|Home | About | Journals | Submit | Contact Us | Français|
Four genome wide linkage scans for diabetic nephropathy have mapped susceptibility loci to chromosome 18q22.3-23 in the region of the carnosinase genes, CNDP1 and CNDP2. CNDP1 has been associated with diabetic nephropathy in Europeans and European Americans, but not African Americans. Individuals homozygous for a five tri-nucleotide repeat allele (5L; D18S880) are protected from diabetic nephropathy. We identified 64 variants after sequencing the exons, promoter, and 3′ UTR of CNDP1 and CNDP2 in African American and European American DNA samples. After scanning 44 of these variants, extensive genotyping of 12 SNPs and D18S880 was performed in 1025 African American cases with type 2 diabetes (DM)-associated end-stage renal disease (ESRD) and 1064 African American non-diabetic non-nephropathy controls to assess whether the carnosinase genes influence risk for DM-ESRD in African Americans. Evidence of association with DM-ESRD was seen with 2 SNPs: rs6566810 and rs4892247; 3 two-marker haplotypes: rs6566810 and rs17089362, rs17089362 and rs890336, and rs890334 and rs12717111 (global empirical p=0.0034, 0.0275, and 0.0002 respectively) and 3 three-marker haplotypes: rs6566810, rs17089362, and rs890336; rs890335, rs890334, and rs12717111; and rs890334, rs12717111, and D18S880 (global empirical p=0.0074, 1.5E-05, and 0.0032 respectively). The risk haplotypes (rs6566810, rs17089362 [A,T] and rs6566810, rs17089362, rs890336 [A,T,C]) were most strongly associated with DM-ESRD among African Americans in the non 5L-5L group. Variants in the carnosinase genes appear to contribute to diabetic nephropathy susceptibility in African Americans. Protection from diabetic nephropathy afforded by 5L-5L homozygosity in CNDP1 may be masked by the effects of additional risk haplotypes in CNDP1 and CNDP2.
As the number of individuals with type 2 diabetes mellitus (DM) has more than doubled in the past twenty-five years, the number with diabetic complications has also grown. In the United States, DM is the leading cause of end-stage renal disease (ESRD), accounting for 44.8% of incident cases (U.S. Renal Data System 2008). Familial clustering of diabetes-associated ESRD suggests a genetic component exists (Freedman et al. 1995; Quinn et al. 1996; Seaquist et al. 1989). Genome-wide linkage scans for diabetic nephropathy (DN) have provided consistent evidence for a DN susceptibility locus at chromosome 18q22.3-23 in Turkish and Pima Indian families (Vardarli et al. 2002), 18q21.1-23 in African American families (Bowden et al. 2004), and 18q22.3 in European American and American Indian families (Iyenger et al. 2007).
The carnosinase genes (carnosinase dipeptidase 1, CNDP1, and carnosinase 2, CNDP2) are located on chromosome 18 at 18q22.3. CNDP1 encodes a secreted serum carnosinase that degrades carnosine specifically, while CNDP2 encodes a nonspecific dipeptidase (Teufel et al. 2003). Carnosine (β-alanyl-L-histidine) has been shown to play a role as a free oxygen radical scavenger, having anti-glycating effects, inhibiting formation of advanced glycation end products (AGEs) (Hipkiss et al. 1998; Boldyrev 2000), and has been reported to influence glucose metabolism (Sauerhofer et al. 2007). D18S880, a microsatellite marker encoding a tri-nucleotide repeat in the signal peptide sequence of exon 2 of CNDP1, has previously been shown to be associated with DN in Europeans (Janssen et al. 2005) and European Americans (EA) (Freedman et al. 2007). These consistent observations in European-derived populations suggest a significant role for CNDP1 with risk for DN, but there have not been detailed studies in other ethnic groups, specifically African Americans. This is of interest since we previously reported evidence for linkage to DN on chromosome 18q in African American (AA) families (Bowden et al. 2004), suggesting a possible role for CNDP1 in AA DN.
This study was conducted under Institutional Review Board approval from Wake Forest University School of Medicine and adhered to the tenets of the Declaration of Helsinki. Identification, clinical characteristics, and recruitment of AA and EA patients and controls have been previously described (Yu et al. 1996). Briefly, 1025 unrelated AA patients with type 2 DM were recruited from dialysis facilities. Individuals with a history of ketoacidosis, or who developed diabetes prior to the age of 25 years and received continuous insulin therapy since diagnosis were assumed to have type 1 diabetes and were excluded. Type 2 DM was diagnosed in African Americans who reported developing DM after the age of 25 years and who did not receive only insulin therapy since diagnosis. Cases had DM diagnosed at least 5 years before initiating renal replacement therapy, background or greater diabetic retinopathy, and/or ≥100 mg/dl proteinuria on urinalysis in the absence of other causes of nephropathy. Another 1064 unrelated AA controls and 39 unrelated EA controls without a current diagnosis of diabetes or renal disease were recruited from the community and internal medicine clinics. All DM-ESRD cases and non-diabetic controls were born in North Carolina, South Carolina, Georgia, Tennessee, or Virginia. DNA extraction was performed using the PureGene system (Gentra Systems, Minneapolis, MN). DNA was also obtained from 44 Yoruba Nigerians from the National Institute of General Medical Sciences (NIGMS) Human Variation Collection (Coriell Cell Repositories, Camden, NJ).
The tri-nucleotide repeat polymorphism, D18S880, in exon 2 of CNDP1 was genotyped by fragment length analysis on an ABI Prism DNA Analyzer 3700 (Applied Biosystems Incorporated, Foster City, CA) in a manner similar to that described by Janssen et al. (Janssen et al. 2005). Fragment length was determined using ABI Prism GeneMapper software v3.0 (Applied Biosystems, Inc.). 56 duplicate samples were run for molecular quality control purposes. Of the replicate pairs, 100% were concordant.
The exons of CNDP2 and CNDP1 were amplified using intronic primers; 1000bp of the promoter regions and 3′ UTRs of CNDP2 and CNDP1 were also amplified. Primer sequences are available on request. All products were sequenced in 3 AAs and 3 EAs with two copies of the 5L allele, and in 3 AAs and 3 EAs with one or no copies of the 5L allele using BigDye Terminator v. 1.1 Cycle Sequencing Kits (Applied Biosystems Inc., Foster City, CA) and an Applied Biosystems 3730×l DNA Analyzer (Applied Biosystems Inc., Foster City, CA). Sequences were viewed and compared using Sequencher software version 4.6 (Gene Codes Corporation, Ann Arbor, MI). SNPs identified from sequencing were genotyped using the MassARRAY system from Sequenom, Inc. (Sequenom, San Diego, CA) (Buetow et al. 2001) in 380 AA case subjects and 364 AA control subjects. SNPs CNDP2 rs6566810, rs17089362, rs890336, rs15165, rs890335, rs890334, rs12717111, rs12717112; CNDP1 rs4892247, rs11659237, rs4891564, and rs2887 demonstrated initial evidence of association in these 380 cases and 364 controls and were then genotyped in the full set of 1025 AA DM-ESRD cases and 1064 AA non DM-ESRD controls. Primer sequences are available on request. Genotyping success rates were >92.1% in case subjects and >92.5 % in control subjects. Concordance among blind duplicate samples was 100%. SNPs rs6566810, rs17089362, rs890335, 890334, and rs12717111 were sequenced in 26 AA controls and 22 AA DM-ESRD cases to verify the accuracy of the genotype calls. The region was PCR amplified and sequenced using BigDye Terminator v. 1.1 Cycle Sequencing Kits (Applied Biosystems Inc, Foster City, CA) and an Applied Biosystems 3730×l DNA Analyzer (Applied Biosystems Inc, Foster City, CA). Sequences were viewed and compared using Sequencher software version 4.6 (Gene Codes Corporation, Ann Arbor, MI). Primer sequences are available on request.
Seventy Ancestral Informative Markers (AIMs) were genotyped in 1025 AA case subjects, 1064 AA control subjects, 44 Yoruba Nigerians, and 39 EAs using Illumina's Custom Genotyping Service or the MassARRAY system (Sequenom, San Diego, CA) (Buetow et al. 2001).
The percentage of individuals homozygous for the protective 5 leucine repeat at D18S880 (5L-5L, CNDP1 Mannheim) was determined in all subjects. The allelic and genotypic frequencies were compared between the case and control groups using the statistical analysis program CLUMP (Sham and Curtis 1995). CLUMP uses a Monte Carlo approach to assess significant departure of observed values in a contingency table from the expected values.
SNPs were tested for departure from Hardy Weinberg proportions using an exact test of HWE proportions for the combined group of cases and controls, and then for cases only, and controls only (Wigginton et al. 2005). Those SNPs out of HWE were noted, but still included for the genotypic analysis. Haplotype block structure was established using Haploview 4.1 (Barrett et al. 2005), defining blocks using the method from Gabriel et al. (Gabriel et al. 2002).
Unadjusted measures of linkage disequilibrium (LD) and association were assessed using the software SNPGWA (Harley et al. 2008; http://www.phs.wfubmc.edu). SNPGWA computes LD statistics, D′ and r2, for each pair of tandem SNPs. SNPGWA also performs multiple tests of association including the overall 2-degree of freedom test (genotype), dominant model, recessive model, additive model (Cochran-Armitage trend test), and the corresponding lack of fit to the additive model. Odds ratios, 95% confidence intervals, and p-values were computed for each model of association. Haplotypic association was also calculated for sets of tandem SNPs using the expectation-maximization (EM) algorithm. Tests were first performed using 1,000 permutations. Permutations were increased to 10,000-1,000,000 where association tests indicated possible significance (empirical p-value < 0.10). Individual two-marker and three-marker haplotype frequencies, p-values, odds ratios, and 95% confidence intervals were computed using the program Dandelion (http://www.phs.wfubmc.edu). Dandelion is unable to calculate odds ratios when the haplotype frequency is less than or equal to 0.001 in cases and/or controls. SNPs rs12717112 and rs4891564 were removed from haplotype analysis due to inconsistence with HWE in the combined population and low MAF respectively.
Ancestral allele frequencies were estimated from the results of the AIMs genotyped in the Yoruba Nigerians and the European Americans. Individual ancestral proportions were generated for each subject using FRAPPE (Tang et al. 2005), an EM algorithm, under a two-population model. The individual ancestral estimates were used as covariates in the association analyses with the program SNPADDMIX, a module of SNPGWA. SNPADDMIX uses logistic regression tests of additive, dominate, recessive, and two degree of freedom genetic models in order to include adjustments for individual estimates of African ancestry.
D18S880 was converted to a biallelic marker for analysis in Haploview, SNPGWA, SNPADDMIX, and Dandelion. The following alleles were classified as the major allele: 4L, and 5L; while, any other allele (6L, 7L, and 8L) was classified as the minor allele. The study population was also stratified by their D18S880 genotype: subjects with the 4L-4L, 4L-5L, and 5L-5L genotypes were in the 5L-5L only group; whereas, subjects with one or no copies of the 5L allele were in the non 5L-5L group.
All 12 exons of the CNDP2 gene and all 12 exons of the CNDP1 gene were sequenced using intronic primers in 6 AA and 6 EA DNAs enriched for the 5L-5L genotype. Additionally, 1000 base pairs (bp) of the promoter region, and the 3′UTR were sequenced in the same individuals. Sixty four variants were identified: 13 coding and 51 non-coding. Results are shown in Table 1.
Forty four variants identified from sequencing and the previously associated marker D18S880 were initially tested for association in 380 AA DM-ESRD cases and 364 AA non-diabetic non-nephropathy controls (Table 2). There were two CNDP2 SNPs that were monomorphic in AAs: 1358 G>A (R399K), and rs3206931 and four CNDP1 SNPs that were monomorphic in AAs: IVS4-122 G>A, rs4891562, IVS9+5 G>A, and IVS9+63 G>A. Four CNDP2 SNPs showed nominal evidence of association under the 2 degree of freedom (2 df) test (p-values 0.002 to 0.074; Table 2). There was also evidence of haplotypic association across CNDP2. Six 2-marker haplotypes and eight 3-marker haplotypes were nominally associated (p-values 0.001 to 0.069; Table 2). In CNDP1, there was evidence of association with three SNPs (p-values 0.028 to 0.064; Table 2) and nominal evidence of association with three 2-marker haplotypes and one 3-marker haplotype (p-values 0.057 to 0.077; Table 2). The associations observed in CNDP1 were at the 3′ end of the gene spanning intron 9 to the 3′ UTR.
CNDP1 and CNDP2 polymorphisms demonstrating nominal evidence of association with DM-ESRD from this initial analysis in 380 DM-ESRD cases and 364 controls were carried forward for detailed analysis in a larger African American sample composed of 1025 AA cases with DM-ESRD and 1064 AA non-diabetic non-nephropathy controls. Characteristics of cases and controls are summarized in Table 3. Cases were more often female (62%) than controls (54%). Although the control group was younger, controls were older (mean age 51.4 years) than the age at DM diagnosis in DM-ESRD cases (mean age at DM 41.1 years). The estimated mean proportion of African ancestry was 0.823 in cases and 0.797 in controls.
Table 4 contains allele and genotype frequencies at the D18S880 marker. In addition to the previously reported alleles for D18S880 (5L, 6L, and 7L by Jannsen et al. and Freedman et al.) a 4L repeat and an 8L repeat were infrequently observed in AAs. As previously reported, significant differences in allele frequency were not seen in AA cases with DM-ESRD versus controls (p=0.641; Table 4A). No significant difference in frequency of the 5L-5L genotype was detected between groups (p=0.842; Table 4B).
In addition to D18S880, 12 SNPs were successfully genotyped in the 1025 AA DM-ESRD cases and 1064 AA controls. Two of these SNPs (rs15165 and rs12717112) were inconsistent with Hardy Weinberg Equilibrium (HWE) proportions (p=0.0046 and 0.0049, respectively) in cases and SNP rs6566810 was inconsistent in controls (p=0.0055) (Supplementary Table 1). In the combined population, rs12717112 was inconsistent with expected HWE proportions (p=0.0008; Supplementary Table 1). Minor allele frequencies (MAF), and genotype frequencies and counts for each SNP are shown in Supplementary Table 1. One SNP, rs4891564 had MAFs below 5% in both cases and controls (Supplementary Table 1).
Admixture-adjusted genotypic results for SNPs associated with DM-ESRD under the 2 df test are shown in Supplementary Table 2. Evidence for association with DM-ESRD was observed with one CNDP2 SNP rs6566810 and one CNDP1 SNP rs4892247 (p=0.0049 and 0.0345 respectively; Supplementary Table 2). Results are further detailed in Table 5. SNP rs6566810 was associated with DM-ESRD in the entire population (p=0.0015 recessive, odds ratio [OR] (95% CI) 2.29 (1.37-3.81)); as was rs4892247 (p=0.0146 recessive, OR (95%CI) 0.73 (0.57-0.94)). SNP rs6566810 was out of HWE in the control population (p=0.0055; Supplementary Table 1). The association tests performed are robust to HWE departures, given correct genotyping. In order to ensure that this slight departure was not due to an error in genotyping, genotype calls were confirmed by direct DNA sequencing of 26 controls and 22 cases. Genotypes from DNA sequencing were 100% concordant with those from the Sequenom MassArray system.
When the population was stratified by genotype at the D18S880 marker, there was no evidence of association with rs6566810 and DM-ESRD in the 5L-5L group (p=0.1156, 2 df test; Table 5). However, rs6566810 remained associated in the non 5L-5L group (p=0.0496, 2 df test) and was associated in the recessive model (p=0.0152, OR (95% CI) 2.14 (1.16-3.96); Table 5). SNP rs4892247, was trending to association in the 5L-5L group (p=0.0587, 2 df test) and was significantly associated under the dominant model (p=0.0181, OR (95% CI) 0.70 (0.53-0.94)). In the non 5L-5L group, rs4892247 was associated with DM-ESRD (p=0.0485 2 df test; and p=0.0166 recessive, OR (95% CI) 0.69 (0.50-0.93); Table 5).
Haplotype analyses of all two-marker and three-marker haplotypes are shown in Supplementary Table 3. In all subjects, there were three two-marker haplotypes that were significantly associated with DM-ESRD: rs6566810 and rs17089362, rs17089362 and rs890336, and rs890334 and rs12717111 (global empirical p=0.0034, 0.0275, and 0.0002 respectively; Table 6). Additionally, there were three three-marker haplotypes that were significantly associated with DM-ESRD in all subjects: rs6566810, rs17089362, and rs890336; rs890335, rs890334, and rs12717111; and rs890334, rs12717111, and D18S880 (global empirical p=0.0074, 1.5E-05, and 0.0032 respectively; Table 6). In the non 5L-5L homozygotes, two two-marker haplotypes were associated: rs6566810 and rs17089362, and rs890334 and rs12717111, (global empirical p=0.0424, and 0.0071 respectively; Table 6), and two three-marker haplotypes were associated: rs890335, rs890334, and rs12717111; and rs890334, rs12717111, and D18S880 (global empirical p= 0.0012, and 0.0222 respectively; Table 6). In contrast, when limited to 5L-5L homozygotes no significant evidence of haplotypic association was seen (Table 6).
The individual haplotype frequencies for all subjects, the 5L-5L only group, and the non 5L-5L group are shown in Table 7. For two of the associated two-marker haplotypes (rs6566810, rs17089362; and rs890334, rs12717111) and all of the associated three-marker haplotypes the most associated individual haplotypes were rare, occurring in 1% or less of the total population. In order to confirm that these haplotypes did exist and were not due to genotyping error, SNPs rs6566810, rs17089362, rs890335, rs890334, and rs12717111 were sequenced in 26 controls and 22 cases. All genotypes obtained from DNA sequencing exactly matched those from the Sequenom MassArray system.
In all subjects, two-marker haplotypes rs6566810, rs17089362 [A,T] and rs17089362, rs890336 [T,C] were associated with risk (p-values, OR (95% CI) = 0.008, 3.21 (1.67-6.17) and 0.038, 1.37 (1.11-1.68) respectively; Table 7A). Three-marker haplotype rs6566810, rs17089362, rs890335 [A,T,C]; was also associated with risk (p-values, OR (95% CI) = 0.0004, 9.11 (3.12-26.59); Table 7A). However, two-marker haplotype rs890334, rs12717111 [A,A] and three-maker haplotypes rs890335, rs890334, rs12717111 [A,A,A] and rs890334, rs12717111, D18S880 [A,A,5L-5L] were protective (p-values, OR (95% CI) = 0.006, 0.16 (0.05-0.47); 0.002, N/A and 0.0155, N/A respectively; Table 7A).
Upon stratifying the population by D18S880 genotype, only one risk haplotype remained associated in the 5L-5L only group (rs6566810, rs17089362, rs890336 [A,T,C] p-value, OR (95% CI) = 0.041, 4.93 (1.44-16.79); Table 7B). While the protective haplotypes were not associated, they all trended towards association (p-values = 0.084-0.092; Table 7B). In the non 5L-5L group two risk haplotypes remained associated: rs6566810, rs17089362 [A,T] and rs6566810, rs17089362, rs890336 [A,T,C] (p-values, OR (95% CI) = 0.021 4.07 (1.61-10.30), and 0.004 N/A respectively, Table 7C). There was also evidence of association with the three protective haplotypes in the non 5L-5L group: rs890334, rs12717111 [A,A]; rs890335, rs890334, rs12717111 [A,A,A]; and rs890334, rs12717111, D18S880 [A,A,5L-5L] (p-values, OR (95% CI) = 0.028, 0.17 (0.05-0.63); 0.011, N/A; 0.039, N/A respectively, Table 7C).
There was evidence of high linkage disequilibrium (LD) in both carnosinase genes (Supplementary Table 3). Four SNPs in CNPD2 were in high LD with pairwise D′ values of 0.994 (r2=0.345) between rs15165 and rs890335, D′=0.996 (r2=0.980) between rs890335 and rs890334, and D′=0.946 (r2=0.888) between rs890334 and rs12717111 (Supplementary Table 3). Similarly, three SNPs in CNDP1 were in high LD with pairwise D′ values of 0.935 (r2=0.419) between SNPs rs4892247 and rs11659237, and D′=0.997 (r2=0.868) between rs11659237 and rs2887 (Supplementary Table 3). D18S880, in exon 2 of CNDP1, was in low LD with the two nearest markers rs12717111 (D′=0.235; r2=0.006) and rs4892247 (D′=0.054; r2=0.003) (Supplementary Table 3).
Genome wide scans for linkage have demonstrated that a gene(s) contributing to DN maps to chromosome 18q22.3-23 in Turkish, Pima Indian, EA, and American Indian families (Vardarli et al. 2002; Iyengar et al 2007) and 18q21.1-23 in AA families (Bowden et al. 2004). The carnosinase genes, CNDP1 and CNDP2, reside on chromosome 18 at 18q22.3. CNDP1 was associated with DN in a small European population (Janssen et al. 2005) and a larger EA population (Freedman et al. 2007). Janssen et al. found the shortest allelic form observed in European-derived populations with 5 leucine repeats (5L; the CNDP1 Mannheim allele) significantly more often in diabetic controls without DN, compared to cases with DN. The observed association was only seen in those individuals homozygous for the CNDP1 Mannheim allele. The 5L-5L genotype was also associated with lower serum carnosinase concentrations (Janssen et al. 2005). Freedman et al. observed that the 5L-5L genotype was significantly more frequent in non-diabetic controls than subjects with DN-ESRD. The number of leucine repeats in this signal peptide was associated with secretion of serum carnosinase in Cos-7 transfected cells. There was a significant increase in the level of secreted carnosinase in cells containing constructs with six, seven, and eight leucine repeats compared to those with four and five leucine repeats (Riedel et al. 2007). Recently, Krolewski et al. claimed to have excluded the carnosinase genes and D18S880 as a cause of type 1 DN (Wanic et al. 2008). While association was not observed with D18S880 frequencies between three study groups (diabetic subjects with normoalbuminuria, diabetic cases with proteinuria, and diabetic cases with ESRD), there was modest association observed in the promoter region of CNDP1 when comparing normoalbuminuric controls versus ESRD cases (3 promoter SNPs had p=0.01) (Wanic et al. 2008).
We further investigated CNDP1 and the adjacent CNDP2 gene by genotyping 12 additional SNPs in 1025 AA type 2 DM-ESRD cases and 1064 AA non-diabetic controls. Evidence of association was detected with two SNPs under the 2 df test in all subjects. SNP rs6566810, which was associated with risk, remained associated in the non 5L-5L group but not in the 5L-5L only group. SNP rs4892247, which was associated with protection against DM-ESRD, also remained associated in the non 5L-5L group and was trending towards association in the 5L-5L only group under the 2 df test. Genotypic associations were seen after adjusting for African ancestry.
Significant evidence of association was also seen with three two-marker haplotypes and three three-marker haplotypes; however, most of the associated haplotypes were rare. We saw association with both risk and protective haplotypes. Three risk haplotypes; rs6566810, rs17089362 [A,T]; rs17089362, rs890336 [T,C] and rs6566810, rs17089362, rs890336 [A,T,C] were associated in all subjects. After splitting the population by 5L-5L status, two of the risk haplotypes (rs6566810, rs17089362 [A,T] and rs6566810, rs17089362, rs890336 [A,T,C]) remained associated in the non 5L-5L group; whereas only one risk haplotype remained associated in the 5L-5L only group (rs6566810, rs17089362, rs890336 [A,T,C]). There were also three protective haplotypes associated in all subjects: rs890334, rs12717111 [A,A]; rs890335, rs890334, rs12717111 [A,A,A]; and rs890334, rs12717111, D18S880 [A,A,5L-5L]. All three of these protective haplotypes remained associated in the non 5L-5L group, and were trending towards association in the 5L-5L only group.
D18S880 was not associated with DM-ESRD in our African-American population under both allelic and genotypic models. However, when the population was stratified by number of leucine repeats at D18S880, the risk allele and haplotypes often remain associated in the non 5L-5L group, and not in the 5L-5L only group. With the protective allele and haplotypes there is still evidence of association in the non 5L-5L group, and weaker evidence of association (or a trend) in the 5L-5L only group. This suggests that the risk haplotypes may be masked by the protection afforded by the 5L-5L genotype and that additional SNPs and haplotypes could be protective. However, as previously reported, D18S880 is not contained in any haplotype block (Freedman et al. 2007). In addition, D18S880 was not in LD with any other marker genotyped in our population (Supplementary Table 3). The identification of risk haplotypes for DM-ESRD that may counteract the protection afforded by 5L-5L homozygosity is a potentially important observation. This finding could partially explain why African American 5L-5L homozygotes do not appear to be protected from DM-ESRD, in contrast to Europeans and EAs.
Our study design prevents us from distinguishing whether the observed associations were with type 2 DM, diabetic nephropathy, or both. However, the high rates of nephropathy in the diabetic AA population make it difficult to recruit large numbers of diabetic subjects lacking nephropathy after 10 years diabetes duration. Recruitment of a large African American type 2 diabetes non-nephropathy control group will be an important step in confirming that our results are associated with susceptibility to diabetic nephropathy and not diabetes per se. The previous association with D18S880 in Europeans and EAs was seen with DN and not diabetes (Janssen et al. 2005; Freedman et al. 2007). CNDP1, which encodes a serum secreted dipeptidase (Teufel et al. 2003), is present in human kidney and expressed at significantly greater levels in podocytes and renal parietal epithelial cells in subjects with DN (Janssen et al. 2005). CNDP2, which encodes a non-specific dipeptidase, is ubiquitously expressed. However, there is stronger expression in human kidney and liver (Teufel et al. 2003). Therefore, variants in CNDP2 may also play a role in DN.
This study represents a comprehensive evaluation of both carnosinase genes, CNDP1 and CNDP2, in a large African American population with type 2 diabetes associated ESRD. While we did not detect association with D18S880, we did find association with variants in CNDP2 and other variants in CNDP1 with DM-ESRD. It was necessary to split the AA study group into potentially protected 5L-5L homozygotes and less protected non 5L-5L homozygotes in order to detect additional SNPs and haplotypes that may confer risk for diabetic nephropathy in this ethnic group. We conclude that both CNDP1 and CNDP2 play roles in DN susceptibility in the high risk African American population.
This research was supported in part by NIH grants RO1 DK066358 (DWB), R01 DK053591 (DWB), RO1 DK070941 (BIF), and in part by the General Clinical Research Center of the Wake Forest University School of Medicine grant M01 RR007122.