Molecular Genetics and Metabolism 106 (2012) 359–365

Contents lists available at SciVerse ScienceDirect

Molecular Genetics and Metabolism

j ourna l homepage: www.e lsev ie r .com/ locate /ymgme

Allelic variations in superoxide dismutase-1 (SOD1) gene and renal andcardiovascular morbidity and mortality in type 2 diabetic subjects

Ana Luísa Neves a,b, Kamel Mohammedi a,c, Nathalie Emery a,d, Ronan Roussel c,d,e, Frédéric Fumeron a,d,Michel Marre a,c,d, Gilberto Velho a,⁎a INSERM, Research Unit 695, Paris, Franceb University of Porto (U.Porto), Faculty of Medicine, Porto, Portugalc Assistance Publique Hôpitaux de Paris, Bichat Hospital, Department of Diabetology, Endocrinology and Nutrition, Paris, Franced Univ. Paris Diderot, Sorbonne Paris Cité, UFR de Médecine, Paris, Francee INSERM, Research Unit 872, Paris, France

Abbreviations: ACE, angiotensin converting enzymeeGFR, estimated glomerular filtration rate; HR, hazard rtive oxygen species; SNP, single nucleotide polymormutase; UAE, urinary albumin excretion at baseline; UT⁎ Corresponding author at: INSERM U-695, Univer

Médecine Bichat, 16 rue Henri Huchard, 75018 Paris, FrE-mail address: gilberto.velho@inserm.fr (G. Velho).

1096-7192/$ – see front matter © 2012 Elsevier Inc. Alldoi:10.1016/j.ymgme.2012.04.023

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 21 February 2012Received in revised form 25 April 2012Accepted 25 April 2012Available online 2 May 2012

Keywords:Diabetic nephropathyCardiovascular complicationsSudden deathGenetic epidemiologyOxidative stressSuperoxide dismutase

Background: Oxidative stress is involved in the pathophysiology of renal and cardiovascular complications ofdiabetes. Superoxide dismutase (SOD) enzymes play a major role in detoxification of reactive oxygen speciesand protection against oxidative stress. Associations of SOD1 gene variants with diabetic nephropathy werereported in patients with type 1 diabetes. We investigated associations of allelic variations in SOD1 genewith nephropathy and cardiovascular complications in patients with type 2 diabetes.Methods: Seven SNPs in SOD1 region were analyzed in 3744 type 2 European Caucasian diabetic patients fromthe DIABHYCAR (a 6-year prospective study) and DIABHYCAR_GENE cohorts. Odds ratios or hazard ratios forprevalence and incidence of diabetic nephropathy and cardiovascular events were estimated.Results: We observed an association of rs1041740 with the prevalence of microalbuminuria at baseline(OR 1.51, 95% CI 1.10–2.10, p=0.01). No association with the incidence of renal events (doubling of theserum creatinine levels or the requirement of hemodialysis or renal transplantation) or cardiovascular events(myocardial infarction or stroke) was observed during follow-up. However, three variants were associated

with increased risk of death from cardiovascular causes (sudden death, fatal myocardial infarction or stroke)during the follow-up: rs9974610 (HR 0.64, 95% CI 0.46–0.88, p=0.005), rs10432782 (HR 1.71, 95% CI1.16–2.48, p=0.007) and rs1041740 (HR 1.78, 95% CI 1.10–2.78, p=0.02).Conclusions: Our results are consistent with a major role for SOD1 in the mechanisms of cardiovascular pro-tection against oxidative stress in type 2 diabetic subjects.

© 2012 Elsevier Inc. All rights reserved.

1. Introduction

Diabetic patients have a 3-fold higher risk than nondiabetic individ-uals of developing atherosclerosis and its clinical complications such asstroke, myocardial infarction, and peripheral vascular disease [1]. Car-diovascular disease accounts for up to 80% of the deaths of type 2 dia-betic patients [2] and sudden death occurs frequently among diabeticpatients [3]. Diabetic nephropathy is a leading cause of renal failure[4,5] and is associated with increased risk of cardiovascular morbidityand mortality in type 1 and in type 2 diabetic patients [6,7].

; ANOVA, analysis of variance;atio; OR, odds ratio; ROS, reac-phism; SOD, superoxide dis-R, untranslated region.sité Paris Diderot, Faculté deance. Fax: +33 1 57 27 75 01.

rights reserved.

Reactive oxygen species (ROS), including free radicals such as super-oxide and nonradical species such as hydrogen peroxide, are producedcontinuously in all cells as part of the normal cellular metabolism [8].The major source of intracellular ROS is the mitochondrial respiratorychain, which produces large amounts of superoxide radicals [9]. Oxida-tive stress occurs when production of ROS exceeds local antioxidant ca-pacity. In this situation, there is increased oxidation of proteins, lipids,carbohydrates andDNA, that can result in tissue and organ damage. Hy-perglycemia increases the production of ROS and causes oxidativestress [10,11]. Oxidative stress influencesmultiple pathways implicatedin diabetic nephropathy [12,13]. There is also compelling evidence thatoxidative stress is associatedwith themetabolic syndrome and its com-ponents [14] and that it plays a key role in the pathophysiology of sev-eral cardiovascular diseases, including hypertension, myocardialinfarction, stroke, and heart failure [15,16].

Superoxide dismutases (SOD) are a class of enzymes that catalyzethe dismutation of superoxide into oxygen and hydrogen peroxide[17]. They play an important role in antioxidant mechanism in nearlyall cells exposed to oxygen. Three isoforms of SOD are expressed in

360 A.L. Neves et al. / Molecular Genetics and Metabolism 106 (2012) 359–365

humans, encoded by different genes. SOD1 is mainly located in the cy-toplasm, SOD2 in the mitochondria and SOD3 is extracellular. SOD1 ac-counts for ~85% of the total cellular SOD activity of most mammaliancells and is highly active in the human kidney and in the vascular wall[18,19].

Associations of SOD1 gene variants with diabetic nephropathyhave been observed in patients with type 1 diabetes [20,21], but nodata regarding associations with nephropathy and cardiovascularcomplications in type 2 diabetes is available. In this study, weassessed the impact of SOD1 allelic variation in the developmentand progression of diabetic nephropathy and in the morbidity andmortality of cardiovascular disease in individuals with type 2 diabetesfollowed prospectively for renal and cardiovascular events.

2. Methods

2.1. Participants

We studied unrelated French type 2 diabetic patients from theDIABHYCAR (n=3137) and the DIABHYCAR_GENE (n=607) cohorts.DIABHYCAR was a 6-year clinical trial conducted in men and womenwith type 2 diabetes selected on the basis of persistent microalbuminuria(urinary albumin excretion, UAE=20–200 mg/l) or macroalbuminuria(UAE>200mg/l) without renal failure (plasma creatinine b150 μmol/l)at baseline. The trial testedwhether a lowdose of ramipril, an angiotensinconverting enzyme (ACE) inhibitor, able to reduceUAEwould also reducecardiovascular and/or renal events such as myocardial infarction, stroke,acute heart failure, end-stage renal failure, and cardiovascular death. Forthe purpose of the trial, a renal event was defined as the doubling of theserumcreatinine levels or the requirement of hemodialysis or renal trans-plantation during follow-up. Myocardial infarction was diagnosed as theoccurrence of at least 2 out of 3 of the following criteria: constrictivechest pain lasting 20 min or longer, increased serum creatinine phospho-kinase and/or troponine levels, or typical electrocardiographic changes.Sudden death was defined as death occurring instantaneously or within1 h after the onset of new cardiac symptoms (arrhythmia, myocardialinfarction) or non-witnessed death, when the body was found and nocause of death could be discovered. Fatal stroke was not included in thisgroup. Results were negative regarding the drug effect and were publi-shed previously [22,23]. The DIABHYCAR_GENE cohort was recruitedconcomitantly to DIABHYCAR and included men and women withT2DM presenting with normal UAE (UAEb20 mg/l) at baseline, andwho remained normoalbuminuric at the end of the follow-up. An inde-pendent committee reviewed all case records from both cohorts to vali-date selection criteria, to grade the renal involvement of each patient,and to adjudicate the clinical events during follow-up [22]. Urinary albu-min was measured by nephelometry [24]. Estimation of the glomerularfiltration rate (eGFR) was computed with the Modification of Diet inRenal Disease (MDRD) formula [25]. Participants gave written informedconsent and study protocols were approved by the ethics committee ofAngers University Hospital.

2.2. DNA studies

The SOD1 gene is located on chromosome 21q22.11 and has a ge-nomic size of 9307 bp. It consists of five exons, with untranslatedregions (UTR) in exons 1 and 5, separated by four introns. The 5′flanking region containing transcription factor binding sites andknown regulatory elements extends up to −500 bp [26]. Seven SNPsin the SOD1 region were analyzed [26]: rs9974610 (~13.6 kb 5′ fromtranscription start site), rs2173962 (~9.9 kb 5′ from transcriptionstart site), rs10432782 (intron 2), rs2070424 (intron 3), rs1041740/rs17880196 (intron 4), rs17880135 (~0.7 kb 3′ from the end of exon5/UTR) and rs202449 (~5.0 kb 3′ from the end of exon 5/UTR). TheSNPs were chosen in HapMap (public release #23) on the basis ofgiving information on ~90% of the allelic variation of SNPs with minor

allele frequency ≥5% at r2>0.8 in haplotype blocks containing SOD1.Genotypes were determined by an Assay by Design (ABD) kit fromApplied Biosystems (rs17880135) or by competitive allele-specificPCR genotyping system assays (KASPar, Kbioscience, Hoddeston, UK).Genotyping success rate was >95%. Genotyping was repeated in 5% ofsubjects with 100% coherence. All genotypes were in Hardy–Weinbergequilibrium.

2.3. Statistical analysis

Results are expressed as mean±SD except when stated otherwise.Differences between groups were assessed by Pearson's chi-squaredtest and by ANOVA. When ANOVA was significant, comparisonsbetween pairs were made using the Tukey–Kramer HSD test.Allelic associations with renal or cardiovascular traits were assessedby regression models. Adjustments for clinical and biological parame-ters were carried out by including these parameters as covariables inthe regression model. Cox proportional hazards survival regressionanalyses were used to examine the effect of explanatory variables ontime-related survival (or disease-free) rates in prospective analyses.Logistic regression analyses were used for cross-sectional analyses.Hazard ratios or odds ratios, respectively, with their 95% confidence in-tervals were computed in these analyses for the minor alleles. Datawere log-transformed for the analyses when the normality of the distri-butionwas rejected by the Shapiro–WilkW test. Correction formultiplecomparisons due tomultiple SNP testing took into account the effectivenumber of independent tests (Meff) based on the degree of linkage dis-equilibrium between SNPs [27,28]. The adjusted significance thresholdwas determined by dividing Meff into the nominal significance thresh-old (p=0.05). Thus, p≤0.02 was considered significant, unless statedotherwise. The power to detect associations of the SNPs with diabeticnephropathy at baseline and with incidence of renal events, cardiovas-cular events and cardiovascular death during follow-up was 0.95, 0.67,0.70 and 0.74 respectively, for odds ratio or hazard ratio equal or higherthan 1.5 and alpha=0.02. Statistics were performed with the JMP soft-ware (SAS Institute Inc., Cary, NC).

3. Results

3.1. SOD1 variants and UAE status at baseline

In a first step, participants were divided into 3 groups according toUAE status at baseline: normal UAE (DIABHYCAR_GENE cohort),microalbuminuria and macroalbuminuria (DIABHYCAR cohort, bothgroups). Characteristics of participants are shown in Table 1. Subjectswith microalbuminuria or with macroalbuminuria were younger, hada shorter duration of diabetes and were more often of male sex thansubjects with normal UAE. Individuals with macroalbuminuria ascompared to those with normal UAE had higher body mass index(BMI), higher blood pressure levels, increased levels of HbA1c, totalcholesterol and triglycerides, and lower levels of HDL cholesterol,while individuals with microalbuminuria presented with intermedi-ate values of these parameters. Genotype frequencies according toUAE status at baseline are shown in Table 2. We observed an associa-tion of the minor T-allele of rs1041740 with microalbuminuria (oddsratio 1.51, 95% CI 1.10–2.10, p=0.01) in a dominant model, adjustedfor sex, age, BMI, duration of diabetes, HbA1c and presence of arterialhypertension. No association of the variant with macroalbuminuriawas observed. Allele and genotype frequencies of the other six SNPswere similar in the 3 groups of subjects.

3.2. SOD1 variants and incidence of renal events during follow-up

Next, we assessed the impact of allelic variations on the renal out-comes of the original DIABHYCAR study. A renal event defined as thedoubling of the serum creatinine levels or the requirement of

Table 1Clinical characteristics according to UAE status at baseline.

Normal UAE Microalbuminuria Macroalbuminuria p

N 607 2392 745Male sex (%) 47 74 71 b0.0001Age (year) 68±7 66±8a 66±8a b0.0001Duration of diabetes (year) 12±7 10±8a 11±8a,b b0.0001BMI (kg/m2) 28.6±4.8 29.2±4.5a 29.8±4.8a,b b0.0001HbA1c (%) 7.7±1.5 7.8±1.7 8.1±1.8a,b b0.0001Arterial hypertension (%) 53 55 65 b0.0001Systolic blood pressure (mm Hg) 141±13 144±14a 147±14a,b b0.0001Diastolic blood pressure (mm Hg) 79±7 82±8a 83±9a,b b0.0001Urinary albumin excretion (mg/l) 7±5 70±44a 758±973a,b b0.0001eGFR (ml/min) 76±19 78±20a 75±25 b0.0001Total cholesterol (mmol/l) 5.7±1.1 5.7±1.0 6.0±1.2a,b b0.0001HDL cholesterol (mmol/l) 1.36±0.34 1.32±0.35a 1.31±0.38a 0.006Triglycerides (mmol/l) 1.6±0.9 2.1±1.3a 2.5±1.7a,b b0.0001

Data expressed as mean±SD. Statistics are Pearson's chi-squared test and ANOVA with log-transformed data. pb0.05 is significant. Tukey Kramer HSD test following ANOVA: sig-nificantly different from values in subjects with normal UAE (a) or with microalbuminuria (b). Arterial hypertension: systolic blood pressure (SBP) >140 mm Hg and/or diastolicblood pressure (DBP) >90 mm Hg; or SBP and DBP below these values in the presence of antihypertensive medication and history of hypertension. Estimated glomerular filtrationrate (eGFR) calculated with the MDRD method. Normal UAE: DIABHYCAR_GENE cohort; microalbuminuria and macroalbuminuria: DIABHYCAR cohort.

361A.L. Neves et al. / Molecular Genetics and Metabolism 106 (2012) 359–365

hemodialysis or renal transplantation occurred in 77 cases (2.46%)during follow-up. Patients who presented a renal event as comparedto those who did not present a renal event during follow-up had atbaseline higher levels of HbA1c, UAE, creatinine and triglycerides,and presented a higher prevalence of arterial hypertension and previ-ous history of myocardial infarction (Table 3). They had a lower BMIand lower levels of eGFR and HDL cholesterol. We have not observedany association of the SNPs with the incidence of renal events duringfollow-up neither in univariate analyses (data not shown) nor incomplex models (Table 4). An alternative definition of renal eventas the decline of eGFR below 60 ml/min yielded similarly negative re-sults (data not shown). Moreover, baseline and follow-up levels of

Table 2Genotype frequencies of SOD1 variants according to UAE status at baseline.

Normal UAE(n=607)

Microalbuminuria(n=2392)

Macroalbuminuria(n=745)

rs9974610AA 0.656 0.675 0.640AG 0.313 0.292 0.327GG 0.031 0.033 0.033

rs2173962AA 0.907 0.922 0.916AG 0.087 0.076 0.081GG 0.006 0.002 0.003

rs10432782TT 0.797 0.795 0.781GT 0.188 0.192 0.204GG 0.015 0.013 0.015

rs2070424AA 0.866 0.867 0.859AG 0.134 0.128 0.135GG 0.000 0.005 0.006

rs1041740CC 0.458 0.425 0.464CT 0.431 0.454 0.421TT 0.111 0.121 0.115

rs17880135TT 0.891 0.891 0.895TG 0.105 0.107 0.102GG 0.004 0.002 0.003

rs202449TT 0.697 0.695 0.686TA 0.269 0.275 0.283AA 0.034 0.030 0.031

SNPs are sorted in 5′ to 3′ order. Odds ratio (OR) for microalbuminuria ormacroalbuminuriamodel. ORs were computed in logistic regression analyses, and adjusted for sex, age, BMI, d

UAE, creatinine and eGFRwere not significantly different across geno-types (data not shown).

3.3. SOD1 variants and incidence of cardiovascular events duringfollow-up

The incidence of cardiovascular events during the follow-up ofDIABHYCAR was 7.8%. Cardiovascular events comprised 95 cases ofmyocardial infarction and158 cases of cerebrovascular accident (stroke)reported in 245 subjects. Main baseline characteristics of subjects withor without a cardiovascular event during the follow-up are shown inTable 3. No association of the SNPs with the incidence of cardiovascular

OR (95% C.I.) formicroalbuminuria

p OR (95% C.I.) formacroalbuminuria

p

0.82 0.09 1.02 0.91(0.66–1.03) (0.77–1.33)

0.88 0.51 0.89 0.63(0.61–1.29) (0.57–1.42)

1.04 0.76 1.19 0.29(0.80–1.37) (0.86–1.64)

1.05 0.75 1.07 0.74(0.77–1.43) (0.73–1.56)

1.51 0.01 0.93 0.58(1.10–2.10) (0.72–1.21)

0.99 0.94 0.89 0.59(0.71–1.39) (0.58–1.35)

1.01 0.92 1.03 0.86(0.81–1.27) (0.77–1.36)

as compared to normal UAE determined for the minor alleles in a dominant (Xm vsMM)uration of diabetes, HbA1c and presence of arterial hypertension. p≤0.02 is significant.

Table 3Clinical characteristics at baseline according to the incidence of renal events, cardiovascular events or cardiovascular death during follow-up.

Renal events p Cardiovascular events p Cardiovascular death p

No Yes No Yes No Yes

N 3060 77 2892 245 2921 216Male sex (%) 73% 74% 0.89 73% 77% 0.13 73% 72% 0.69Age (year) 66±8 67±8 0.15 65±8 68±8 b0.0001 65±8 70±9 b0.0001BMI (kg/m2) 29.4±4.6 28.5±5.3 0.04 29.5±4.6 28.6±4.2 0.004 29.5±4.6 28.8±4.8 0.03Duration of diabetes (year) 10±8 13±9 0.04 10±8 11±8 0.05 10±8 12±7 b0.0001HbA1c (%) 7.8±1.7 8.5±2.1 0.007 7.8±1.7 8.3±2.0 0.0009 7.8±1.7 8.2±2.0 0.02Arterial hypertension (%) 56% 68% 0.04 56% 58% 0.50 55% 69% b0.0001UAE (mg/l) 216±604 769±1023 b0.0001 214±493 414±1441 0.0004 210±497 487±1494 b0.0001Creatinine (μmol/l) 89±20 103±41 b0.0001 89±20 94±20 0.0003 89±20 96±21 b0.0001eGFR (ml/min) 78±21 68±21 b0.0001 78±22 73±19 0.001 78±22 70±18 b0.0001Total cholesterol (mmol/l) 5.79±1.07 5.85±1.24 0.98 5.78±1.07 5.95±1.07 0.02 5.78±1.07 6.03±1.12 0.001HDL cholesterol (mmol/l) 1.32±0.35 1.21±0.34 0.002 1.32±0.36 1.28±0.31 0.25 1.32±0.36 1.28±0.33 0.18Triglycerides (mmol/l) 2.20±1.40 2.85±1.97 0.0003 2.22±1.43 2.19±1.21 0.83 2.22±1.43 2.35±1.24 0.02Previous MI (%) 5% 12% 0.03 5% 10% 0.005 5% 13% b0.0001

Data expressed as mean±SD. Statistics are Pearson's chi-squared test and ANOVA with log-transformed data. Arterial hypertension: systolic blood pressure (SBP) >140 mm Hgand/or diastolic blood pressure (DBP) >90 mm Hg; or SBP and DBP below these values in the presence of antihypertensive medication and history of hypertension. Estimated glo-merular filtration rate (eGFR) calculated with the MDRD method. pb0.05 is significant.

362 A.L. Neves et al. / Molecular Genetics and Metabolism 106 (2012) 359–365

events during follow-up was observed, neither in univariate analyses(data not shown) nor in complexmodels (Table 5). Stratification by car-diovascular phenotype (myocardial infarction or cerebrovascular acci-dent) yielded similarly negative results.

3.4. SOD1 variants and cardiovascular mortality during follow-up

Death attributed to cardiovascular causes occurred in 216 partici-pants (6.9%) during the follow-up of DIABHYCAR. Cardiovasculardeath comprised 13 cases of fatal myocardial infarction, 31 cases of

Table 4Genotype frequencies of SOD1 variants and incidence of a renal event during follow-up.

No renal event(n=3060)

Renal event(n=77)

HR(95% C.I.)

p

rs9974610AA 0.665 0.726 0.68 0.16AG 0.301 0.260 (0.40–1.13)GG 0.034 0.037

rs2173962AA 0.921 0.918 1.13 0.77AG 0.077 0.068 (0.52–2.96)GG 0.002 0.014

rs10432782TT 0.793 0.726 1.42 0.19GT 0.193 0.260 (0.83–2.35)GG 0.014 0.014

rs2070424AA 0.866 0.808 1.59 0.14AG 0.128 0.192 (0.85–2.77)GG 0.006 0.000

rs1041740CC 0.434 0.446 0.93 0.77CT 0.446 0.459 (0.58–1.48)TT 0.120 0.095

rs17880135TT 0.890 0.933 0.58 0.21TG 0.107 0.067 (0.20–1.31)GG 0.003 0.000

rs202449TT 0.692 0.730 0.83 0.47TA 0.277 0.243 (0.48–1.36)AA 0.031 0.027

SNPs are sorted in 5′ to 3′ order. Hazard ratios (HR) for the minor allele in a dominantmodel (Xm vs MM) determined in Cox proportional hazards survival regressionanalyses. HR adjusted for sex, age, BMI, duration of diabetes, HbA1c, total cholesterol,HDL cholesterol, triglycerides and urinary albumin excretion levels, presence of arterialhypertension and history of previous myocardial infarction at baseline and for thetreatment group during follow-up in the original DIABHYCAR study. p≤0.02 issignificant.

fatal cerebrovascular accident, 137 cases of sudden death and 35cases of cardiovascular death of other causes (congestive heart fail-ure, arrhythmias). Subjects who died of cardiovascular causes duringfollow-up as compared to the other participants were older and had alonger duration of diabetes at baseline (Table 3). They had higherlevels of HbA1c, total cholesterol, triglycerides, creatinine, UAE, andpresented a higher prevalence of arterial hypertension and previoushistory of myocardial infarction. They had lower BMI and lower levelsof eGFR. Cox proportional hazards survival regression analyses(Table 5 and Fig. 1) showed an association with decreased risk of car-diovascular death for the minor allele of rs9974610 (hazard ratio0.64, 95% CI 0.46–0.88, p=0.005) and with increased risk of cardio-vascular death for the minor alleles of rs10432782 (hazard ratio1.71, 95% CI 1.16–2.48, p=0.007) and rs1041740 (hazard ratio 1.78,95% CI 1.10–2.78, p=0.02). A nominal association (p=0.03) with in-creased risk was also observed for the minor allele of rs2173962.

4. Discussion

We have observed a modest association of rs1041740 of SOD1withthe prevalence of incipiens nephropathy (microalbuminuria) at base-line in a cohort of type 2 diabetic subjects followed prospectively forrenal and cardiovascular events. No allelic association was observedwith the prevalence of established nephropathy (macroalbuminuria)at baseline nor with the evolution of renal function during follow-up.Interestingly, no allelic association of SOD1was observed with the inci-dence of cardiovascular events (myocardial infarction or stroke), butthree variants (rs9974610, rs10432782 and rs1041740) were associat-ed with the risk of cardiovascular death during the follow-up.

A role for SOD1 in the mechanisms of protection against diabetic ne-phropathy is supported both by studies in animalmodels and by geneticstudies in type 1 diabetic patients. Overexpression of human SOD1 inmurine models of diabetes is associated with reduced renal cell injuryand reduced albuminuria, suggesting that increased SOD1activity atten-uates diabetic nephropathy [29,30]. Consistent with these observations,diabetic nephropathydevelops faster in Sod1 knocked-outmice [31]. Re-garding the genetic studies in type 1 diabetic patients, we have previ-ously observed in the prospective SURGENE cohort strong associationsof the minor T-allele of rs1041740 with the prevalences of incipiensand established/advanced nephropathy at baseline (OR 5.75 and 8.95,respectively), with the incidence of incipiens nephropathy duringfollow-up (HR 1.46) and with decreased eGFR throughout the study[21]. This association was confirmed in the cross-sectional study ofGENEDIAB andGenesis, two independent cohorts of type 1 diabetic sub-jects [21]. Moreover, Al Kateb and coworkers observed associations of

Table 5Genotype frequencies of SOD1 variants and incidence of cardiovascular events and cardiovascular death during follow-up.

Cardiovascular events Cardiovascular death

No (n=2892) Yes (n=245) HR (95% C.I.) p No (n=2921) Yes (n=216) HR (95% C.I.) p

rs9974610AA 0.667 0.661 1.06 0.67 0.663 0.725 0.64 0.005AG 0.300 0.309 (0.80–1.39) 0.305 0.230 (0.46–0.88)GG 0.033 0.030 0.032 0.045

rs2173962AA 0.921 0.924 0.87 0.58 0.922 0.858 1.80 0.03AG 0.077 0.076 (0.51–1.38) 0.076 0.135 (1.06–2.90)GG 0.002 0.000 0.002 0.007

rs10432782TT 0.791 0.792 0.99 0.95 0.794 0.699 1.71 0.007GT 0.195 0.195 (0.71–1.35) 0.192 0.287 (1.16–2.48)GG 0.014 0.013 0.014 0.014

rs2070424AA 0.865 0.861 1.07 0.72 0.867 0.838 1.38 0.18AG 0.130 0.131 (0.73–1.53) 0.127 0.162 (0.85–2.12)GG 0.005 0.008 0.006 0.000

rs1041740CC 0.434 0.436 0.92 0.70 0.433 0.430 1.78 0.02CT 0.445 0.458 (0.59–1.37) 0.449 0.408 (1.10–2.78)TT 0.121 0.106 0.118 0.162

rs17880135TT 0.891 0.889 1.07 0.75 0.894 0.891 0.85 0.50TG 0.106 0.111 (0.69–1.58) 0.103 0.109 (0.51–1.35)GG 0.003 0.000 0.003 0.000

rs202449TT 0.695 0.667 1.06 0.66 0.693 0.678 1.25 0.19TA 0.275 0.295 (0.80–1.40) 0.276 0.294 (0.89–1.72)AA 0.030 0.038 0.031 0.028

SNPs are sorted in 5′ to 3′ order. Hazard ratios (HR) for the minor allele in a dominant model (Xm vs MM) except for rs1041740 (recessive mm vs MX model) determined in Coxproportional hazards survival regression analyses. HR adjusted for sex, age, BMI, duration of diabetes, HbA1c, total cholesterol, HDL cholesterol, triglycerides (cardiovascular deathonly) and urinary albumin excretion levels, presence of arterial hypertension (cardiovascular death only) and history of previous myocardial infarction at baseline and for the treat-ment group during follow-up in the original DIABHYCAR study. p≤0.02 is significant.

363A.L. Neves et al. / Molecular Genetics and Metabolism 106 (2012) 359–365

the minor G-allele of rs17880135 with the incidence of both persistentmicroalbuminuria and severe nephropathy in patients from the DCCT/EDIC study [20]. Moreover, rs202446 (which is in linkage disequilibriumwith rs17880135)was associatedwith persistentmicroalbuminuria in afamily-based subset of the study [20]. We have confirmed the strong

Fig. 1. Rates of survival (Kaplan–Meier curves) from cardiovascular causes of deathduring the follow-up of DIABHYCAR according to SOD1 rs9974610 and rs10432782 ge-notypes. Data presented in a dominant model for the minor alleles.

association of the minor allele of rs17880135 with incipiens (OR 7.53),established (OR 6.04) and advanced nephropathy (OR 10.03) in cross-sectional analyses of GENEDIAB andGenesis cohorts [21]. As the statisti-cal power of our present study was adequate, it seems plausible that theweaker association of SOD1 variantswith diabetic nephropathy in type 2as compared to type 1 diabetic subjects has a biological rather than apurely statistical basis. It is tempting to speculate that the increased car-diovascularmortality associatedwith SOD1 variantsmight havemaskedthe renal effect of the variants in type 2 diabetic subjects. For instance,the T-allele of rs1041740 was associated both with diabetic nephropa-thy in type 1 diabetic subjects and with cardiovascular death in type 2diabetic subjects. In this regard, the frequency of the risk allele of all var-iants associated with cardiovascular death in the present study waslower in the DIABHYCAR cohort of type 2 diabetic subjects than in thecohorts of younger type 1 diabetic subjects where we observed associa-tions of SOD1with diabetic nephropathy [21].

Several lines of evidence demonstrate that oxidative stress plays arole in the pathogenesis of atherosclerosis. Increased production ofROS impairs endothelial function and endothelium-dependent vaso-dilation in humans, notably by inactivation of NO [32]. Antioxidantenzymes such as the SODs scavenge ROS and inhibit NO degradationin the vascular wall [19]. Oxidative stress also induces cell prolifera-tion, hypertrophy, apoptosis and inflammation in endothelial andsmooth cells of the vascular wall and in myocardial cells [19,32].Moreover, oxidative stress is associated with the metabolic syndromeand with many risk factors for atherosclerosis such as hypertension,dyslipidemia, obesity and diabetes [14]. It contributes to developmentof cardiovascular diseases, including stroke, myocardial infarction,angina pectoris, and heart failure [15,16,33].

To our knowledge, ours is thefirst study to assess the impact of SOD1variants on cardiovascular disease. In the light of the observed associa-tions of SOD1 variantswith the risk of death from cardiovascular causes,we have no definite explanation for the lack of association with

364 A.L. Neves et al. / Molecular Genetics and Metabolism 106 (2012) 359–365

myocardial infarction and/or stroke. Cases of sudden death and cases offatalmyocardial infarction or stroke comprised the largemajority of thecases of cardiovascular death. Our results are suggestive that the allelicvariations of SOD1modulate severe and/or acute cardiovascular events.Studies in animal models support a role for SOD1 in the mechanisms ofprotection against cardiovascular diseases and acute cardiovascularevents. Cardiac graft experiments in mice showed that overexpressionof human SOD1 in donor hearts attenuates apoptosis and the inflamma-tory response of ischemia–reperfusion injury during cardiac graft, miti-gates the development of graft coronary artery disease and promotescardiomyocyte survival [34]. Consistent with these observations,targeted inactivation of mouse Sod1 gene in donor hearts resulted ingrafts more susceptible to ischemia–reperfusion injury as compared towild-type donor hearts [35]. In addition, following 30 min of experi-mental ischemia, mouse with Sod1 inactivated grafts presented signifi-cantly higher myocardial necrosis and lower myocardial performance[35]. Moreover, a recent study in a rat model of cardiac arrest showedthat intravenous injection of PEP-1–SOD1 fusion protein significantlyattenuated cerebral ischemia–reperfusion damage [36]. These resultsin rodent experimental models suggest that SOD1 is an importantdefense element during ischemia in the heart and the brain.

The genetic mechanism behind the allelic associations that wehave observed is unclear. All SNPs that were studied are located out-side coding, splice or known regulatory regions of SOD1 [26]. Nonehave obvious functional properties that could predict deleterious ef-fects on enzyme function, which suggests that these associationsmay reflect linkage disequilibrium with one or more functional vari-ants elsewhere in the SOD1 region. Alternatively, it is not possible toexclude that our findings only reflect type 1 error (false positive re-sults) due to population stratification. However, we believe this ex-planation is unlikely. A type 1 error due to population stratificationis less likely to occur in a prospective study, much less prone to selec-tion bias, than in case control studies. Moreover, association with therisk of cardiovascular death was observed for three independent tagSNPs. Finally, our results are in agreement with data from the litera-ture that showed that SOD1 has a protective role against oxidativestress in the kidney and the heart.

In conclusion, we have observed an association of rs1041740 ofSOD1 with the prevalence of microalbuminuria, and associations ofrs9974610, rs10432782 and rs1041740 with the risk of death fromcardiovascular causes in patients with type 2 diabetes. Our resultsare consistent with a major role for SOD1 in the mechanisms of car-diovascular protection against oxidative stress in type 2 diabetic sub-jects. These results need to be confirmed in independent cohorts ofdiabetic and non diabetic subjects. Further studies are also neededto identify the functional variants that modulate renal and cardiovas-cular risk in type 1 and type 2 diabetic subjects.

Acknowledgments

ALN was supported by a scholarship awarded by the PortugueseSociety of Hypertension. This work was supported by a grant fromthe Association Diabète Risque Vasculaire (ADRV), France. The analy-sis and interpretation of the data have been done without the partic-ipation of these organizations.

References

[1] W.B. Kannel, D.L. McGee, Diabetes and cardiovascular disease. The Framinghamstudy, JAMA 241 (1979) 2035–2038.

[2] S.M. Haffner, S. Lehto, T. Ronnemaa, K. Pyorala, M. Laakso, Mortality from coro-nary heart disease in subjects with type 2 diabetes and in nondiabetic subjectswith and without prior myocardial infarction, N. Engl. J. Med. 339 (1998)229–234.

[3] B. Balkau, X. Jouven, P. Ducimetiere, E. Eschwege, Diabetes as a risk factor for sud-den death, Lancet 354 (1999) 1968–1969.

[4] B. Stengel, S. Billon, P.C. Van Dijk, K.J. Jager, F.W. Dekker, K. Simpson, J.D. Briggs,Trends in the incidence of renal replacement therapy for end-stage renal diseasein Europe, 1990–1999, Nephrol. Dial. Transplant. 18 (2003) 1824–1833.

[5] C.A. Jones, A.S. Krolewski, J. Rogus, J.L. Xue, A. Collins, J.H. Warram, Epidemic ofend-stage renal disease in people with diabetes in the United States population:do we know the cause? Kidney Int. 67 (2005) 1684–1691.

[6] P. Rossing, P. Hougaard, K. Borch-Johnsen, H.H. Parving, Predictors of mortality ininsulin dependent diabetes: 10 year observational follow up study, BMJ 313(1996) 779–784.

[7] C.T. Valmadrid, R. Klein, S.E. Moss, B.E. Klein, The risk of cardiovascular diseasemortality associated with microalbuminuria and gross proteinuria in personswith older-onset diabetes mellitus, Arch. Intern. Med. 160 (2000) 1093–1100.

[8] I.S. Young, J.V. Woodside, Antioxidants in health and disease, J. Clin. Pathol. 54(2001) 176–186.

[9] D.M. Guidot, J.M. McCord, R.M. Wright, J.E. Repine, Absence of electron transport(Rho 0 state) restores growth of a manganese-superoxide dismutase-deficientSaccharomyces cerevisiae in hyperoxia. Evidence for electron transport as a majorsource of superoxide generation in vivo, J. Biol. Chem. 268 (1993) 26699–26703.

[10] M. Brownlee, Biochemistry and molecular cell biology of diabetic complications,Nature 414 (2001) 813–820.

[11] J.L. Evans, I.D. Goldfine, B.A. Maddux, G.M. Grodsky, Oxidative stress and stress-activated signaling pathways: a unifying hypothesis of type 2 diabetes, Endocr.Rev. 23 (2002) 599–622.

[12] M. Brownlee, The pathobiology of diabetic complications: a unifying mechanism,Diabetes 54 (2005) 1615–1625.

[13] J.M. Forbes, M.T. Coughlan, M.E. Cooper, Oxidative stress as a major culprit in kid-ney disease in diabetes, Diabetes 57 (2008) 1446–1454.

[14] E. Hopps, D. Noto, G. Caimi, M.R. Averna, A novel component of the metabolic syn-drome: the oxidative stress, Nutr. Metab. Cardiovasc. Dis. 20 (2010) 72–77.

[15] D.D. Heistad, Oxidative stress and vascular disease: 2005 Duff lecture,Arterioscler. Thromb. Vasc. Biol. 26 (2006) 689–695.

[16] J.L. Mehta, N. Rasouli, A.K. Sinha, B. Molavi, Oxidative stress in diabetes: a mech-anistic overview of its effects on atherogenesis and myocardial dysfunction,Int. J. Biochem. Cell Biol. 38 (2006) 794–803.

[17] I. Fridovich, Superoxide radical and superoxide dismutases, Annu. Rev. Biochem.64 (1995) 97–112.

[18] S.L. Marklund, Extracellular superoxide dismutase and other superoxide dis-mutase isoenzymes in tissues from nine mammalian species, Biochem. J. 222(1984) 649–655.

[19] F.M. Faraci, S.P. Didion, Vascular protection: superoxide dismutase isoforms in thevessel wall, Arterioscler. Thromb. Vasc. Biol. 24 (2004) 1367–1373.

[20] H. Al-Kateb, A.P. Boright, L. Mirea, X. Xie, R. Sutradhar, A. Mowjoodi, B. Bharaj, M.Liu, J.M. Bucksa, V.L. Arends, M.W. Steffes, P.A. Cleary, W. Sun, J.M. Lachin, P.S.Thorner, M. Ho, A.J. McKnight, A.P. Maxwell, D.A. Savage, K.K. Kidd, J.R. Kidd,W.C. Speed, T.J. Orchard, R.G. Miller, L. Sun, S.B. Bull, A.D. Paterson, Multiple su-peroxide dismutase 1/splicing factor serine alanine 15 variants are associatedwith the development and progression of diabetic nephropathy: the DiabetesControl and Complications Trial/Epidemiology of Diabetes Interventions andComplications Genetics Study, Diabetes 57 (2008) 218–228.

[21] K. Mohammedi, S. Maimaitiming, N. Emery, N. Bellili-Munoz, R. Roussel, F.Fumeron, S. Hadjadj, M. Marre, G. Velho, Allelic variations in superoxidedismutase-1 (SOD1) gene are associated with increased risk of diabetic nephrop-athy in type 1 diabetic subjects, Mol. Genet. Metab. 104 (2011) 654–660.

[22] M. Lievre, M. Marre, G. Chatellier, P. Plouin, J. Reglier, L. Richardson, F. Bugnard, D.Vasmant, The non-insulin-dependent diabetes, hypertension, microalbuminuriaor proteinuria, cardiovascular events, and ramipril (DIABHYCAR) study: design,organization, and patient recruitment. DIABHYCAR Study Group, Control. Clin.Trials 21 (2000) 383–396.

[23] M. Marre, M. Lievre, G. Chatellier, J.F. Mann, P. Passa, J. Menard, Effects of low doseramipril on cardiovascular and renal outcomes in patients with type 2 diabetesand raised excretion of urinary albumin: randomised, double blind, placebo con-trolled trial (the DIABHYCAR study), BMJ 328 (2004) 495.

[24] M. Marre, J.P. Claudel, P. Ciret, N. Luis, L. Suarez, P. Passa, Laser immuno-nephelometry for routine quantification of urinary albumin excretion, Clin.Chem. 33 (1987) 209–213.

[25] A.S. Levey, J.P. Bosch, J.B. Lewis, T. Greene, N. Rogers, D. Roth, A more accuratemethod to estimate glomerular filtration rate from serum creatinine: a new pre-diction equation. Modification of diet in renal disease study group, Ann. Intern.Med. 130 (1999) 461–470.

[26] P. Milani, S. Gagliardi, E. Cova, C. Cereda, SOD1 transcriptional and posttranscrip-tional regulation and its potential implications in ALS, Neurol. Res. Int. 2011(2011) 458427.

[27] J.M. Cheverud, A simple correction for multiple comparisons in interval mappinggenome scans, Heredity 87 (2001) 52–58.

[28] D.R. Nyholt, A simple correction for multiple testing for single-nucleotide poly-morphisms in linkage disequilibrium with each other, Am. J. Hum. Genet. 74(2004) 765–769.

[29] P.A. Craven,M.F.Melhem, S.L. Phillips, F.R. DeRubertis, Overexpression of Cu2+/Zn2+

superoxide dismutase protects against early diabetic glomerular injury in transgenicmice, Diabetes 50 (2001) 2114–2125.

[30] F.R. DeRubertis, P.A. Craven, M.F. Melhem, E.M. Salah, Attenuation of renal injury indb/db mice overexpressing superoxide dismutase: evidence for reduced superoxide–nitric oxide interaction, Diabetes 53 (2004) 762–768.

[31] F.R. DeRubertis, P.A. Craven, M.F. Melhem, Acceleration of diabetic renal injury inthe superoxide dismutase knockout mouse: effects of tempol, Metabolism 56(2007) 1256–1264.

365A.L. Neves et al. / Molecular Genetics and Metabolism 106 (2012) 359–365

[32] Y. Higashi, K. Noma, M. Yoshizumi, Y. Kihara, Endothelial function and oxidativestress in cardiovascular diseases, Circ. J. 73 (2009) 411–418.

[33] T. Heitzer, T. Schlinzig, K. Krohn, T. Meinertz, T. Munzel, Endothelial dysfunction,oxidative stress, and risk of cardiovascular events in patients with coronary arterydisease, Circulation 104 (2001) 2673–2678.

[34] M. Tanaka, G.K. Mokhtari, R.D. Terry, L.B. Balsam, K.H. Lee, T. Kofidis, P.S. Tsao,R.C. Robbins, Overexpression of human copper/zinc superoxide dismutase(SOD1) suppresses ischemia–reperfusion injury and subsequent development

of graft coronary artery disease in murine cardiac grafts, Circulation 110(2004) II-200–II-206.

[35] T. Yoshida, N. Maulik, R.M. Engelman, Y.S. Ho, D.K. Das, Targeted disruption of themouse Sod I gene makes the hearts vulnerable to ischemic reperfusion injury,Circ. Res. 86 (2000) 264–269.

[36] Y.E. Zhang, S.Z. Fu, X.Q. Li, P. Chen, J.L. Wang, J. Che, J.M. Tang, S.Y. Chen, J.N. Wang,PEP-1–SOD1 protects brain from ischemic insult following asphyxial cardiac ar-rest in rats, Resuscitation 82 (2011) 1081–1086.