Central fat accumulation is a strong risk factor for type 2 diabetes. Genome-wide association studies have identified numerous loci associated with body fat distribution. The objectives of the current study are to examine whether genes in genetic loci linked to fat distribution can be linked to fat cell size and number (morphology) and/or adipose tissue function. We show, in a cohort of 114 women, that almost half of the 96 genes in these loci are indeed associated with abdominal subcutaneous adipose tissue parameters. Thus, adipose mRNA expression of the genes is strongly related to adipose morphology, catecholamine-induced lipid mobilization (lipolysis), or insulin-stimulated lipid synthesis in adipocytes (lipogenesis). In conclusion, the genetic influence on body fat distribution could be mediated via several specific alterations in adipose tissue morphology and function, which in turn may influence the development of type 2 diabetes.

Adipose tissue is distributed in many different body regions. Central body fat accumulation is governed by genetic factors (1) and is, independently of total adiposity, associated with metabolic conditions such as type 2 diabetes and cardiovascular disease (2). A recent genome-wide association (GWA) meta-analysis identified 68 genetic loci that associate with body fat distribution, primarily waist-to-hip ratio (WHR) but also related traits such as waist circumference (3). Bioinformatics analysis of the genes in the vicinity of the 68 genetic loci implicated adipose pathways in the regulation of fat distribution, and several of the candidate genes are indeed expressed at the mRNA level in adipose tissue. However, direct evidence for these loci and/or genes being involved in adipocyte function is lacking in most cases.

Fat mass is determined by adipose tissue morphology (i.e., size and number of adipocytes) (4), and both parameters contribute to the total weight of specific human fat depots, thereby regulating fat distribution (5). In addition, alterations in adipocyte function can also influence adipose distribution. The lipid droplet constitutes >95% of the fat cell volume. As reviewed (6), impaired ability of the major lipolysis-regulating hormones, catecholamines, to stimulate the hydrolysis of adipocyte lipids (lipolysis) may cause retention of lipids in fat cells and thereby contribute to fat mass expansion. Insulin is a major anabolic hormone and controls lipid accumulation in fat cells. The ability of insulin to stimulate fat synthesis through glucose conversion into lipids (lipogenesis) could therefore also be of importance. Here, we investigated whether these adipocyte phenotypes (adipose morphology, lipolysis, and lipogenesis) could be linked to abdominal subcutaneous white adipose tissue (WAT) expression of 96 candidate genes in the vicinity of the recently described genetic loci for body fat distribution (3). This centrally located adipose region contributes to variations in WHR and in other parameters of body fat distribution.

The study included 114 Swedish women without diabetes who were recruited from the general adult population in the Stockholm (Sweden) area (Table 1). They displayed a large interindividual variation in BMI and WHR. The study was approved by the regional ethics board, and written informed consent was obtained from each subject. They were investigated in the morning after an overnight fast, when abdominal subcutaneous WAT was obtained by fine-needle aspiration (7). WAT gene expression was assessed by Gene 1.0 or 1.1 ST Affymetrix arrays, as described (8). Between-array results were normalized in Affymetrix Power Tools.

Table 1

Characteristics of 114 women

MeanRange
Age, years 43 22–72 
BMI, kg/m2 34 20–53 
WHR 0.93 0.74–1.09 
WAT morphology, pL 19 −400 to +494 
Isoprenaline activation of lipolysis, ratio 8.7 1.4–66.5 
Insulin stimulated lipogenesis, nmol of glucose/2 h/107 adipocytes 4.71 0.12–46.19 
MeanRange
Age, years 43 22–72 
BMI, kg/m2 34 20–53 
WHR 0.93 0.74–1.09 
WAT morphology, pL 19 −400 to +494 
Isoprenaline activation of lipolysis, ratio 8.7 1.4–66.5 
Insulin stimulated lipogenesis, nmol of glucose/2 h/107 adipocytes 4.71 0.12–46.19 

Fat cells were isolated from adipose tissue and incubated in vitro, exactly as described (8), to determine lipolysis (glycerol release) and lipogenesis (incorporation of radiolabeled glucose into fat cell lipids). The cells were incubated in the absence or presence of increasing concentrations of isoprenaline (a synthetic catecholamine) or insulin, and the results at maximum effective concentrations were used. The isoprenaline effect was measured as the ratio of glycerol release without the drug. On a separate aliquot of isolated fat cells, their mean volume was measured and set in relation to the total body fat content, exactly as described in detail (10), to obtain an adipose tissue morphology value. The measured volume was compared with the calculated expected volume given the subject’s total fat mass. The difference between measured and calculated fat cell volume is the morphology value.

View Large

Adipocytes isolated from abdominal subcutaneous WAT were used to measure activation of lipolysis by isoprenaline (a synthetic catecholamine) and insulin-stimulated lipogenesis, as described (9), and data distribution is summarized in Table 1. By comparing the size of the adipocytes with the total amount of body fat, the morphology of WAT could be quantitatively determined (Δ values), as described in detail (10). Negative Δ values indicate hyperplasia (many small adipocytes), and positive Δ values indicate hypertrophy (few but large adipocytes); the distribution of values is reported in Table 1. Adipose hypertrophy is linked to insulin resistance and the development of type 2 diabetes, as discussed (10).

Linear regression analysis showed age did not influence the examined parameters. As expected, an inverse relationship was found between BMI and insulin-stimulated lipogenesis (r = 0.27) or catecholamine-stimulated lipolysis (r = 0.28, P ≤ 0.004 for both) but, in accordance with previous studies (10), not with adipose morphology (r = 0.18, P = 0.13). Therefore, further analyses on lipolysis and lipogenesis were adjusted for BMI.

First, we analyzed all 22,168 probe sets on the arrays for association with WHR and the adipose phenotypes in R software using the SAM Bioconductor package (11). Among significant genes, we focused on GWA candidates for body fat distribution; that is, all genes that were localized within 250 kb of the lead single nucleotide polymorphism in the loci associated with body fat distribution listed by Shungin et al. (3) in their Tables 1, 2, and 3. This included all primary candidate genes in 49 genetic loci associated with WHR adjusted for BMI (WHRadjBMI) in the aforementioned GWA study (3). We also included additional candidate genes that were obtained by expression quantitative trait loci (eQTL) analysis or given high priority according to bioinformatics analysis and all primary candidate genes in 19 loci associated with waist and hip circumference measures. Three genes were not represented with probes on the Affymetrix microarrays. The remaining 96 genes are listed in Supplementary Table 1. The subset of genes whose expression was associated with WHR and with adipose phenotypes according to SAM analysis is shown in Supplementary Table 2.

We next extended the analysis to take into account possible confounders. To this end, multiple regression analysis was used to investigate the associations between the expression of the 96 genes and WHR or the adipocyte phenotypes, taking into account the use of different array batches and, when appropriate, BMI. Because 96 genes were investigated, we considered a P value of < 0.0005 as statistically significant. There was a strong overlap between the results obtained by SAM and those obtained by multiple regression, although somewhat more genes were significant with SAM. We therefore focused on results obtained by regression analysis, which takes potential confounders into account.

As observed in Table 2, only two genes—CALCRL and FAM13A—were correlated with WHR and WHRadjBMI, which is in accordance with previous WAT expression data on a few of the WHR-associated genes (12). CALCRL and FAM13A were negatively associated with adipose morphology, whereas CPEB4, HLA-DRA, MSC, and PLCG2 correlated positively with this phenotype. Thus, WHR-associated genes may influence adipocyte size and number in abdominal subcutaneous WAT in opposite ways, resulting in different body fat distribution. Eight candidate genes for body fat distribution in Table 2 correlated with catecholamine-stimulated lipolysis, where all but one (CBX3) showed a positive association. None of these genes associated with adipose morphology, indicating that some genes linked to body fat distribution may influence central fat accumulation by specifically enhancing (in most cases) or suppressing the lipolytic activity of adipocytes.

Table 2

Relationship between WAT gene expression and adipose morphology, isoprenaline-stimulated activation of lipolysis in adipocytes, or WHR

Gene IDWAT morphology
Lipolysis
WHR
Standard coefficientP valueStandard coefficientP valueStandard coefficientP value
CALCRL§□ −0.44 <0.0001 — — −0.39 <0.0001 
CBX3 — — −0.77 0.0001 — — 
CPEB4§ 0.59 <0.0001 — — — — 
EYA2 — — 0.43 <0.0001 — — 
FAM13A§□ −0.46 <0.0001 — — −0.49 <0.0001 
FGF2# — — — — −0.65 <0.0001 
HLA-DRA§ 0.36 0.0003 — — 0.41 <0.0001 
HMGXB4 — — 0.56 0.0002 — — 
MSC 0.57 <0.0001 — — — — 
PLCG2 0.60 0.0002 — — — — 
PPARG# — — 0.76 <0.0001 — — 
PTPDC1 — — 0.53 <0.0001 — — 
SFXN2 — — 0.61 <0.0001 — — 
TBX15# — — 0.56 <0.0001 — — 
VPS53 — — 0.69 0.0005 — — 
Gene IDWAT morphology
Lipolysis
WHR
Standard coefficientP valueStandard coefficientP valueStandard coefficientP value
CALCRL§□ −0.44 <0.0001 — — −0.39 <0.0001 
CBX3 — — −0.77 0.0001 — — 
CPEB4§ 0.59 <0.0001 — — — — 
EYA2 — — 0.43 <0.0001 — — 
FAM13A§□ −0.46 <0.0001 — — −0.49 <0.0001 
FGF2# — — — — −0.65 <0.0001 
HLA-DRA§ 0.36 0.0003 — — 0.41 <0.0001 
HMGXB4 — — 0.56 0.0002 — — 
MSC 0.57 <0.0001 — — — — 
PLCG2 0.60 0.0002 — — — — 
PPARG# — — 0.76 <0.0001 — — 
PTPDC1 — — 0.53 <0.0001 — — 
SFXN2 — — 0.61 <0.0001 — — 
TBX15# — — 0.56 <0.0001 — — 
VPS53 — — 0.69 0.0005 — — 

Multiple regression analysis was used with array batch and BMI (the latter only for lipolysis and WHR) as cofactors. For adipose morphology, negative standard coefficient values indicate a stimulatory effect of gene expression on hyperplasia and positive values indicate a stimulatory effect on hypertrophy. Lipolysis values were 10-log transformed. P ≤ 0.0005 was considered to be statistically significant because 96 genes on the arrays were investigated in total.

§Only expression data in adipose tissue reported in the literature.

#Functional data in adipocytes/adipose tissue reported in the literature.

□Also significantly correlated with WHRadjBMI, which is an adjustment described before (3).

View Large

As demonstrated in Table 3, a surprisingly large number of genes correlated with insulin-stimulated lipogenesis, 14 genes in a positive fashion and 21 in an inverse fashion. Apparently, regulation of insulin-induced lipid synthesis (stimulation or inhibition) in adipocytes is a very important pathway in WAT that may link candidate genes to central fat distribution. Similar to lipolysis and morphology, the lipogenesis-associated genes seem specific for this adipocyte process because only 4 (CBX3, HLA-DRA, PLCG2, and PTPDC1) overlapped with the 15 genes associated with the other examined adipose phenotypes. These genes, linked to insulin function, may have a particular importance for the type 2 diabetes risk associated with central fat accumulation because insulin resistance is a key feature in this diabetic condition.

Table 3

Relationship between WAT gene expression and insulin-stimulated lipogenesis in abdominal subcutaneous adipocytes

Gene IDStandard coefficientP valueGene IDStandard coefficientP value
ARL15§ −1.11 <0.0001 NUDT6 −0.71 0.0002 
BCL2# 0.58 0.0005 NKX2–6 0.92 <0.0001 
CBX3 0.98 <0.0001 PDXDC1§ −0.49 0.0004 
CEP250 0.67 <0.0001 PIGC§ 0.75 0.0001 
CMIP§ −0.86 <0.0001 PLCG2 −0.72 <0.0001 
DNM3§ −0.62 0.0002 PLXND1# −1.12 <0.0001 
FGF2# 0.71 <0.0001 PTPDC1 −0.51 0.0001 
FGFR4# 0.96 <0.0001 RFX7 −0.99 <0.0001 
GMDS −0.80 <0.0001 SFXN2 −0.57 0.0003 
GORAB −0.61 <0.0001 SPATA5 −0.71 <0.0001 
GRB14# −0.39 <0.0001 TBX15# −0.55 0.0002 
HLA-DRA§ 0.38 0.0003 TFPI§ −0.82 <0.0001 
HOXA11 0.62 <0.0001 UQCC 0.38 0.0001 
ITGB6 0.77 <0.0001 VPS53 −0.89 <0.0001 
KIAA1731 −0.73 <0.0001 WARS2§ −0.60 <0.0001 
KLF13 −0.56 0.0003 ZBTB7B 0.63 <0.0001 
KLF14# 0.61 <0.0001 ZNF664§ 0.54 0.0001 
MAP3K1# −0.51 <0.0001 — — — 
Gene IDStandard coefficientP valueGene IDStandard coefficientP value
ARL15§ −1.11 <0.0001 NUDT6 −0.71 0.0002 
BCL2# 0.58 0.0005 NKX2–6 0.92 <0.0001 
CBX3 0.98 <0.0001 PDXDC1§ −0.49 0.0004 
CEP250 0.67 <0.0001 PIGC§ 0.75 0.0001 
CMIP§ −0.86 <0.0001 PLCG2 −0.72 <0.0001 
DNM3§ −0.62 0.0002 PLXND1# −1.12 <0.0001 
FGF2# 0.71 <0.0001 PTPDC1 −0.51 0.0001 
FGFR4# 0.96 <0.0001 RFX7 −0.99 <0.0001 
GMDS −0.80 <0.0001 SFXN2 −0.57 0.0003 
GORAB −0.61 <0.0001 SPATA5 −0.71 <0.0001 
GRB14# −0.39 <0.0001 TBX15# −0.55 0.0002 
HLA-DRA§ 0.38 0.0003 TFPI§ −0.82 <0.0001 
HOXA11 0.62 <0.0001 UQCC 0.38 0.0001 
ITGB6 0.77 <0.0001 VPS53 −0.89 <0.0001 
KIAA1731 −0.73 <0.0001 WARS2§ −0.60 <0.0001 
KLF13 −0.56 0.0003 ZBTB7B 0.63 <0.0001 
KLF14# 0.61 <0.0001 ZNF664§ 0.54 0.0001 
MAP3K1# −0.51 <0.0001 — — — 

Multiple regression was used with array batch and BMI as cofactors. Lipogenesis values were 10-log transformed. See legend to Table 2 for further details.

§Only expression data in adipose tissue are reported in the literature.

#Functional data in adipocytes/adipose tissue are reported in the literature.

View Large

We used three batches of arrays to determine gene expression; however, results were similar within each of the batches. Combining the Gene 1.0 and 1.1 ST Affymetrix array results had a negligible effect on the findings. Data in Tables 2 and 3 were validated by real-time quantitative PCR (Supplementary Table 3) using RNA from subcutaneous WAT obtained from a previously examined cohort of 55 women (9). We chose five genes displaying among the strongest correlations in the aforementioned tables and examined the relationship between WAT morphology, WHR, or insulin-stimulated lipogenesis and gene expression. Seven of the eight examined relationships showed a statistically significant relationship in agreement with the data presented in Tables 2 and 3; the eighth relationship was directionally consistent in the validation cohort.

To further explore the function of genetic variants associated with body fat distribution, we performed cis-eQTL analysis of the genes in Tables 2 and 3. We related WAT expression of 32 genes to the genotype of 26 single nucleotide polymorphisms representing different genetic loci. We confirmed eQTLs for ZNF664 (rs4765219) and CMIP (rs2925979) and identified a new eQTL for PDXDC1 (rs4985155) (Supplementary Table 4) (3), three genes whose expression is associated with insulin-stimulated lipogenesis. The limited number of significant associations might be due to limited statistical power because only data from ∼100 subjects were available for eQTL analysis.

We finally investigated the expression pattern of the genes in Tables 2 and 3 during differentiation of human subcutaneous progenitor cells to adipocytes (Table 4) using previously published gene expression array data (9). There was no clear pattern for genes associated with lipolysis or lipogenesis (values not shown). However, five of the six morphology-linked genes in Table 1 showed a significant up- or downregulation during differentiation (CPEB4, PLCG2, HLA-DRA, FAM13A, and MSC). This indicates that some genes associated with body fat distribution may influence WAT morphology via effects on adipose differentiation.

Table 4

Expression of genes associated with adipose morphology during adipocyte differentiation in 12 subjects

GeneProbe setDay 4Day 8Day 12Day 12 vs. 4P (day 4 vs. 12)
CPEB4 8110055 162 ± +22 229 ± +73 273 ± +97 1.69 0.0020 
PLCG2 7997453 45 ± +11 32 ± +10 36 ± +13 0.80 0.0160 
HLA-DRA 8118548 351 ± +210 490 ± +367 601 ± +479 1.71 0.0210 
FAM13A 8101728 185 ± +56 244 ± +73 264 ± +128 1.43 0.0370 
MSC 8151334 494 ± +208 296 ± +107 321 ± +153 0.65 0.0440 
CALCRL 8057578 60 ± +49 90 ± +45 86 ± +41 1.43 0.1500 
GeneProbe setDay 4Day 8Day 12Day 12 vs. 4P (day 4 vs. 12)
CPEB4 8110055 162 ± +22 229 ± +73 273 ± +97 1.69 0.0020 
PLCG2 7997453 45 ± +11 32 ± +10 36 ± +13 0.80 0.0160 
HLA-DRA 8118548 351 ± +210 490 ± +367 601 ± +479 1.71 0.0210 
FAM13A 8101728 185 ± +56 244 ± +73 264 ± +128 1.43 0.0370 
MSC 8151334 494 ± +208 296 ± +107 321 ± +153 0.65 0.0440 
CALCRL 8057578 60 ± +49 90 ± +45 86 ± +41 1.43 0.1500 

Values are mean + SD. Expression on day 4 vs. 12 was compared by paired t test.

View Large

When data are taken together, 45 of 96 of the genes associated with genetic loci for body fat distribution show a relationship between WAT gene expression and different aspects of adipocyte function linked to subcutaneous abdominal fat accumulation. Most of these genes were associated with one specific adipocyte phenotype (morphology, lipolysis, or lipogenesis), which points to a very distinct role for each gene. Thus, our findings suggest a central role for adipocytes in linking genetic variants with central fat mass distribution. Admittedly, only a few genes showed correlation with WHR. One plausible explanation is that fat distribution is the far away clinical end point of multiple pathways partly differing between subjects (i.e., a multifactorial etiology involving heterogeneity between subjects). Specific genes can then show a stronger association with intermediate phenotypes than with the clinical outcome. It should also be clarified that the genetic association between WHR and the genetic loci studied here becomes strong when very large numbers of individuals are included. Because our present cohort included only 114 subjects, it is quite expected that the association with WHR becomes nonsignificant.

What is known about the role of the examined genes in fat cells? For each of the genes presented in Tables 2 and 3, we searched for previous publications related to the topic on PubMed (http://www.ncbi.nlm.nih.gov/pubmed). We used the key words “adipocyte,” “fat cell,” “adipose tissue,” “lipolysis,” and “lipogenesis.” The genes were categorized as follows: no publications relevant for the key words, expression data available, or functional data available. The bioinformatics analysis revealed that the mechanism of action on adipocytes is unknown and/or that only gene expression data are available for most of the described genes (Tables 2 and 3). Therefore, the molecular and causal aspects need to be investigated much further. Earlier GWA studies suggest important sex dimorphism for WHR-associated genetic loci (13,14). Because we only studied women, the possible influence of sex remains to be established. Finally, the functional analyses were performed on adipocytes isolated from the centrally located subcutaneous fat. Thus, the effect of the examined gene loci on visceral, gluteal, or other WAT regions with regard to metabolic function or morphology is not known at present. Unfortunately, it is not possible for ethical reasons to study deep fat depots in the present clinical setting. Nevertheless, our findings support the hypothesis that the genetic influence on body fat distribution in humans is governed to a large extent governed by the morphology and metabolic function of adipocytes. Whether this also involves the endocrine role of adipocytes remains to be established.

Acknowledgments. The authors greatly appreciate the skillful technical support from research nurses Katarina Hertel and Yvonne Widlund and laboratory technicians Eva Sjölin, Kerstin Wåhlen, and Elisabeth Dungner (Department of Medicine, Huddinge, Karolinska Institutet, Stockholm, Sweden).

Funding. This study was supported by the Strategic Research Programme (SRP) in Diabetes at Karolinska Institutet; Center of Innovative Medicine; the Swedish Research Council for Health, Working Life and Welfare; Stockholm Läns Landsting; Novo Nordisk Foundation; European Association for the Study of Diabetes/Eli Lilly Foundation; the Swedish Diabetes Foundation; and the European Union/European Federation of Pharmaceutical Industries and Associations Innovative Medicines Initiative Joint Undertaking (European Medical Information Framework grant No. 115372). R.J.S. is supported by the SRP in Diabetes at Karolinska Institutet.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. I.D. and P.A. designed the study. I.D., R.J.S., and P.A. conducted the data analysis. I.D. and H.G. were responsible for genotyping. M.R. conducted the bioinformatics analysis. D.B. performed initial normalization analyses of microarrays. I.D., M.R., and P.A. contributed to study subject recruitment. P.A. wrote first version of the manuscript. All authors contributed to subsequent writing and approved the final version of the manuscript. I.D. and P.A. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Schleinitz
 
D
,
Böttcher
 
Y
,
Blüher
 
M
,
Kovacs
 
P
.
The genetics of fat distribution
.
Diabetologia
 
2014
;
57
:
1276
–
1286
[PubMed]
Google Scholar
Crossref
Search ADS
 
2.
Pischon
 
T
,
Boeing
 
H
,
Hoffmann
 
K
, et al.
General and abdominal adiposity and risk of death in Europe
.
N Engl J Med
 
2008
;
359
:
2105
–
2120
[PubMed]
Google Scholar
Crossref
Search ADS
 
3.
Shungin
 
D
,
Winkler
 
TW
,
Croteau-Chonka
 
DC
, et al.;
ADIPOGen Consortium
;
CARDIOGRAMplusC4D Consortium
;
CKDGen Consortium
;
GEFOS Consortium
;
GENIE Consortium
;
GLGC
;
ICBP
;
International Endogene Consortium
;
LifeLines Cohort Study
;
MAGIC Investigators
;
MuTHER Consortium
;
PAGE Consortium
;
ReproGen Consortium
.
New genetic loci link adipose and insulin biology to body fat distribution
.
Nature
 
2015
;
518
:
187
–
196
[PubMed]
Google Scholar
Crossref
Search ADS
 
4.
Prins
 
JB
,
O’Rahilly
 
S
.
Regulation of adipose cell number in man
.
Clin Sci (Lond)
 
1997
;
92
:
3
–
11
[PubMed]
Google Scholar
Crossref
Search ADS
 
5.
Arner
 
P
,
Andersson
 
DP
,
Thörne
 
A
, et al.
Variations in the size of the major omentum are primarily determined by fat cell number
.
J Clin Endocrinol Metab
 
2013
;
98
:
E897
–
E901
[PubMed]
Google Scholar
Crossref
Search ADS
 
6.
Arner
 
P
,
Langin
 
D
.
Lipolysis in lipid turnover, cancer cachexia, and obesity-induced insulin resistance
.
Trends Endocrinol Metab
 
2014
;
25
:
255
–
262
[PubMed]
Google Scholar
Crossref
Search ADS
 
7.
Kolaczynski
 
JW
,
Morales
 
LM
,
Moore
 
JH
 Jr, et al.
A new technique for biopsy of human abdominal fat under local anaesthesia with Lidocaine
.
Int J Obes Relat Metab Disord
 
1994
;
18
:
161
–
166
[PubMed]
Google Scholar
 
8.
Löfgren
 
P
,
Hoffstedt
 
J
,
Näslund
 
E
,
Wirén
 
M
,
Arner
 
P
.
Prospective and controlled studies of the actions of insulin and catecholamine in fat cells of obese women following weight reduction
.
Diabetologia
 
2005
;
48
:
2334
–
2342
[PubMed]
Google Scholar
Crossref
Search ADS
 
9.
Arner
 
E
,
Mejhert
 
N
,
Kulyté
 
A
, et al.
Adipose tissue microRNAs as regulators of CCL2 production in human obesity
.
Diabetes
 
2012
;
61
:
1986
–
1993
[PubMed]
Google Scholar
Crossref
Search ADS
 
10.
Arner
 
E
,
Westermark
 
PO
,
Spalding
 
KL
, et al.
Adipocyte turnover: relevance to human adipose tissue morphology
.
Diabetes
 
2010
;
59
:
105
–
109
[PubMed]
Google Scholar
Crossref
Search ADS
 
11.
Tusher
 
VG
,
Tibshirani
 
R
,
Chu
 
G
.
Significance analysis of microarrays applied to the ionizing radiation response
.
Proc Natl Acad Sci U S A
 
2001
;
98
:
5116
–
5121
[PubMed]
Google Scholar
Crossref
Search ADS
 
12.
Schleinitz D, Klöting N, Lindgren CM, et al. Fat depot-specific mRNA expression of novel loci associated with waist-hip ratio. Int J Obes (Lond) 2014;38:120–125
13.
Heid
 
IM
,
Jackson
 
AU
,
Randall
 
JC
, et al.;
MAGIC
.
Meta-analysis identifies 13 new loci associated with waist-hip ratio and reveals sexual dimorphism in the genetic basis of fat distribution
.
Nat Genet
 
2010
;
42
:
949
–
960
[PubMed]
Google Scholar
Crossref
Search ADS
 
14.
Randall
 
JC
,
Winkler
 
TW
,
Kutalik
 
Z
, et al.;
DIAGRAM Consortium
;
MAGIC Investigators
.
Sex-stratified genome-wide association studies including 270,000 individuals show sexual dimorphism in genetic loci for anthropometric traits
.
PLoS Genet
 
2013
;
9
:
e1003500
[PubMed]
Google Scholar
Crossref
Search ADS
 
© 2016 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered.
2016

Supplementary data

Supplementary Data- pdf file