- Top of page
Mice lacking the integrin αMβ2 (Mac-1, CD11b/CD18) develop an obese phenotype on western diet rich in fat. However, no association has been found between variations in the human genes encoding the integrin αMβ2 and obesity. This study was aimed to investigate the association between a single-nucleotide polymorphism (SNP) (rs235326) in the gene encoding human integrin β2 subunit (ITGB2) with obesity. Our subject cohort comprised 651 people of Japanese ethnicity, of which 274 were Japanese Americans living in Hawaii, and the remaining 377 were native Japanese, two populations in the same genetic background with or without westernized life style. We genotyped the rs235326 polymorphism using a TaqMan assay. In the Japanese-American population, the risk of obesity was found to be 3.29-fold higher (a 95% confidence interval of 1.25–8.67, P = 0.02) in TT homozygotes than in C carriers, using a recessive model and logistic regression analysis that had been adjusted for age. This association was not found in native Japanese individuals. These results indicate that the rs235326 polymorphism in the ITGB2 gene is associated with obesity in Japanese living in the United States whose diet has become “westernized.”
The integrins are a family of αβ heterodimeric transmembrane cell-adhesion receptors (1). Defects in αMβ2 (Mac-1, CD11a/CD18), a leukocyte β2-integrin, or its ligand intercellular adhesion molecule 1 have previously been reported to result in obesity in mice (2). This observation was subsequently supported by the finding that both fat oxidation and insulin metabolism are impaired in integrin β2 subunit (ITGB2) knockout mice (3). Two additional studies further support the involvement of the β2 integrins in energy metabolism. Elevated plasma levels of nonesterified fatty acids are present in individuals with insulin resistance and type 2 diabetes, and cause increased expression of integrin αMβ2 on monocytes, thus promoting their adhesion to endothelial cells (4). Integrin αXβ2 is expressed on adipose tissue macrophages in obese mice but not in lean mice (5). The β2 integrins have thus been postulated to be involved in obesity and regulation of metabolic responses to fasting. However, the influence of variations in the ITGB2 gene upon the development of obesity in humans has not been reported previously.
A combination of environmental and genetic factors can generate a predisposition to obesity. Several genetic variations affecting obesity recently have come to light, including the genes for the β3 adrenergic receptor (6,7), peroxisome proliferator-activated receptor-γ (8), uncoupling proteins (9), and adiponectin (10). On the basis of meta analyses, a Trp64Arg variation in the β3 adrenergic receptor gene has been shown to be associated with obesity regardless of the ethnicity of the individual (7), whereas a relationship between the Pro12Ala polymorphism in the peroxisome proliferator-activated receptor-γ gene and obesity was observed only in whites (8). These results indicate that an association between a genetic variation and obesity is dependent upon both gene-gene and gene-environment interactions in a given population.
We have been analyzing the morbidity levels of diseases that are related to “westernized” life styles among the Japanese communities living in Hawaii since 1970 (11). We have shown that Japanese Americans have largely adopted the local diet (12), and that the prevalence of obesity and type 2 diabetes is much higher among these individuals than among matched populations still living in Japan (13). In this study, we have analyzed the association of one single-nucleotide polymorphism (SNP) in the ITGB2 gene, rs235326, in both Japanese-American and native Japanese populations. Linkage disequilibrium plot (14) of SNPs in the ITGB2 gene is shown as Figure 1, from which the size of ITGB2 and location of rs235326 can be seen. These two groups live in distinct environments but have the same genetic background. Our results disclose an association between the rs235326 SNP and obesity in Japanese Americans.
Figure 1. Linkage disequilibrium (LD) plot of single-nucleotide polymorphisms (SNPs) in integrin β2 subunit (ITGB2) and location of rs235326. Data are downloaded from HapMap database (Rel. 22; Japanese and Chinese) and visualized by Haploview 4.0. SNPs with minor allele frequency ≥0.001 are described. DPI, dots per inch.
Download figure to PowerPoint
A comparison between the clinical characteristics of the Japanese-American and Japanese subjects in this study is shown in Table 1. Because the average age of the Japanese-American population is greater than that of the native Japanese group, clinical data were analyzed using analysis of covariance with age as a covariate. The average weight, BMI, and waist circumference were found to be significantly higher in the Japanese-American group. In addition, the serum insulin levels at 2 h after a 75-g oral glucose-tolerance test was found to be higher, and both diabetes and obesity were more prevalent in the Japanese-American population compared with the Japanese population. As the genetic background is similar for both study groups, these differences are likely to be a consequence of environment and thus could be a reflection of the more “westernized” lifestyle of Japanese Americans. In fact, we have previously investigated the nutrition status of Japanese Americans and have found that their intake of animal fat, saturated fatty acids, and simple carbohydrates (fructose in particular) is far greater than is observed for the typical native Japanese diet. The animal fat intakes of Japanese vs. Japanese Americans are 26.8 g vs. 43.8 g in men and 25.2 g vs. 31.0 g in women, saturated fatty acids are 16.2 g vs. 26.9 g in men and 15.9 vs. 20.0 in women. The intake of simple carbohydrates in Japanese Americans is 1.2–1.6 times higher than that of Japanese (15).
Table 1. Clinical characteristics and comparison between Japanese Americans and Japanese
The genotype frequencies and allelic frequencies of the rs235326 SNP in the two populations are similar (Table 2). There was no evidence of a departure from Hardy-Weinberg equilibrium. We next assessed association of rs235326 with obesity in each population by logistic regression after adjustment for age (Table 3). In the Japanese-American population, a susceptibility to obesity was found to be significantly higher in T carriers by both the additive and recessive models. This was not seen in the Japanese population using any of the three models. In the Japanese-American population, T allele affects on susceptibility to obesity which is not completely abolished by the presence of C allele. The analyses of the association with our sample sizes required odds ratio between 2 and 4 to have ∼80% power in the Japanese American and between 1.5 and 3 in the Japanese population. To understand whether type 2 diabetes affected on this association, we then recalculated the association including morbidity of type 2 diabetes as a covariate, and found that the odds ratios were minimally changed (Table 4). We also examined relations of rs235326 to fasting and 2 h glucose and insulin, and association of type 2 diabetes by analysis of covariance or logistic regression analysis adjusting for age. No significant difference was found in the glucose and insulin levels (data not shown). Because considerable number of patients with type 2 diabetes were on medication, this may be influenced by the treatment. There was no association of rs235326 with type 2 diabetes (odds ratio, 95% confidence interval; 1.51, 0.55–4.14 in recessive model). These results suggest that association of rs235326 with obesity was not influenced by relation of type 2 diabetes to rs235326.
Table 2. Genotype and allele frequency of rs235326
Table 3. Regression analysis of rs235326 and obesity adjusted for age
Table 4. Regression analysis of rs235326 and obesity adjusted for age and morbidity of diabetes
In this study, we find that TT homozygotes for rs235326 have a greater than threefold increased susceptibility to obesity, compared with the C carriers, in a Japanese-American population. This is the first report to describe an effect of an SNP in ITGB2 on obesity in humans; however, this association was not found in the native Japanese population. Thus, effects of T allele appear to be dependent on environmental factors. The most likely responsible factor is the adoption of a westernized diet. The Japanese gene pool appears to have been adapted to starvation during its evolution (16). For example, the prevalence of a gain-of-function allele (Pro 12) of peroxisome proliferator-activated receptor-γ is higher in Japanese compared with that in white populations (96% vs. 80%) (17), and the Arg 64 allele of the β3 adrenergic receptor, which confers a lower resting metabolic rate, is also higher among Japanese compared with that among whites (20% vs. 10%) (17). This suggests that the normal metabolic state among Japanese is likely to be more sensitive to a “westernized” diet and lifestyle and may have favored the manifestation of the effect of the rs235326 SNP on obesity.
Associations between rs235326 and morbidity have been reported for coronary restenosis (18) and myeloperoxidase-antineutrophil cytoplasmic antibodies-positive systemic vasculitis (19), indicating that this synonymous polymorphism is associated with alteration in the functional properties of the β2 integrins. The β2 subunit forms four integrin heterodimers, αLβ2, αMβ2, αXβ2, and αDβ2, all of which are exclusively expressed on leukocytes. Involvement of the β2 integrins in obesity has been demonstrated in mice (2,3,4,20). These lines of evidence in mice are in concert with our current findings that a polymorphism in ITGB2 is associated with obesity.
Because we only evaluated a single polymorphism in the ITGB2 gene, we cannot be certain whether the association identified is due to this sequence variant or to another variant that is in tight linkage disequilibrium with rs235326. rs235326 is a synonymous variant, so any effects would not be likely be due to changes in the function of the β2 protein. However, synonymous SNPs can alter the rate of protein translation, so it is conceivable that this SNP alters steady-state levels of β2 protein. Alternatively, rs235326 may simply be a marker for another variant in or adjacent to ITGB2 or even in a gene on a different chromosome. However, considering the obesity of integrin β2 null mice, this variant should directly affect either β2 protein function or β2 expression.
In conclusion, we have found the influence of the rs235326 SNP in ITGB2 on obesity, an effect that is dependent on environmental factors that differ between Japan and Hawaii. Our current results thus strongly suggest that β2 integrins play an important role in fat metabolism in humans and in mice.