D2 Thr92Ala and PPARγ2 Pro12Ala Polymorphisms Interact in the Modulation of Insulin Resistance in Type 2 Diabetic Patients

Authors


(almaia@ufrgs.br)

Abstract

Type 2 deiodinase (D2) converts T4 into its active metabolite T3, an essential step in thyroid metabolism. A Thr92Ala polymorphism in the gene encoding D2 has been inconsistently associated with insulin resistance (IR). Recently, it was reported that the D2 Thr92Ala (rs225014) and the peroxisome proliferator-activated receptor (PPAR) γ2 Pro12Ala (rs1801282) polymorphisms interact in the modulation of metabolic syndrome in nondiabetic subjects. Here, we investigated the effect of both polymorphisms, isolated or in combination, on IR in patients with type 2 diabetes mellitus (DM2). The D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms were genotyped in 721 DM2 patients. IR was evaluated using the homeostasis model assessment—IR (HOMAIR) index in a subgroup of 246 DM2 subjects. The frequencies of D2 Ala92 and PPARγ2 Ala12 variants were 0.390 and 0.074, respectively. Patients carrying D2 Ala/Ala genotype had a higher fasting plasma insulin and HOMAIR index as compared to patients carrying Thr/Ala or Thr/Thr genotypes (P = 0.022 and P = 0.001, respectively). A significant synergistic effect was observed between D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms on HOMAIR index, with carriers of both D2 Ala/Ala genotype and PPARγ2 Ala12 allele showing the highest HOMAIR values, after adjusting for age, gender, BMI, and use of medication for DM2 (P = 0.010). In conclusion, DM2 patients harboring both D2 Ala/Ala genotype and PPARγ2 Ala12 allele seem to present more severe IR than those with other D2/PPARγ2 genotype combinations. These findings suggest that these polymorphisms interact in the IR modulation, which may constitute a potential therapeutic target.

Introduction

Type 2 diabetes mellitus (DM2) is a heterogeneous group of disorders usually characterized by varying degrees of insulin resistance (IR) and increased blood glucose concentrations (1). Ultimately, IR results either from uninhibited hepatic gluconeogenesis or decreased glucose disposal rate in tissues such as skeletal muscle, adipose tissue, and liver (2). The type 2 iodothyronine deiodinase (D2) catalyzes the intracellular conversion of the prohormone thyroxine (T4) into its active metabolite triiodothyronine (T3), an essential step in thyroid hormone metabolism (3). In humans, the gene that encodes D2 (DIO2) is also expressed in skeletal muscle and adipocytes (4,5), in which thyroid hormone is known to upregulate the expression of the primary glucose transporter 4, and thereby to increase basal and insulin-stimulated glucose uptake (6,7,8).

Previous studies have reported that a DIO2 single-nucleotide polymorphism A/G, in which a threonine (Thr) changes to alanine (Ala) at codon 92 (Thr92Ala), was associated with a ∼20% lower glucose disposal rate in nondiabetic white subjects and a more pronounced IR in DM2 patients (9,10). Moreover, decreased D2 activity has been described in tissue biopsy samples obtained from subjects homozygous for the Ala92 minor allele (10). Nevertheless, other studies failed to demonstrate an association between the D2 Thr92Ala polymorphism and glycemic traits or DM2 (11,12,13).

The peroxisome proliferator-activated receptor (PPAR) γ2 gene encodes a transcription factor involved in the regulation of adipocyte differentiation and intracellular insulin signaling, and it is recognized as playing an important role in determining the risk of IR-related abnormalities (14,15,16). Of interest, in a recent study, a significant interaction was observed between the D2 Thr92Ala polymorphism and the Pro12Ala polymorphism of the PPARγ2 gene (17). This gene-gene interaction was associated with elevated systolic and diastolic blood pressure (BP) and also with an increased risk of metabolic syndrome in 590 nondiabetic subjects from Italy. Among these subjects, carriers of both D2 Ala92 and PPARγ2 Ala12 alleles displayed the most severe phenotypes. This study also reported the presence of a PPAR element in the DIO2 promoter, thus providing biological plausibility for the observed interaction (17).

Based on these observations, we hypothesized that the interaction between the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms might account for part of the conflicting results on the association between the D2 Thr92Ala polymorphism and IR observed in different populations. Therefore, we decided to evaluate the potential synergistic effect of D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms on IR and related characteristics in a sample of white DM2 patients.

Methods and Procedures

Patients and phenotype measurements

The sample population consists of DM2 patients participating in a multicenter study that started recruiting patients in Southern Brazil in 2002. That project aimed to study risk factors for DM2. Initially, it included four centers located at general hospitals in the State of Rio Grande do Sul, namely Grupo Hospitalar Conceição, Hospital São Vicente de Paula, Hospital Universitário de Rio Grande, and Hospital de Clínicas de Porto Alegre. The detailed description of that study can be found elsewhere (18). DM2 was defined by diagnosis of diabetes after the age of 40 years with no use of insulin during the first year after diagnosis and no previous episodes of ketoacidosis (1).

The D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms were analyzed in blood samples from 721 white DM2 patients. IR was evaluated only in 246 patients, including a subgroup of 183 subjects previously reported (10). All patients were of European ancestry (mostly descendants of Portuguese, Spanish, Italians, and Germans). The ethnic group was defined on the basis of self-classification and subjective classification (skin color, nose and lip shapes, hair texture, and family history). A standard questionnaire was used to collect information about age, age at DM2 diagnosis, and drug treatment. All patients underwent physical and laboratory evaluations. They were weighed without shoes and in light outdoor clothes and had their height measured. BMI was calculated as weight (kg)/height (m)2. To calculate waist-to-hip ratio, waist circumference was measured at the narrowest point, as viewed from the front, and hip circumference was measured at the widest point. BP was measured twice after a 5-min rest in the sitting position using a mercury sphygmomanometer (Korotkoff phases I and V). The mean value of two measurements was used to calculate systolic and diastolic BP. Hypertension was defined as BP levels ≥140/90 mm Hg, or if the patient was taking antihypertensive drugs.

The information obtained from the study did not influence the patients' diagnosis or treatment. The protocol was approved by the Hospital ethical committees, and all patients gave their written informed consent.

Biochemical measurements

A serum sample was collected after a 12-h fast. Glucose levels were determined using the glucose oxidase method; creatinine by the Jaffé reaction; glycated hemoglobin by an ion-exchange high-performance liquid chromatography procedure (Merck-Hitachi L-9100 glycated hemoglobin analyzer; Merck, Darmstadt, Germany; reference range: 4.7–6.0%); and total plasma cholesterol, high-density lipoprotein cholesterol, and triglycerides by enzymatic methods. Serum insulin was measured by radioassay (ElecsysR Systems 1010/2010/modular analytics E170; Roche Diagnostics, Indianopolis, IN) in a subgroup of 246 white DM2 patients not receiving insulin therapy and with serum creatinine <114.4 µmol/l (1.5 mg/dl). Insulin sensitivity was estimated by homeostasis model assessment—IR (HOMAIR) index (fasting insulin (micro units per ml) × fasting glucose (mmol/l)/22.5) (19). The mean HOMAIR values of control subjects in our laboratory was 1.84 ± 1.02 (20).

Molecular analyses

DNA was extracted from peripheral blood leukocytes by a standardized salting-out procedure. Both D2 Thr92Ala (rs225014) and PPARγ2 Pro12Ala (rs1801282) polymorphisms were determined using primers and probes contained in the Human Custom TaqMan Genotyping Assay 40× (Assays-By-Design Service; Applied Biosystems, Foster City, CA). Primer and probe sequences used for genotyping the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms were: D2-5′-GGTACCATTGCCACTGTTGTCA-3′ (forward primer), D2-5′-GTCAGGTGAAATTGGGTGAGGAT-3′ (reverse primer), D2-FAM-5′-ATGTCTCCAGTGCAGAA-3′, D2-VIC-5′-CATGTCTCCAGTACAGAA-3′, PPARγ2-5′-GTTATGGGTGAAACTCTGGGAGATT-3′ (forward primer), PPARγ2-5′-GCAGACAGTGTATCAGTGAAGGAAT-3′ (reverse primer), PPARγ2-FAM-5′-CTATTGACGCAGAAAG-3′, PPARγ2-VIC-5′-CTCCTATTGACCCAGAAAG-3′. The reactions were conducted in a 96-well plate, in a total 5-µl reaction volume using 2 ng of genomic DNA, TaqMan Genotyping Master Mix 1× (Applied Biosystems), and Custom TaqMan Genotyping Assay 1×. The plates were then positioned in a real-time PCR thermal cycler (7500 Fast Real PCR System; Applied Biosystems) and heated for 10 min at 95 °C, followed by 50 cycles of 95 °C for 15 s and 63 °C for 1 min. Fluorescence data files from each plate were analyzed using automated allele-calling software (s.d.S 2.1; Applied Biosystems). Only 717 patients were successfully genotyped for both analyzed polymorphisms. All amplification reactions were performed twice. The genotyping success was >95%, with a calculated error rate based on PCR duplicates of 0.01%.

Statistical analyses

Results are expressed as mean ± s.d., % or median (minimum-maximum values). Allelic frequencies were determined by gene counting and departures from the Hardy-Weinberg equilibrium were verified using χ2 tests. Clinical and laboratory characteristics were compared using ANOVA, unpaired Student's t-test or χ2, as appropriate. Multiple linear regression analysis was performed with HOMAIR index (logarithmic) as dependent variable and age, gender, BMI, use of medications for DM2 treatment, and D2 Thr92Ala genotypes (Thr/Thr-Thr/Ala vs. Ala/Ala) as independent variables. Interaction between D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms in modulating fasting insulin and HOMAIR index was tested by general linear model univariate analyses, after adjusting for covariates (age, gender, BMI, and use of medications for DM2). In these interaction analyses, each of the two polymorphisms was modeled as dichotomous variables: D2 Thr/Thr-Thr/Ala vs. Ala/Ala, and PPARγ2 Pro/Pro vs. Pro/Ala-Ala/Ala. Variables with skewed distribution were logarithmically transformed before analyses. Bonferroni corrections were used to account for multiple comparisons carried out on the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms. All analyses were performed in SPSS version 16.0 (SPSS, Chicago, IL). P < 0.05 was considered statistically significant.

Results

The characteristics of the 721 DM2 patients belonging to the present study were as follows. The mean age was 58.7 ± 10.5 years, the mean DM2 duration was 12.1 ± 9.4 years, the mean glycated hemoglobin was 7.35 ± 2.03%, and the mean BMI was 28.8 ± 5.1 kg/m2. Males comprised 48.1% (n = 347) of the sample, and 66.2% (n = 477) of all patients had arterial hypertension. Eighty-four patients (11.7%) were treated by diet alone. Sulfonylureas alone were used by 31.5% (n = 227), metformin alone by 33.7% (n = 243), and the combination of both by 14.4% (n = 104) of the patients. Sixty-three patients (8.7%) were treated by other combinations of metformin and/or sulfonylureas with insulin. Clinical and laboratory characteristics of the whole sample stratified by sex can be found on Table 1. Systolic BP, BMI, total cholesterol, and high-density lipoprotein-cholesterol were significantly higher in females; whereas waist-to-hip ratio and serum creatinine were higher in males (Table 1).

Table 1.  Clinical and laboratory characteristics of DM2 patients grouped according to sex
inline image

A subgroup of 246 DM2 patients who were not receiving insulin therapy and had serum creatinine levels <114.4 µmol/l was further analyzed for IR measurements. This subgroup was representative of the whole sample: the mean age was 59.2 ± 9.9 years (P = 0.503 for the comparison with the whole sample), the mean DM2 duration was 12.5 ± 8.5 years (P = 0.582), the mean glycated hemoglobin was 7.5 ± 1.7% (P = 0.222), and the mean BMI was 29.0 ± 4.9 kg/m2 (P = 0.893). Males comprised 45.3% (n = 111) of the sample (P = 0.459), and 70.3% (n = 173) of these patients had arterial hypertension (P = 0.687).

Of the 714 patients genotyped for the D2 Thr92Ala polymorphism, 268 (37.5%) patients were homozygous for the Thr92 allele (Thr/Thr), 336 (47.1%) were heterozygous (Thr/Ala), and 110 (15.4%) were homozygous for the Ala92 allele (Ala/Ala). Genotypes were in agreement with those predicted by the Hardy-Weinberg equilibrium (P = 0.821). The Ala92 allele frequency was 0.39. Table 2 summarizes the clinical and laboratory data of the patients grouped according to the D2 Thr92Ala polymorphism. The median fasting plasma insulin level in Ala/Ala subjects was significantly higher than in patients with the Thr/Ala or Thr/Thr genotypes (P = 0.022), whereas fasting plasma glucose levels did not differ significantly among these genotypes (P = 0.147). However, when taking into consideration a Bonferroni threshold of 0.0031 (P = 0.05 divided by 16 variables analyzed in Table 2), median fasting insulin levels were not statistically different among D2 Thr92Ala genotypes. Patients carrying the Ala/Ala genotype had a higher HOMAIR index when compared with patients carrying Thr/Ala or Thr/Thr genotypes (P = 0.001), and this remained statistically significant after Bonferroni correction. The mean age, DM2 duration, proportion of males, systolic and diastolic BP, BMI, use of medication for DM2, creatinine levels, and metabolic control were not significantly different among the three different genotypes (Table 2).

Table 2.  Clinical and laboratory characteristics of DM2 patients grouped according to D2 Thr92Ala genotype
inline image

Assuming a recessive model of inheritance, patients with Thr/Ala and Thr/Thr genotypes were grouped and compared with patients carrying the Ala/Ala genotype. Fasting plasma insulin levels were higher in the Ala/Ala group compared with Thr/Ala-Thr/Thr group (median 116.7 pmol/l (minimum 22.2–maximum 418.0) vs. 78.4 (6.9–312.5); P = 0.001)). The HOMAIR index also was higher in the Ala/Ala group compared with the Thr/Ala-Thr/Thr group (6.6 (0.9–35.2) vs. 3.6 (0.3–19.5); P = 0.00001)). Because insulin sensitivity is known to be influenced by multiple independent factors, a multiple linear regression analysis was performed with HOMAIR index (log10 HOMAIR) as dependent variable. The Ala/Ala genotype remained significantly associated with HOMAIR index after controlling for age, gender, BMI, and use of medication for DM2 (standardized coefficient β for Ala/Ala genotype = −0.275, 95% confidence interval: −1.011 to −0,304; P = 0.00001).

Six-hundred and nineteen patients (85.9%) carried the PPARγ2 Pro/Pro, 97 (13.4%) the Pro/Ala, and only 5 (0.7%) the Ala/Ala genotype. Genotypes were in Hardy-Weinberg equilibrium (P = 0.610). The Ala12 allele frequency was 0.074. Table 3 illustrates the clinical and laboratory characteristics of patients grouped according to the presence of the PPARγ2 Ala12 allele (Pro/Pro vs. Pro/Ala-Ala/Ala). Taking into consideration a Bonferroni threshold of P < 0.0031, the PPARγ2 Pro12Ala polymorphism had no statistically significant effect on fasting plasma insulin levels, HOMAIR index or other variable illustrated on Table 3.

Table 3.  Clinical and laboratory characteristics of DM2 patients subdivided according to PPARγ2 Pro12Ala genotypes
inline image

Table 4 depicts the interaction analyses between D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms on plasma insulin levels and HOMAIR index in a subgroup of 246 DM2 patients evaluated for these features. In addiction, the F-statistic values obtained from GLM analyses used for studying the effect of D2 Thr92Ala/PPARγ2 Pro12Ala interaction on IR can be found on Table 5. As expected, D2 Ala/Ala carriers showed higher HOMAIR values as compared to patients with D2 Thr/Thr-Thr/Ala genotypes, and this result was independent of the PPARγ2 genotype (P = 0.009 in the PPARγ2 Pro/Pro group and P = 0.039 in the PPARγ2 Ala12 group) (Table 4). However, a significant gene-gene interaction was observed between the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms in modulating HOMAIR index (age-, gender-, BMI-, and use of medication-adjusted P value for interaction = 0.010): patients carrying both D2 Ala/Ala genotype and PPARγ2 Ala12 allele showed a higher HOMAIR index (8.5 (1.3–28.2)) as compared to patients carrying other D2/PPARγ2 genotype combinations (3.7 (0.3–19.5) in the D2 Thr/Thr—PPARγ2 Pro/Pro group, 5.8 (0.9–35.2) in the D2 Ala/Ala—PPARγ2 Pro/Pro group, and 3.5 (0.5–17.2) in the D2 Thr/Thr—PPARγ2 12Ala group (Table 4 and Figure 1). It is worth mentioning that the mean age, proportion of males, BMI, use of medication for DM2 and plasma insulin levels were not significantly different among the three different genotypes (Table 4). Besides, the percentages of these 246 patients who were using metformin and/or sulfonylurea drugs were not significantly different from those percentages observed for the whole DM2 sample (data not shown).

Table 4.  Interaction analyses between D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms on insulin resistance
inline image
Table 5.  General linear models (GLM) results for the interaction analyses between D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms on fasting insulin levels and HOMAIR
inline image
Figure 1.

HOMAIR index, calculated as described in Methods and Procedures, in DM2 patients according to D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms. Results are expressed as median (percentiles 25–75%). Circles represent outlier values. Age-, gender-, BMI-, and use of medication for DM2-adjusted interaction P value (general linear model univariate analysis) = 0.010. Ala, alanine; D2, type 2 deiodinase; DM2, type 2 diabetes mellitus; HOMAIR, homeostasis model assessment—IR; PPARγ2, peroxisome proliferator-activated receptor γ2; Pro, proline; Thr, threonine.

Discussion

Previous studies have shown contradictory results on the role of the D2 Thr92Ala polymorphism and IR (9,10,11,12,13). The reasons for the discrepant findings are still a matter of debate. Here, we demonstrate a significant gene-gene interaction between D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms in modulating HOMAIR index, with patients carrying both D2 Ala/Ala genotype and PPARγ2 Ala12 allele showing a higher HOMAIR index as compared to patients with other D2 Thr92Ala/PPARγ2 Pro12Ala genotype combinations. This result adds to the understanding of the molecular gene-gene interactions controlling IR and might partially explain some of the conflicting results on this issue.

The D2 Ala92 minor allele has been reported to be biologically less active than the common Thr92 variant (10). In humans, skeletal muscle and adipose tissue are the main sites of insulin-dependent glucose disposal (21). It has been hypothesized that, due to decreased D2 activity, carriers of the D2 Ala92 allele in homozygosis would have lower D2-generated T3 effect in skeletal muscle and adipocytes, which, consequently, could create a state of relative intracellular hypothyroidism, decreasing the expression of genes involved in energy use, such as glucose transporter 4, and thus leading to IR (6,7,22). However, the mechanism of reduced D2 activity in D2 Ala/Ala subjects is not clear. Although, a decrease in D2 velocity has been associated with the Ala/Ala genotype, two studies failed to identify any significant changes in the biochemical properties of the mutant D2 enzyme in vitro, suggesting that this polymorphism could be only a marker for abnormal DIO2 gene expression (10,23). Thus, it seems reasonable to speculate that the lower D2 activity is the result of linkage disequilibrium between the Thr92Ala polymorphism and a functional polymorphism elsewhere in the DIO2 gene. It is worth mentioning that the D2 Ala92 variant was also previously associated with an increased risk for symptomatic osteoarthritis (24), mental retardation in the iodine-deficient areas of China (25), Graves' disease (26), and arterial hypertension (27).

DM2 comprises a multifactorial group of disorders where both environmental factors and multiple genes contribute to its pathogenesis (28). Genetic studies have identified 20 common polymorphisms associated with DM2 (29). Take into account that these 20 loci explain only 6% of the DM2 heritability, and that the median odds ratio (OR) obtained for each described polymorphism is 1.14, it is not unexpected that some genetic studies fail to show association with this disease, even when it actually exists (28).

Even though the D2 Thr92Ala polymorphism was reported as being associated with lower glucose disposal rate in nondiabetic white subjects (9) and with a more pronounced IR in DM2 patients (10), the studies of Grarup et al. (11), Maia et al. (12), and Mentuccia et al. (13) failed to demonstrate an association between this polymorphism and IR or DM2. The study of Grarup et al. (11), in Danish subjects with (n = 1058) or without DM2 (n = 4770), was the only one of these three studies with an adequate power (β > 80%) to detect an OR of 1.14, the median OR obtained for most of the genetic polymorphisms associated with DM2. Maia et al. (12) analyzed 1631 participants from the Offspring Cohort of the Framingham Heart Study (Framingham, MA). DM2 comprised only 10.4% of this population, and the power calculation was based on an OR of 1.9, which seems to be an overestimation. If we consider a median OR of 1.14, the power of this study to detect an association between the Thr92Ala polymorphism and DM2 is <80%. Mentuccia et al. (13) studied 1,268 subjects of the Old Order of Amish (Amish, PA). Of these subjects, 1,103 individuals were considered as nondiabetic controls and 117 as DM2 patients, also having a power <80% to detect an OR of 1.14. In addition, the contradictory results on the association between the D2 Thr92Ala polymorphism and DM2 and/or IR can also be explained for different study designs or different genetic backgrounds of the analyzed populations.

PPARγ is member of a family of ligand-activated nuclear receptor and transcription factors, and it plays an important role in adipocyte differentiation, lipid storage, and insulin sensitivity (30). It is also the target molecule for thiazolidinediones, agents that enhance sensitivity to insulin in vivo (30). The common Pro12Ala polymorphism is the result of a CCA to GCA mutation in codon 12 of exon B, which encodes the amino-terminal domain defining the adipocyte-specific PPARγ2 isoform (14,16). The Pro12Ala polymorphism is a functional variant and exhibits in vitro reduced binding to DNA and modest impairment of transcriptional activity (15). The Ala12 minor allele is associated with a reduced risk of DM2 (14,16,31,32), and it has been associated with lower insulin concentrations, a crude indication of greater insulin sensitivity (32). However, a meta-analysis of 57 studies containing data related to IR of cohorts with normal or impaired glucose tolerance revealed no significant effect of the Pro12Ala polymorphism on diabetes-related traits across all studies (33). Statistically significant effects were only detectable in selected groups. For example, in the white subgroup, BMI was greater in individuals carrying the Pro/Ala-Ala/Ala genotypes compared with the Pro/Pro genotype, and HOMAIR index was significantly higher in the Pro/Pro individuals compared with Pro/Ala-Ala/Ala individuals of the obese subgroup of subjects, indicating greater insulin sensitivity in obese carriers of the Ala12 allele (33). Nevertheless, in our study, we did not observe any statistically significant association between the PPARγ2 Pro12Ala polymorphism and fasting plasma insulin levels, HOMAIR index, or other glycemic related variables.

Of interest, recently, Fiorito et al. (17), studying a health white population (Gargano, Italy), reported a significant interaction between the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms on the modulation of systolic and diastolic BP and metabolic syndrome. They also identified a putative peroxisome proliferator response element sequence at nucleotides −488 to −504 in the promoter region of human DIO2 gene, which, similarly to the canonical peroxisome proliferator response element motif, specifically binds to the PPARγ2/retinol X receptor-α complex, thus suggesting a potential role of PPARγ2 in the regulation of DIO2 gene expression. A possible interaction between these two polymorphisms is further supported by the present study: in the presence of the PPARγ2 Ala12 allele, the D2 Ala/Ala genotype was associated with an increase in HOMAIR index when compared to patients carrying other D2/PPARγ2 genotype combinations (Table 4 and Figure 1). It is noteworthy that Fiorito et al. (17) found no significant effect of the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms, isolated or in combination, on HOMAIR index. Nevertheless, these authors compared HOMAIR index between D2 Thr/Thr subjects vs. Thr/Ala-Ala/Ala subjects, and this could have diluted the effect of the D2 Ala/Ala genotype on IR, given that subjects carrying the D2 Thr/Ala genotype seems to have similar HOMA-IR index as compared to subjects with the Thr/Thr genotype (9,10).

The mechanism by which the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms interact is still not explained. But, since PPARγ2 is exclusively expressed in adipocytes, the mechanism through which the PPARγ2 Pro12Ala polymorphism interacts with the D2 Thr92Ala polymorphism, influencing the risk to IR, must originate in adipose tissue (34). It is worth mentioning that Grosovsky et al. found that pioglitazone treatment led to a 1.6–1.9-fold increase (5 or 10 µmol/l, respectively) in primary human skeletal myocyte D2 activity, even in the absence of insulin, sustaining the possibility that endogenous or exogenous PPARγ ligands could exert their metabolic effects in part by altering thyroid hormone signaling (35). This effect was also seen with other PPARγ agonists, including ciglitazone and troglitazone (≅1.6-fold increase for both). It remains to be determined whether the pioglitazone effect observed in this study is purely transcriptional or whether D2 activity is altered via a posttranscriptional mechanism, e.g., an effect on D2 ubiquitination (35).

Some factors unrelated to the D2 Thr92Ala and PPARγ2 polymorphisms could have interfered with the findings of the present study. First, medications for DM2 treatment could have played a role because some are known to affect insulin sensitivity. However, we minimized such a possibility by excluding insulin-treated subjects from the group of 246 patients analyzed for IR, and the proportion of patients on metformin and/or sulfonylureas was similar among different genotype groups. Furthermore, D2 Ala12 variant remained significantly associated with HOMAIR index in a linear regression model analysis that included use of medication for DM2 as an independent variable. In addition, none of these 246 patients were using thiazolidinediones drugs, agents that change insulin sensitivity. Second, IR was assessed by calculation of HOMAIR index rather than the reference method, the euglycemic-hyperinsulinemic clamp (19,36). Although HOMAIR index is only an estimate of IR, it is simple to perform and shows a good correlation with the reference method (19,36), the euglycemic clamp and, therefore, is considered a good approach for cohort and epidemiological studies (37). Third, we cannot rule out the possibility of stratification bias in our sample, even though we analyzed only self-defined white subjects, thus reducing the risk of false positive/negative associations due to this bias. In addition, the 246 DM2 patients analyzed for IR were representative of the whole sample (n = 721) with respect to age, DM2 duration, GHb, gender, BMI, and presence of arterial hypertension. The frequencies of the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms also were similar between these 246 DM2 patients and the whole sample (data not shown). Finally, our results could represent a type 1 error. However, the P value for HOMAIR index obtained in the interaction analysis between the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms remains significant at the 0.01 level. This finding, along with the scientific plausibility of an association, provides evidence against type 1 error.

In summary, the D2 Thr92Ala and PPARγ2 Pro12Ala polymorphisms might interact in the modulation of IR in white DM2 patients. Possibly due to low levels of D2 activity in adipocytes and, consequently, a marked reduced T3 biological effect, patients with both D2 Ala/Ala genotype and PPARγ2 Ala12 allele showed the worst IR phenotype. This gene-gene interaction could explain part of the conflicting results reporting associations of the D2 Thr92Ala or PPARγ2 Pro12Ala polymorphism with IR. On the other hand, we cannot exclude an effect of other variables such as dietary influence, different study designs or different genetic backgrounds of the studied populations on the results. Although further in vivo and in vitro studies are needed to evaluate the functional and epidemiological importance of this interaction, these findings are exciting from both biological and clinical perspectives due to their potential target therapeutic implications. One remarkable aspect is that most of the genetic loci so far associated with DM2 seem to affect insulin secretion but not insulin sensitivity, such is this case. In the context of the current pandemics of obesity and obesity related-IR, the identification of genetic polymorphism interactions that contribute to increase IR is an important step for a better understanding of the molecular mechanisms of this disease.

ACKNOWLEDGEMENT

This study was partially supported by grants from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), and Fundo de Incentivo à Pesquisa e Eventos (FIPE) at Hospital de Clínicas de Porto Alegre. The authors thank Prof Dr Luis Henrique Canani and Prof Dr Jorge Luiz Gross for sharing with us the clinical data regarding the DM2 patients.

DISCLOSURE

The authors declared no conflict of interest.

Ancillary