A Disease Haplotype for Advanced Nephropathy in Type 2 Diabetes at the ACE Locus
Previous investigations of the ACE gene as a susceptibility factor for diabetic nephropathy have primarily focused on its insertion/deletion (Ins/Del) polymorphism. In a departure from these earlier studies, we used three tagging markers (A-5466C, T-3892C, and Ins/Del) at the ACE locus to test for disease haplotype associations. A case-control study design was used where case subjects were type 2 diabetic patients with advanced diabetic nephropathy, as indicated by the presence of proteinuria or chronic renal failure/end-stage renal disease, while control subjects were normoalbuminuric, despite >6 years of diabetes. None of the individual markers showed significant disease association when considered on their own. However, haplotype analyses revealed a near doubling in the prevalence of the A.T.D risk haplotype in case subjects (0.136) compared with control subjects (0.075) (P = 0.009), thus providing first evidence for a disease haplotype for advanced diabetic nephropathy at the ACE locus.
- ACR, albumin-to-creatinine ratio
- CRF, chronic renal failure
- ESRD, end-stage renal disease
- Ins/Del, insertion/deletion
- MAF, minor allele frequency
Among all the genes that have been evaluated as possibly conferring susceptibility to advanced diabetic nephropathy, ACE is the most scrutinized to date (1). Studies of ACE have primarily focused on examining whether the disease associated with the insertion/deletion (Ins/Del) of a 287-bp Alu sequence located in intron 16 (2). Emphasis has been placed on this particular polymorphism since its D-allele has been linked to high serum ACE levels, while the reverse has been reported for the I-allele (3). However, after the initial report by Marre et al. (4) who reported a protective effect of the II genotype on diabetic nephropathy, conflicting reports have since emerged over the past decade (1). A recent meta-analysis of 47 studies totaling 14,727 subjects found a modest association between the II genotype and risk reduction for diabetic nephropathy (1). Still, the magnitude of this association was limited, and this posed difficulty in its replication across individual studies.
The 28-kb ACE locus is characterized by low haplotype diversity and can be described by three frequent haplotypes: A, B, and C (5). Haplotypes A and B, which comprise different alleles at all common polymorphic sites, are associated with low and high circulating ACE activity, respectively. An ancestral 5′ recombination event in the chromosomal region between intron 5 and exon 8 has been advocated to explain the origin of haplotype C, which resembles haplotype A at the polymorphic sites in the 5′ region of the gene and resembles the B haplotype in the 3′ region (5). Subsequent study revealed an additional haplotype called D (identified with the new polymorphisms found downstream of the ACE transcribed sequences and sharing alleles with haplotype B and A), and this is consistent with a second ancestral 3′ recombination breakpoint ∼16 kb downstream from the intron 5 and exon 8 breakpoint (6). These two breakpoints are thought to flank the gene interval that harbors most of the genetic variation that determines ACE activity, with haplotypes B, C, and D having higher ACE activity than haplotype A (6). Aside from this, however, there is evidence for the presence of two quantitative trait loci acting additively on ACE levels (7), and a second interval that consists of part of ACE, including its promoter, has been mapped upstream of the 5′ recombination breakpoint. One may speculate that in certain pathological conditions, genetic variations in the promoter of the ACE may modulate gene expression in a tissue-specific manner. Of particular relevance to diabetes, it has been demonstrated that hyperglycemia increases ACE mRNA levels in primary cultures of rat mesangial cells in comparison with normal glucose conditions (8).
In view of the complexities modulating ACE expression, the case for a thorough haplotype analysis of ACE for association with diabetic nephropathy cannot be underscored. In this study, we utilized tagging markers to capture ACE haplotype diversity. These markers and their haplotypes were then tested for association with advanced diabetic nephropathy using a case-control study consisting of Caucasians with type 2 diabetes.
RESEARCH DESIGN AND METHODS
The Joslin Clinic provides care for about 16,000 patients with diabetes. The majority of these patients are Massachusetts residents referred to the clinic within 5 years of the diagnosis of diabetes, and most remain under the care of the clinic for life. Certain demographic and clinical information and most laboratory results are computerized and available for research purposes. The computer databases were used to identify patients eligible for our genetic studies. Between 1995 and 1999, a random sample of individuals with type 2 diabetes were recruited for the study of the genetics of nephropathy from among patients attending the Joslin Clinic. Diabetes has been classified as type 2 if it was diagnosed between ages 30 and 64 years and was treated for at least 2 years with diet or oral hypoglycemic agents. Only Caucasian patients aged <75 years at enrollment were eligible for the study.
Diagnosis of diabetic nephropathy.
Diabetic nephropathy was determined on the basis of the medical records of the Joslin Clinic (supplemented with records of other non-Joslin physicians, if necessary) and results of serum creatinine and routine urinary analyses, including measurements of the albumin-to-creatinine ratio (ACR) (9). Patients were classified as control subjects if they had type 2 diabetes with duration >6 years, and the ACR (in milligrams per gram) was <17 (for men) or <25 (for women) in at least two of the last three urine specimens spanning a 2-year interval preceding the examination for this study. Patients with microalbuminuria or intermittent proteinuria were not included in this study. Patients were considered case subjects if they had persistent proteinuria or if they had end-stage renal disease (ESRD) due to diabetic nephropathy. Persistent proteinuria was defined as two of three successive urinalyses positive by either reagent strip (>2+ on Multistix; Bayer, Elkhart, IN) or an ACR >250 (men) or >355 mg/g (women). Patients with persistent proteinuria and serum creatinine >2.0 mg/dl were considered case subjects with chronic renal failure (CRF).
Examination of study participants.
The Committee on Human Subjects of the Joslin Diabetes Center approved the protocols and informed consent procedures for this study. All patients randomly selected for the genetic studies were examined at the clinic. After consenting to participate in the study, each subject had a standardized physical examination and provided a diabetes history regarding its diagnosis, treatment, and complications. Each individual provided a blood sample for biochemical measurements and DNA extraction. Patient medical records were thoroughly reviewed to minimize the possibility of the presence of nondiabetic kidney disease, and patients were also directly questioned regarding whether they were ever diagnosed for nondiabetic kidney disease by doctors. A total of 291 (83%) patients of 350 selected case subjects with nephropathy were examined, and 167 (80%) patients of 210 selected control subjects were examined and had genomic DNA extracted and genotyped for all three ACE tagging markers.
Selection and genotyping of ACE tagging markers.
To explore genetic variation in the ACE locus, genotype data were downloaded from the CardioGenomics database (http://cardiogenomics.med.harvard.edu/home). The selection of tagging markers was performed by running the tagger program implemented in Haploview (www.broad.mit.edu). The criteria for r2 was set at >0.8. This meant that any marker that was not eventually chosen as a tagging marker was already strongly correlated with at least one of the tagging markers with r2 > 0.8. Genetic variation within the quantitative trait loci flanked by the two ancestral recombination breakpoints could be described by Ins/Del polymorphism (rs13447447). Ten frequent SNPs with minor allele frequency (MAF) >0.1 (rs8077276, rs4277405, rs4459609, rs1800764, rs4291, rs4293, rs4295, rs4298, rs4305, and rs4424958) located upstream of T1237C (rs4309), which flanks the 3′ side of the 5′ ancestral recombination breakpoint, could be tagged by two SNPs, A-5466C (rs4459609) and T-3892C (rs1800764). The structure of the common haplotypes A, B, and C could thus be resolved using tagging markers A-5466C, T-3892C, and Ins/Del.
The Ins/Del polymorphism was genotyped using PCR (25 μl reaction volume) on 20 ng of genomic DNA using 0.6 units of Taq polymerase in the presence of 1.5 mmol/l MgCl2 for 40 cycles with an annealing temperature of 68°C using primers CTGGAGACCACTCCCATCCTTTCT(F) and GATGTGGCCATCACATTCGTCAGAT(R). The I- and D-alleles were observed as bands of 490 and 190 bp, respectively. As preferential amplification of the D-allele can result in the mistyping of ID as DD, additional PCRs were performed using I-allele–specific primers TGGGACCACAGCGCCCGCCACTAC(F) and TCGCCAGCCCTCCCATGCCCATAA (R) with PCR annealing temperature of 70°C, yielding a 335-bp product. The T-3892C polymorphism was amplified using primers TAGTGTATATAGGGCTTGGTACTTTC(F) and TTAGATATGACACCAAAAGCATAAG (R) with annealing temperature of 58°C. This was followed by a second round of PCR using nested primers ATAGTGTATATAGGGCTTGGTAC(F) and AGAAGATATTTGCAAAGTATGTACTG(R) with annealing temperature of 60°C, giving a 114-bp product. In the presence of the C-allele, digestion with PstI resulted in fragments of 90 and 24 bp, while in the presence of the T-allele, the 114-bp product remained uncut. Marker A-5466C was amplified using ACCACAGATTCCGATGCCGC(F) and reverse CAAACCCTTTCTCTCCAGTGCTCAG(R) primers. Genotyping was next accomplished by template-directed dye-terminator incorporation (AycloPrime-FP SNP Detection System; Perkin-Elmer, Boston, MA) using probe GTGGGTTGACCTTGGCTGGGCATA with fluorescence polarization detected on a Wallac VICTOR2 Multilabel Plate Reader (Perkin-Elmer), used according to the manufacturer’s instruction.
Data on the study groups were compared using χ2 for categorical data. Student’s t test and the nonparametric Wilcoxon’s rank-sum test were used for continuous variables as appropriate (SAS system for Windows version 6.12; SAS Institute, Cary, NC). Estimation of haplotype frequencies and statistical significance of haplotype association with disease was evaluated according to Schaid et al. (10) using Haplo.stat software. An empirical P value of <0.05 was considered statistically significant. At α = 0.05, our study of each tagging markers has at least 80% power to detect an odds ratio (OR) of 1.5.
Selected clinical characteristics of the type 2 diabetic subjects are shown in Table 1. Age at diagnosis of diabetes, BMI, and glycemic control (as assessed by HbA1c values) were comparable between the two groups (P = NS), as was the sex proportions in each group (P = NS). Known duration of diabetes was significantly longer among case than control subjects (P < 0.0001). Case subjects had higher systolic and diastolic blood pressure compared with control subjects (P < 0.01), and they were more frequently treated with antihypertensive drugs (P < 0.0001). About one-half of the case subjects (52.6%) had already developed CRF/ESRD as a result of diabetic nephropathy at the time of enrollment in this study.
Single marker and haplotype analyses.
Genotype distributions for the three tagging markers did not deviate significantly from Hardy-Weinberg equilibrium when case and control subjects were analyzed together or as separate groups (P = NS). None of the markers differed significantly between case and control subjects in terms of overall allele and genotype distributions (P = NS) (Table 2). Regarding Ins/Del, homozygosity for the I-allele was not significantly associated with a reduced risk of diabetic nephropathy compared with D-allele carriage (OR 0.73 [95% CI 0.43–1.25]). However, this OR was similar to the pooled OR (0.68 [0.39–1.16]) reported for association with advanced diabetic nephropathy among type 2 diabetic Caucasians in a recent meta-analysis (1). Consideration of potentially important covariates, including sex, blood pressure, and diabetes duration, did not reveal any association between any of the tagging markers and diabetic nephropathy (data not shown). Genotype distributions were similar in case subjects regardless of whether they were proteinuric or had developed CRF/ESRD (data not shown).
In contrast, haplotype analyses using all three tagging markers, A-5466C, T-3892C, and Ins/Del, revealed that while there were three major D-allele–bearing haplotypes, only one (A.T.D) was more prevalent in case compared with control subjects (Table 3). This haplotype was nearly twice as common in case (0.136) as in control subjects (0.075), a difference that was statistically highly significant (P = 0.009). Two-marker haplotype analyses involving Ins/Del revealed that the A.D haplotype (formed with A-5466C) and the T.D haplotype (formed with T-3892C) was more common in case than control subjects with haplotype-specific P values of 0.064 and 0.005, respectively. Thus, all three tagging SNPs appear necessary to define the A.T.D risk haplotype. Haplotype distribution was comparable between case subjects who were proteinuric or have already developed CRF/ESRD (P = NS) (data not shown).
We performed a comprehensive haplotype analysis to evaluate whether ACE is indeed a susceptibility gene for diabetic nephropathy in patients with type 2 diabetes. Our approach represents a significant departure from studies reported over the last decade in that those studies were restricted only to the ACE Ins/Del marker located in intron 16 (1). Importantly, the salient finding in this study was that while the genotype distribution of ACE Ins/Del was not discrepant between case and control subjects, subsequent haplotype analysis using additional tagging markers yielded evidence for such a disease association.
A compelling rationale for emphasizing the Ins/Del polymorphism in past association studies has been that this marker accounted for half the variance of serum enzyme levels (3). Specifically, the D-allele has been linked with high serum ACE levels, while the converse is true for the I-allele. More recent studies have, however, demonstrated that a variety of D-allele haplotypes are linked to high ACE levels (11). Notably, our findings revealed that only one haplotype showed a significant disease association with diabetic nephropathy. At the same time, the single I-allele–bearing haplotype was slightly less frequent in case than control subjects, although this difference was not statistically significant (Table 3).
The implication of a specific D-allele haplotype for risk of diabetic nephropathy is clearly intriguing. Although it remains to be verified experimentally, one may speculate that this risk haplotype defines a specific combination of genetic variations that control ACE gene expression in particular cell types, possibly in the kidney. Another possibility is that this haplotype may be in tight linkage disequilibrium with a yet-to-be-identified functional polymorphism that directly influences gene expression. Of potential relevance is emerging evidence supporting the presence of a local renal renin-angiotensin system with documented expression of renin, angiotensinogen, angiotensin I and II, and angiotensin II type 1 receptors (8,12–15). In the case of ACE, an interesting observation is the reported aberrant expression of this gene in the glomerular tuft of the human diabetic kidney, whereas normal expression of ACE is typically confined to the brush border regions of renal proximal tubules (16). Albumin-stimulated expression of the type II receptor for the profibrotic cytokine transforming growth factor-β in proximal tubular cells has also been shown to be inhibited by losartan, an angiotensin receptor blocker (17).
Finally, a few possible limitations to our study may be contemplated. The first limitation is that among the case subjects, there could be some subjects who have chronic renal disease not attributed to diabetes. However, short of examining renal biopsies, diagnosis of such cases is difficult to establish even if retinopathy and hypertension are present. In view of the highly significant association between the A.T.D haplotype and diabetic nephropathy (P = 0.009), the impact of case heterogeneity, which would tend to diminish evidence of disease association, must be arguably limited. The second limitation relates to the differences in certain clinical characteristics (specifically hypertension) between case and control subjects. While this could be construed as inadequate matching, these clinical differences are expected since established diabetic nephropathy is characterized by the presence of hypertension. Matching case and control subjects based on this characteristic would result in findings that are difficult to interpret in the context of our current understanding of the etiology of diabetic nephropathy. The third limitation is the medium size of our study and the lack of confirmation of our findings in another population. While both points are valid, scarcity of DNA collections from other case-control studies on genetic determinants of diabetic nephropathy in type 2 diabetes prevents us from replicating our findings.
In conclusion, we have provided first evidence for the presence of a disease haplotype for diabetic nephropathy at the ACE locus. This genetic association was detected despite the absence of any significant finding when Ins/Del or any of the other tagging markers were evaluated on their own. Our study highlights the importance of considering the effects of allelic heterogeneity when investigating susceptibility genes for diabetic nephropathy and other complex diseases.
This research was supported by grants DK053534 and DK058549 from the U.S. National Institutes of Health and NMRC/0850/2004 and NMRC/1018/2005 from the National Medical Research Council, Singapore.
We thank Adam Smiles (Joslin Diabetes Center) for development and maintenance of genetic and phenotypic databases and Tan K.W. (National University of Singapore) for technical assistance in the laboratory.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received April 13, 2006.
- Accepted June 21, 2006.