Association of the Pro12Ala Polymorphism in the PPAR-γ2 Gene With 3-Year Incidence of Type 2 Diabetes and Body Weight Change in the Finnish Diabetes Prevention Study

The study population, inclusion and exclusion criteria, and power calculation have been described earlier in detail (3,22). In short, the DPS is a multicenter study with five participating centers in Finland. The diagnosis of diabetes and other categories of glucose intolerance were based on the criteria adopted by the World Health Organization in 1985 (23). IGT was defined as fasting plasma glucose <7.8 mmol/l and a 2-h plasma glucose 7.8–11.0 mmol/l (oral glucose tolerance test [OGTT] 75 g). Altogether 522 overweight subjects (BMI 31.1 ± 4.6 kg/m²) aged 40–68 years and with IGT were randomly allocated to one of the two treatment modalities, the intensive diet and exercise intervention group or the control group. The screening for the Pro12Ala polymorphism of the PPAR-γ2 gene was performed in 490 (161 men and 329 women) subjects.

The intervention program has been described previously (3,22). Briefly, subjects in the intervention group were given individually tailored dietary advice aimed at reducing weight and the intake of total and saturated fat and increasing the intake of dietary fiber. Furthermore, subjects in the intervention group were individually guided to increase their level of physical activity. The control group received general information on the benefits of weight reduction, physical activity, and healthy diet.

The study protocol was approved by the ethics committee of the National Public Health Institute in Helsinki, Finland. All subjects gave written informed consent.

A medical history and a physical examination were done on a yearly basis. In this report, measurements at baseline and at the 3-year examination were used. Waist circumference was measured midway between the lowest rib and iliac crest and hip circumference over the great trochanters, with 0.5-cm precision with the subjects in a standing position. BMI was calculated as weight (kg) divided by height (m) squared (2). The relative weight change during the 3-year follow-up was calculated as follows: ([weight_{3-year follow-up} − weight_baseline]/weight_baseline) × 100. In a 2-h OGTT, samples for glucose and insulin were taken before (0 min) and 120 min after a glucose load (75 g). Plasma glucose was measured at each center by standard methods, and the measurements were standardized by the central laboratory in Helsinki (3). Serum insulin was determined with a radioimmunoassay (Pharmacia, Uppsala, Sweden). The homeostasis model assessment for insulin resistance (HOMA-IR) was calculated using the following formula: fasting plasma glucose (mol/l) × fasting serum insulin (μU/l)/22.5.

The Pro12Ala polymorphism of the PPAR-γ2 gene was determined by the PCR-SSCP (single-strand conformation polymorphism) method as previously reported (17).

The significance of differences in allele frequencies was analyzed using two-tailed Fisher’s exact test in the StatXact-4 program version 4.0.1 (Cytel Software Corporation, Cambridge, MA). Otherwise, the data were analyzed using the SPSS/WIN program version 9.0 (SPSS, Chicago). χ² analysis was applied to test for the significance of differences in sex distribution among genotypes. Normal distribution of variables was checked with the Kolmogorov-Smirnov (Lilliefors) test, and logarithmic transformation was used for those not normally distributed. The differences among the genotypes were evaluated by the univariate ANOVA (general linear model [GLM]) with adjustment for age, sex, and BMI, when appropriate. The differences in the 3-year weight change among genotypes were assessed by GLM with glucose tolerance status at the 3-year examination (IGT versus type 2 diabetes) as a covariate. Logistic regression analysis was used to assess whether the Ala12 allele predicted the development of type 2 diabetes. In the logistic regression analysis, genotypes were encoded as 0 = Pro12Pro and 1 = Pro12Ala or Ala12Ala and study groups were encoded as 1 = intervention group and 2 = control group. Weight change used in the logistic regression analysis was calculated as weight (kg)_{3 year} − weight (kg)_baseline. In subjects whose diabetes was diagnosed before the 3-year examination, the baseline weight was subtracted from the weight measured at the visit when the diabetes diagnosis was done. In addition, age and sex were considered in the logistic regression model together with genotype, study group, and weight change, but they were not significantly related to the risk of type 2 diabetes (data not shown), and therefore they were omitted from the final logistic regression model. A P value <0.05 was considered statistically significant. Data are presented as means ± SD, unless otherwise indicated.

In the intervention group, the Pro12Pro genotype was found in 163 subjects, the Pro12Ala genotype was found in 79 subjects, and 6 subjects were homozygous for the Ala12 allele. In the control group, 174 subjects had the Pro12Pro genotype, 61 had the Pro12Ala genotype, and 7 had the Ala12Ala genotype. Observed genotype frequencies of the polymorphism were in agreement with Hardy-Weinberg expectations. The frequency of the Ala12 allele did not differ between the intervention group and the control group (0.183 vs. 0.155, respectively, P = 0.268).

At baseline, the Pro12Ala polymorphism was not associated with body weight, BMI, waist and hip circumferences, waist-to-hip ratio, HOMA-IR, 2-h plasma glucose, or fasting and 2-h serum insulin concentrations (Table 1). However, subjects with the Ala12Ala genotype tended to have higher fasting plasma glucose concentrations than those with other genotypes.

By 3 years of intervention, 19 subjects in the intervention group and 50 subjects in the control group had developed type 2 diabetes (Fig. 1B). When the association between the Pro12Ala polymorphism and the incidence of type 2 diabetes was considered in the logistic regression model, the odds of the development of type 2 diabetes for those with the Ala12 allele was over twofold compared with the odds of the development of type 2 diabetes for subjects with the Pro12Pro genotype (Table 2). Similarly, the odds of the development of type 2 diabetes was over twofold for subjects in the control group compared with subjects in the intervention group. A higher weight at the baseline examination as well as weight gain during the follow-up were also significant predictors for the development of type 2 diabetes (Table 2, Model 3).

The logistic regression analysis was also carried out separately in the intervention group and in the control group. When the weight change during the follow-up was included in the model, in the control group the odds of the development of type 2 diabetes for those with the Ala12 allele was 2.36-fold compared with the odds of the development of type 2 diabetes for subjects with the Pro12Pro genotype (95% CI 1.21–4.60). In the intervention group, the Ala12 allele was not significantly associated with an increased risk of type 2 diabetes, even though the risk was almost twice as high as in those with the Pro12Pro genotype (odds ratio = 1.90, 95% CI 0.70–5.18).

Next we studied whether the association of the Ala12 allele with type 2 diabetes risk was modified by the degree of obesity. Therefore, DPS study subjects were divided into the two groups according to median of the baseline BMI. The first group, a lower BMI group (LBMI, mean BMI 27.7 ± 1.6 kg/m², n = 244), consisted of subjects whose baseline BMI was below or equal to the median of the whole study group baseline BMI. The second group, a higher BMI group (HBMI, mean BMI 34.7 ± 4.0 kg/m², n = 246), consisted of subjects whose baseline BMI was over the median of the whole study group baseline BMI. The Ala12 allele was a significant predictor of the development of type 2 diabetes in the LBMI group only (Table 3).

Finally, the influence of the Pro12Ala polymorphism on 3-year weight change was examined. In the intervention group, subjects with the Ala12Ala genotype lost more weight than subjects with the Pro12Pro genotype (−8.3 ± 7.3% vs. −3.4 ± 5.7%, P = 0.043) and the weight loss in subjects with the Pro12Ala genotype (−4.0 ± 6.0%) was close to that in subjects with the Pro12Pro genotype (Fig. 1A). In contrast, in the control group, 3-year weight changes were relatively small and the Pro12Ala polymorphism was not associated with the weight changes (−1.2 ± 6.1% for the Pro12Pro genotype, 0.2 ± 6.9% for the Pro12Ala genotype, and 1.0 ± 6.9% for the Ala12Ala genotype, P = 0.733) (Fig. 1A).

The present study is the first to examine the association of the Pro12Ala polymorphism of the PPAR-γ2 gene with the incidence of type 2 diabetes in a high-risk IGT population in a longitudinal study design. Interestingly, we found that in the IGT subjects, the Ala12 allele may predispose to the development of type 2 diabetes. In line with our study, the Ala12 allele has been reported to associate with type 2 diabetes in Canadian Oji-Cree women (24). Similarly, Evans et al. (25) found an association between the Ala12 allele and type 2 diabetes. Hasstedt et al. (26) studied members of familial type 2 diabetic kindreds and reported that the proportion of individuals with type 2 diabetes increased with the number of Ala12 alleles, although the result was not statistically significant.

The Ala12 allele has been associated with high insulin sensitivity (12,27,28), but there are also studies in which this association has not been found (13,26,29). The PPAR-γ gene is also expressed in pancreatic β-cells (30), and therefore the Ala12 allele could interfere with β-cell function. Indeed, type 2 diabetic patients with the Ala12 allele might have a reduced capacity for insulin secretion (31). In addition, Stefan et al. (29) reported that in healthy normal weight carriers of the Ala12 allele, second-phase insulin secretion after the experimental elevation of serum free fatty acids was decreased compared with that in subjects with the Pro12Pro genotype. In the present study, the Ala12 allele was associated with the type 2 diabetes risk particularly in less obese IGT subjects, in whom the reason for IGT could be related to disturbances in the insulin secretion rather than impaired insulin action. Thus, it is tempting to speculate that the Ala12 allele may predispose to insulin deficiency in the high-risk IGT population and therefore contribute to the development of type 2 diabetes.

In the present study, the Ala12 allele was not significantly associated with the incidence of type 2 diabetes in the intervention group, but a small number of cases might explain this. It has also been proposed that PPAR-γ2 is one of the mediators of the gene-environment interactions (32) and, likewise, it could be possible that the beneficial changes in diet, increased physical activity, and concomitant weight loss, as well as some gene-nutrient and gene-exercise interactions, protected subjects with the Ala12 allele in the intervention group from the development of type 2 diabetes in the present study. For example, subjects in the intervention group reduced the intake of total and saturated fat and increased the level of physical activity (3), the factors known to improve insulin sensitivity (33,34). Therefore, we suggest that weight reduction together with other beneficial lifestyle changes improved insulin sensitivity and reduced the β-cell stress and, consequently, alleviated the deleterious effect of the Ala12 allele on the development of type 2 diabetes in the intervention group. In accordance with this presumption, there is evidence from the Troglitazone in Prevention of Diabetes Study that an induced β-cell rest may protect from the development of type 2 diabetes in high-risk individuals (35). In addition, Nicklas et al. (36) found that postmenopausal women with the Ala12 allele had a greater increase in insulin sensitivity and fasting carbohydrate oxidation after a 6-month hypocaloric diet compared with women with the Pro12Pro genotype in spite of similar amounts of weight loss. However, women with the Ala12 allele had a greater decrease in fasting lipid oxidation, and they regained more weight during 1-year follow-up compared with women with the Pro12Pro genotype (36). This may imply that although subjects with the Ala12 allele may in some circumstances lose weight more easily during active weight reduction program, they may regain the weight more easily after treatment as well.

In contrast to our results, the Ala12 allele has also been supposed to be slightly protective against type 2 diabetes (10,12, 21,27,31). The reason for the discrepancy between these studies and the present study may partly be related to differences in study designs. All earlier studies have been cross-sectional, and the frequency of the Ala12 allele has been compared between the groups of type 2 diabetic patients and subjects with normal glucose tolerance. In the present study, the incidence of type 2 diabetes was evaluated after the 3-year follow-up in an IGT study population known to be at an increased risk for type 2 diabetes. Thus, our study population does not represent the general population, and our results can be generalized only to those subjects with the IGT. Several of the studies investigating the relationship between the Pro12Ala polymorphism and type 2 diabetes have been carried out in Japanese Americans (12) or in Japanese subjects (27,31). In the Japanese population, the frequency of the Ala12 allele is over four times lower than in the Finnish population (12,17, 21), and this complicates the comparisons of the results obtained in these two populations. Moreover, as discussed above, the degree of obesity, the composition of diet, and the level of physical activity may modify the effect of the Ala12 allele on the risk of type 2 diabetes, and differences in these lifestyle factors between the present study and previous studies may also partly explain the divergent results.

Although the size of the DPS population was sufficient for the original intervention study (3), the number of subjects with the Ala12Ala genotype was only moderate for the association study described here. Thus, carriers of the Ala12Ala genotype were combined with subjects with the Pro12Ala genotype in the logistic regression analysis in the present study. In other words, the logistic regression model utilized in this study assumed dominant effect of the Ala12 allele over the Pro12 allele. However, it would be interesting to investigate the association between the Ala12 allele and the risk of type 2 diabetes in other large prospective studies using a statistical model supposing also a recessive mode of inheritance for the Ala12 allele.

In summary, in the present study, the Ala allele in codon 12 of the PPAR-γ2 gene was associated with the development of type 2 diabetes in the high-risk IGT population. However, among those who followed an intensive diet and exercise program, subjects with the Ala12Ala genotype lost more weight than subjects having the Pro12Pro genotype, and none of subjects with the Ala12Ala genotype developed type 2 diabetes in this group. Therefore, the weight loss, together with other beneficial lifestyle changes, might reverse to some extent the diabetogenic impact of the Ala12 allele. Because our study is the first longitudinal study investigating the association between the Ala12 allele and the risk of type 2 diabetes, there is a need to confirm the present findings in other prospective studies.

This study was supported by grants from the Finnish Academy (nos. 38387 and 46558 to J.T.; no. 40758 to M.U.), the EVO-fund of the Kuopio University Hospital (no. 5106 to M.U.), the Ministry of Education, the Finnish Diabetes Research Foundation, and the European Union (no. QLRT 1999-00674 to M.L.).

The authors thank Pirjo Halonen, MSc, for her expert guidance in statistical analyses and Minna Kiuttu for her skilful technical assistance.