Gene variants in the FTO gene are associated with adiponectin and TNF-alpha levels in gestational diabetes mellitus

The prevalence of obesity and overweight is increasing around the world [1]. Across the Organisation for Economic Co-operation and Development (OECD) countries, Mexico is the country with the second highest obesity rate among adults and a greater proportion of this population are women [2]. In 2012, among women of childbearing age, 18% were obese [3]. Obesity in women of reproductive age is a known risk factor for numerous health problems, including gestational diabetes mellitus (GDM), pregnancy induced hypertension, preeclampsia, caesarean delivery, post-partum hemorrhage, and delivery of large-for-gestational-age infants [4, 5]. Excessive gestational weight has also been associated with adverse health outcomes for both the mother and her child [6].

Obesity is a neuroendocrine disorder in which both genetic predisposition and environmental factors act in concert [7]. In recent years, genome-wide association studies (GWAS) have identified >90 loci containing single-nucleotide-polymorphisms (SNPs), many in intronic regions, associated with human obesity [8, 9]. The strongest genetic association with risk to polygenic obesity is in SNPs in intron 1 and 2 of the FTO (fat mass and obesity associated) gene [10, 11].

The precise mechanism responsible for FTO’s effect on obesity in humans has been elusive. It has been suggested that the obesity-associated FTO SNPs may affect obesity through greater food intake and increased hunger/lowered satiety and through modulating adiposity factors. The FTO SNP rs9939609 has been found to be significantly associated with levels of circulating ghrelin and leptin [12, 13], which are key mediators of ingestive behavior, and has been significantly associated with serum adiponectin, an antihyperglycemic, antiatherogenic, and anti-inflammatory adipokine [14]. Additionally, FTO SNP rs9939609 has been related to systemic inflammation [15], which is implicated in insulin resistance and diabetogenesis [16].

Increasing evidence suggests that the FTO SNPs influence the expression of genes proximal to FTO in the locus, namely, RPGRIP1L, IRX3, and IRX5 [17‐19]. A recent study identified rs1421085 as a causal variant to obesity, which influences the expression of IRX3 and IRX5 in preadipocytes, changing human adipocytes from fuel utilization through enhanced thermogenesis to the function as substrate storage [20].

FTO variants rs9939609 and rs8050136 also confer higher risk for type 2 diabetes (T2D) in an adiposity-dependent manner and have been associated with body mass index (BMI) and GDM in Caucasians women [21‐25]. The present study investigates the polymorphisms rs8050136, rs9939609, and rs1421085 of the FTO gene in women with gestational diabetes and their associations with maternal pre-pregnancy weight and BMI, gestational weight gain and mediators of insulin resistance in GDM like leptin, adiponectin, ghrelin and tumor necrosis factor-alpha (TNF-alpha), compared with healthy pregnant controls.

Table 1 displays the anthropometric and biochemical characteristics from control and GDM subjects. Age, weight, BMI, glucose, insulin, HOMA-IR and triglycerides were significantly higher in the GDM women compared with controls. GDM patients also had a high prevalence of diabetes in their family history. Gestational weight gain, total cholesterol, leptin, adiponectin, ghrelin and TNF-alpha were not significantly different between groups.

All genotypes in both groups were in HW equilibrium (P > 0.05). We estimated linkage disequilibrium (LD) among the three studied variants: rs8050136 was in complete linkage disequilibrium with rs9939609 and these SNPs and rs1421082 were in a tight LD block (D′ 0.94 and γ² 0.89).

The genotype and allele frequencies of the three studied polymorphisms were not significantly different between the groups (Table 2). As a result of the small number of homozygotes for the rare alleles, we compared the common allele homozygotes and the carriers of the rare allele in a dominant model.

In GDM group, women carrying the A risk allele of rs8050136 and the A risk allele of rs9939609 had increased plasma TNF-alpha, and decreased adiponectin levels; these associations remained significant after adjusting for age and pre-gestational body weight. Moreover, C risk allele of rs1421085 was also associated with both increased weight gain during pregnancy and with TNF-alpha (Table 3). In contrast, no significant associations of FTO rs8050136, rs9939609, and rs1421085 were found with any of the studied biomarkers analyzed in control group (data not shown).

In the logistic regression analysis, the association between GDM and maternal age as well as weight was significant (OR 1.24, 95% CI 1.15–1.35, and P = 0.001; OR 1.08, 95% CI 1.05–1.12, and P = 0.001, respectively).

GDM accounts for one of the most frequent gestational complications, affecting 4–18% of all pregnancies worldwide, depending on the GDM criteria used [29]. In recent decades, the number of patients with GDM has increased significantly; the main factors for this rise are the increasing prevalence of overweight and obesity in both developed and developing countries [30, 31]. Likewise, some other risk factors for GDM include age, ethnic origin and a family history of T2D [32]. In this study, GDM patients were reported to be associated with older age, higher BMI and greater parental history of T2D, when compared to healthier pregnant women.

Additionally, evidence is accumulating that susceptibility to GDM has a genetic component [25, 33]. FTO has been reported to be associated with obesity, and recent studies have shown the association of some FTO SNPs with GDM, like rs8050136 or rs9939609 [23‐25]; however, to our knowledge, no studies on rs1421085 have been previously reported. In the present study, we examined the association of these genetic variants in FTO gene in women with GDM.

We found no association between the FTO SNPs rs8050136 and rs9939609 with GDM, in accordance with Cho et al. and Fabrico et al. [34, 35]. However, our results differed from three studies that examined rs9939609, which demonstrated an association among the A-alleles and the risk for GDM [23‐25]. We hypothesized that the A-alleles’ low frequency added to the small sample size of this study could be the major factors for the divergence of our results. It has been suggested that the effects of all three SNPs in the FTO gene may be apparent only in populations with higher minor allele frequency [36]. The frequencies of the rare A-alleles observed for the GDM group were lower than those reported for Europeans and Euro-Brazilians [23‐25, 35]. On the other hand, these frequencies are similar to those reported for Korean women [34]. To date, most of the genetic association studies have been performed in white populations.

Several genetic variants of FTO have been related to body weight, hip and waist circumference and BMI in numerous populations of European origin [37‐39]. Nevertheless, we did not find any association between the FTO SNPs and the pre-pregnancy anthropometric characteristics analyzed. Our results differed from those of Lawlor et al. and Gaillard et al. which demonstrated an association among the risk variants in rs9939609, and rs8050136 and pre-pregnancy weight and BMI [40, 41]. However, we observed that the risk allele C in SNP rs1421085 was associated with greater gestational weight gain in GDM group. Nevertheless, given the composite nature of gestational weight gain, further examination of this outcome might be undertaken by examining the association of this variant with maternal fat stores and fetal adiposity. It should be pointed out that this variant was identified recently as a causal variant in preadipocytes, where the risk allele affects the basic anabolic function of the adipocyte, shifting it from substrate storage to fuel dissipation through increased mitochondrial uncoupling [20].

The genetic variant rs9939609 evaluated in the present study has been involved in appetite ratings through regulation of circulating ghrelin and leptin [12, 13]. Our data did not show these associations; however, we measured total ghrelin, which includes both acylated and desacylated forms. Measuring the acylated form of this hormone, might have been an advantage if we consider that ghrelin is biologically active only in the acylated form [42]. Regarding leptin, the association of FTO polymorphisms with this adipokine appears to be mediated via increased adiposity, as in some studies the association disappears when correcting by BMI [43, 44]. In our study, we did not find an association of FTO polymorphisms with BMI.

In addition, considerable data support a role for adipocyte-derived cytokines, such as adiponectin, and TNF-alpha in the pathogenesis of GDM [45‐47]. Therefore, we undertook to investigate the influence of FTO genetic variants on maternal adipokines. In the present study, no statistical differences were found at these hormone levels between GDM and controls. However, interestingly higher levels of pro-inflammatory TNF-alpha and lower levels of adiponectin were associated with the FTO rs8050136 and rs9939609 risk allele A in GDM group. Both associations remained significant after adjusting for age, maternal pre-gestational body weight and insulin therapy. Our results are mainly consistent with those of De Luis et al., who also observed that the minor A allele of the FTO rs9939609 was significantly associated with lower serum adiponectin concentrations independently of potential confounders including adiposity [14]. Regarding the association of the FTO SNPs with systemic inflammation, a German study has shown that rs9939609 A-allele was associated with increased levels of C-reactive protein (CRP) [15]. These findings were significant, independent of the adiposity level. However, a recent study by Zimmermann et al. in a population of Danish men representing a broad range of BMI did not find an independent association between the obesity associated FTO rs9939609 variant and systemic inflammation [43]. To our knowledge no previous studies have investigated the relationship between FTO SNPs and TNF-alpha, only one study to date has examined the association between FTO mRNA expression and expression of TNF-alpha in subcutaneous adipose tissue (SAT), and found that SAT FTO mRNA expression was related to SAT TNF-alpha expression, linking adipose tissue FTO expression and a proinflammatory molecule implicated in insulin resistance [48]. Several studies have suggested that increased levels of TNF-alpha may be related to the development of GDM independent of BMI [49]. TNF-alpha impairs insulin signalling by increasing serine phosphorylation of insulin receptor substrate (IRS)-1, and diminishing insulin receptor (IR) tyrosine kinase activity [50].

This study is subject to certain limitations, the first being the differences in age and weight between women with GDM and women with normal pregnancy. A control group consisting of age- and BMI-matched pregnant women without GDM might have been more suitable. Second, our study was underpowered to detect associations of FTO SNPs with GDM, probably because of their low frequencies. Finally, for adiposity, we only considered BMI, a surrogate for body fatness, and body fat percentage is a more accurate measurement.

The present study suggests that the FTP SNPs rs8050136, rs9939609, and rs1421085 are not a major genetic regulator in the etiology of GDM in the studied ethnic group. However, these SNPs were associated with serum adiponectin and TNF-alpha concentrations in GDM subjects. Additional larger studies are required to examine the association of FTO genetic variants in Mexican women.