Defective insulin secretion is a key feature in the

Transcription

1 Brief Genetics Report olymorphisms in the SLC2A2 (GLUT2) Gene Are Associated With the Conversion From Impaired Glucose Tolerance to Type 2 Diabetes The Finnish Diabetes revention Study Olli Laukkanen, 1 Jaana Lindström, 2 Johan Eriksson, 2 Timo T. Valle, 2 Helena Hämäläinen, 3 irjo Ilanne-arikka, 4 Sirkka Keinänen-Kiukaanniemi, 5,6 Jaakko Tuomilehto, 2,7 Matti Uusitupa, 8 and Markku Laakso 1 Impaired insulin secretion is a fundamental defect in type 2. The aim of this study was to investigate whether single nucleotide polymorphisms (SNs) in the genes regulating insulin secretion (SLC2A2 [encoding GLUT2], GCK, TCF1 [encoding HNF-1 ], HNF4A, GI, and GL1R) are associated with the conversion from impaired glucose tolerance (IGT) to type 2 in participants of the Finnish Diabetes revention Study. With the exception of SLC2A2, other genes were not associated with the risk of type 2. All four SNs of SLC2A2 predicted the conversion to, and rs5393 (AA genotype) increased the risk of type 2 in the entire study population by threefold (odds ratio 3.04, 95% CI , 0.008). The risk for type 2 in the AA genotype carriers was increased in the control group (5.56 [ ], 0.003) but not in the intervention group. We conclude that the SNs of SLC2A2 predict the conversion to in obese subjects with IGT. Diabetes 54: , 2005 From the 1 Department of Medicine, University of Kuopio, Kuopio, Finland; the 2 Department of Epidemiology and Health romotion, Diabetes and Genetic Epidemiology Unit, National ublic Health Institute, Helsinki, Finland; the 3 Research Department, Social Insurance Institution, Turku, Finland; the 4 Finnish Diabetes Association and Department of Internal Medicine, Tampere University Hospital, Tampere, Finland; the 5 Department of ublic Health Science and General ractice, University of Oulu, Oulu University Hospital, Oulu, Finland; the 6 Department of Sport Medicine, Oulu Deaconess Institute, Oulu, Finland; the 7 Department of ublic Health, University of Helsinki, Helsinki, Finland; and the 8 Department of Clinical Nutrition, University of Kuopio, Kuopio, Finland. Address correspondence and reprint requests to rof. Markku Laakso, MD, Department of Medicine, University of Kuopio, Kuopio, Finland. markku.laakso@kuh.fi. Received for publication 9 February 2005 and accepted in revised form 18 April DS, Diabetes revention Study; GI, glucose-dependent insulinotropic polypeptide; GL-1, glucagon-like peptide 1; HNF-1, hepatocyte nuclear factor 1 ; IGT, impaired glucose tolerance; SN, single nucleotide polymorphism by the American Diabetes Association. 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. Defective insulin secretion is a key feature in the pathogenesis of type 2 in addition to insulin resistance. Therefore, variants in the genes regulating insulin secretion are plausible candidate genes for type 2. In the present study, single nucleotide polymorphisms (SNs) in five genes regulating -cell function were evaluated as risk factors for type 2. GLUT2 is a high-capacity facilitative glucose transporter expressed in liver, kidney, intestine, and pancreatic -cells (1). In the pancreas, GLUT2 regulates entry of glucose into the pancreatic cell and thus insulin secretion. SNs of the SLC2A2 gene (encoding GLUT2) have been associated with the risk of in some (2,3) but not in most (4 11) of the studies. Glucokinase phosphorylates glucose to glucose-6-phosphate and plays a key role in the regulation of insulin secretion (12). The G-30A polymorphism in the pancreatic -cell specific promoter of the glucokinase (GCK) gene has been associated with reduced -cell function (13), impaired glucose tolerance (14), and type 2 (15). Hepatocyte nuclear factor 1 (HNF-1 ) is an important transactivating factor that is involved in pancreatic -cell glucose sensing because it increases transcription of many genes participating in the insulin secretion process. The G319S polymorphism of the TCF1 gene, encoding HNF-1, has been associated with type 2 in Oji-Cree (16). However, there is no definite evidence that TCF1 is a major contributor to type 2. HNF4A constitutes a part of a network of transcription factors, including HNF- 1, controlling gene expression in pancreatic -cells and liver. Recent studies have indicated that SNs in HNF4A are associated with the risk of type 2 (17 19) Incretins, glucagon-like polypeptide-1 (GL-1), and glucose-dependent insulinotropic polypeptide (GI) are substances that are responsible for fast insulin response to ingested glucose (20). In patients with type 2, the secretion of GL-1 from intestinal mucosa and the responsiveness of the pancreatic -cells to GI are diminished DIABETES, VOL. 54, JULY 2005

2 O. LAUKKANEN AND ASSOCIATES TABLE 1 Subjects with the specific risk genotype versus subjects without the risk genotype who converted to type 2 Total population (n 479) Intervention group (n 241) Control group (n 238) SLC2A2 rs5393 AA vs. C allele 64 vs. 8 (17.0 vs. 7.8) vs. 4 (8.9 vs. 8.2) vs. 4 (25.5 vs. 7.4) rs5394 CC vs. T allele 64 vs. 8 (16.5 vs. 8.8) vs. 4 (8.5 vs. 9.5) vs. 4 (24.9 vs. 8.2) rs5400 (T110I) CC vs. T allele 61 vs. 11 (17.1 vs. 9.0) vs. 4 (9.3 vs. 6.9) vs. 7 (25.3 vs. 10.9) rs5404 (T198T) GG vs. A allele 64 vs. 8 (16.5 vs. 8.8) vs. 4 (8.5 vs. 9.8) vs. 4 (25.0 vs. 8.0) GCK rs (G-30A) GG vs. A allele 56 vs. 13 (16.0 vs. 11.0) vs. 3 (9.0 vs. 5.3) vs. 10 (23.1 vs. 16.4) TCF1 rs (I27L) AA vs. C allele 27 vs. 45 (14.0 vs. 15.7) vs. 12 (9.4 vs. 8.3) vs. 33 (18.6 vs. 23.4) rs (S487N) GG vs. A allele 35 vs. 37 (14.7 vs. 15.4) vs. 10 (9.4 vs. 8.1) vs. 27 (19.8 vs. 23.1) HNF4A rs CC vs. T allele 51 vs. 21 (16.5 vs. 12.4) vs. 5 (10.8 vs. 5.4) vs. 16 (21.7 vs. 20.8) rs GG vs. A allele 51 vs. 21 (16.6 vs. 12.3) vs. 5 (11.0 vs. 5.3) vs. 16 (21.6 vs. 21.1) rs GG vs. A allele 52 vs. 20 (14.6 vs. 16.1) vs. 6 (8.6 vs. 9.0) vs. 14 (20.4 vs. 24.6) GI rs (G103S) CC vs. T allele 20 vs. 52 (15.7 vs. 14.8) vs. 13 (12.7 vs. 7.3) vs. 39 (18.8 vs ) GL-1R rs (G168S) GG vs. A allele 38 vs. 34 (15.9 vs. 14.2) vs. 10 (8.4 vs. 9.1) vs. 24 (25.0 vs. 18.5) rs (L260F) CC vs. A allele 21 vs. 51 (15.4 vs. 14.9) vs. 18 (4.5 vs. 10.3) vs. 33 (25.7 vs. 19.6) Data are n (%), unless otherwise indicated. No data on the association of SNs of incretin hormones, or their receptor genes, with the risk of type 2 are available. In this study, we investigated whether SNs in the SLC2A2, GCK, TCF1, HNF4A, GI, and GL-1 receptor (GL-1R) genes predict the conversion from impaired glucose tolerance (IGT) to type 2 in participants of the Finnish Diabetes revention Study (DS). RESEARCH DESIGN AND METHODS The main objective of the Finnish DS was to investigate whether lifestyle intervention influences the conversion to type 2 during a 3-year follow-up (21). A total of 522 middle-aged (mean age 55 years) and overweight (BMI 25kg/m 2 ) Finnish subjects with IGT (22) participated in the study and were randomized to an intervention or control group. The intervention group was given intensive and individualized nutritional counseling as well as individual advice to increase physical activity and to reduce weight, whereas the control group was given only general advice. An oral glucose tolerance test was performed at each annual follow-up visit, and the diagnosis of was confirmed with a second test. The study protocol was approved by the ethics committee of the National ublic Health Institute in Helsinki, and all study subjects provided written informed consent. Weight change was calculated from the baseline value to the last weight measurement available, which varied from 1 to 3 years based on a new diagnosis of before the 3-year follow-up visit. For those who did not convert to, weight change was calculated as a difference in weight between baseline and 3 years. DNA analysis. DNA sample was available from 507 subjects (intervention group 259 and control group 248 subjects). SNs of SLC2A2 (promoter SNs rs5393 and rs5394 and exon SNs rs5400 [T110I] and rs5404 [T198T]), TCF1 (rs and rs ), HNF4A (rs , rs , and rs ), GI (rs ), and GL-1R (rs and rs ) were genotyped using the TaqMan Allelic Discrimination Assays (Applied Biosystems). Genotyping reaction was amplified on a GeneAmp CR system 2700 (95 C for 10 min, followed by 40 cycles of 95 C 15 s and 60 C 1 min), and fluorescence was detected on an ABI rism 7000 sequence detector (Applied Biosystems). SN rs (G-30A) of GCK was genotyped with restriction fragment length polymorphism method using Alw21I restriction enzyme. The primer and probe sequences are available from the authors by request. Statistical analysis. All data were analyzed with the SSS/Win programs (version 10.0, SSS, Chicago, IL). Results are given as means SD or percentages. Variables, which were not normally distributed, were logarithmically transformed before statistical analyses. ANOVA was used to compare three groups, and Student s t test was used for independent samples or 2 test to compare two groups. Nonparametric Mann-Whitney U test and Kruskal- Wallis H test were applied to compare changes in weight and glucose levels (%). Logistic regression analysis was performed to evaluate whether the SNs investigated predicted the conversion to type 2. RESULTS The genotype distributions of SLC2A2, GCK, TCF1, HNF4A, GI, and GL-1R did not differ between the intervention and control groups, and they were in Hardy- Weinberg equilibrium. Altogether, 72 of 479 subjects whose DNA and follow-up data were available converted from IGT to type 2 during the 3-year follow-up. SNs rs5393 and rs5400 of SLC2A2 were associated with the conversion to type 2 DIABETES, VOL. 54, JULY

3 GENES REGULATING INSULIN AND TYE 2 DIABETES TABLE 2 SNs in SLC2A2 as predictors for the development of type 2 (logistic regression analysis) OR Total population Intervention group Control group 95% CI OR 95% CI OR 95% CI rs5393 (AA vs. C allele)* Model 1 (univariate) Model 2 (multivariate) rs5394 (CC vs. T allele)* Model 1 (univariate) Model 2 (multivariate) rs5400 (CC vs. T allele)* Model 1 (univariate) Model 2 (multivariate) rs5404 (GG vs. A allele)* Model 1 (univariate) Model 2 (multivariate) *rs5393 genotypes were coded as 0 the C allele, 1 the AA genotype; rs5394 genotypes 0 T allele, 1 CC genotype; rs5400 genotypes 0 T allele, 1 CC genotype; rs5404 genotypes 0 A allele, 1 GG genotype. Adjusted for age, sex, weight at baseline, and weight change. in the entire study population ( and 0.031, respectively) (Table 1). In the control group, all four SNs of SLC2A2 predicted the development of type 2 ( values from to 0.017). In contrast, no differences were found in the conversion to type 2 between the genotypes in the intervention group. SNs of other genes investigated or their haplotypes (HNF4A, data not shown) did not predict the conversion to type 2. Logistic regression analyses showed that the risk genotypes of the four SNs of SLC2A2 were associated with a two- to threefold risk of developing type 2 in the entire study population and a four- to fivefold risk in the control group (Table 2). Moreover, subjects carrying the risk genotypes of SLC2A2 in the control group had increased plasma glucose levels both at fasting and at 120 min in an oral glucose tolerance test. The difference was most prominent for rs5404 (fasting: GG % vs. A allele %, 0.018; 120 min: GG % vs. A allele %, 0.015). No differences in insulin levels, insulin secretion (homeostasis model assessment of insulin secretion), or weight change among the genotypes were found (data not shown). SNs in other genes investigated were not associated with differences in baseline characteristics in the study groups. Next, we tested the effect of the intervention on the conversion to type 2 in carriers and noncarriers of the risk genotypes of SLC2A2 (Table 3). The subjects possessing the risk genotypes of all four SNs had a 3.5-fold risk of developing type 2 if they were in the control group, compared with that in the intervention group ( 0.001). No differences were found between the study groups in the conversion to type 2 in noncarriers of the risk genotypes. All SNs of SLC2A2 investigated were in strong linkage disequilibrium with each other (D 0.98, 0.001), and they formed seven haplotypes (ACCG, ACTG, ACTA, CCTG, CCTA, CTTG, and CTTA). None of these haplotypes increased the risk of type 2 beyond that of an individual SN. The CTTA haplotype protected from the development of type 2 in the control group (odds ratio [OR] 0.27 [95% CI ], 0.016). In our recent study, we found that AT-sensitive K channel genes, ABCC8 (encoding sulfonylurea receptor 1, SUR1) and KCNJ11 (encoding Kir6.2), predicted the conversion from IGT to type 2 (23). Since these genes regulate insulin secretion, we investigated whether they have an additive effect with rs5393 on the risk of type 2. We found that the conversion to type 2 was significantly increased in the presence of both the AA genotype of rs5393 of SLC2A2 (the SN which was most strongly associated with type 2 risk) and the GG genotype of rs of ABCC8 (Fig. 1). Subjects having both the SLC2A2 and ABCC8 risk genotypes had a 6.5-fold risk for the conversion to type 2 compared with those having neither of the risk genotypes (OR 6.49 [95% CI ], 0.012). Instead, there was no significant TABLE 3 The effect of intervention on the conversion to (%) and OR calculated by logistic regression analysis according to the SNs of SLC2A2 Control Intervention OR (95% CI) Control Intervention OR (95% CI) rs5393 AA genotype C allele ( ) ( ) rs5394 CC genotype T allele ( ) ( ) rs5400 CC genotype T allele ( ) ( ) rs5404 GG genotype A allele ( ) ( ) DIABETES, VOL. 54, JULY 2005

4 O. LAUKKANEN AND ASSOCIATES FIG. 1. type 2 according to SN rs5393 of the GLUT2 (SLC2A2) gene and SN rs of the SUR1 (ABCC8) gene (A) and according to SN rs5393 of the GLUT2 (SLC2A2) gene and SN rs5219 of the Kir6.2 (KCNJ11) gene (B). GLUT2 (AA), SUR1 (GG), and Kir6.2 (23K allele) are high-risk genotypes. additive effect between SLC2A2 and KCNJ11 (rs5219 SN) on the risk of type 2 ( 0.062). DISCUSSION This is the first report to demonstrate in a prospective study setting that in subjects with IGT, SNs in SLC2A2 are associated with a threefold risk for developing type 2. This increase in risk was observed in the entire study population and in the control group. Furthermore, the carriers of the high-risk genotypes benefited more from intervention than noncarriers. The SNs in the GGK, TCF1, HNF4A, GI, and GL1R genes did not affect the conversion to type 2. The majority of previous studies have failed to show that SLC2A2 is associated with type 2 (4 11), and only two studies have shown a positive association (2,3). In U.K. Europid subjects, SNs in SLC2A2 were associated with type 2 (2). In that study, the risk was associated with a minor allele of rs5400 (T110I) and rs5404 (T198T), whereas in our study the risk was attributable to the common homozygotes. We do not have any explanation for this discrepancy, but it may indicate differences in study populations. We gave our results without adjustment for the number of tests undertaken, which may lead to false-positive results due to multiple testing. However, the SNs of SLC2A2 were consistently associated with the conversion to type 2 even after adjustment for confounding factors (Table 2), quite similar to the results reported in a previous study (2). Our finding that the SNs of SLC2A2 were associated with the risk for type 2 in the control group, but not in the intervention group, is very similar to our previous results on the ABCC8 gene (23). Furthermore, SNs in SLC2A2 and ABCC8 contributed to the risk of type 2 independently of weight change. Risk genotypes of these genes had an additive effect on the conversion to type 2, and a similar nonsignificant trend was observed with respect to SLC2A2 and KCNJ11 (Fig. 1). Our results may indicate that subjects with risk genotypes in this network of genes regulating insulin secretion (SLC2A2, ABCC8, and KCNJ11) are at particularly high risk for type 2. In the present study, we did not find any differences in fasting and 2-h insulin levels among the genotypes of SLC2A2. In the Finnish DS, insulin secretion is influenced by the lifestyle intervention, which improved insulin sensitivity (24). Because the effect of SNs of SLC2A2 was observed only in the control group, this may imply that the improvement in insulin sensitivity by lifestyle changes is also beneficial to the preservation of insulin secretion in -cells. We did not find any association of the SNs of GCK, TCF1, HNF4A, GI, or GL-1R with type 2. However, negative findings do not exclude the possibility that these genes may contribute to the risk of in other populations. SNs in HNF4A have been shown to be associated with type 2 in other populations (18,19) and in one previous study on Finns (17). Different study designs and the relatively small sample size of the DS may explain discrepant findings. For example, to replicate a positive association of SNs of HNF4A and the conversion to type 2 in the present study, a sample size of 1,400 subjects would have been needed. In conclusion, the high-risk genotypes of SLC2A2 predicted the conversion to type 2 in Finnish subjects with IGT. We also demonstrated that carriers of these risk genotypes benefited from the lifestyle intervention. This study gives further evidence that genes regulating insulin secretion are important for the risk of type 2 also in obese subjects with IGT, generally assumed to have the primary defect in insulin action. ACKNOWLEDGMENTS This study was financially supported by grants from the Academy of Finland (38387 and to J.T., to M.U.), the EVO fund of the Kuopio University Hospital (5106 to M.U., 5194 to M.L.), the Ministry of Education, the Finnish Diabetes Research Foundation, the Juho Vainio Foundation, the Yrjö Jahnsson Foundation, and the European Union (EUGENE2, LSHM-CT to M.L.). We thank Teemu Kuulasmaa, MSc, for his help in statistical analyses of the data and Kaija Eirola and Leena Uschanoff for their help in genotyping the subjects. REFERENCES 1. Mueckler M: Facilitative glucose transporters. Eur J Biochem 219: , Barroso I, Luan J, Middelberg R, Harding AH, Franks W, Jakes RW, Clayton D, Schafer AJ, O Rahilly S, Wareham NJ: Candidate gene association study in type 2 indicates a role for genes involved in beta-cell function as well as insulin action. LoS Biol 1:41 55, Alcolado JC, Baroni MG, Li SR: Association between a restriction fragment length polymorphism at the liver/islet cell (GluT 2) glucose transporter and familial type 2 (non-insulin-dependent) mellitus. Diabetologia 34: , Matsutani A, Koranyi L, Cox N, ermutt MA: olymorphisms of GLUT2 and DIABETES, VOL. 54, JULY

5 GENES REGULATING INSULIN AND TYE 2 DIABETES GLUT4 genes: use in evaluation of genetic susceptibility to NIDDM in blacks. Diabetes 39: , atel, Bell GI, Cook JT, Turner RC, Wainscoat JS: Multiple restriction fragment length polymorphisms at the GLUT2 locus: GLUT2 haplotypes for genetic analysis of type 2 (non-insulin-dependent) mellitus. Diabetologia 34: , Li SR, Alcolado JC, Stocks J, Baroni MG, Oelbaum RS, Galton DJ: Genetic polymorphisms at the human liver/islet glucose transporter (GLUT2) gene locus in Caucasian and West Indian subjects with type 2 (non-insulindependent) mellitus. Biochim Biophys Acta 1097: , Baroni MG, Alcolado JC, ozzilli, Cavallo MG, Li SR, Galton DJ: olymorphisms at the GLUT2 (beta-cell/liver) glucose transporter gene and non-insulin-dependent mellitus (NIDDM): analysis in affected pedigree members. Clin Genet 41: , Tanizawa Y, Riggs AC, Chiu KC, Janssen RC, Bell DS, Go R, Roseman JM, Acton RT, ermutt MA: Variability of the pancreatic islet beta cell/liver (GLUT 2) glucose transporter gene in NIDDM patients. Diabetologia 37: , Shimada F, Makino H, Iwaoka H, Miyamoto S, Hashimoto N, Kanatsuka A, Bell GI, Yoshida S: Identification of two novel amino acid polymorphisms in beta-cell/liver (GLUT2) glucose transporter in Japanese subjects. Diabetologia 38: , Matsubara A, Tanizawa Y, Matsutani A, Kaneko T, Kaku K: Sequence variations of the pancreatic islet/liver glucose transporter (GLUT2) gene in Japanese subjects with noninsulin dependent mellitus. J Clin Endocrinol Metab 80: , Møller AM, Jensen NM, ildal J, Drivsholm T, Borch-Johnsen K, Urhammer SA, Hansen T, edersen O: Studies of genetic variability of the glucose transporter 2 promoter in patients with type 2 mellitus. J Clin Endocrinol Metab 86: , Matschinsky F, Liang Y, Kesavan, Wang L, Froguel, Velho G, Cohen D, ermutt MA, Tanizawa Y, Jetton TL, et al: Glucokinase as pancreatic beta cell glucose sensor and gene (Review). J Clin Invest 92: , Stone LM, Kahn SE, Fujimoto WY, Deeb SS, orte D Jr: A variation at position-30 of the -cell glucokinase gene promoter is associated with reduced -cell function in middle-aged Japanese-American men. Diabetes 45: , Yamada K, Yuan X, Ishiyama S, Ichikawa F, Koyama KI, Koyanagi A, Koyama W, Nonaka K: Clinical characteristics of Japanese men with glucokinase gene -cell promoter variant. Diabetes Care 20: , März W, Nauck M, Hoffmann MM, Nagel D, Boehm BO, Koenig W, Rothenbacher D, Winkelmann BR: G(-30)A polymorphism in the pancreatic promoter of the glucokinase gene associated with angiographic coronary artery disease and type 2 mellitus. Circulation 109: , Hegele RA, Zinman B, Hanley AJ, Harris SB, Barrett H, Cao H: Genes, environment and Oji-Cree type 2. Clin Biochem 36: , Silander K, Mohlke KL, Scott LJ, eck EC, Hollstein, Skol AD, Jackson AU, Deloukas, Hunt S, Stavrides G, Chines S, Erdos MR, Narisu N, Conneely KN, Li C, Fingerlin TE, Dhanjal SK, Valle TT, Bergman RN, Tuomilehto J, Watanabe RM, Boehnke M, Collins FS: Genetic variation near the hepatocyte nuclear factor-4 alpha gene predicts susceptibility to type 2. Diabetes 53: , Love-Gregory LD, Wasson J, Ma J, Jin CH, Glaser B, Suarez BK, ermutt MA: A common polymorphism in the upstream promoter region of the hepatocyte nuclear factor-4 gene on chromosome 20q is associated with type 2 and appears to contribute to the evidence for linkage in an Ashkenazi Jewish population. Diabetes 53: , Weedon MN, Owen KR, Shields B, Hitman G, Walker M, McCarthy MI, Love-Gregory LD, ermutt MA, Hattersley AT, Frayling TM: Common variants of the hepatocyte nuclear factor-4 2 promoter are associated with type 2 in the U.K. population. Diabetes 53: , Vilsbøll T, Holst JJ: Incretins, insulin secretion and type 2 mellitus. Diabetologia 47: , Tuomilehto J, Lindström J, Eriksson JG, Valle TT, Hämäläinen H, Ilanne- arikka, Keinänen-Kiukaanniemi S, Laakso M, Louheranta A, Rastas M, Salminen V, Uusitupa M: revention of type 2 mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 344: , World Health Organization: Diabetes Mellitus: Report of a WHO Study Group. Geneva, World Health Org., 1985 (Tech. Rep. Ser. no. 727) 23. Laukkanen O, ihlajamäki J, Lindström J, Eriksson J, Valle TT, Hämäläinen H, Ilanne-arikka, Keinänen-Kiukaanniemi S, Tuomilehto J, Uusitupa M, Laakso M: olymorphisms of the SUR1 (ABCC8) and Kir6.2 (KCNJ11) genes predict the conversion from impaired glucose tolerance to type 2 : The Finnish Diabetes revention Study. J Clin Endocrinol Metab 89: , Uusitupa M, Lindi V, Louheranta A, Salopuro T, Lindstrom J, Tuomilehto J for the Finnish Diabetes revention Study Group: Long-term improvement in insulin sensitivity by changing lifestyles of people with impaired glucose tolerance: 4-year results from the Finnish Diabetes revention Study. Diabetes 52: , DIABETES, VOL. 54, JULY 2005