Minor Effect of GLUT1 Polymorphisms on Susceptibility to Diabetic Nephropathy in Type 1 Diabetes

Individuals with type 1 diabetes were recruited for the study from among patients attending the Joslin Clinic in the decade 1991–2000. Diabetes was classified as type 1 if it was diagnosed before age 30 years and continuous treatment with insulin began within 1 year of diagnosis. During 1991–1993, we enrolled a 50% sample of Joslin Clinic attendees with type 1 diabetes in the age range 15–44 years (n = 1,598) in the Natural History of Microalbuminuria Study (17). On the basis of multiple measurements of the albumin-to-creatinine ratio (ACR) in random urine specimens during a 2-year observation period, patients were classified as having normoalbuminuria (n = 1,080), persistent microalbuminuria (n = 312), or advanced nephropathy (persistent proteinuria or end-stage renal disease [ESRD], n = 206) (17).

As unrelated control subjects for genetic studies on diabetic nephropathy, we selected Caucasian patients with normoalbuminuria and type 1 diabetes duration ≥15 years. Among the 1,080 normoalbuminuric participants in the Natural History of Microalbuminuria Study, 420 met these additional criteria, and we examined 307 (73%) of them.

As unrelated subjects for the genetic studies on diabetic nephropathy, we selected Caucasian patients with persistent proteinuria or ESRD. We were able to enroll and examine 152 of the 206 patients with advanced nephropathy identified in the Natural History Study described above. In addition to these subjects, we enrolled and examined 201 (86%) of 235 Caucasian patients with type 1 diabetes and proteinuria or ESRD attending the clinic between 1994 and 2000 who had not been included in the 50% random sample screened in 1991–1993 for the Natural History Study. The two groups of subjects did not differ with regard to sex, age at examination, duration of diabetes, or values of HbA_1c, total serum cholesterol, or serum triglycerides. Therefore, we pooled the two groups, bringing the total number of unrelated subjects with advanced diabetic nephropathy to 352.

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. For the family-based study, parents provided blood samples for DNA extraction. The Committee on Human Subjects of the Joslin Diabetes Center approved the protocol and informed consent procedures.

The diabetic nephropathy status of each patient was determined on the basis of questionnaires, medical records of the Joslin Clinic (supplemented with records of other physicians if necessary), and measurements of the ACR. Methods for measuring the ACR have been described previously (17). Patients were classified as control subjects (n = 307) if they had diabetes duration ≥15 years and the ACR (in mg/g) was <17 (men) or <25 (women) in at least two of the last three urine specimens. Patients with microalbuminuria or intermittent proteinuria were excluded from the study. Patients were considered case subjects (n = 352) if they had persistent proteinuria or if they had ESRD due to diabetic nephropathy. Persistent proteinuria was defined as two of three successive urinalyses positive by reagent strip (≥2+ on Multistix; Bayer Corporation, Diagnostics Division, Elkhart, IN) or by an ACR (in mg/g) ≥250 (men) or ≥ 355 (women). At the time of the first examination (1992–2000), case subjects were divided into those with persistent proteinuria (n = 260) and those with ESRD (n = 92). Those with proteinuria were followed until the end of July 2001, by which time ESRD had developed in 60 of them, i.e., renal dialysis had been initiated. These were considered incident cases of ESRD. The individuals who had begun dialysis or received a renal transplant before the first examination were considered prevalent cases of ESRD. For prevalent or incidence cases of ESRD, the year when dialysis began was used to calculate the duration of diabetes. In total, we had 152 case subjects with ESRD (92 prevalent and 60 incident case subjects). For the present study DNA was available for 230 control individuals with normoalbuminuria and 262 case subjects with persistent proteinuria (n = 151) or ESRD (n = 111).

The DNA sequence of the human GLUT1 has not yet been published but was obtained by querying GenBank (available online at http://www.ncbi.nlm.nih.gov/). A recent entry (GenBank accession no. NT_004852, 16 April 2001 release) containing human genomic sequence was found that contained the entire annotated GLUT1 gene (located on human chromosome 1p).

Enhancer-1 and -2 have been described for mouse GLUT1 (GenBank accession no. D10230 and D10231) (14). In addition, a separate enhancer element (denoted enhancer-3) has been described for the rat homolog (GenBank accession no. U82754) (15). To determine which part of the human gene exhibited sequence similarity to these three enhancers, we used the PipMaker program (available online at http://nog.cse.psu.edu/cgi-bin/pipmaker?basic), which identifies conserved segments between pairs of DNA sequence. The presence of transcription factor binding sites in DNA sequence was detected using the TESS program to query the TRANSFAC transcription factor database (available online at http://www.cbil.upenn.edu/cgi-bin/tess/tess?RQ=SEA-FR-Query).

The 0.3-kb known human gene promoter (18) was amplified by PCR from genomic DNA of eight individuals. The putative human enhancer-1 (located ∼3.6 kb upstream of exon 1), enhancer-3 (located ∼6.6 kb upstream of exon 1), and enhancer-2 (located in intron 2) were likewise amplified. These PCR products were purified and then sequenced using an ABI 377 automated DNA sequencer machine in conjunction with dye terminator chemistry in accordance with the manufacturer’s protocol (Perkin Elmer).

The XbaI SNP represents a G-to-T transversion in intron 2 of GLUT1 (Fig. 1). This was amplified from genomic DNA as a 1.1-kb PCR product using primers TGTGCAACCCATGAGCTAA (F′) and CCTGGTCTCATCTGGATTCT (R′) (16,19). PCR (25 μl reaction volume) was performed on 20 ng of genomic DNA using 0.6 units of Taq polymerase (PGC Scientific) with above primers in the presence of 1.5 mmol/l MgCl₂ for 30 cycles. Cycling parameters were: denaturation at 95°C for 45 s, annealing at 55°C for 45 s, and extension at 72°C for 90 s, with final extension at 72°C for 10 min. PCR products were then digested with XbaI restriction endonuclease at 37°C for >3 h before fractionation on a 1.2% agarose gel. When the restriction site was present [denoted XbaI(+) allele], two fragments (0.9 and 0.2 kb) were visualized. In the absence of this site [denoted XbaI(-) allele], the 1.1-kb product remained intact.

The HaeIII SNP is a common T-to-C transition (Ala¹⁵→Ala¹⁵) in exon 2 of the GLUT1 gene (Fig. 1) (16). This was amplified from genomic DNA as a 173-bp PCR product using primers CTCCCAGACACGCCTATAACAGT (F′) and GGCTGGTGTCCATAAGCCAACG (R′). PCR was performed as above in the presence of 1.5 mmol/l MgCl₂ using Taq polymerase for 40 cycles with the following parameters: denaturation at 95°C for 45 s, annealing at 66°C for 45 s, and extension at 72°C for 60 s, with final extension at 72°C for 10 min. PCR products were digested with HaeIII enzyme at 37°C for >3 h and then electrophoresed in a 2.8% agarose gel. The HaeIII(+) allele was observed as two fragments of 138 and 35 bp after digestion, whereas the HaeIII(-) allele was seen as the intact 173-bp product.

The SNP (denoted enhancer-1 SNP) found in putative enhancer-1 represents a C-to-T transition (Fig. 1). This SNP was amplified from genomic DNA as a 635-bp product using primers AAGATAAAGAAGGGATGCCG (F′) and CAGCATTCAGCTTGAACGAC (R′) by PCR as described above, with an annealing temperature of 61°C. The PCR product was then subjected to allele-specific oligonucleotide (ASO) hybridization (20) using the following probes: GGAACCCATCACTGTGT and GGAATCCATCACTGTGT.

Two SNPs (denoted enhancer-2 SNP 1 and enhancer-2 SNP 2) were found in close proximity (162 bp apart) within the putative enhancer-2 of GLUT1 (Fig. 1). Both SNPs were amplified as a 403-bp product using the same pair of primers CGATGATGGAAATGCGAG (F′) and ACAAGGCACACGCAGTGA (R′), with an annealing temperature of 61°C. Detection of enhancer-2 SNP 1 was achieved using ASO hybridization with probes TCCCATGCATTGGAGGA and TCCCGTGCATTGGAGGA. Likewise, enhancer-2 SNP 2 was performed using ASO probes TTCTCGGGAGGAAGGCT and TTCTTGGGAGGAAGGCT.

An SNP (C to T transition) was located in putative enhancer-3 (Fig. 1). This polymorphism was amplified as a 525-bp fragment using primers AGACCTCCAGAAAACAATACG (F′) and GATCCAGCAACAGTACAAGC (R′), with annealing temperature of 58°C. Alleles were discriminated using ASO probes GATTCAGGTGAGGTAAC and GATTTAGGTGAGGTAAC.

Altogether, 93 and 89% of the 492 patients included in this case-control study were genotyped for the XbaI SNP and enhancer-1 SNP, respectively; the remainder could not be determined due to technical difficulties with PCR amplification. Of the patients, >98% were successfully genotyped for the remaining four SNPs studied.

Haplotype frequencies were estimated using the EH program (21). Categorical data were compared using χ² test (two-sided). For continuous variables, ANOVA was used in cases of multiple group comparisons, whereas Student’s t test (two-sided) was used for pairwise comparisons. The Breslow-Day test was used to test for homogeneity of odds ratios (ORs). P values <0.05 were considered statistically significant.

The population attributable-risk percent (PAR%) was calculated according to the following formula: PAR% = [(RR − 1)/(RR − 1 + 1/P_e)] × 100 (22). RR was estimated from the OR. The proportion of population who have exposure (P_e) to XbaI(-)XbaI(-) or enhancer-2 SNP 1 AA risk genotypes was based on the frequency of these genotypes in our control group (see results). This latter approximation is valid, considering that homozygosity for XbaI(-) occurred at a comparable frequency among control subjects in other independent studies conducted on type 1 diabetic Caucasians (12,13).

Two regions of GLUT1 were identified that showed substantial sequence similarity to enhancer-1 and enhancer-2 of the mouse homolog according to the PipMaker program (the sequence alignments are listed in the online appendix at http://diabetes.diabetesjournals.org). A 634-bp sequence (putative enhancer-1) found ∼3.6 kb upstream of exon 1 was similar (65% identity over 645 bp) to the 610-bp enhancer-1, which is located 2.7 kb upstream of the mouse gene (Fig. 1) (14). Another sequence of 997-bp (putative enhancer-2) was likewise observed to share sequence similarity (46% identity over 1,159 bp) to the 1,342-bp mouse enhancer-2 (14). Like the mouse enhancer-2, this putative human enhancer is also located in intron 2 of GLUT1 (Fig. 1). In addition, a 367-bp putative enhancer (denoted enhancer-3) located ∼6.6 kb upstream of exon 1 was identified (Fig. 1). This cis-element exhibited significant sequence similarity (66% identity over 368 bp) to a 666-bp enhancer present at ∼6 kb upstream of the rat GLUT1 gene major transcription start site (15).

Sequencing of putative enhancer regions of GLUT1 amplified from genomic DNA revealed four common SNPs. Of these, one is located in enhancer-1 and consists of a C-to-T transition (termed enhancer-1 SNP) (Fig. 1). Two SNPs were found in putative enhancer-2 and comprised an A-to-G transition (enhancer-2 SNP 1) and a C-to-T transition (enhancer-2 SNP 2) (Fig. 1). The fourth SNP (termed enhancer-3 SNP) is a C-to-T transition and resides in enhancer-3 (Fig. 1). No polymorphisms were found in the GLUT1 promoter region.

Of the many potential transcription sites detected in the vicinity of the enhancer SNPs, some were predicted with greater certainty than others (i.e., above default secondary threshold in the TESS program). Of these better predictions, a site for the transcription factor CytR was predicted with both C and T alleles of enhancer-1 SNP. On the other hand, sites for GATA-1 and HSTF appeared to be specific for the C allele. The CytR and GATA-1 sites were detected in the mouse but not the rat homolog. An upstream stimulatory factor (USF) binding site was detected with both A and G alleles of enhancer-2 SNP 1, but it was absent in the mouse homolog. Likewise, sites for MGF and LyF-1 were detected with both C and T alleles of enhancer-2 SNP 2 but were not predicted in the mouse homolog. In the case of enhancer-3 SNP, a deltaEF site was predicted with both C and T alleles. However, predicted sites for NF-X3, tal-1, E12, and TFE3-S were specific for the C allele, whereas sites for Cdx-1 and MBF-1 were confined to the T allele. None of these enhancer-3 SNP-associated sites were conserved in the rat homolog.

Selected clinical characteristics of control and case subjects are presented in Table 1. Diabetes duration and sex ratio were similar between the two groups. Values for HbA_1c, serum total cholesterol, and triglycerides were significantly elevated in case subjects compared with control subjects at the time of examination (P < 0.01 for each parameter). Case subjects were slightly younger than control subjects when diabetes was diagnosed (P = 0.03).

All six SNPs were examined for association with diabetic nephropathy using a case-control study design. Genotype frequencies for all polymorphisms did not deviate significantly from Hardy-Weinberg equilibrium when case and control subjects were analyzed together or as separate groups.

The overall genotype distribution of the XbaI SNP was not significantly different between case and control subjects (Table 2). However, homozygosity for the XbaI(-) allele was associated with a slightly increased risk of diabetic nephropathy compared with the other genotypes combined (OR 1.83 [95%CI 1.01–3.33]). A significant difference in overall genotype distribution between case and control subjects was observed for enhancer-2 SNP 1 (P = 0.036) (Table 2). Specifically, there was an excess of the AA genotype among case subjects (10.7%) compared with control subjects (4.8%). These homozygotes were at a significantly higher risk of diabetic nephropathy compared with the AG and GG genotypes combined (OR 2.38 [95% CI 1.16–4.90]). The genotype distribution of both XbaI SNP and enhancer-2 SNP 1 were similar among case subjects when examined as subcategories based on whether these patients were proteinuric or had developed ESRD.

The effect of glycemic control and duration of diabetes on the association between these SNPs and diabetic nephropathy was examined. Stratifying according to the median HbA_1c value of 8.5%, the association was seen in both resulting strata (i.e., with HbA_1c ≤8.5% or with HbA_1c >8.5%). Similarly, stratifying by the median duration of diabetes (25 years) revealed that the observed associations were also independent of this parameter. In addition, ANOVA revealed that XbaI SNP and enhancer-2 SNP 1 genotypes did not affect HbA_1c values and systolic and diastolic blood pressure values in either control or case subjects.

Strong linkage disequilibrium exists between enhancer-2 SNP 1 and the XbaI SNP. Specifically, 100% of enhancer-2 SNP 1 AA homozygotes were also homozygous for the XbaI(-) allele. Because the XbaI SNP is more polymorphic than enhancer-2 SNP 1, the enhancer-2 SNP 1 AA genotype was detected in 67% of XbaI(-) homozygotes (data not shown). Frequency of haplotypes formed from these two markers was estimated using the EH program. As with allele frequencies of the individual markers, overall haplotype distribution did not differ between case and control subjects (P = NS). However, patients who were homozygous for risk alleles at both XbaI SNP and enhancer-2 SNP 1 [i.e., homozygosity for XbaI(-)/A haplotype] were at an increased risk of diabetic nephropathy compared with those without this genotype (OR 2.40 [95% CI 1.13–5.07]) (data not shown). This OR is similar to that found when either SNP was considered alone. This suggested an absence of a combined effect between these two SNPs on the development of diabetic nephropathy.

A significant association (as reflected by OR) thus exists between either XbaI SNP or enhancer-2 SNP 1 and diabetic nephropathy in Caucasians with type 1 diabetes. To gauge the magnitude of the effect of these GLUT1 polymorphisms in conferring risk of this complication in our population, the PAR% was calculated. Using the procedure described in research design and methods, we estimated PAR% for XbaI(-)XbaI(-) homozygosity to be 6.73%, whereas that for enhancer-2 SNP 1 AA homozygosity was very similar at 6.21%.

The four remaining SNPs in GLUT1 were also investigated in the case-control study (Fig. 1). No association was observed between these markers and the development of diabetic nephropathy (Table 2).

The present study was carried out to comprehensively investigate the role of GLUT1 in conferring genetic susceptibility to nephropathy in type 1 diabetic patients. Previous studies have shown that polymorphisms located in exons of this gene do not alter the amino acid sequence of the protein (16,23). Thus, if GLUT1 does contribute to genetic susceptibility, the causative polymorphism(s) is likely to be located in a regulatory portion of the gene.

Aside from the basal promoter (18), the regulatory regions of human GLUT1 gene have not been well characterized. To circumvent this knowledge deficit, we identified putative enhancer regions by sequence similarity to known enhancer regions of the mouse (14) and rat genes (15). Common polymorphisms identified in these enhancer regions were then examined for association with diabetic nephropathy using a case-control study design. In addition, we took advantage of our large patient collection to test whether the GLUT1 XbaI SNP was associated with the development of this microvascular complication because conflicting results have so far been reported for this genetic marker (9–13).

Our study demonstrated that among Caucasian type 1 diabetic patients, XbaI(-)XbaI(-) homozygotes were at a slightly increased risk of diabetic nephropathy compared with patients with other genotypes. This result is in apparent agreement with the findings of a small British study by Hodgkinson et al. (12), which found homozygosity for XbaI(-) to be associated with diabetic nephropathy (OR 3.74 [95%CI 1.39–10.05], our calculation). This OR, though approximately twofold higher than that observed in our study, was not statistically different (Breslow-Day test for homogeneity of ORs, P = NS). On the other hand, this British study may have provided a less accurate estimate of the strength of the association because it consisted of only 44 control and 70 case subjects. In comparison, a total of 207 control and 249 case subjects were genotyped for the XbaI SNP in our study. It is reasonable that our four-times larger data set should yield a more reliable estimate of the genotype distribution. This should also make our study less prone to detecting spurious associations that may inadvertently occur in smaller studies, especially when the gene effect is minor.

Genotype data for the XbaI SNP in our patient sample was also compared with a recent Danish study that was smaller but still comparable in size (192 control and 175 case subjects). This study, conducted on type 1 diabetic Caucasians, did not detect any association between this SNP and diabetic nephropathy (13). Because our study found the association between this SNP and diabetic nephropathy to be small, failure of the Danish study to observe this association might have occurred by chance. Alternatively, it might be contemplated that GLUT1 confers variable genetic risk to diabetic nephropathy in different Caucasian populations.

Aside from investigating XbaI SNP, a novel finding in our study is the identification of enhancer-2 SNP 1. Like the XbaI SNP, this new SNP is also located in intron 2, at a distance <2 kb from the former polymorphism. Results of our case-control study revealed that homozygosity for the A allele of enhancer-2 SNP 1 was significantly associated with diabetic nephropathy. Strong linkage disequilibrium was observed between this marker and XbaI SNP, and this likely accounted for the latter being associated with diabetic nephropathy.

However, in contrast to XbaI SNP, enhancer-2 SNP 1 could be the causative polymorphism directly influencing genetic susceptibility to diabetic nephropathy because it resides in a putative enhancer region that has potential to modulate gene expression. In this regard, the murine homologues of human enhancer-2 (and enhancer-1) have been previously demonstrated to confer insulin-responsiveness to reporter gene constructs when stably transfected into NIH/3T3 HIR3.5 cells (24). Also, enhancer-2 SNP 1 is located within a predicted binding site for USF, which is a transcription factor that confers responsiveness to stimulation by insulin and glucose (25,26). Thus, it might be hypothesized that in diabetic patients with the enhancer-2 SNP 1 AA risk genotype, high intracellular glucose levels may accumulate in mesangial cells in response to insulin and hyperglycemia (27). Excessive intracellular glucose levels may then activate various pathogenetic mechanisms that exacerbate mesangial expansion and glomerulosclerosis in diabetic nephropathy (6,28,29). Because these enhancer elements have so far been identified based on similarity between human and rodent genomic sequences, the effect of these putative enhancers on gene transcription could be examined in future studies. In addition, the effect of allelic variants of enhancer-2 SNP 1 on enhancer function may be tested to provide insight into the regulation of GLUT1.

In conclusion, our study confirms that SNPs at the GLUT1 locus are associated with susceptibility to diabetic nephropathy in type 1 diabetes. Although these SNPs confer a considerable personal risk to diabetic nephropathy, they account for a limited proportion of cases among type 1 diabetic patients.

This research was supported by National Institutes of Health Grant DK41526. D.N. is a recipient of the National Science & Technology Board/Singapore Economic Development Board Biomedical Sciences Fellowship. L.C. is a recipient of a fellowship from Juvenile Diabetes Foundation International (3-2001-343).