The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Interaction of DIO2 T92A and PPARγ2 P12A Polymorphisms in the Modulation of Metabolic Syndrome
Version of Record online: 6 SEP 2012
2007 North American Association for the Study of Obesity (NAASO)
Volume 15, Issue 12, pages 2889–2895, December 2007
How to Cite
Fiorito, M., Torrente, I., De Cosmo, S., Guida, V., Colosimo, A., Prudente, S., Flex, E., Menghini, R., Miccoli, R., Penno, G., Pellegrini, F., Tassi, V., Federici, M., Trischitta, V. and Dallapiccola, B. (2007), Interaction of DIO2 T92A and PPARγ2 P12A Polymorphisms in the Modulation of Metabolic Syndrome. Obesity, 15: 2889–2895. doi: 10.1038/oby.2007.343
- Issue online: 6 SEP 2012
- Version of Record online: 6 SEP 2012
- Received for review February 20, 2007, Accepted in final from April 25, 2007
- abdominal obesity;
- blood pressure;
- genetic susceptibility;
- insulin resistance;
- coronary artery disease
Type 2 iodothyronine deiodinase (DIO2) converts thyroid prohormone tetraiodothyronine into the biologically active triiodothyronine hormone, which increases insulin sensitivity at the skeletal muscle level. The DIO2 T92A polymorphism modulates deiodinase activity and has been inconsistently associated with insulin resistance. Also, the P121A polymorphism of the peroxisome proliferator-activated receptor (PPAR) γ2 gene, which encodes a transcription factor involved in insulin signaling, has been inconsistently associated with insulin resistance. This study was aimed at evaluating the combined effect of DIO2 T92A and PPARγ2 P12A polymorphisms on insulin resistance-related features in 590 non-diabetic whites. A significant gene-gene interaction was observed in the modulation of systolic (p = 0.01) and diastolic (p = 0.02) blood pressure and metabolic syndrome (p = 0.02), with carriers of both DIO2 A92 and PPARγ2 A12 variants showing the worst phenotype. This latter interaction was also shown by multifactor dimensionality reduction analysis (p = 0.0045). A peroxisome proliferator response element in the DIO2 promoter was identified by in silico analysis and confirmed by in vitro gel shift mobility assay, thus providing a biological plausibility for the observed gene-gene interaction. If confirmed in other populations, comprising several thousand individuals, these data may help identify individuals at risk for insulin resistance-related abnormalities.
Insulin resistance and the related metabolic syndrome (MS)1 impose a tremendous burden on health care systems (1, 2). Both environmental and genetic factors may be pathogenic for MS, with the latter being mostly unknown (3, 4). Thyroid hormones play a positive role in insulin action by up-regulating the expression of glucose transporter, Glut-4 (glucose transporter 4), in skeletal muscle (5). As a consequence, genes encoding modulators of thyroid hormone metabolism and effect are plausible candidates for susceptibility to MS and related abnormalities. By converting the prohormone tetraiodothyronine into the biologically active hormone triiodothyronine (T3), type 2 iodothyronine deiodinase (DIO2) is a key protein in the homeostatic system that controls intracellular thyroid hormone effects (6).
Several studies have investigated the role of the DIO2 missense T92A polymorphism on insulin resistance-related traits, but results have been conflicting (7, 8, 9). According to the polygenic model proposed for complex traits, many genes are likely to be simultaneously involved, interacting with each other, in modulating insulin resistance and related abnormalities (4). Among the several genes so far investigated, the P121A polymorphism of the peroxisome proliferator-activated receptor (PPAR) γ2 gene, which encodes a transcription factor involved in the regulation of adipocyte differentiation and intracellular insulin signaling (10, 11), has been recognized as playing a major role in determining the risk of insulin resistance-related phenotypes.
In the present study, we evaluated the combined effect of DIO2 T92A and PPARγ2 P12A polymorphisms on several features related to the MS in a white population from Gargano (Centre East Coast, Italy). Of the whole sample comprising 590 subjects (whose clinical features are reported in Table 1), 238 individuals (40.3%) were carrying the TT, 280 (47.5%) the TA, and 72 (12.2%) the AA DIO2 genotype. The percentages were, therefore, in Hardy-Weinberg equilibrium (HWE). The T92A polymorphism had no effect on insulin resistance and related variables when considering either a genetic additive model (i.e., by comparing TT vs. TA vs. AA individuals; data not shown), a recessive model (i.e., by comparing TT + TA vs. AA individuals, data not shown), or a dominant model [i.e., by comparing TT vs. TA + AA (designated XA) individuals, Table 2A].
|Age (yrs)||36.5 ± 11.8|
|BMI (kg/m2)||25.4 ± 4.4|
|Waist (cm)||82.7 ± 12.7|
|SBP (mm Hg)||114 ± 13|
|DBP (mm Hg)||76 ± 9|
|HDL-cholesterol (mg/dL)||53 ± 13|
|Triglycerides (mg/dL)*||73.0 (27 to 489)|
|Glucose (mg/dL)||89 ± 8|
|HOMAIR*||1.4 (0.3 to 8.7)|
|MS score||0.79 ± 0.9|
|DIO2 T92A||PPARγ2 P12A|
|Age (yrs)||36.2 ± 12||36.7 ± 12||36.6 ± 12||35.6 ± 12|
|BMI (kg/m2)||25.4 ± 4.7||25.3 ± 4.1||0.74||25.2 ± 4.3||26.3 ± 5.1||0.015|
|Waist (cm)||83.5 ± 14||82.1 ± 12||0.16||82.4 ± 13||84.2 ± 14||0.11|
|SBP (mm Hg)||113 ± 13||114 ± 13||0.37||114 ± 13||113 ± 14||0.49|
|DBP (mm Hg)||76 ± 9||76 ± 9||0.78||76 ± 9||76 ± 8||0.92|
|HDL-cholesterol (mg/dl)||53 ± 13||53 ± 13||0.79||53 ± 13||53.0 ± 14||0.80|
|Triglycerides (mg/dl)*||75 (31 to 489)||72 (27 to 307)||0.19||72 (27 to 489)||76 (35 to 307)||0.13|
|Glucose (mg/dl)||89 ± 8||90 ± 10||0.53||90 ± 9||89 ± 9||0.66|
|HOMAIR*||1.5 (0.3–7.9)||1.4 (0.3 to 8.7)||0.73||1.4 (0.3 to 8.7)||1.5 (0.4 to 5.4)||0.32|
|MS score||0.84 ± 1.0||0.75 ± 0.9||0.25||0.78 ± 0.9||0.87 ± 1.1||0.36|
We then stratified the whole population according to gender (men/women) or BMI or age (in both cases, above and below the median value). No difference across different strata was found in any variable tested, with the exception of triglyceride levels, which were lower in subjects carrying the XA genotype among younger, but not older, subjects (data not shown).
Of the whole sample, 521 individuals (88.3%) were carrying the PP, 68 (11.5%) the PA, and only one (0.2%) the AA PPARγ2 genotype. The percentages were, therefore, in HWE. With the exception of BMI, which turned out to be significantly higher in PA and AA individuals considered together and designated as XA, the P12A polymorphism had no effect on insulin resistance and typical related variables (Table 2B). Also, no association was observed between both gene polymorphisms and glomerular filtration rate or albumin-to-creatinine ratio (data not shown).
Among subjects with the PPARγ2 XA genotype, DIO2 XA carriers had a significantly higher systolic blood pressure (SBP) and diastolic blood pressure (DBP) than DIO2 TT individuals (Table 3A). Also, the MS score tended to be higher in DIO2 XA vs. TT individuals, although the difference was not significant with the present sample size (Table 3A). In contrast, among subjects with the PPARγ2 PP genotype, no differences were observed between DIO2 XA and TT individuals (Table 3B). In fact, the MS score tended to be lower, rather than higher, in XA vs. TT individuals, although not significantly with the present sample size (Table 3B). A significant interaction was observed between the DIO2 T92A and the PPARγ2 P12A polymorphism in modulating SBP (age- and gender-adjusted p value for interaction = 0.01), DBP (adjusted p value = 0.02), and the MS score (adjusted p value = 0.02). Of note, a recent paper (12) described an association of DIO2 T92A with hypertension in individuals from the U.S. and Europe, thus suggesting that the effect of this polymorphism on blood pressure may be a generalized phenomenon.
|PPARγ2 × 12A||PPARγ2 P12P|
|Age (yrs)||34.7 ± 11||36.2 ± 12.3||36.5 ± 12||36.7 ± 12|
|BMI (kg/m2)||25.6 ± 5.5||26.9 ± 4.7||0.42||25.4 ± 4.6||25.1 ± 4.1||0.46||0.2|
|Waist (cm)||83.0 ± 15||85.1 ± 13||0.71||83.6 ± 14||81.7 ± 12||0.09||0.3|
|SBP (mm Hg)||108.0 ± 12||116 ± 14||0.01||114 ± 13||114 ± 13||0.99||0.01|
|DBP (mm Hg)||72 ± 8||78 ± 8||0.02||77 ± 10||76 ± 9||0.28||0.02|
|HDL-cholesterol (mg/dl)||55 ± 15||51 ± 14||0.31||53 ± 13||53.3 ± 13.3||0.88||0.3|
|Triglycerides (mg/dl)*||76 (41 to 195)||76 (35 to 307)||0.89||75 (31 to 489)||71 (27 to 300)||0.14||0.5|
|Glucose (mg/dl)||89 ± 8||89 ± 10||0.92||89 ± 9||90 ± 10||0.54||0.9|
|HOMAIR*||1.7 (0.5 to 4.3)||1.4 (0.4 to 5.4)||0.61||1.4 (0.3 to 7.9)||1.4 (0.3 to 8.7)||0.87||0.57|
|MS score||0.61 ± 0.9||1.05 ± 1.2||0.13||0.87 ± 1.0||0.71 ± 0.90||0.06||0.02|
In addition, we also explored the effects of gene-gene interaction in determining the absence/presence of the MS, as defined according to the Adult Treatment Panel III (13), using the multifactor dimensionality reduction (MDR), a model-free, non-parametric method for detecting multilocus genotype combinations that predict disease risk for common, complex disease (14). The two-locus MDR model, comprising the DIO2 T92A and the PPARγ2 P12A polymorphism, had a high prediction accuracy (training accuracy = 0.615; testing accuracy = 0.617; p = 0.0045), thus suggesting an epistatic interaction between the two loci in determining the absence/presence of the MS.
To test for biological mechanisms underlying the observed interaction between DIO2 and PPARγ2, we checked for the presence of a peroxisome proliferator response element (PPRE) in the promoter region of the human DIO2 gene. In fact, we identified a putative PPRE sequence from nucleotides −488 to −504 in the DIO2 gene promoter, similar to the canonical PPRE motif (Figure 1) potentially able to bind heterodimers formed by PPARγ2 and its partner, the retinoid X receptor (RXR) α (15). Then, to determine whether the PPARγ2/RXRα complex, in fact, binds to the putative DIO2-PPRE, a gel mobility shift assay was performed. The labeled oligonucleotide containing the putative DIO2-PPRE motif was incubated with the PPARγ2/RXRα complex. As shown in Figure 2, mobility of the labeled probe was clearly shifted (lane 2), indicating the binding of the PPARγ2/RXRα complex to the putative site. The specificity of band shifting was indicated by band disappearance in the presence of an excess of unlabeled DIO2-PPRE oligonucleotides (Figure 2, lane 3) and the lack of band shifting when a non-PPRE sequence (GATA binding element) was used (Figure 2, lanes 8 and 9). As positive controls, both the non-canonical PPRE consensus located in the promoter of the human aquaporin adipose (AQPap) gene (16) (Figure 2, lanes 4 and 5) and the canonical PPRE consensus (Figure 2, lanes 6 and 7) were used.
This study indicates that the DIO2 T92A and PPARγ2 P12A polymorphisms interact in modulating some phenotypes related to insulin resistance, namely blood pressure and the MS score. Although we acknowledge that the sample size of subgroups carrying the different genotypes is quite small and does not allow firm conclusions to be drawn, data we report here are in line with the proposed scenario of a polygenic control of insulin resistance and related abnormalities. Further studies on several thousand individuals from different geographic regions are certainly needed to confirm our present data. Compatible with the gene-gene interaction we observed in the association study is our finding of a putative PPRE sequence at nucleotide −488 to −504 in the promoter region of human DIO2 gene, which, similarly to the canonical (AGGTCA N AGGTCA), specifically binds to the PPARγ2/RXRα complex, thus suggesting that DIO2 gene transcription and expression may be mediated by PPARγ2. Several other in vitro experiments (including the stimulation with PPARγ2 synthetic agonists of cells endogenously expressing DIO2) are needed to confirm that the nucleotide sequence we described is, in fact, a PPRE with a functional role on the modulation of gene transcription. The DIO2 A92 variant has been reported to be biologically less active than the common T92 variant (8). It has been hypothesized that, due to a decreased deiodinase activity, carriers of the DIO2 A92 variant are characterized by lower T3 effect on target tissue, which, in turn, reduces the expression of the Glut-4 (glucose transporter 4) glucose transporter (17, 18). In fact, in our sample, the deleterious effect on insulin resistance of the DIO2 A92 variant (which has a reduced deiodinase activity) is appreciated only in individuals carrying also the PPARγ2 A12 variant, which has a reduced transacting activity (Table 3A). Therefore, subjects carrying both the DIO2 A92 and the PPARγ2 A12 variants might be those with very low levels of both DIO2 tissue expression and activity and, consequently, a marked reduced T3 biological effect.
In conclusion, in the white population of the Gargano area, the DIO2 T92A and the PPARγ2 P12A polymorphisms interact in modulation features related to the MS. Further studies in larger samples comprising several thousand individuals from different geographic regions possibly investigating also the concomitant role of other potential genes are likely to contribute to a better understanding of the role of the DIO2 gene in contributing to the polygenic control of this metabolic abnormality.
Research Methods and Procedures
The sample studied was recruited in the Gargano area (Centre East Coast Italy) as part of an ongoing project on the genetics of insulin resistance. Selection criteria were as follows: fasting plasma glucose < 6.1 mM, no medications known to affect glucose and lipid metabolism, and absence of systemic diseases. Enrolled subjects (all white) underwent physical examination including measurement of height and weight. Blood pressure was measured twice with subjects in the sitting position, after at least 5 minutes of rest, using an appropriately sized cuff. DBP was recorded at the disappearance of Korotkoff sound (Phase V). Venous blood of all fasting subjects was sampled from an antecubital vein for the measurements of glucose, insulin, total and high-density lipoprotein (HDL)-cholesterol, triglycerides, and creatinine, as elsewhere described (18). The homeostasis model assessment of insulin resistance (HOMAIR) index was calculated as fasting serum insulin (milliunits per milliliter)/fasting plasma glucose (millimolar)/22.5, as elsewhere described (19). Glomerular filtration rate was estimated by serum creatinine with the abbreviated formula of the Modification Diet and Renal Disease: [186 × (serum creatinine)−1.154 × (age)−0.203 × (0.742 if a woman] (19). Urinary albumin excretion was measured on morning sterile urine and reported as albumin-to-creatinine ratio (milligrams per millimoles) (19). Clinical features of the MS were considered according to Adult Treatment Panel III suggestions (13). Based on the number of criteria present in each subject, an individual MS-related score was assigned ranging from 0 to 5 (13). Individual with scores of 3 to 5 were considered affected with MS (13). The study was performed according to the Helsinki declaration, and the protocol was approved by the local ethics committee. All subjects provided written informed consent.
Genomic DNA was extracted from peripheral blood according to standard procedures. The T92A DIO2 polymorphism was screened with the ABI PRISM SNaPshot Multiplex Kit (4323151; Applied Biosystems, Foster City, CA), based on the dideoxy single-base extension of an unlabeled primer. The polymerase chain reaction product (255 base pairs) was amplified using the sense primer (5′-CTCAGGGCTGGCAAAGTCAAG) and the antisense primer (5′-CCACACTCTATTAGAGCCAATTG) and purified by digestion with ExoI and shrimp alkaline phosphatase enzymes. The primer extension reaction was performed with fluorescently labeled dideoxy-nucleotides and the primer, 5′-TACCATTGCCACTGTTGTCACCTCCTTCTG. The primer-extension reaction product was electrophoresed on an ABI PRISM3100 Genetic Analyzer after purification by digestion with a calf intestine alkaline phosphatase enzyme. Data were analyzed with GeneScan Analysis Software, version 3.1, and GeneScan-120 LIZ size standard analysis parameter (Applied Biosystems). The PPARγ2 P12A polymorphism was screened by SnaPshot analysis with dideoxy single-base extension of an unlabeled primer, according to the manufacturer's instructions (ABI PRISM) (SNaPshot Multiplex Kit 4323151) as previously described (20).
Electrophoretic Mobility Shift Assays (EMSAs)
EMSA was performed as described previously, with slight modifications (21). The full lengths of PPARγ2 (NM_015869) and RXRα (NM_002957) genes were subcloned into two different pcDNA3.1 vectors (Invitrogen), and the corresponding proteins were synthesized in vitro using the TNT T7 Quick Coupled Transcription/Translation System (Promega, Madison, WI). The translation products were checked by electrophoresis on a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis. A double-strand oligonucleotide spanning from nucleotides −488 to −523 of the human deiodinase type 2 promoter was 32P-radiolabeled using tetraiodothyronine polynucleotide kinase (Promega) and [γ-32P]ATP. The in vitro translation product (2.5 µL) was incubated in 20 µL of 1× buffer containing 20 mM Tris-HCl (pH 7.6), 75 mM KCl, bovine serum albumin (2 µg), glycerol (10%), and poly dI-dC (0.3 µg) for 10 minutes at room temperature with or without 25 pmol of unlabeled oligonucleotides (competitors). Subsequently, 54 cpm of labeled oligonucleotides were added to both reaction mixes for 30 minutes at room temperature. The DNA-protein complexes were resolved from the free probe by electrophoresis on a 4% polyacrylamide gel in 0.5× Tris/borate/EDTA buffer. The gels were dried and autoradiographed at −80 °C. The oligonucleotides used in the gel mobility assay were as follows: DIO2-PPRE, 5′-GA-CTCAAAAAAATTTAAAGGACCTAGGTCACAGAT-3′; AQP-PPRE (PPRE of AQPap gene), 5′-GCTGCTCCTGCTCCTCCAGGGGAGAGGTCAGTAAG-3′; PPRE (PPRE consensus), 5′-AACTAGGTCAAAGGTCA-3′; and GATA, 5′-CACTTCTTAACAGCAGAAAGTCTTAACTCT-3′ (Figure 1).
Values are mean ± standard deviation or median (range). Comparisons between groups were tested by unpaired Student's t test or Mann-Whitney (as appropriate) and among groups by one-way ANOVA. Mean values, after adjusting for covariates, were evaluated by analysis of covariance. HWE was evaluated by χ2 test. Interaction between gene polymorphisms in modulating different variables was tested by general linear model univariate analysis. Variables not normally distributed, such as triglycerides, were logarithmically transformed before analysis. A p value < 0.05 was considered significant. All analyses were performed by the SPSS/PC statistical program (version 12.0 for Windows; SPSS, Inc., Chicago, IL).
MDR analysis includes a combined cross-validation/permutation testing procedure. Cross-validation divides the data into a training set and a testing set. With 10-fold cross-validation, the data are divided into 10 equal parts, and the model is developed on 9 of the 10 parts (training set) and then tested on the 1 remaining part (testing set). This is repeated for each possible 9 of 10 and 1 of 10 data parts, and the resulting 10 testing accuracies are averaged. Models that are true-positives are likely to generalize to independent data sets and will have estimated testing accuracies of >0.5. Permutation testing was performed to assess the probability of obtaining a testing accuracy as large as or larger than that observed in the original data given the null hypothesis of no association being true. This is carried out by randomizing the case-control labels 1000 times and repeating the MDR analysis on each randomized data set. This process yields an empirical distribution of testing accuracies under the null hypothesis that is, in turn, used to calculate a p value. The MDR analysis was carried out using version 1.0.0 of the open-source MDR software package (available free at http:www.epistasis.org).
The authors acknowledge the financial support of the Italian Ministry of Health (Ricerca Corrente 2004).
Nonstandard abbreviations: MS, metabolic syndrome; T3, triiodothyronine; DIO2, type 2 iodothyronine deiodinase; PPAR, peroxisome proliferator-activated receptor; HWE, Hardy-Weinberg equilibrium; SBP, systolic blood pressure; DBP, diastolic blood pressure; MDR, multifactor dimensionality reduction; PPRE, peroxisome proliferator response element; RXR, retinoid X receptor; AQPap, aquaporin adipose; HDL, high-density lipoprotein; HOMAIR, homeostasis model assessment of insulin resistance; EMSA, electrophoretic mobility shift assay.
- 13Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults (2001) Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP). Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA. 285: 2486–2497.