type 2 diabetes
| Characteristic | n = 176 |
|---|---|
| Sex (female/male) | 70/106 |
| Age, years (range) | 67.4 ± 9.5 (38–92) |
| Proportion of elderly participants † (%) | 70.5 |
| Duration of diabetes (years) | 15.2 ± 10.7 |
| BMI (kg/m2) | 24.9 ± 4.7 |
| HbA1c (%) | 8.7 ± 1.7 |
| Fasting plasma glucose (mg/dL) | 141.8 ± 38.3 |
| Serum albumin (g/dL) | 4.0 ± 0.4 |
| eGFR (mL/min/1.73 m2) | 69.1 ± 21.1 |
| NDR/SDR/>pre‐PDR (n) | 131/20/25 |
| Treatment for diabetes (n) | |
| Insulin | 104 |
| Biguanide | 67 |
| DPP4 inhibitor | 62 |
| Sulfonylurea/glinide | 35 |
| GLP‐1 receptor agonist | 29 |
| SGLT2 inhibitor | 23 |
| Thiazolidine | 6 |
Aged ≥65 years.
BMI, body mass index; DPP4, dipeptidyl peptidase‐4; eGFR, estimated glomerular filtration rate; GLP‐1, glucagon‐like peptide‐1; HbA1c, hemoglobin A1c; NDR, no diabetic retinopathy; PDR, proliferative diabetic retinopathy; SDR, simple diabetic retinopathy; SGLT2, sodium–glucose cotransporter 2.
Measurements of body composition
A bioelectrical impedance analysis (BIA) was carried out for the measurement of total and segmental body composition using a commercial device (InBody770, Inbody Japan, Tokyo, Japan; Appendix S1. supplementary methods). Total lean mass (TLM) was defined as the sum of the soft lean mass except for lipids and bone minerals in the whole body, and appendicular lean mass (ALM) was estimated as the sum of the soft lean mass of both arms and both legs. Body fat mass (BFM) was defined as the sum of the lipids in the whole body. Body resistance (R) was used to estimate the SMM in the whole body, according to a previously described formula (Appendix S1. Supplementary methods). 10
Genetic analysis
Genomic DNA was extracted from peripheral leukocytes of participants using a standard phenol‐chloroform method, after obtaining informed consent, and stored at 4°C at the Kindai University Faculty of Medicine, Osaka, Japan. Five single‐nucleotide polymorphisms (SNPs) previously associated with lean mass were genotyped: rs2943656 in IRS1, rs9991501 in HSD17B11, rs2287926 in VCAN, rs4842924 in ADAMTSL3 and rs9936385 in FTO. 9 Genotyping was carried out using the TaqMan SNP genotyping assay according to the manufacturer's instructions (Applied Biosystems, Tokyo, Japan; Appendix S1; supplementary methods). Allele frequencies were estimated by direct counting.
Statistical analysis
The relationship between body composition and other continuous variables was analyzed by a simple linear regression analysis (least squares method). A multiple linear regression analysis with the forward‐backward stepwise selection method was carried out to identify variables that were independently associated with a change in indices of body composition. The χ2‐test and Fisher's exact probability test were used to determine the significance of differences in the distributions of the number of participants and alleles. Statistical significance was defined as P < 0.05.
Results
Correlations between age and body compositions in type 2 diabetes patients
The relationships between indices of body composition and age are shown in Fig. 1. TLM (r = −0.24, P = 0.001), ALM (r = −0.24, P = 0.001) and SMM (r = −0.38, P = 1.8 × 10−7) were significantly negatively correlated with age (Table 2). BFM did not show a significant correlation with age (r = 0.005, P = 0.95, Appendix S2; supplementary results).
Figure 1.
Correlations between indices of body composition and age. (a) Total lean mass (kg/m2) and age (years); r = −0.24, P = 0.001. (b) Appendicular lean mass (kg/m2) and age (years); r = −0.24, P = 0.001. (c) Skeletal muscle mass index (kg/m2) and age (years); r = −0.38, P = 1.8 × 10−7. (d) Body fat mass index (kg/m2) and age (years). Open circles: women, closed squares: men.
Table 2.
Simple regression analysis of indices for body compositions in type 2 diabetes patients
| Total lean mass | Appendicular lean mass | Skeletal muscle mass | Body fat mass | |||||
|---|---|---|---|---|---|---|---|---|
| r | P | r | P | r | P | r | P | |
| Age | −0.24 | 0.001 | −0.24 | 0.001 | −0.38 | 1.8 × 10−7 | 0.005 | NS |
| Duration of diabetes | 0.018 | NS | 0.024 | NS | 0.054 | NS | 0.068 | NS |
| BMI | 0.68 | 1.6 × 10−25 | 0.64 | 1.3 × 10−21 | 0.66 | 3.1 × 10−23 | 0.91 | 2.0 × 10−70 |
| HbA1c | 0.029 | NS | 0.029 | NS | 0.025 | NS | 0.098 | NS |
| Serum albumin | 0.074 | NS | 0.074 | NS | 0.10 | NS | 0.018 | NS |
| Total lean mass | – | 0.97 | 3.8 × 10−108 | 0.86 | 6.3 × 10−54 | 0.41 | 2.1 × 10−8 | |
| Appendicular lean mass | 0.97 | 3.8 × 10−108 | – | 0.88 | 6.1 × 10−57 | 0.35 | 1.7 × 10−6 | |
| Skeletal muscle mass | 0.86 | 6.3 × 10−54 | 0.88 | 6.1 × 10−57 | – | 0.31 | 3.5 × 10−5 | |
| Body fat mass | 0.41 | 2.1 × 10−8 | 0.35 | 1.7 × 10−6 | 0.31 | 3.5 × 10−5 | – | |
Total n = 176.
BMI, body mass index; HbA1c, hemoglobin A1c; NS, not significant.
Correlations between clinical variables and body composition
To elucidate the clinical determinants of each body composition index other than age and sex, a simple regression analysis was carried out (Table 2). The duration of diabetes, hemoglobin A1c levels and serum albumin levels did not show significant correlations with any index of body composition (TLM, ALM, SMM and BFM). Body mass index (BMI) showed a highly significant correlation with body composition (TLM: r = 0.68, P = 1.6 × 10−25; ALM: r = 0.64, P = 1.3 × 10−21; SMM: r = 0.66, P = 3.1 × 10−23; BFM: r = 0.91, P = 2.0 × 10−70). TLM (r = 0.41, P = 2.1 × 10−8), ALM (r = 0.35, P = 1.7 × 10−6) and SMM (r = 0.31, P = 3.5 × 10−5) showed significant positive correlations with BFM.
Multiple linear regression analysis of body composition with clinical variables, including diabetes medications
To further study the clinical determinants of body composition, including medications for diabetes treatment, a multiple linear regression analysis was carried out (Table S1). BMI was excluded as an independent variable owing to multicollinearity with body composition (tolerance <0.1). Sex, age and body composition indices (BFM for TLM/ALM/SMM and SMM for BFM) were significantly correlated with indices of body composition as independent predictors. With respect to medications for diabetes treatment, insulin/sulfonylurea/glinides, biguanide and thiazolidine were rejected as independent predictors for all indices of body composition. Sodium–glucose cotransporter 2 inhibitor treatment was not correlated with TLM (β = 0.083, not significant [NS]), ALM (β = 0.086, NS) or SMM (β = 0.085, NS). Dipeptidyl peptidase‐4 (DPP4) inhibitor treatment was not correlated with ALM (β = 0.069, NS). Glucagon‐like peptide‐1 (GLP‐1) receptor agonist treatment was significantly correlated with BFM (β = 0.14, P = 0.025).
Contribution of genetic loci and clinical parameters to lean mass
To determine the contribution of genetic loci to lean mass, including clinical determinants as independent variables, five loci (IRS1, HSD17B11, VCAN, ADAMTSL3 and FTO), previously identified by genome‐wide association studies in healthy individuals, were genotyped for all participants (Table S2). 9 No variants were observed for HSD17B11 rs9991501 in the study participants. Therefore, these four SNPs were included as independent variables in a multiple regression analysis for body composition. Among the medications for diabetes treatment, insulin/SU/glinide, biguanide, and thiazolidine were excluded as independent variables based on the multiple linear regression analysis of the clinical determinants of body composition.
As the strongest determinant of body composition, sex showed a significant positive correlation with TLM (β = 0.58, P = 2.4 × 10−20), ALM (β = 0.62, P = 3.6 × 10−23) and SMM (β = 0.58, P = 2.3 × 10−22), indicating that the mass was higher in men than in women (Table 3). In contrast, sex showed a significant negative correlation with BFM (β = −0.58, P = 3.6 × 10−16), indicating that the mass was lower in men than in women. As the second strongest determinant, BFM showed a significant positive correlation with TLM (β = 0.58, P = 2.4 × 10−20), ALM (β = 0.56, P = 4.1 × 10−19) and SMM (β = 0.48, P = 1.8 × 10−10), and SMM showed a significant positive correlation with BFM (β = 0.67, P = 1.7 × 10−16), indicating that it was the strongest determinant of BFM. Age showed a strong negative correlation with TLM (β = −0.21, P = 0.000095), ALM (β = −0.21, P = 0.0001) and SMM (β = −0.35, P = 1.3 × 10−10), indicating that lean mass, as well as SMM, declined with age. In contrast, BFM showed a significant positive correlation with age (β = 0.25, P = 0.00016), indicating that BFM increases with age.
Table 3.
Multiple regression analysis of indices of body composition in type 2 diabetes including genetic loci as independent variables
| Objective variables |
Sex 0: Female 1: Male |
Body composition | Age |
IRS1 rs2943656 0:A/A, G/A 1: G/G |
ADAMTSL3 rs4842924 0:C/C, T/C 1: T/T |
FTO rs9936385 0:C/C, T/C 1: T/T |
Drug 0: No 1: Yes |
|||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| β | P | β | P | β | P | β | P | β | P | β | P | β | P | |
|
Total lean mass R 2 = 0.54 |
0.58 | 2.4 × 10−20 | Body fat mass | −0.21 | 0.000095 | — | −0.08 | NS | — | SGLT2i | ||||
| 0.58 | 2.4 × 10−20 | 0.08 | NS | |||||||||||
|
Appendicular lean mass R 2 = 0.56 |
0.62 | 3.6 × 10−23 | Body fat mass | −0.21 | 0.0001 | — | −0.13 | 0.017 | — | SGLT2i | ||||
| 0.56 | 4.1 × 10−19 | 0.09 | NS | |||||||||||
|
Skeletal muscle mass R 2 = 0.59 |
0.58 | 2.3 × 10−22 | Body fat mass | −0.35 | 1.3 × 10−10 | 0.11 | 0.026 | −0.09 | NS | 0.08 | NS | SGLT2i | ||
| 0.48 | 1.8 × 10−10 | 0.08 | NS | |||||||||||
|
Body fat mass R 2 = 0.42 |
−0.58 | 3.6 × 10−15 | Skeletal muscle mass | 0.25 | 0.00016 | — | 0.15 | 0.0095 | — | GLP‐1ra | ||||
| 0.67 | 1.7 × 10−16 | 0.14 | 0.016 | |||||||||||
Total n = 176. The independent variables for each model are listed below. No linear relationship was observed between the independent variables in all models. Smooth muscle mass (SMM) index, total lean mass index, or appendicular lean mass index as dependent variables: age, sex, fat mass index, sodium–glucose cotransporter 2 inhibitor (SGLT2i), dipeptidyl peptidase‐4 inhibitor, glucagon‐like peptide‐1 receptor agonist (GLP‐1ra), ADAMTSL3 rs4842924, VCAN rs2287926, IRS1 rs2943656 and FTO rs9936385. Fat mass index as dependent variables: age, sex, SMM index, SGLT2i, dipeptidyl peptidase‐4 inhibitor, GLP‐1ra, ADAMTSL3 rs4842924, VCAN rs2287926, IRS1 rs2943656 and FTO rs9936385. NS, not significant.
As for the genetic determinants of body composition, IRS1 rs2943656 showed a significant positive correlation with SMM (β = 0.11, P = 0.026), suggesting that participants with the G/G genotype had a higher mass than that of individuals with other genotypes. ADAMTSL3 rs4842924 was significantly negatively correlated with ALM (β = −0.13, P = 0.017), suggesting that participants with the T/T genotype had a lower mass than that of individuals with other genotypes. In contrast, rs4842924 showed a significant positive correlation with BFM (β = 0.15, P = 0.0095), suggesting that mass was higher in participants with the T/T genotype than in those with other genotypes. With respect to medications, GLP‐1 receptor agonist (GLP‐1ra) treatment showed a positive correlation with BFM (β = 0.14, P = 0.016).
To further evaluate the differences in genetic factors contributing to body composition, the participants were stratified by age (elderly: ≥65 years, non‐elderly: <65 years) for multiple regression analysis. We observed that IRS1 rs2943656 significantly correlated with SMM (β = 0.23, P = 0.008) and BFM (β = −0.26, P = 0.02) in non‐elderly individuals (Table S3), but not in elderly individuals (Table S4); ADAMTSL3 rs4842924 significantly correlated with ALM (β = −0.23, P ≤ 0.01) and SMM (β = −0.20, P = 0.02) in non‐elderly individuals, but not in elderly individuals; and VCAN rs2287926 significantly correlated with BFM (β = 0.30, P = 0.007) in non‐elderly individuals, but not in elderly individuals.
Difference in SMM between IRS1 rs2943656 genotypes
To validate the observed associations between SNPs and body composition in type 2 diabetes patients, SMM, ALM and BFM were evaluated in a stratified analysis (Fig. 2). For IRS1 rs2943656 (A), the SMM was significantly higher in participants with the G/G genotype than in those with other genotypes (A/A + G/A genotypes) in women (8.6 ± 1.0 vs 8.0 ± 1.0 kg/m2, P < 0.05) and men (9.8 ± 1.2 vs. 9.2 ± 0.8 kg/m2, P < 0.05). For ADAMTSL3 rs4842924, BFM (B) in women tended to be lower for the C/C + T/C genotypes than the T/T genotype (8.0 ± 3.9 vs 9.9 ± 4.6, P = 0.08), with no difference in men (6.9 ± 2.7 vs 7.1 ± 2.6, NS). When the BFM was divided by SMM for adjustment owing to the large standard deviation and strong correlation between these parameters, a similar pattern was observed (C; women: 0.94 ± 0.39 vs 1.13 ± 0.45, P = 0.08, men: 0.70 ± 0.26 vs 0.73 ± 0.24, NS). For ADAMTSL3 rs4842924 (D), ALM did not differ significantly between participants with C/C + T/C genotypes and those with the T/T genotype in women (6.3 ± 1.0 vs 6.3 ± 1.0, NS) or men (7.5 ± 1.0 vs 7.3 ± 1.1, NS).
Figure 2.
Comparison of body composition among genotypes for lean mass. (a) Skeletal muscle mass index stratified by IRS1 rs2943656 genotypes (A/A + G/A vs G/G). (b) Body fat mass index stratified by ADAMTSL3 rs4842924 genotypes (C/C + T/C vs T/T). (c) Body fat mass/skeletal muscle mass stratified by ADAMTSL3 rs4842924 genotypes (C/C + T/C vs T/T). (d) Appendicular muscle mass stratified by ADAMTSL3 rs4842924 genotypes (C/C + T/C vs T/T).
Discussion
To adjust for confounding factors in genetic association studies of body composition in type 2 diabetes patients, clinical determinants of lean mass, SMM and BFM were analyzed by a multiple regression analysis in the present study. Among all independent predictors of body composition, the strongest determinant was sex and the second strongest was the body composition itself. Aging‐related losses were consistently observed for TLM, ALM and SMM, but aging‐related gain was observed for BFM.
With respect to medications for diabetes, only the GLP‐1ra was positively correlated with BFM, suggesting that patients treated with GLP‐1ra have a greater BFM. GLP‐1ra induces a reduction in appetite and deceleration of gastric emptying, leading to a reduction in bodyweight. 11 The positive correlation between GLP‐1ra use and BFM, therefore, could be explained by the preferential administration of GLP‐1ra to obese patients for the purpose of weight reduction. This speculation was supported by the observation of a higher BMI in participants treated with the GLP‐1ra compared with those not having received this treatment (26.9 vs 24.5, P = 0.01, Student's t‐test).
After the initial screening for clinical determinants of body composition, the contribution of genetic loci to body compositions was assessed by a multiple regression analysis. IRS1 rs2943656 was significantly correlated with SMM, but not with TLM, ALM or BFM. The allele for increased SMM (i.e. the G allele) was consistent with previous results for healthy individuals. 12 A similar tendency was observed in the comparison of SMM between IRS1 rs2943656 genotypes. TLM consists of SMM and internal organs of the whole body, and ALM mainly consists of SMM of the arms and legs. A recent report showed that the SMM measured by BIA (InBody770) is closely correlated to the skeletal muscle index measured by computed tomography at the level of the third lumbar vertebra. 13 Taken together, these observations suggest that there is a strong correlation between the IRS1 polymorphism and SMM in the body truncus, rather than that in the arms and legs. Insulin receptor substrate 1 (IRS1) is highly expressed in skeletal muscle and adipocytes, and is a key molecule in the insulin/insulin‐like growth factor‐1 signaling pathway, which plays a critical role in the induction of muscle hypertrophy, as well as in the blockade of muscle atrophy. 14 , 15 The rs2943656 SNP is located within a regulatory region (promoter) of IRS1. A search for transcription factor binding sites (http://tfbind.hgc.jp/) predicted that the G allele of rs2943656 disrupts the binding motif for interferon regulatory factor 1 and MYB (MYB proto‐oncogene, transcription factor), and creates an alternative binding motif for v‐MYB (myeloblastosis viral oncogene homolog [Avian]). An expression quantitative trait loci analysis showed that rs2943656 affects the genotype‐dependent differences in the expression of IRS1, with the lowest expression for the G/G genotype in subcutaneous and omental adipose tissues (GTEx project, https://www.gtexportal.org/home/), suggesting that the polymorphism influences gene expression in skeletal muscle. In healthy individuals, a previous study reported that the lean mass‐increasing allele of IRS1 rs2943656 (i.e. the G allele) is associated with a reduced BMI and fat mass. 12 The present data obtained from individuals with type 2 diabetes showed a similar tendency, with a significant correlation observed between IRS1 rs2943656 and BFM in non‐elderly individuals (the G allele served as a fat mass‐reducing allele; Table S3). The opposite effect of the G allele on muscle and fat mass suggests that IRS1 shows tissue‐specific regulation. Despite the high expression of IRS1 in adipose tissue, we did not detect any evidence for a correlation between BFM and IRS1 rs2943656 in the multiple regression analysis. The degradation of the IRS1 protein by post‐transcriptional modification has been observed in the adipose tissue in a mouse model of type 2 diabetes, but not in control mice, resulting in severe insulin resistance in type 2 diabetes. 16 In clear contrast to adipose tissue, the phosphorylation of IRS1 in skeletal muscle was not uniformly altered in patients with type 2 diabetes, despite the severe insulin resistance in skeletal muscle, suggesting that the post‐transcriptional modification of IRS1 differs between adipose tissue and skeletal muscle in type 2 diabetes. 17
ADAMTSL3, a member of the ADAMTS (a disintegrin and metalloprotease with thrombospondin motifs)‐like subfamily, is a component of the extracellular matrix and is ubiquitously expressed, including in adipose tissue and skeletal muscle. 18 Genetic association studies have consistently shown that ADAMTSL3 polymorphisms are associated with adult height 19 , 20 and lean mass 9 , 12 in humans. An animal study also reported that SNPs in the ADAMTSL3 are significantly correlated with body measurement traits, especially body size. 21 The multiple regression analysis in the present study showed that ADAMTSL3 rs4842924 was significantly correlated with BFM and ALM. However, the relationship was in the opposite direction; the T allele increased fat mass and the C allele increased lean mass, consistent with previous results for healthy individuals. 12 We further found that BFM in individuals with the T/T genotype tends to be higher in women than in men (P = 0.08, Fig. 1b,c). Further studies with larger sample sizes are required to confirm this finding. Although ADAMTSL3 rs4842924 is an intronic polymorphism, the expression quantitative trait loci data showed that the locus affects the expression levels of ADAMTSL3, with the lowest expression for the T/T genotype in skeletal muscle and adipose tissue, suggesting a direct link between this polymorphism and the regulation of gene expression.
Stratified analysis by age group showed that the contributions of IRS1 and ADAMTSL3 to body composition were concentrated in non‐elderly individuals (Table S3), but not in elderly individuals (Table S4), suggesting that genetic factors predominantly correlate with body composition in younger individuals rather than elderly individuals with type 2 diabetes. As aging is one of the most significant factors affecting body composition, the contribution of genetic factors might be masked by aging‐related factors in elderly individuals. These data in turn suggest that pharmacological and non‐pharmacological interventions, such as resistance exercise and nutritional support, might improve body composition to help prevent sarcopenia and frailty in elderly individuals with type 2 diabetes.
The present study had some limitations. The BIA using InBody770 tended to overestimate the total and appendicular SMM in comparison with results obtained by dual‐energy X‐ray absorptiometry. 22 The genetic loci for lean mass, however, were identified by the assessment of body composition by either dual‐energy X‐ray absorptiometry or BIA. 9 , 12 Longitudinal studies might show alternative clinical variables, such as glycemic control and drugs for diabetes treatment, involved in changes in body composition in type 2 diabetes. 23 Further functional analyses of IRS1 and ADAMTSL3 are required to show a direct link between these genetic loci and phenotypes. The present study did not show the association of VCAN and FTO loci with body composition in type 2 diabetes patients, which might show the similarities (IRS1 and ADAMTSL3) and differences (VCAN and FTO) in genetic factors that contribute to lean mass between healthy individuals and individuals with type 2 diabetes; however, the further analysis might need to exclude the roles of VCAN and FTO loci for body compositions. In particular, our stratified age group analysis showed a significant correlation between VCAN rs2287926 and BFM in non‐elderly individuals with type 2 diabetes. As SNPs in FTO have been reported to be associated with early‐onset extreme obesity, the contribution of FTO loci to body composition might be attenuated in non‐obese participants with type 2 diabetes. 24
In summary, the present study clarified the contribution of genetic factors, IRS1 and ADAMTSL3, to interindividual variation in body composition in type 2 diabetes patients, independent of key clinical variables, such as sex, age and body composition itself. These data improve our understanding of the etiology and determinants of sarcopenia and obesity in type 2 diabetes patients, facilitating the establishment of effective methods for the prediction, prevention, and intervention of sarcopenia and frailty in diabetes patients.
Disclosure statement
The authors declare no conflict of interest.
Supporting information
Appendix S1 Supplementary methods
Appendix S2 Supplementary results
Table S1 Multiple regression analysis of indices for body compositions in type 2 diabetes patients including medications as independent variables.
Table S2 Genotype and allele frequencies of single‐nucleotide polymorphisms in participants with type 2 diabetes.
Table S3 Multiple regression analysis of body composition indices in non‐elderly individuals with type 2 diabetes.
Table S4 Multiple regression analysis of body composition indices in elderly individuals with type 2 diabetes.
Acknowledgements
We thank Ms Yayoi Kibayashi, Ms Shie Hayase and Ms Mariko Shiota for their skillful technical assistance. This study was supported in part by Grants‐in‐Aid for Scientific Research (C) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (17K09852, SN; 18K08493, NB; 18K08530, HI). We thank Editage (www.editage.com) for English language editing.
Minohara T, Noso S, Babaya N, et al. Associations between genetic loci related to lean mass and body composition in type 2 diabetes. Geriatr. Gerontol. Int. 2021;21:932–938. 10.1111/ggi.14259
[Correction added on 19 August 2021, after first online publication: The word “patients” has been removed from the title.]
References
- 1. Deshmukh AS. Insulin‐stimulated glucose uptake in healthy and insulin‐resistant skeletal muscle. Horm Mol Biol Clin Investig 2016; 26: 13–24. [DOI] [PubMed] [Google Scholar]
- 2. Wagenknecht LE, Langefeld CD, Scherzinger AL et al. Insulin sensitivity, insulin secretion, and abdominal fat: the insulin resistance atherosclerosis study (IRAS) family study. Diabetes 2003; 52: 2490–2496. [DOI] [PubMed] [Google Scholar]
- 3. Müller MJ, Lagerpusch M, Enderle J, Schautz B, Heller M, Bosy‐Westphal A. Beyond the body mass index: tracking body composition in the pathogenesis of obesity and the metabolic syndrome. Obes Rev 2012; 13 (suppl 2): 6–13. [DOI] [PubMed] [Google Scholar]
- 4. Al‐Sofiani ME, Ganji SS, Kalyani RR. Body composition changes in diabetes and aging. J Diabetes Complications 2019; 33: 451–459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Kim TN, Choi KM. Sarcopenia: definition, epidemiology, and pathophysiology. J Bone Metab 2013; 20: 1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Sugimoto K, Tabara Y, Ikegami H et al. Hyperglycemia in non‐obese patients with type 2 diabetes is associated with low muscle mass: the multicenter study for clarifying evidence for sarcopenia in patients with diabetes mellitus. J Diabetes Investig 2019; 10: 1471–1479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Arden NK, Spector TD. Genetic influences on muscle strength, lean body mass, and bone mineral density: a twin study. J Bone Miner Res 1997; 12: 2076–2081. [DOI] [PubMed] [Google Scholar]
- 8. Hsu FC, Lenchik L, Nicklas BJ et al. Heritability of body composition measured by DXA in the diabetes heart study. Obes Res 2005; 13: 312–319. [DOI] [PubMed] [Google Scholar]
- 9. Zillikens MC, Demissie S, Hsu YH et al. Large meta‐analysis of genome‐wide association studies identifies five loci for lean body mass. Nat Commun 2017; 8: 80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Janssen I, Heymsfield SB, Baumgartner RN, Ross R. Estimation of skeletal muscle mass by bioelectrical impedance analysis. J Appl Physiol 2000; 89: 465–471. [DOI] [PubMed] [Google Scholar]
- 11. Nauck MA, Vilsbøll T, Gallwitz B, Garber A, Madsbad S. Incretin‐based therapies: viewpoints on the way to consensus. Diabetes Care 2009; 32 (suppl 2): S223–S231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Karasik D, Zillikens MC, Hsu YH et al. Disentangling the genetics of lean mass. Am J Clin Nutr 2019; 109: 276–287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Kim EY, Kim SR, Won DD, Choi MH, Lee IK. Multifrequency bioelectrical impedance analysis compared with computed tomography for assessment of skeletal muscle mass in primary colorectal malignancy: a predictor of short‐term outcome after surgery. Nutr Clin Pract 2020; 35: 664–674. [DOI] [PubMed] [Google Scholar]
- 14. Kilpeläinen TO, Zillikens MC, Stančákova A et al. Genetic variation near IRS1 associates with reduced adiposity and an impaired metabolic profile. Nat Genet 2011; 43: 753–760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Glass DJ. Skeletal muscle hypertrophy and atrophy signaling pathways. Int J Biochem Cell Biol 2005; 37: 1974–1984. [DOI] [PubMed] [Google Scholar]
- 16. Wang Y, Nishina PM, Naggert JK. Degradation of IRS1 leads to impaired glucose uptake in adipose tissue of the type 2 diabetes mouse model TALLYHO/Jng. J Endocrinol 2009; 203: 65–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Karlsson HKR, Kasahara A, Ikeda M et al. Quantitative phosphoproteomic analysis of IRS1 in skeletal muscle from men with normal glucose tolerance or type 2 diabetes: a case‐control study. Metabolism 2021; 118: 154726. [DOI] [PubMed] [Google Scholar]
- 18. Hall NG, Klenotic P, Anand‐Apte B, Apte SS. ADAMTSL‐3/punctin‐2, a novel glycoprotein in extracellular matrix related to the ADAMTS family of metalloproteases. Matrix Biol 2003; 22: 501–510. [DOI] [PubMed] [Google Scholar]
- 19. Lango Allen H, Estrada K, Lettre G et al. Hundreds of variants clustered in genomic loci and biological pathways affect human height. Nature 2010; 467: 832–838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. N'Diaye A, Chen GK, Palmer CD et al. Identification, replication, and fine‐mapping of loci associated with adult height in individuals of african ancestry. PLoS Genet 2011; 7: e1002298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Liu Y, Zan L, Zhao S, Xin Y, Jiao Y, Li K. Molecular characterization, expression pattern, polymorphism and association analysis of bovine ADAMTSL3 gene. Mol Biol Rep 2012; 39: 1551–1560. [DOI] [PubMed] [Google Scholar]
- 22. Lee SY, Ahn S, Kim YJ et al. Comparison between dual‐energy X‐ray absorptiometry and bioelectrical impedance analyses for accuracy in measuring whole body muscle mass and appendicular skeletal muscle mass. Nutrients 2018; 10: 738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Sugimoto K, Ikegami H, Takata Y et al. Glycemic control and insulin improve muscle mass and gait speed in type 2 diabetes: the MUSCLES‐DM study. J Am Med Dir Assoc 2021; 22: 834–838. [DOI] [PubMed] [Google Scholar]
- 24. Hinney A, Nguyen TT, Scherag A et al. Genome wide association (GWA) study for early onset extreme obesity supports the role of fat mass and obesity associated gene (FTO) variants. PLoS One 2007; 2: e1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1 Supplementary methods
Appendix S2 Supplementary results
Table S1 Multiple regression analysis of indices for body compositions in type 2 diabetes patients including medications as independent variables.
Table S2 Genotype and allele frequencies of single‐nucleotide polymorphisms in participants with type 2 diabetes.
Table S3 Multiple regression analysis of body composition indices in non‐elderly individuals with type 2 diabetes.
Table S4 Multiple regression analysis of body composition indices in elderly individuals with type 2 diabetes.