UCP2 plays an important role in the development and treatment of overweight and obesity, as an obesity candidate gene involved in the regulation of glucose/lipid metabolism and energy homeostasis. UCP2 affects the susceptibility to obesity and obesity-related diseases by decreasing the activity or expression of these UCPs, thereby increasing oxidative phosphorylation coupling to reduce energy expenditure. Thus, the expression/activity levels of UCP2 influence the relationship between UCP2 gene polymorphisms and obesity [26]. The 3'UTR region of the UCP2 Ins/Del polymorphic gene, which is only 158 bp away from the transcription termination codon, may function by participating in mRNA processing or transcriptional stability, and lower transcript stability may result in lower UCP2 protein translocation [11]. These three polymorphisms are currently the most studied in the UCP2 gene: 45bp Ins/Del in exon 8, missense variant in exon 4 (rs660339, Ala55Val, C/T), and one in the promoter region (rs659366, 866G/A) [12]. The association of one of these 45bp Ins/Del polymorphisms with overweight and obesity remains highly controversial. Therefore, we systematically summarized 20 eligible case-control studies using the largest sample size available for this meta-analysis to assess the association between UCP2 45bp Ins/Del with genetic susceptibility and overweight and obesity and guide future studies.
The analysis showed that the UCP2 45bp Ins/Del gene polymorphism was significantly associated with genetic susceptibility to overweight and obesity only in the recessive model (OR = 1.24, 95%CI = 1.07–1.43, P = 0.004), but race-specific subgroup analysis showed that the UCP2 45bp Ins/Del gene polymorphism was significantly associated with genetic susceptibility to overweight and obesity in Asian populations in the allelic (OR = 1.18, 95%CI = 1.02–1.36, P = 0.027), dominant (OR = 1.20, 95%CI = 1.02–1.41, P = 0.030), and heterozygote (OR = 1.19, 95%CI = 1.01–1.41, P = 0.043) models, and no significant association was found in Caucasian populations.
Brondani's study showed similar findings to this study that polymorphism is significantly associated with increased BMI in Asian populations [27]. However, in a study on Caucasian populations, the findings of Brondani et al. contradicted our findings by suggesting that the UCP2 45bp Ins/Del gene polymorphism was significantly associated with obesity in Caucasian populations [28]. Comparative results from other studies have shown that individuals carrying the II genotype and Ins allele are at higher risk of obesity than those carrying the DD genotype compared to other genotypes or alleles [17, 21, 29]. Moreover, Ins allele carriers have a higher BMI in some populations [30]. Conversely, Zhang et al. and Surniyantoro et al. showed that 45bp Ins/Del gene polymorphism is not a risk factor for overweight and obesity [24, 31]. The nutritional characteristics of populations may influence the relationship between genetic variation and obesity, and food and cultural habits and environmental factors differ for each race, which may influence the association between UCP polymorphisms and obesity [24, 32]. Therefore, future studies on obesity gene polymorphisms should consider environmental factors and dietary habits. Moreover, there are numerous genes associated with obesity [33], and each population has different genetic variants in its gene pool; these factors may also play a role in obesity. Kring et al. concluded that there is a lack of significant correlation between genetic variants and BMI because obesity is a mixed phenotype, and several other proxies for overweight and obesity are available, such as waist circumference, waist circumference for a given BMI, sagittal abdominal diameter, and waist-to-hip ratio [15], which should be considered in future studies. Some researchers have suggested that the effect of insertion polymorphisms may be age-related and associated with late-onset obesity [8]. Therefore, we conducted a subgroup analysis of the study population divided into child and adult groups for age and only found significant associations in the recessive (OR = 1.19, 95% CI = 1.01–1.41, P = 0.011) and homozygote (OR = 1.24, 95% CI = 1.04–1.48, P = 0.017) models in the adult group, while the other groups did not show significant associations. A study by Gul et al. concluded that UCP2 exon 8 Ins/Del had no significant effect on the risk of obesity in adolescents, which is similar to our findings, and that low HDL cholesterolemia may be associated with the Ins allele [12]. Gender-stratified UCP2 45bp Ins/Del gene polymorphism analysis by Surniyantoro et al. and Papazoglou et al. showed that gene polymorphisms had opposite effects on male and female populations, with the II genotype and I allele leading to reduced UCP2 expression and increased body weight in the male population. Conversely, in the female population, the UCP2 45bp Ins/Del gene polymorphism is recessively associated with obesity[22, 24].
Heterogeneity between studies is a matter of concern, and high heterogeneity was observed in all our analyses of genetic models of the association of the UCP2 45bp Ins/Del gene polymorphism with overweight and obesity. After stratifying the analysis by ethnicity, heterogeneity was significantly lower in the analysis of Asian populations; however, significant heterogeneity remained in the analysis of Caucasian populations, suggesting some other confounding factors in the study on Caucasian populations. The sources of heterogeneity were worth exploring; therefore, a regression analysis was implemented, including covariates such as age, sex, and sample size. However, none of these covariates individually or jointly explained the observed heterogeneity. Study quality, general characteristics of participants, representation of participants, gene-environment interactions, and genotyping methods may contribute to heterogeneity. Without additional information on the metabolic and clinical characteristics of the articles analyzed, the role of these factors in the sources of heterogeneity is difficult to describe accurately.
Therefore, the results of the current meta-analysis should be interpreted with caution. First, the definition of cases is not uniform across studies; in some studies, cases are defined as obese populations, morbidly obese populations, or mixed populations of overweight and obesity. Moreover, in some relevant studies, the threshold values of overweight or obesity were different. Second, due to technical limitations, we only retrieved studies published in English and Chinese, which may lead to certain omissions. Lastly, our analysis revealed significant heterogeneity among studies on Caucasian populations.