The cell death–inducing DFFA (DNA fragmentation factor-α)-like effector A (CIDEA) gene is implicated as an important regulator of body weight in mice and humans and is therefore a candidate gene for human obesity. Here, we characterize common CIDEA gene polymorphisms and investigate them for association with obesity in two independent Swedish samples; the first comprised 981 women and the second 582 men. Both samples display a large variation in BMI. The only detected coding polymorphism encodes an exon 4 V115F amino acid substitution, which is associated with BMI in both sexes (P = 0.021 for women, P = 0.023 for men, and P = 0.0015 for joint analysis). These results support a role for CIDEA alleles in human obesity. CIDEA-deficient mice display higher metabolic rate, and the gene cross-talks with tumor necrosis factor-α (TNF-α) in fat cells. We hypothesize that CIDEA alleles regulate human obesity through impact on basal metabolic rate and adipocyte TNF-α signaling.

The cell death–inducing DFFA (DNA fragmentation factor-α)-like effector A (CIDEA) gene has been implicated as an important regulator of body weight in mice and humans and is therefore a candidate gene for human obesity. Several factors support that CIDEA is involved in the regulation of body weight. CIDEA-null mice are resistant to diet-induced obesity and diabetes and display higher metabolic rate and lipolysis in brown adipose tissue than their wild-type littermates (1). In obese women undergoing weight reduction through low-calorie diet, mRNA for CIDEA is the highest upregulated in subcutaneous adipose tissue among 8,000 investigated genes (2). Furthermore, CIDEA mRNA is downregulated in subcutaneous adipose tissue of obese subjects (3).

Obesity is the strongest risk factor for the development of type 2 diabetes, and there is a marked parallel increase in the prevalence of these two disorders in most countries. Although a genetic impact on common obesity is established, underlying susceptibility genes are largely unknown. The CIDEA gene is encoded on human chromosome 18p11.21. This region is established among Caucasians as a susceptibility locus for type 2 diabetes in connection with obesity (4). In addition, Chagnon et al. (5) has reported linkage and association of polymorphisms in the melanocortin receptor 5 (MC5R) gene on chromosome 18p11.2 with BMI. As far as we know, there is no publication of CIDEA gene polymorphisms and the gene promoter has not been characterized.

Based on the above, we consider CIDEA a strong candidate gene for obesity and here characterize common gene polymorphisms and analyze them for association with obesity in two Swedish samples.

By sequencing exons and surrounding regions of the CIDEA gene in 24 obese women and 2.2 kb upstream of exon 1 in 48 subjects, we detected 16 single nucleotide polymorphisms (SNPs) (Fig. 1). The only detected coding SNP, C19787G→T, encodes an exon 4 V115F amino acid substitution. This polymorphism is located in a nonconserved region of the protein (Fig. 1). Since the targets for CIDEA action are unknown and overexpression of CIDEA leads to marked apoptosis (6), it is currently difficult to elucidate the consequence of this amino acid substitution. In addition, we detected seven common and two rare (allele frequency <5% in sequenced women) SNPs in the 5′ region as well as three common and three rare SNPs in introns (Fig. 1).

Initially, we genotyped six SNPs with rare allele frequency >5% in our main case-control material, which was composed of 547 obese and 434 nonobese women recruited in Stockholm, Sweden. All seven common 5′ SNPs were in strong, or perfect, linkage disequilibrium; therefore, no more than two of those were genotyped. Rare SNPs were excluded due to the low power to detect phenotypic association for such SNPs. The genotyping success rate for each SNP was >97% (Table 1). Genotype distributions for all SNPs were in Hardy Weinberg equilibrium (results not shown). In the initial genotyping of 524 women, SNP C81172Gins and rs7230534 as well as C19660G→A and rs8090997 were in strong linkage disequlibrium (D′ 1.0, log likelihood ratio 21, and D′ 0.91, log likelihood ratio 7, respectively) and present on the same haplotypes (Table 2). C81172Gins and C19660G→A genotypes therefore provided little unique information and were not genotyped further. For the six genotyped SNPs, allele frequencies did not differ between obese and nonobese women (Table 1). However, there was a nominal significant difference in BMI between genotypes at C19787G→T, the GG having the highest BMI and the GT having intermediate values (P = 0.021) (Fig. 2).

To extend the analysis, we next genotyped a second independently sampled dataset consisting of 352 obese and 230 nonobese Swedish men for four common CIDEA SNPs, excluding C81172Gins and C19660G→A. Because differences in obesity phenotype between sexes (e.g., men are more prone to develop abdominal obesity) as well as sex-specific hormonal effects may constitute confounders in the analysis, and sex-specific loci for complex disease have been identified, we considered it appropriate to do sex-specific analysis. In this dataset, obesity was associated with the C19787G→T G allele (odds ratio [OR] 1.32 [95% CI 1.03–1.69], P = 0.027 [one sided]) (Table 1). In addition, again there was a nominal significant difference in BMI between C19787G→T genotypes (P = 0.023) (Fig. 2). Analysis of C19787G→T in men and women jointly gave a P value of 0.0015 (Fig. 2).

Using SNPHAP and attached programs, the CIDEA gene SNPs were estimated to define eight haplotypes with frequency >2% (Table 2) that comprised two haploblocks (Fig. 1). In the joint analysis of women and men, no haplotype was associated with obesity (results not shown).

The most important result of this study is the nominal association between BMI and the CIDEA V115F genotype in independent samples of Swedish women and men. Although these results will not remain significant if multiple testing of different loci is taken into account, the relevance of CIDEA gene alleles for obesity is supported by the similar phenotypic associations observed in two independent samples. Because the obese/nonobese phenotype is defined by BMI, it is not necessary to correct for multiple testing regarding these two investigated phenotypes. In addition, previous results with CIDEA in knock-out mice, human obesity, and humans undergoing weight reduction make the CIDEA gene, and in particular the only identified nonsynonomous SNP, a priori a strong candidate for common obesity (13). This should reduce the need for multiple comparison corrections. In addition, when the two materials are combined, the P value will still be significant when multiple comparisons are taken into account. We believe it is valid to combine the female and male cohorts because our results do not support a sex-specific effect on CIDEA. We consider the observed association of CIDEA V115F genotype with BMI, although preliminary, interesting and important to follow up. CIDEA-deficient mice display higher metabolic rate (1). In human fat cells, the gene cross talks with tumor necrosis factor-α (TNF-α), which is an established candidate gene for obesity and influences lipolysis (3). We hypothesize that CIDEA alleles regulate human obesity through impact on basal metabolic rate, TNF-α signaling, and lipid metabolism in adipocytes.

To identify common CIDEA gene polymorphisms, we sequenced the five gene exons according to Genbank NT_010859, exon-intron borders, and 260 base pairs downstream of the 3′ end in 24 obese women of Swedish descent. This sample provides 91% power to detect an SNP with rare allele frequency (<5%) [1 − (0.952)24]. In addition, 2.2 kb upstream of exon 1 (Genbank AP005264) were sequenced in 48 subjects. Newly detected SNPs were labeled according to position in the Genbank sequences above. The Staden software was used for sequence assembly, and all sequences were scored manually for the presence of SNPs.

Two independently sampled datasets of adult subjects were used to investigate CIDEA impact on obesity. The major sample comprised 434 nonobese (aged 40 ± 10 years, BMI 24 ± 3 kg/m2) and 547 obese (aged 41 ± 11 years, BMI 39 ± 5 kg/m2) women. The second sample comprised 230 nonobese (aged 43 ± 15 years, BMI 24 ± 2 kg/m2) and 352 obese (aged 47 ± 11 years, BMI 40 ± 5 kg/m2) men. Obese men had higher waist circumference (129 ± 12 vs. 115 ± 12 cm), serum insulin (26 ± 16 vs. 17 ± 10 mU/l), and plasma triglycerides (2.2 ± 1.6 vs. 1.7 ± 0.9 mmol/l) than obese women. All individuals were recruited in Stockholm, Sweden, and were of Nordic origin for at least two generations. The obese subset included 77 women and 125 men with diagnosed hypertension, 32 women and 55 men with type 2 diabetes, and 3 women and 15 men with dyslipidemia. Otherwise, the subjects were healthy according to self-report and free of medication. Women and men were separately consecutively recruited in order to study candidate genes for obesity in men and women, respectively. Individuals came to the laboratory in the morning after an overnight fast for determination of body height and weight as well as waist circumference using standardized procedures. A venous blood sample was obtained for extraction of DNA. The study was approved by the committee on ethics at Karolinska Institute. It was explained in detail to each subject, and his/her informed consent was obtained.

The DNA samples were genotyped using Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (SEQUENOM, San Diego, CA) as described (7). Linkage disquilibrium between adjacent markers and haplotypes was estimated using SNPHAP and attached programs (available at http://archimedes.well.ox.ac.uk/pise/snphap-simple.html).

Subjects were grouped according to the World Health Organization definition of obesity (obese, BMI >30 kg/m2; and nonobese, BMI ≤30 kg/m2). Frequencies of genotypes, alleles, and haplotypes in obese and lean groups were analyzed by Fisher’s exact test. Analyses in men that were extensions of nominal significant associations obtained in women were considered one-sided hypotheses; we therefore display one-sided P values. The nonparametric Kruskal-Wallis and the parametric ANCOVA with age as a covariate were used to compare BMI between genotypes. Data are presented as means ± SD unless otherwise indicated.

FIG. 1.

CIDEA gene SNPs and homology comparison in the exon 4 region harboring the V115F substitution. Polymorphisms were detected by sequencing exons and surrounding regions in women (n = 24–48). Arrows, haploblock inferred by SNPHAP and attached program. *Allele frequency <5% in sequenced women.

FIG. 1.

CIDEA gene SNPs and homology comparison in the exon 4 region harboring the V115F substitution. Polymorphisms were detected by sequencing exons and surrounding regions in women (n = 24–48). Arrows, haploblock inferred by SNPHAP and attached program. *Allele frequency <5% in sequenced women.

Close modal
FIG. 2.

Association of BMI with C19787G→T genotype. The major sample (middle graph) of age-matched women (n = 981), the second sample (left graph) of age-matched men (n = 582), and the joint samples (right graph) were analyzed for association between BMI and C19787G→T genotype using Kruskal-Wallis test. Very similar results were obtained with ANCOVA with age as covariate. Values are means ± SE. *P = 0.021, **P = 0.023, ***P = 0.0015.

FIG. 2.

Association of BMI with C19787G→T genotype. The major sample (middle graph) of age-matched women (n = 981), the second sample (left graph) of age-matched men (n = 582), and the joint samples (right graph) were analyzed for association between BMI and C19787G→T genotype using Kruskal-Wallis test. Very similar results were obtained with ANCOVA with age as covariate. Values are means ± SE. *P = 0.021, **P = 0.023, ***P = 0.0015.

Close modal
TABLE 1

Association of CIDEA gene SNPs with obesity*

SNPNon obese women (n = 434)Obese women (n = 547)Success rate (%)Allele frequency (%)
Rs7230534 TT 245 (58) CT 156 (37) CC 23 (5) TT 300 (57) CT 190 (36) CC 34 (7) 97 24 
C81172Gins GG 154 (58) Gdel 101 (38) deldel 10 (4) GG 134 (54) Gdel 100 (40) deldel 14 (6) 98 24 
Rs8090997 GG 123 (29) AG 218 (51) AA 88 (20) GG 137 (25) AG 285 (53) AA 117 (22) 99 47 
C19660G→A GG 161 (61) AG 83 (31) AA 22 (8) GG 150 (61) AG 86 (35) AA 11 (4) 98 23 
C19699C→T CC 229 (54) CT 180 (42) TT 16 (4) CC 301 (55) CT 203 (37) TT 41 (8) 99 27 
C19787G→T GG 171 (41) GT 203 (48) TT 45 (11) GG 236 (44) GT 251 (47) TT 46 (9) 97 33 
 Non obese men (n = 230)
 
  Obese men (n = 352)
 
    
Rs7230534 TT 134 (59) CT 85 (37) CC 8 (4) TT 212 (61) CT 123 (35) CC 15 (4) 99 22 
Rs8090997 GG 53 (23) AG 123 (53) AA 54 (24) GG 86 (24) AG 181 (52) AA 84 (24) 100 50 
C19699C→T CC 139 (60) CT 83 (36) TT 8 (4) CC 192 (55) CT 137 (39) TT 23 (6) 100 24 
C19787G→T GG 91 (40) GT 106 (47) TT 30 (13) GG 163 (47) GT 156 (44) TT 31 (9) 99 33 
SNPNon obese women (n = 434)Obese women (n = 547)Success rate (%)Allele frequency (%)
Rs7230534 TT 245 (58) CT 156 (37) CC 23 (5) TT 300 (57) CT 190 (36) CC 34 (7) 97 24 
C81172Gins GG 154 (58) Gdel 101 (38) deldel 10 (4) GG 134 (54) Gdel 100 (40) deldel 14 (6) 98 24 
Rs8090997 GG 123 (29) AG 218 (51) AA 88 (20) GG 137 (25) AG 285 (53) AA 117 (22) 99 47 
C19660G→A GG 161 (61) AG 83 (31) AA 22 (8) GG 150 (61) AG 86 (35) AA 11 (4) 98 23 
C19699C→T CC 229 (54) CT 180 (42) TT 16 (4) CC 301 (55) CT 203 (37) TT 41 (8) 99 27 
C19787G→T GG 171 (41) GT 203 (48) TT 45 (11) GG 236 (44) GT 251 (47) TT 46 (9) 97 33 
 Non obese men (n = 230)
 
  Obese men (n = 352)
 
    
Rs7230534 TT 134 (59) CT 85 (37) CC 8 (4) TT 212 (61) CT 123 (35) CC 15 (4) 99 22 
Rs8090997 GG 53 (23) AG 123 (53) AA 54 (24) GG 86 (24) AG 181 (52) AA 84 (24) 100 50 
C19699C→T CC 139 (60) CT 83 (36) TT 8 (4) CC 192 (55) CT 137 (39) TT 23 (6) 100 24 
C19787G→T GG 91 (40) GT 106 (47) TT 30 (13) GG 163 (47) GT 156 (44) TT 31 (9) 99 33 

Data are genotype n (%) unless otherwise indicated.

*

Obesity defined as BMI >30 kg/m2 and nonobese as BMI ≤30 kg/m2.

Obesity is marginally associated with the C19787G→T G allele (OR 1.32 [95% CI 1.03–1.69], P = 0.027 [one sided]).

TABLE 2

CIDEA gene haplotypes

HaplotypeFrequency (%)
TGAACG 21 
CCGGCT 14 
TGAGTG 14 
TGGGCG 11 
TGGGTG 10 
TGAGCT 
TGGGCT 
CCGGCG 
HaplotypeFrequency (%)
TGAACG 21 
CCGGCT 14 
TGAGTG 14 
TGGGCG 11 
TGGGTG 10 
TGAGCT 
TGGGCT 
CCGGCG 

SNPs in order: rs7230534, C81172Gins, rs8090997, C19660G→A, C19699C→T, and C19787G→T. Estimated haplotype frequencies are based on genotypes from n = 981 women (n = 524 for C81172Gins and C19660G→A) using SNPHAP. Similar haplotype distribution was seen in men. Haplotypes with frequency >2% are shown.

1.
Zhou Z, Yon Toh S, Chen Z, Guo K, Ng CP, Ponniah S, Lin SC, Hong W, Li P: Cidea-deficient mice have lean phenotype and are resistant to obesity.
Nat Genet
35
:
49
–56,
2003
2.
Dahlman I, Linder K, Arvidsson Nordstrom E, Andersson I, Liden J, Verdich C, Sorensen TIA, Arner P: Changes in adipose tissue gene expression by energy-restricted diets in obese women.
Am J Clin Nutr
81
:
1275
–1285,
2005
3.
Arvidsson Nordström E, Ryden M, Backlund EC, Dahlman I, Kaaman M, Blomqvist L, Cannon B, Nederggard J, Arner P: A human-specific role of cell death–inducing DFFA (DNA fragmentation factor-α)-like effector A (CIDEA) in adipocyte lipolysis and obesity.
Diabetes
54
:
1726
–1734,
2005
4.
Parker A, Meyer J, Lewitzky S, Rennich JS, Chan G, Thomas JD, Orho-Melander M, Lehtovirta M, Forsblom C, Hyrkko A, Carlsson M, Lindgren C, Groop LC: A gene conferring susceptibility to type 2 diabetes in conjunction with obesity is located on chromosome 18p11.
Diabetes
50
:
675
–680,
2001
5.
Chagnon YC, Chen WJ, Perusse L, Chagnon M, Nadeau A, Wilkison WO, Bouchard C: Linkage and association studies between the melanocortin receptors 4 and 5 genes and obesity-related phenotypes in the Quebec Family Study.
Mol Med
3
:
663
–673,
1997
6.
Inohara N, Koseki T, Chen S, Wu X, Nunez G: CIDE, a novel family of cell death activators with homology to the 45 kDa subunit of the DNA fragmentation factor.
Embo J
17
:
2526
–2533,
1998
7.
Dahlman I, Eriksson P, Kaaman M, Jiao H, Lindgren CM, Kere J, Arner P: alpha(2)-Heremans-Schmid glycoprotein gene polymorphisms are associated with adipocyte insulin action.
Diabetologia
47
:
1974
–1979,
2004