Advanced glycation end-products (AGEs) may play an important role in the pathogenesis and progression of cardiovascular and renal complications of diabetes. Four putative AGE receptors (RAGEs), AGE-R1, AGE-R2, and AGE-R3 have been described. In this study, we scanned the sequence of the genes encoding these AGE receptors in 48 patients with type 1 diabetes and investigated the identified polymorphisms (n = 19) in 199 type 1 diabetic patients with nephropathy and 193 type 1 diabetic patients without nephropathy. Overall, none of the polymorphisms was strongly associated with nephropathy. The minor allele of a polymorphism located in the promoter region of the RAGE gene (C-1152A) conferred a weak protective effect (P < 0.05) and was associated with a longer duration of nephropathy-free diabetes (P = 0.08).
Microvascular lesions and accelerated atherosclerosis are the major causes of morbidity and early mortality in diabetes (1). There are data showing that genetic factors may influence the risk of developing both micro- and macrovascular complications in diabetic patients (2,3,4,5). The mechanisms by which chronic hyperglycemia exerts its adverse effects are not completely understood, but the role of advanced glycation end-products (AGEs), the reactive derivatives of nonenzymatic glucose-protein, and glucose-lipid condensation reaction are now well established (6). In vivo and in vitro studies indicate that AGEs have a role in the pathogenesis of diabetic nephropathy and the progression of renal failure (7). AGEs have a number of deleterious effects, including cross-linking of proteins, modification of matrix components, platelet aggregation, defective vascular relaxation, abnormal lipoprotein metabolism, and induction of cytotoxic pathways that may affect the pathophysiology of the vascular wall (7). The inhibition of AGE formation by aminoguanidine treatment improves the microvascular lesions found in diabetic animals within both the glomerulus and the retina (6) and prevents diabetes-induced arterial wall protein cross-linking (8).
AGE-specific cellular receptor (AGE-R) complexes have been identified on different cell types, including endothelial cells, smooth muscle cells, lymphocytes, and monocytes, where they mediate AGE removal as well as multiple biological responses, including the triggering of inflammatory genes. They include (see Table 1 for nomenclature) the AGE receptor (RAGE) (9), the Dolichyl-diphosphooligosaccharide-protein glycosyltransferase(AGE-R1) (10), the Protein kinase C substrate, 80KH phosphoprotein (AGE-R2) (11), and Galectin-3 (AGE-R3) (12). Recent studies in the NOD mouse further demonstrated that increased levels of AGE in kidneys were associated with reduced AGE-R sites on mesangial cells; this suggests that impaired kidney AGE-R function may contribute to renal disease in this type of mouse (13).
Because polymorphisms in genes encoding RAGEs might affect the genetic susceptibility to vascular complications of diabetes, we screened the four AGE-R genes for sequence variation and investigated the identified variants in type 1 diabetic patients with and without nephropathy.
To identify polymorphisms, DNA from 48 type 1 diabetic patients with or without nephropathy was amplified by polymerase chain reaction (PCR). A total of 62 pairs of primers were designed to generate overlapping fragments of <300 bp, spanning each exon of the four genes as well as their promoter and intronic sequences flanking exons (Table 1). Mutation detection was performed using single- strand conformational polymorphism (SSCP) analysis and direct sequencing of the allelic PCR fragments. A few polymorphisms may not have been detected by the SSCP method we used. SSCP variants were detected in the four genes, five in RAGE, five in AGE-R1, four in AGE-R2, and five in AGE-R3. Sequencing of the PCR fragments revealed the presence of missense mutations in exon 3 of RAGE (G82S), exon 15 of AGE-R2 (I452V), and exon 3 of AGE-R3 (P64H and T98P). Silent mutations were identified in exon 1 of RAGE (A2A), exon 3 of AGE-R1 (G120G), and exon 15 of AGE-R2 (L445 l). Five polymorphisms were found in the 5′ untranslated regions of RAGE and AGE-R3 genes, and the remaining ones were found in introns or 3′ untranslated regions (Table 2). All polymorphisms were further genotyped using allele-specific amplification in two groups of Caucasian type 1 diabetic patients from Denmark, matched for age, sex, and duration of diabetes. One group consisted of subjects with nephropathy (n = 199), and the other group consisted of subjects with longstanding diabetes without nephropathy (n = 193). Genotype and minor allele frequencies for the two groups of patients are reported in Tables 3 and 4 (when two polymorphisms were in complete concordance, only one of them was entered in the table). In patients with or without nephropathy, there was no significant deviation from Hardy-Weinberg equilibrium for any genotype. A weak association was observed for the C-1152A polymorphism located in the promoter region of the RAGE gene; the minor allele at this locus is less frequent in patients with than without nephropathy (odds ratio of heterozygotes 0.49 [0.25–0.94]; P = 0.031). In addition, a larger number of the glutamic acid repeats in the AGE-R2 gene was associated with the presence of nephropathy; long alleles corresponding to 10, 11, and 12 repeats were more frequent in the presence of nephropathy than were the short 8 and 9 repeat alleles (P = 0.041). A possible protective effect of the RAGE-1152A allele was consistent with a trend for longer duration of diabetes before onset of nephropathy in C-1152A heterozygotes (21.4 ± 9.6 years) as compared with C-1152C homozygotes (17.8 ± 7.3 years) (P = 0.08). In patients with nephropathy, RAGE-1152A heterozygotes had a higher mean level of plasma HDL cholesterol (n = 16; mean 1.75 ± 0.12) than did RAGE-1152C homozygotes (n = 172; 1.44 ± 0.04) (P = 0.02); a similar result was observed for plasma apolipoprotein (apo)-A1 (2.00 ± 0.09 and 1.72 ± 0.03, respectively; P = 0.005). No such association was observed in patients without nephropathy. Plasma levels of triglycerides, total cholesterol, apoB, LDL cholesterol, and lipoprotein-a did not differ according to RAGE-1152 genotypes (data not shown).
Polymorphisms in the RAGE gene have recently been identified and shown to not be associated with type 2 diabetes and macrovascular complications (14) or with diabetic microangiopathy (15). However, in these studies, only the coding regions of the RAGE gene were screened for mutations. Because RAGE expression and induction of cellular oxidant stress are closely linked in diabetic tissue, where AGEs accumulate, polymorphisms in the regulatory regions are therefore candidates for microvascular disease. We identified a C-1152A polymorphism in the promoter region of the RAGE gene. This region overlaps with the 3′ untranslated region of the PBX2 gene on chromosome 6 (16). A PBX2 pseudogene (φPBX2), which has a highly conserved sequence, is present on chromosome 3 (17); this could result in the erroneous assignment of the C-1152A polymorphism to the RAGE gene promoter. However, two arguments rule out this possibility: 1) the 3′ primer (ataggatcgggggctctgag) used to amplify the promoter region of the RAGE gene for genotyping the C-1152A and T-388A polymorphisms was chosen outside the region homologous to φPBX2; and 2) complete concordance, based on the observation of 346 common homozygotes and 46 heterozygotes, was observed between the C-1152A polymorphism and the RAGE G+196A polymorphism located downstream of the RAGE gene in a region not homologous to φPBX2. This implies that the C-1152A polymorphism is specific for the RAGE gene. Our results show that the C-1152A polymorphism is weakly associated with the presence of nephropathy in type 1 diabetes. This association might be spurious because of the number of statistical tests performed and should be confirmed in another study. If the association is real, the mechanism by which this C-1152A allele could exert its protective role is not clear. The region –1543/–587 has been identified as important in mediating promoter responsiveness to lipopolysaccharide stimulus, probably because of the presence of two necrosis factor-κB–like binding sites (18). However, the role of negative regulatory elements within this promoter, which explains variable basal expression and inducibility in diverse cell types, is not elucidated. The C-1152A polymorphism creates a potential Islet-1 transcription factor–like binding site that could modify RAGE expression. Interestingly, the apoA1 and HDL cholesterol levels are higher in the C-1152A carriers among diabetic patients with nephropathy, thus supporting a protective role of this polymorphism against deleterious effects of RAGE-AGEs interaction on lipid profile. Indeed, it has been shown that glycation of the protein component of HDL induces a greater transfer of cholesteryl ester from HDL to lighter-density lipoproteins (19). Further studies are necessary to understand better the function of the RAGE promoter. In addition, the association observed with the C-1152A polymorphism could be caused either by its complete concordance with the G+196A polymorphism located in the 3′ untranslated region of the gene (whose function is unknown) or by another polymorphism not yet identified in the HLA region on chromosome 6. Although no association or linkage with diabetic complications per se has been identified with markers located in the HLA region, putative loci for metabolic factors potentially affecting the occurrence of vascular complications, such as triglyceride levels or fasting glucose, have been described in Finnish families (20) and in Pima Indians (21).
The role of the glutamic acid repeat in exon 11 of the AGE-R2 gene in the pathogenesis of nephropathy is unknown. Little is known about the structure/function relationship of that acidic protein. However, similar polymorphic glutamic acid repeat elements have recently been described in the α2B-adrenergic receptor gene and have been shown to be associated with reduced basal metabolic rate in obese subjects (22). Because these repeat elements were shown to be important for agonist-dependent receptor desensitization, allelic proteins might have different properties. Further studies are necessary to understand fully the role of AGE-R2.
We identified several polymorphisms in the genes encoding putative AGE receptors, and none of these polymorphisms was strongly associated with diabetic nephropathy. However, the C-1152 polymorphisms located in the promoter of the RAGE gene should be investigated further in relation to this pathology.
RESEARCH DESIGN AND METHODS
The patients participating in this study have been previously described in detail (23). All albuminuric type 1 diabetes patients (n = 242) attending the outpatient clinic at the Steno Diabetes Center, Gentofte, Denmark, in 1993 who were >18 years of age and had their glomerular filtration rate measured during the same year were invited to participate in the study; 199 patients were included. A total of 193 type 1 diabetic patients with persisting normoalbuminuria recruited from our outpatient clinic and matched for sex, age, and duration of diabetes served as control subjects.
Genomic DNA was prepared from leukocytes. Primers were designed to span all exon and exon-intron boundary sequences as well as promoter regions from RAGE and AGE-R3. Mutation detection was performed using radioactive SSCP analysis with electrophoresis of samples on 6% acrylamide/7.5% glycerol nondenaturing gels at 1,200 V for 5 h at room temperature. Only this assay condition for SSCP analysis was used because we previously observed (24) that using three other conditions did not improve the sensitivity of the technique, which we estimate may be as high as 90% (25). DNA from patients presenting different SSCPs was reamplified by PCR, and subsequent sequencing in both directions using the big-dye sequencing kit (Pharmacia) was performed. Sample electrophoresis was performed on an automated DNA sequencer (ABI). Genotyping of relevant polymorphisms in 392 subjects was performed using allele-specific oligonucleotides (ASO). Primer sequences, ASOs, and assay conditions are available on our web site at http://genecanvas.idf.inserm.fr. For biallelic polymorphisms, comparison of genotype frequencies between patients with or without nephropathy was performed using a χ2 test with 2DF, except when one genotype was absent (1 df). For the two repeat polymorphisms, the differences in allele frequencies were considered. For RAGE/C-1152A where a difference was significant at P < 0.05, a logistic regression (SAS Proc Logist) adjusted for duration of diabetes was performed to calculate the odds ratio and 95% CI for the complication. Calculations were performed with SAS (Cary, NC).
Address correspondence and reprint requests to Dr. François Cambien, INSERM U525-SC7, 17 rue du Fer a moulin, 75005 Paris, France. E-mail: email@example.com.
Received for publication 20 June 2000 and accepted in revised form 23 January 2001
AGE, advanced glycation end-product; AGE-R, AGE-specific cellular receptor; apo, apolipoprotein; ASO, allele-specific oligonucleotides; PCR, polymerase chain reaction; RAGE, AGE receptor; SSCP, single-strand conformational polymorphism.