Multiple superoxide dismutase 1/splicing factor serine alanine 15 variants are associated with the development and progression of diabetic nephropathy: the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Genetics study

Abstract

Background: Despite familial clustering of nephropathy and retinopathy severity in type 1 diabetes, few gene variants have been consistently associated with these outcomes.

Research design and methods: We performed an individual-based genetic association study with time to renal and retinal outcomes in 1,362 white probands with type 1 diabetes from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study. Specifically, we genotyped 1,411 SNPs that capture common variations in 212 candidate genes for long-term complications and analyzed them for association with the time from DCCT baseline to event for renal and retinal outcomes using multivariate Cox proportion hazards models. To address multiple testing and assist interpretation of the results, false discovery rate q values were calculated separately for each outcome.

Results: We observed association between rs17880135 in the 3' region of superoxide dismutase 1 (SOD1) and the incidence of both severe nephropathy (hazard ratio [HR] 2.62 [95% CI 1.64-4.18], P = 5.6 x 10(-5), q = 0.06) and persistent microalbuminuria (1.82 [1.29-2.57], P = 6.4 x 10(-4), q = 0.46). Sequencing and fine-mapping identified additional SOD1 variants, including rs202446, rs9974610, and rs204732, which were also associated (P < 10(-3)) with persistent microalbuminuria, whereas rs17880135 and rs17881180 were similarly associated with the development of severe nephropathy. Attempts to replicate the findings in three cross-sectional case-control studies produced equivocal results. We observed no striking differences between risk genotypes in serum SOD activity, serum SOD1 mass, or SOD1 mRNA expression in lymphoblastoid cell lines.

Conclusions: Multiple variations in SOD1 are significantly associated with persistent microalbuminuria and severe nephropathy in the DCCT/EDIC study.

FIG. 1

Log SOD1 gene expression in EBV cells (cycle threshold number compared with b-actin; A and B), log serum SOD1 mass (ng/ml) (C and D), and log serum SOD activity (units/ml) (E and F) by genotype at rs17880135, separately for DCCT/EDIC probands (A, C, and E) and nondiabetic relatives (B, D, and F). The box and whisker plots have lines at the median, boxes at the inter-quartile values, whiskers at the nearest value not beyond 1.5 times the inter-quartile range, and lines beyond the whiskers are outliers. There are no significant differences by genotype for SOD1 expression for the probands (P = 0.42) or relatives (P = 0.60). For SOD1 mass, there was no difference between those homozygous for the minor allele (C/C) and those homozygous for the major allele (A/A) (P = 0.23 and 0.12, probands and relatives, respectively), but there was a borderline result for A/C vs. A/A (P = 0.08) in probands, but not relatives (P = 0.16). For SOD activity, there were no significant differences by genotype in either relatives or probands (P > 0.6). The P values reported for mRNA are for multiple linear regression models with additive genotype coding and adjustment for the matching variables sex, age, A1C, and AER or cystatin for the proband and relatives, respectively. The P values for SOD1 mass and SOD activity are from unconditional logistic regression with adjustment for the matching variables. The results for the same individuals by genotype at rs17881180, rs202446, rs9974610, and rs204732 are in online appendixes 30 and 31.

FIG. 2

Localization of SOD1 in normal human kidney by immunohistochemistry. A: The expression in cortex (top half) is considerably stronger than in the medulla (bottom half). B: In the cortex, SOD1 is expressed in glomeruli and proximal and distal tubules in both cytoplasm and nuclei. Vessels (arrow) show only weak expression in endothelial cells. C: In the medulla, there is weak expression in the loops of Henle and in collecting ducts (arrows), both nuclear and cytoplasmic. D: Within glomeruli, the strongest expressers of SOD1 are podocytes (P) and parietal epithelial cells (PE), whereas glomerular endothelial cells (E) show variable expression. Mesangial cells (M) are uniformly negative, and vascular endothelial cells (VE) show weak expression. (Please see http://dx.doi.org/10.2337/db07-1059 for a high-quality digital representation of this figure.)

FIG. 3

Distribution of haplotype frequencies across the SOD1 locus in a worldwide study. Stacked bars represent haplotype frequencies for the four SOD1 SNPs in order: rs17881180-rs4998557-rs1041740-rs17880135. Individual marker allele frequencies are provided in online appendix 20. The respective ALFRED site (http://alfred.med.yale.edu/) UIDs are SI003835T, SI003836U, SI003837V, and SI003838W; the four-site haplotype UID is SI006518U. Exact allele frequencies for each site and population and for the haplotypes in each population and descriptions of the populations, specific sample of each, and sample sizes can be found in ALFRED and online appendix 20.