Interaction between the uncoupling protein 2 −866G>A gene variant and cigarette smoking to increase oxidative stress in subjects with diabetes

Methods and results Amongst 453 Caucasian diabetic men there was a significant interaction ( p = 0.001) between genotype and smoking in determining plasma Total AntiOxidant Status (TAOS). Current smokers with the −866AA genotype had the lowest TAOS (indicating higher oxidative stress) of all subjects (AA vs. GG: 32.00 ± 17.4% vs. 45.8 ± 12.6%, p = 0.04). In a sub-sample of 20 subjects (10 GG, 10 AA) matched for baseline characteristics, plasma markers of oxidative stress in current smokers were significantly higher in AA compared to GG subjects (TAOS 36.8 ± 9.5% vs. 51.4 ± 9.5%, p = 0.04; F 2 -isoprostanes 1133.6 ± 701.2 pg ml −1 vs. 500.8 ± 64.7 pg ml −1 , p = 0.04). Conclusions This study demonstrates an interaction between the UCP2 −866G>A variant and smoking to increase oxidative stress in vivo . Keywords UCP2 Genetics Coronary heart disease Diabetes mellitus Inflammation Oxidative stress Smoking Introduction The generation of reactive oxygen species (ROS) is implicated in the pathogenesis of coronary heart disease (CHD), being increased in subjects with atherosclerosis or its classical risk factors [1–3] . Uncoupling proteins (UCPs 1–3) [4] dissipate the inner mitochondrial membrane proton electrochemical gradient that drives ATP synthesis [5] . UCP1 expression is restricted to brown adipose tissue (BAT) [6] , whilst UCP3 is predominantly expressed in skeletal muscle [7] . UCP2 is likely to be the ancestral UCP, being expressed ubiquitously [8–10] and sharing 59% and 73% sequence homology with UCP1 [8,9] and UCP3 [7] respectively. The mitochondrial location and electrochemical actions of UCP2 make it a plausible negative regulator of ROS production [11,12] . In support, decreased UCP2 expression (through endothelial anti-sense strategies [10] or macrophage gene-deletion [13] ) increases ROS generation, while bone marrow transplant from UCP2 (gene) knockout donor mice demonstrate evidence of increased oxidative stress and lesion size in atherosclerotic-prone mice [14] . To maintain homeostasis, not only is UCP2 expression induced by extra-mitochondrial sources of ROS [15–17] , but also by intramitochondrial superoxide [18] , thus protecting against further ROS generation. A common functional variant exists in the promoter of the human UCP2 gene (−866G>A) [19] , with the A-allele construct being associated with lower UCP2 expression [20] in transient transfection experiments in non-pancreatic β-cell lines. We have previously studied the association of this variant in relation to plasma total antioxidant status (TAOS) in diabetic men and with prospective CHD risk in healthy middle-aged men [21] . The A allele was associated with significantly lower plasma TAOS (i.e. increased oxidative stress) and increased F 2 -isoprostanes (i.e. a direct measure of lipid peroxidation) in the diabetic subjects with CHD, and with a highly significant doubling in prospective CHD risk (OR: 2.08 [95%CI: 1.49–2.86] in otherwise healthy men [21] . The risk associated with this genotype was substantially increased by the presence of conventional risk factors (diabetes, hypertension and obesity) which are also associated with increased oxidative stress. Since cigarette smoking is a potent stimulator of ROS production [22] , the aim of our current study was to examine the interaction between the UCP2 −866G>A gene variant and cigarette smoking in the previously published sample of men with diabetes [21] . As oxidative stress is increased in the presence of atherosclerosis [3] , we reasoned that it would also be essential to study the association after grouping the subjects by CHD status. Methods Subjects Ethical approval was granted by the institutional ethical committee and all subjects gave written informed consent before recruitment. Patients were recruited from the University College London Diabetes and Cardiovascular Study (UDACS), described elsewhere [21] . Briefly, this comprises 1011 consecutive subjects recruited from the diabetes clinic at University College London Hospitals (UCLH) NHS Trust between the years 2001–2. All patients had diabetes according to WHO criteria [23] . The presence of CHD was recorded if any patient had positive coronary angiography/angioplasty, coronary artery bypass, cardiac thallium scan, exercise tolerance test, myocardial infarction or symptomatic/treated angina. Any individual who was asymptomatic or had negative investigations was categorised as ‘no CHD’. Analyses were confined to Caucasian men, as in our previous study. Smoking status was defined as current, never or ex-smokers. The latter comprised of those who had stopped smoking for greater than 12 months. Current smokers included subjects who had stopped smoking within a 12 month period. None of the subjects were knowingly taking any form of vitamin supplementation. Plasma samples were collected within a 12 month period and stored immediately at −80 °C. Smoking history was available on 453 of the Caucasian men recruited in UDACS with genotype data (93.4% of total Caucasian men). Of these, 191 were never smokers, 177 ex-smokers and 85 current smokers. Plasma TAOS was measured in 438 of the 453 subjects. Plasma markers of oxidative stress Plasma total anti-oxidant status (TAOS), which is inversely related to oxidative stress, was measured by Sampson's modification of Laight's photometric microassay [24] , using 2.5 μl citrated plasma samples in 96-well ELISA plates. The TAOS of plasma was determined by its capacity to inhibit the peroxidase-mediated formation of the 2,2-azino-bis-3-ethylbensthiazoline-6-sulfonic acid (ABTS+) radical. There are two arms to the assay, a control arm and test arm. In the control arm phosphate buffered saline is used instead of plasma. The difference in absorbance (control absorbance minus test absorbance) divided by the control absorbance (expressed as a percentage) was used to represent the percentage inhibition of the reaction. The inter- and intra-assay coefficients of variation were 10.1% and 4.3% respectively. Previously, we have shown that baseline plasma TAOS is associated with prospective risk and has a good correlation with plasma F 2 -isoprostanes [25] . Plasma total F 2 -isoprostanes were measured using gas chromatography and mass spectroscopy in 20 UCP2 −866G>A homozygous subjects (5 GG and 5 AA current smokers and 5 GG and 5 AA non-smokers), as previously described [26,27] , with the mean for each group closely matched for baseline characteristics including drug treatment. In brief, 500 pg of [ 2 H 4 ]-8-iso-prostaglandin F 2α (PGF 2α ) ( Cayman Co., Ann Arbor, MI, USA ) were added as internal standards to 500 μl of plasma followed by alkaline hydrolysis for 1 h (37 °C). To extract F 2 -isoprostanes, the pH was adjusted to 3, and the samples were extracted on a tC18 solid-phase extraction cartridge ( Waters, UK ), converted to the pentaflurobenzyl ester, purified by thin-layer chromatography and analyzed as the trimethysilyl ether. Detection was performed by selected ion monitoring gas chromatography negative ion chemical ionization/mass spectrometry. Samples were analyzed as the pentafluorobenzyl ester trimethysilyl ether derivatives monitoring the M-181 ions, m/z 569 for endogenous F 2 -isoprostanes, and m/z 573 for [ 2 H 4 ]-8-iso-PGF 2α . The F 2 -isoprostanes elute as a series of chromatographic peaks over approximately 20 s, and quantification is based on the primary peak that co-elutes with the deuterated 8-iso-PGF 2α internal standard. All results were calculated by reference to the deuterated 8-iso-PGF 2α internal standard. The inter assay CV is 7%. Results were calculated by reference to deuterated 8-iso-PGF 2α internal standards. UCP2 −866G>A genotyping Genotypes were determined using leucocyte DNA polymerase chain reaction amplification (PCR) using published primers digested with Mlu1 as described [19] and products were resolved by MADGE [28] and confirmed by two independent technicians blind to subject outcome, with discrepancies resolved by repeat genotyping. Statistical analysis Analysis was performed using SPSS (version 10.1, SPSS Inc., Chicago). Data are reported for those individuals amongst whom high-throughput genotyping was available. Results are presented as mean ± standard deviation or median (interquartile range). Geometric mean and approximate standard deviation is shown for plasma F 2 -isoprostane. Deviations from Hardy–Weinberg equilibrium were considered using chi-squared tests. Allele frequencies are shown with the 95% confidence interval. Analysis of variance (ANOVA) was used to assess the association between genotypes and baseline characteristics on normally distributed data or after appropriate transformation (log or square root). The relationships between baseline parameters and plasma TAOS were tested by Pearson rank correlation co-efficient. ANCOVA was performed to test the association between genotype and TAOS after adjustment for potential confounders using multiple regression analysis to obtain a residual. Chi-squared tests were used to compare differences in categorical variables by genotype and smoking status. In all cases a p -value of less than 0.05 was considered statistically significant. The interaction between genotype and smoking status in determining plasma TAOS was performed using linear regression where the combined effects were compared to the individual effects of these variables (likelihood ratio test). This analysis has been run with and without the inclusion of potential confounders. Results Plasma TAOS, UCP2 genotype and smoking status The baseline characteristics of the subjects grouped by smoking status are summarized in Table 1 . As previously described [21] , plasma TAOS (measured in 438/453 subjects) was independent of pharmacotherapy, but correlated positively with plasma HDL-cholesterol, and negatively with triglyceride, glucose and HbA 1c (correlation co-efficient r = 0.1, −0.1, and −0.1; all p < 0.05). As shown in Table 1 , current smokers had an 8% lower plasma TAOS relative to never smokers ( p = 0.04). When current smokers were compared to ex- and never-smokers grouped together, the relative difference in plasma TAOS was 7% ( p = 0.05). Interestingly, plasma TAOS in the ex-smokers was midway between that for the current and ex-smokers ( Table 1 ). The genotype distribution for the UCP2 −866G>A variant was in Hardy–Weinberg equilibrium (χ 2 = 0.40, p = 0.526) with a G allele frequency of 0.66 [0.63–0.69] and an A allele frequency of 0.34 [0.31–0.37]. Of note, when the genotype distribution was compared according to smoking history, the frequency of the AA genotype was lower in the current smokers (7.1%) compared to the never smokers and ex-smokers (12%; p = 0.05). No association was seen between genotype and plasma TAOS in the overall group (GG: 44.2 ± 12.8% vs. GA: 42.1 ± 13.7% vs. AA: 40.8 ± 11.8%; p = 0.06). Furthermore, there was no difference in any other baseline characteristics (including treatment or duration of diabetes) by UCP2 genotype (data not shown). After adjustment of plasma TAOS for the above correlates (HDL-cholesterol, triglyceride, glucose and HbA 1c ), the association between genotype and plasma TAOS remained essentially unchanged (linearity between groups p = 0.03, ANOVA; p = 0.09). In the whole sample, both before and after adjustment for potential confounders, there was a highly significant interaction between smoking status (never, ex-smokers and current smokers) and genotype (GG, GA, AA) in determining plasma TAOS, ( p = 0.001 and p < 0.001 after adjustment for HDL-cholesterol, triglyceride, glucose and HbA 1c ) as illustrated in Fig. 1 . Subjects without CHD In subjects without CHD ( n = 350), no association between smoking status and plasma TAOS was observed (never smokers vs. ex-smokers vs. current smokers: 44.5 ± 12.0% vs. 43.6 ± 13.0% vs. 41.0 ± 15.4%, linearity between groups p = 0.08, ANOVA p = 0.18). Furthermore, no association was seen between genotype and plasma TAOS (GG: 45.0 ± 12.9% vs. GA 42.0 ± 14.12% vs. AA 43.5 ± 8.7%, ANOVA p = 0.16, after adjustment p = 0.10). However, there was a highly significant interaction between smoking status (never, ex-smokers and current smokers) and genotype (GG, GA, AA) in determining plasma TAOS, p = 0.003 (after adjustment of TAOS p = 0.005). Subjects with CHD In subjects with CHD ( n = 103), plasma TAOS was lower compared to those without (CHD 40.9 ± 13.1% vs. No CHD 43.4 ± 13.1%, p = 0.05). There was no association between smoking status and TAOS (never smokers vs. ex-smokers vs. current smokers: 41.7 ± 13.1% vs. 41.2 ± 12.3% vs. 36.7 ± 16.7%, ANOVA p = 0.50). However as previously described [21] , there was a significant association between genotype and plasma TAOS (GG vs. GA vs. AA: 41.6 ± 12.4% vs. 42.4 ± 12.3% vs. 30.1 ± 16.1%, ANOVA p = 0.02, after adjustment p = 0.03). No interaction was seen between smoking status (never, ex-smokers current smokers) and genotype (GG, GA, AA) in determining plasma TAOS ( p = 0.06, after adjustment p = 0.05). Plasma F 2 -isoprostanes, UCP2 genotype and smoking status To investigate the above association further, twenty men from UDACS, free from CHD were selected for homozygosity for the UCP2 −866G>A variant and closely matched for baseline characteristics ( Table 2 ). Within this sample, plasma TAOS was not different between never smokers compared to current smokers (smokers 44.0 ± 11.8% vs. non-smokers 47.9 ± 8.3%, p = 0.44). However, plasma total F 2 -isoprostanes were significantly higher (56%) in smokers (693.9 ± 380.6 pg ml −1 ) compared to non-smokers (446.2 ± 124.5 pg ml −1 ), p = 0.04. With respect to genotype, plasma TAOS was not different in the whole sample (AA 42.1 ± 9.9% vs. GG 49.3 ± 9.6%, p = 0.13) nor in the non-smokers (AA 48.7 ± 6.2% vs. GG 47.2 ± 10.4%, p = 0.82). However in smokers a significant difference was observed with AA smokers having lower TAOS (AA 36.8 ± 9.5% vs. GG 51.4 ± 9.5%, p = 0.04), as shown in Fig. 2 a. In accordance with this, plasma F 2 -isoprostane concentration was elevated in AA smokers (AA 1133.6 ± 701.2 pg ml −1 vs. GG 500.8 ± 64.7 pg ml −1 , p = 0.04), but not in AA non-smokers (AA 463.5 ± 96.3 pg ml −1 vs. GG 432.7 ± 150.5 pg ml −1 , p = 0.74; Fig. 2 b). In this small sample, there was no evidence of an interaction between genotype and smoking status in determining F 2 -isoprostanes ( p = 0.11) nor plasma TAOS ( p = 0.07). Discussion This study provides evidence of an interaction between the UCP2 −866G>A gene variant and cigarette smoking in determining plasma levels of oxidative stress. In subjects free from manifest CHD, smoking and the AA genotype alone were not associated with change in a plasma marker of oxidative stress, however in combination, there was an interaction resulting in increased oxidative stress. As observed, mean plasma TAOS was lowest in current smokers, intermediate in ex-smokers and highest in never-smokers. Subjects with the UCP2 −866AA genotype who were current smokers, had the lowest plasma TAOS of all groups tested (approximately 30% lower than AA never-smokers and 20% lower than AA ex-smokers). Within the current smokers, there was a linear association between plasma TAOS and genotype, such that with each additional −866A allele, plasma TAOS decreased ( Fig. 1 ). This association was also observed in the ex-smokers, except that the difference in plasma TAOS by genotype was of a smaller magnitude. In the never smokers, no significant difference was observed by genotype on plasma TAOS. In line with these observations, within the small subset of men without CHD, matched for baseline characteristics (including age, glycaemic control and treatment), UCP2 AA smokers had the lowest mean plasma TAOS and greatest mean F 2 -isoprostane concentrations levels in the groups studied. We have previously reported that the AA genotype is associated with increased oxidative burden in diabetes subjects with CHD, but not in those without CHD [21] . In the present study, the interaction between genotype and smoking was evident in the group without CHD. This might be explained by the fact that CHD (with its cluster of associated risk factors) [3] , would be expected to be more pro-oxidant than cigarette smoking in subjects free from manifest CHD, and thus the interaction between genotype and smoking would be overwhelmed in subjects with CHD. The current work supports our previous findings that the UCP2 −866A variant is associated with increased oxidative stress and with prospective CHD risk [21] , and further supports a role for UCP2 (and hence the mitochondrial electron transport chain) in the regulation of ROS generation, and highlights its potential impact upon CHD risk. Our current study demonstrated that in the presence of an environmental stimulant of oxidative stress (cigarette smoking), the −866A allele is associated with increased plasma ROS burden compared to the wild type −866G allele. Previously, in vitro studies have shown that UCP2 gene expression is induced by both extra- and intramitochondrial sources of ROS [16–18] , and selective down-regulation of UCP2 increases endothelial cell ROS generation [10] . This may explain the dependence of the observed genotypic effect on cigarette smoking (or on the presence of CHD). Therefore, under conditions of increased oxidative stress (e.g. smoking, diabetes, CHD), increased UCP2 expression should prove vasculo-protective [13,17] and may be an important target for therapy. In support of this, UCP2 protects against atherosclerosis in LDL-receptor deficient mice [14] . Promoter constructs of the −866A allele are associated with greater repression of transcription in somatic non-β cells [20] . It is likely, therefore, that the UCP2 −866A allele is associated with lower inducible UCP2 expression within the vasculature or circulating immune cells. As such, one would anticipate the A allele to be associated with increased oxidative stress and higher risk of CHD as demonstrated in these studies. In summary, diabetes mellitus and cigarette smoking are two major risk factors for CHD and both are associated with increased oxidative stress [3,24,26] . This data supports the view that UCP2 is an integral component of antioxidant defence, particularly in a hostile pro-oxidant environment, and this gene variant is associated with a major effect in determining UCP2 activity. The observation that the AA subjects only have reduced plasma TAOS in smokers or subjects with CHD is a clear example of gene-environment interaction. Further prospective study is required to explore the interaction between this gene variant and cigarette smoking in determining premature CHD risk in relation to biochemical markers of oxidative stress. Acknowledgements JWS and the recruitment of UDACS was supported by a clinical training fellowship from Diabetes UK (BDA: RD01/0001357). The British Heart Foundation supported SSD and SEH: FS/2001044, RG2000015. References [1] H. Cai D.G. Harrison Endothelial dysfunction in cardiovascular diseases: the role of oxidant stress Circ Res 87 2000 840 844 [2] G.M. Chisolm D. Steinberg The oxidative modification hypothesis of atherogenesis: an overview Free Radic Biol Med 28 2000 1815 1826 [3] D. Harrison K.K. Griendling U. Landmesser B. Hornig H. Drexler Role of oxidative stress in atherosclerosis Am J Cardiol 91 2003 7A 11A [4] F. Bouillaud E. Couplan C. Pecqueur D. Ricquier Homologues of the uncoupling protein from brown adipose tissue (UCP1): UCP2, UCP3, BMCP1 and UCP4 Biochim Biophys Acta 1504 2001 107 119 [5] K.D. Garlid M. Jaburek P. Jezek Mechanism of uncoupling protein action Biochem Soc Trans 29 2001 803 806 [6] C.S. Lin M. Klingenberg Isolation of the uncoupling protein from brown adipose tissue mitochondria FEBS Lett 113 1980 299 303 [7] O. Boss S. Samec A. Paoloni-Giacobino C. Rossier A. Dulloo J. Seydoux Uncoupling protein-3: a new member of the mitochondrial carrier family with tissue-specific expression FEBS Lett 408 1997 39 42 [8] C. Fleury M. Neverova S. Collins S. Raimbault O. Champigny C. Levi-Meyrueis Uncoupling protein-2: a novel gene linked to obesity and hyperinsulinemia Nat Genet 15 1997 269 272 [9] R.E. Gimeno M. Dembski X. Weng N. Deng A.W. Shyjan C.J. Gimeno Cloning and characterization of an uncoupling protein homolog: a potential molecular mediator of human thermogenesis Diabetes 46 1997 900 906 [10] C. Duval A. Negre-Salvayre A. Dogilo R. Salvayre L. Penicaud L. Casteilla Increased reactive oxygen species production with antisense oligonucleotides directed against uncoupling protein 2 in murine endothelial cells Biochem Cell Biol 80 2002 757 764 [11] A. Negre-Salvayre C. Hirtz G. Carrera R. Cazenave M. Troly R. Salvayre A role for uncoupling protein-2 as a regulator of mitochondrial hydrogen peroxide generation Faseb J 11 1997 809 815 [12] L. Casteilla M. Rigoulet L. Penicaud Mitochondrial ROS metabolism: modulation by uncoupling proteins IUBMB Life 52 2001 181 188 [13] D. Arsenijevic H. Onuma C. Pecqueur S. Raimbault B.S. Manning B. Miroux Disruption of the uncoupling protein-2 gene in mice reveals a role in immunity and reactive oxygen species production Nat Genet 26 2000 435 439 [14] J. Blanc M.C. Alves-Guerra B. Esposito S. Rousset P. Gourdy D. Ricquier Protective role of uncoupling protein 2 in atherosclerosis Circulation 107 2003 388 390 [15] F.Y. Lee Y. Li H. Zhu S. Yang H.Z. Lin M. Trush Tumor necrosis factor increases mitochondrial oxidant production and induces expression of uncoupling protein-2 in the regenerating mice [correction of rat] liver Hepatology 29 1999 677 687 [16] C. Pecqueur M.C. Alves-Guerra C. Gelly C. Levi-Meyrueis E. Couplan S. Collins Uncoupling protein 2, in vivo distribution, induction upon oxidative stress, and evidence for translational regulation J Biol Chem 276 2001 8705 8712 [17] K.S. Echtay D. Roussel J. St-Pierre M.B. Jekabsons S. Cadenas J.A. Stuart Superoxide activates mitochondrial uncoupling proteins Nature 415 2002 96 99 [18] K.S. Echtay M.P. Murphy R.A. Smith D.A. Talbot M.D. Brand Superoxide activates mitochondrial uncoupling protein 2 from the matrix side. Studies using targeted antioxidants J Biol Chem 277 2002 47129 47135 [19] H. Esterbauer C. Schneitler H. Oberkofler C. Ebenbichler B. Paulweber F. Sandhofer A common polymorphism in the promoter of UCP2 is associated with decreased risk of obesity in middle-aged humans Nat Genet 28 2001 178 183 [20] F. Krempler H. Esterbauer R. Weitgasser C. Ebenbichler J.R. Patsch K. Miller A functional polymorphism in the promoter of UCP2 enhances obesity risk but reduces type 2 diabetes risk in obese middle-aged humans Diabetes 51 2002 3331 3335 [21] S.S. Dhamrait J.W. Stephens J.A. Cooper J. Acharya A.R. Mani K. Moore Cardiovascular risk in healthy men and markers of oxidative stress in diabetic men are associated with common variation in the gene for uncoupling protein 2 Eur Heart J 25 2004 468 475 [22] J.D. Morrow B. Frei A.W. Longmire J.M. Gaziano S.M. Lynch Y. Shyr Increase in circulating products of lipid peroxidation (F2-isoprostanes) in smokers. Smoking as a cause of oxidative damage N Engl J Med 332 1995 1198 1203 [23] K.G. Alberti P.Z. Zimmet New diagnostic criteria and classification of diabetes—again? Diabet Med 15 1998 535 536 [24] M. Sampson N. Gopaul I. Davies D. Hughes M. Carrier Plasma F2 isoprostanes. Direct evidence of increased free radical damage during acute hyperglycaemia in type 2 diabetes Diabetes Care 25 2002 537 541 [25] J.W. Stephens D.R. Gable S.J. Hurel G.J. Miller J.A. Cooper S.E. Humphries Increased plasma markers of oxidative stress are associated with coronary heart disease in males with diabetes mellitus and with 10-year risk in a prospective sample of males Clin Chem 52 2006 446 452 [26] J.D. Morrow L.J. Roberts 2nd Mass spectrometry of prostanoids: F2-isoprostanes produced by non-cyclooxygenase free radical-catalyzed mechanism Methods Enzymol 233 1994 163 174 [27] B. Ferraro F. Galli B. Frei E. Kingdon F. Canestrari C. Rice-Evans Peroxynitrite-induced oxidation of plasma lipids is enhanced in stable hemodialysis patients Kidney Int 63 2003 2207 2213 [28] I.N. Day S.E. Humphries Electrophoresis for genotyping: microtiter array diagonal gel electrophoresis on horizontal polyacrylamide gels, hydrolink, or agarose Anal Biochem 222 1994 389 395