Abstract
OBJECTIVE
The purpose of this study was to examine whether known genetic risk factors for type 1 diabetes (HLA-DRB1, -DQA1, and -DQB1 and insulin locus) play a role in the etiology of diabetic nephropathy.
RESEARCH DESIGN AND METHODS
Genetic analysis of HLA-DRB1, -DQA1, -DQB1 and the insulin gene (INS) was performed in the Genetics of Kidneys in Diabetes (GoKinD) collection of DNA (European ancestry subset), which includes case patients with type 1 diabetes and nephropathy (n = 829) and control patients with type 1 diabetes but not nephropathy (n = 904). The availability of phenotypic and genotypic data on GoKinD participants allowed a detailed analysis of the association of these genes with diabetic nephropathy.
RESULTS
Diabetic probands who were homozygous for HLA-DRB1*04 were 50% less likely to have nephropathy than probands without any DRB1*04 alleles. In heterozygous carriers, a protective effect of this allele was not as clearly evident; the mode of inheritance therefore remains unclear. This association was seen in probands with both short (<28 years, P = 0.02) and long (≥28 years, P = 0.0001) duration of diabetes. A1C, a marker of sustained hyperglycemia, was increased in control probands with normoalbuminuira, despite long-duration diabetes, from 7.2 to 7.3 to 7.7% with 0, 1, and 2 copies of the DRB1*04 allele, respectively. This result is consistent with a protective effect of DRB1*04 that may allow individuals to tolerate higher levels of hyperglycemia, as measured by A1C, without developing nephropathy.
CONCLUSIONS
These data suggest that carriers of DRB1*04 are protected from some of the injurious hyperglycemic effects related to nephropathy. Interestingly, DRB1*04 appears to be both a risk allele for type 1 diabetes and a protective allele for nephropathy.
Type 1 diabetes affects about 1 in every 400 to 600 children and adolescents and is the predominant form of diabetes in young people. Studies have estimated a yearly 3% increase of type 1 diabetes globally (1). Approximately one in three European Americans with type 1 diabetes develops severe nephropathy that leads to end-stage renal diseases (ESRD) (2).
There is strong evidence showing genetic susceptibility in developing diabetic nephropathy (3–5); however, identifying the responsible genetic variants has been limited by modest data collections. The Genetics of Kidneys in Diabetes (GoKinD) study, an initiative supported by the Juvenile Diabetes Research Foundation, the National Institute of Diabetes and Digestive and Kidney Diseases, and the Centers for Disease Control and Prevention (CDC), was created to address this research gap (6). The GoKinD study includes 1,879 individuals with long-term type 1 diabetes, approximately half with nephropathy (defined as persistent proteinuria or ESRD) and half without, and extensive phenotypic data.
The polymorphic HLA genes—DRB1, DQA1, and DQB1—located on chromosome 6 and the insulin gene on chromosome 11 are the major genetic risk factors associated with type 1 diabetes (7,8). Very early small studies looked at HLA in people with diabetic nephropathy, but no reliable associations were seen (9,10). A later larger study also found no significant distribution difference of HLA haplotypes or alleles in people with diabetes and nephropathy versus people with type 1 diabetes but without nephropathy. Additionally, this study examined the role of the –23 INS single nucleotide polymorphism in people with diabetic nephropathy and found no association (10).
RESEARCH DESIGN AND METHODS
The GoKinD DNA samples genotyped were from 1,879 individuals from the U.S. and Canada with long-term (>10 years) diabetes. The analyses for this study were restricted to probands of European ancestry who had no known relationship to another proband. The resulting subset including both singleton and trio probands totaled 829 case probands with persistent proteinuria or ESRD and 904 control probands with normoalbuminuria. The parents of trio probands were not included in this analysis because the genes of interest are risk factors for type 1 diabetes, and transmission was distorted in both case and control patients. Mueller et al. (6) presented a detailed description of the study’s eligibility criteria, ascertainment of participants, and associated clinical data. Participants were recruited at two centers, the Joslin Diabetes Center (JDC) and the George Washington University Biostatics Center (GWU). JDC recruited participants who were already enrolled in previous studies at the Joslin Clinic. GWU worked with Matthews Media Group to identify volunteers who were directed to 1 of 27 clinical centers.
DNA was extracted from whole blood and/or transformed lymphoblasts. Self-reported sex was confirmed using the AmpFlSTR Green I PCR Amplification Kit (Applied Biosystems, Foster City, CA) with modifications. Detailed extraction and quality control methods can be found in the supplementary material of Mueller et al. (6).
Genotyping methods
HLA genotyping methods for DRB1, DQA1, and DQB1 were based on PCR amplification from genomic DNA. The exon(s) evaluated were amplified then followed by florescent-based sequencing. The sequence data were analyzed using MatchTools analysis or Conexio Genomics’s Assign software (Applied Biosystems, Western Australia). DRB1 exon 2 genotyping was performed using an AlleleSEQR DRB1 kit (Atria Genetics; San Francisco, CA) according to the manufacturer’s directions. The DQA1 method, developed at CDC (11,12), evaluated exons 2 and 3 for all participants and only exon 1 in participants with any of the following alleles: *010101, *010102, *010401, *0105, *0302, or *0303. The DQB1 method, also developed at CDC, evaluated polymorphisms in exons 2 and 3. Additional sequence-specific reactions were included to resolve ambiguous combinations, specifically *030101/*0302 versus *0304/*030302, *0202/*030302 versus *0203/*0302, or *0202/*030101 versus *0203/*0304 (Appendix A [available in the online appendix at http://dx.doi.org/10.2337]). The –23 INS single nucleotide polymorphism rs689 was genotyped using 5-μl 5′ nuclease assay reactions on a 7900HT Real-Time PCR System (Applied Biosystems) according to the manufacturer’s instructions. Products were amplified (forward primer, 5′-ctgggctcgtgaagcatgt-3′/reverse primer 5′-gcaggaggcgcatcca-3′) and detected with custom-designed probes (5′-VIC-CTGCCTGTCACCCAG-MGBNFQ-3′ and 5′-6FAM-CTGCCTGTCTCCCA-MGBNFQ-3′, MGBNFQ-minor groove binder/nonfluorescent quencher; Applied Biosystems).
Statistical methods
High-resolution molecular HLA data were grouped into risk categories: DQA1’s two known high-resolution risk alleles, *050101 and *030101, and DQB1’s risk alleles, *020101 and *030201. All other alleles for each gene were grouped and designated as “X.” DRB1*03 and *04 molecular subgroups were combined, and all other DRB1 alleles were designated as “X.” (High-resolution classification and nomenclature is assigned by the World Health Organization Nomenclature Committee and available at http://www.anthonynolan.com/HIG/index.html.) Nonparametric Wilcoxon’s and χ2 tests assessed differences between case and control patients in baseline characteristics and allele and genotype frequencies. Analyses were corrected for covariates that were different between case and control patients including diabetes duration, age at type 1 diabetes diagnosis, sex, recruitment site, and smoking status (Table 1). Stratified analysis of the association according to diabetes duration was performed as suggested by Rogus et al. (13). Linear trend analysis was performed using Chacko’s test for homogeneity against ordered alternatives (14). The enrollment sites for the analyses were defined as site 1 (JDC) and site 2 (GWU-affiliated clinics).
TABLE 1.
Characteristics of the study group according to diabetic nephropathy status
| Case patients | Control patients | P | |
|---|---|---|---|
| n | 829 | 904 | |
| Age at type 1 diabetes diagnosis | 11.8 ± 6.7 | 12.9 ± 7.3 | 0.0042 |
| Age at entry | 43.1 ± 6.9 | 38.3 ± 8.6 | <0.0001 |
| Diabetes duration | 31.2 ± 7.8 | 25.4 ± 7.8 | <0.0001 |
| Male sex (%) | 425 (51.27) | 366 (40.49) | <0.0001 |
| Recruitment site 1 (%) | 452 (54.52) | 481 (53.21) | 0.5832 |
| Smoking status (%) | 411 (49.58) | 292 (32.41) | <0.0001 |
| Insulin –23 SNP (%) | 0.0719 | ||
| A | 1,380 (83.23) | 1,545 (85.45) | |
| T | 278 (16.77) | 263 (14.55) | |
| DQA1 (%) | 0.1635 | ||
| *050101 | 552 (33.29) | 573 (31.69) | |
| *030101 | 463 (27.93) | 558 (30.86) | |
| *X† | 643 (38.78) | 677 (37.44) | |
| DQB1 (%) | 0.0702 | ||
| *020101 | 552 (33.29) | 572 (31.64) | |
| *030201 | 522 (31.48) | 636 (35.18) | |
| *X‡ | 584 (35.22) | 600 (33.19) | |
| DRB1 (%) | 0.0054§ | ||
| *03 | 553 (33.35) | 574 (31.75) | |
| *04 | 593 (35.77) | 740 (40.93) | |
| *X§ | 512 (30.88) | 494 (27.32) |
Data are means ± SD and n (%).
All DQA1 alleles except DQA1*050101 and *030101.
All DQB1 alleles except DQB1*020101 and *030201.
All DRB1 alleles except DRB1*03 and *04. §Bonferroni-corrected criteria for multiple gene testing, P < 0.0125.
RESULTS
Type 1 diabetes susceptibility gene distribution
Proband characteristics are summarized by nephropathy status in Table 1. Case probands were slightly younger than control probands when diagnosed with diabetes. Control probands were older at the time of enrollment and thus had longer diabetes duration. Allele frequencies for the four genes examined are summarized in Table 1. Genotype data from HLA-DRB1, -DQA1, and -DQB1 were grouped into categories known to be associated with risk for type 1 diabetes: *03 and *04 at DRB1, *050101 and *030101 at DQA1, and *020101 and *030201 at DQB1. All other alleles for each gene were grouped and designated as “X.” Allele distributions of DQA1, DQB1, and insulin did not differ between cases and controls; however, the DRB1 04 allele was significantly more frequent in controls (P = 0.0054). The remaining analyses further examined the association of DRB1 with diabetic nephropathy.
HLA-DRB1 genotype distribution
The distribution of the six genotypes, defined by DRB1 alleles *03, *04, and *X, differed significantly (P = 0.0063) between case and control probands (Table 2). The frequency of *04/*04 homozygotes in control probands was almost twice that of case probands and made the largest contribution to the χ2 for the table. The P value decreased to 0.0012 after adjustment for diabetes duration, age at diabetes diagnosis, sex, recruitment site, and smoking status, and the adjusted relative odds illustrate the pattern of association. Since case probands included two phenotypes, ESRD and persistent proteinuria, the analysis was repeated in each phenotype separately. The allele and genotype distributions were quite similar for case probands with proteinuria and case probands with ESRD, and the results for each clinical phenotype remained significant (Table 3). Additionally, there was no difference in ESRD DRB1 genotype distribution compared with control probands when the data were stratified on median ESRD duration (data not shown). These findings are consistent with a protective effect of DRB1*04 alleles.
TABLE 2.
Distribution of DRB1 genotype according to diabetic nephropathy status
| DRB1 genotype | n | Case patients (%) | Control patients (%) | P† | Unadjusted OR | Adjusted OR† |
|---|---|---|---|---|---|---|
| All | 1,733 | 829 | 904 | 0.0063 | ||
| *03/*03 | 149 | 74 (8.93) | 75 (8.30) | 0.959 | 0.836 | |
| *03/*04 | 565 | 259 (31.24) | 306 (33.85) | 0.823 | 0.643 | |
| *03/*X‡ | 264 | 146 (17.61) | 118 (13.05) | 1.202 | 1.113 | |
| *04/*04 | 153 | 55 (6.63) | 98 (10.84) | 0.545 | 0.479 | |
| *04/*X‡ | 462 | 224 (27.02) | 238 (26.33) | 0.915 | 0.804 | |
| *X/*X‡ | 140 | 71 (8.56) | 69 (7.63) | 1.00 | 1.00 |
P = 0.0012 after adjustment for diabetes duration, age at diagnosis, sex, recruitment site, and smoking status. †Odds ratio (OR) adjusted for diabetes duration, age at diagnosis, sex, recruitment site, and smoking status.
All DRB1 genotypes except *03 and *04.
TABLE 3.
DRB1 alleles and genotypes in case patients with ESRD versus control and case patients with proteinuria (no ESRD) versus control patients
| DRB1 allele/genotype | n | ESRD (%) | Control (%) | P | n | Proteinuria (%) | Control (%) | P |
|---|---|---|---|---|---|---|---|---|
| All | 2,892 | 1,084 | 1,808 | 0.0236 | 2,382 | 574 | 1,808 | 0.0403 |
| *03 | 929 | 355 (32.75) | 574 (31.75) | 772 | 198 (34.49) | 574 (31.75) | ||
| *04 | 1,132 | 392 (36.16) | 740 (40.93) | 941 | 201 (35.02) | 740 (40.93) | ||
| *X† | 831 | 337 (31.09) | 494 (27.32) | 669 | 175 (30.49) | 494 (27.32) | ||
| All | 1,446 | 542 | 904 | 0.0096 | 1,191 | 287 | 904 | 0.0131 |
| *Y/*Y‡ | 449 | 187 (34.50) | 262 (28.98) | 366 | 104 (36.24) | 262 (28.98) | ||
| *04/*Y‡ | 862 | 318 (58.67) | 544 (60.18) | 709 | 165 (57.49) | 544 (60.18) | ||
| *04/*04 | 135 | 37 (6.83) | 98 (10.84) | 116 | 18 (6.27) | 98 (10.84) |
Data are n (%).
All DRB1 alleles except *03 and *04.
All DRB1 genotypes except *04.
Early, small studies have reported an association of DR4 antigens and diabetic retinopathy (9), but this association was not seen in later studies (15,16). To determine whether there was a nonrandom distribution of the type 1 diabetes risk alleles by self-reported retinopathy status, the distribution of alleles in the GoKinD case probands with and without retinopathy as well as control probands with and without retinopathy was examined. The distribution of DRB1 alleles was not associated with self-reported retinopathy status in either the case probands (P = 0.2564) or the control probands (P = 0.2096) (Table 4).
TABLE 4.
Distribution of DRB1 alleles according to self-reported retinopathy status in case and control patients
| Case patients with no retinopathy (%) | Case patients with self-reported retinopathy (%) | Control patients with no retinopathy | Control patients with self-reported retinopathy (%) | |
|---|---|---|---|---|
| DRB1 allele | ||||
| *03 | 467 (33.36) | 81 (32.40) | 83 (27.48) | 490 (32.67) |
| *04 | 510 (36.43) | 81 (32.40) | 132 (43.71) | 605 (40.33) |
| *X† | 423 (30.21) | 88 (35.20) | 87 (28.81) | 405 (27.00) |
| P | 0.2564 | 0.2096 | ||
Data are n (%).
All DRB1 alleles except *03 and *04.
Diabetic nephropathy is believed to result from an interaction between cumulative exposure to hyperglycemia and genetic susceptibility factors. Simulation studies have shown that these genetic susceptibility factors show up most strongly in early appearing cases, while they are depleted in long-duration unaffected control patients, which have the greatest cumulative exposure (13). Thus, the association between DRB1*04 genotypes and diabetic nephropathy was examined using duration of diabetes as a surrogate measure of cumulative exposure. Duration was divided at its median (28 years) into “short duration” and “long duration.” Furthermore, DRB1*03 alleles were grouped with *X and designated as “Y,” representing all alleles except *04. In short-duration probands, the pattern of association between DRB1*04 and diabetic nephropathy was similar to the total group and remained significant (P = 0.0241) despite the reduced sample size (Table 4). In long-duration probands the association between DRB1*04 and diabetic nephropathy was stronger (P = 0.0009) than in the total group. Moreover, in this stratum, the frequency of *04 heterozygotes and *04/*04 homozygotes was increased in control probands. Using the *Y/*Y genotype as a reference group, the adjusted odds ratio for the *04/*Y genotype was nearly the same as that for the *04/*04 genotype. Note that in the control probands, the frequency of the *Y/*Y genotype decreased significantly (P = 0.0016) between the short- and long-duration strata, as predicted by the simulation models described above (13). Therefore, the association between DRB1*04 alleles and diabetic nephropathy is most clearly displayed in a comparison of the genotype distribution in short-duration case probands with the long-duration control probands (Table 5).
TABLE 5.
Genotype frequencies according to study group and diabetes duration
| DRB1 genotype | n | Case patients (%) | Control patients (%) | P | Adjusted odds ratio† |
|---|---|---|---|---|---|
| Diabetes duration <28 years | 0.0241‡ | ||||
| *Y/*Y§ | 295 | 102 (38.93) | 193 (32.77) | 1 | |
| *04/*Y§ | 476 | 145 (55.34) | 331 (56.20) | 0.847 | |
| *04/*04 | 80 | 15 (5.73) | 65 (11.04) | 0.395 | |
| Diabetes duration ≥28 years | 0.0009|| | ||||
| *Y/*Y§ | 258 | 189 (33.33) | 69 (21.90) | 1 | |
| *04/*Y§ | 551 | 338 (59.61) | 213 (67.62) | 0.575 | |
| *04/*04 | 73 | 40 (7.05) | 33 (10.48) | 0.533 | |
|
| |||||
| DRB1 genotype | n | Case patients: diabetes duration <28 years (%) | Control patients: diabetes duration ≥28 years (%) | P | Adjusted odds ratio† |
|
| |||||
| <0.0001¶ | |||||
| *Y/*Y§ | 171 | 102 (38.9) | 69 (21.9) | 1 | |
| *04/*Y§ | 358 | 145 (55.3) | 213 (67.6) | 0.522 | |
| *04/*04 | 48 | 15 (5.7) | 33 (10.5) | 0.257 | |
Data are n (%).
Odds ratios adjusted for diabetes duration, age at diagnosis, sex, recruitment site, and smoking status.
Adjusted P = 0.0193 (adjusted for diabetes duration, age at diagnosis, sex, recruitment site, and smoking status).
All DRB1 alleles except *04.
Adjusted P = 0.0035 (adjusted for diabetes duration, age at diagnosis, sex, recruitment site, and smoking status).
Adjusted P = 0.0005 (adjusted for age at diagnosis, sex, recruitment site, and smoking status).
A1C values in control probands
Cumulative diabetes exposure depends on both the duration of diabetes and the intensity of exposure. The intensity of hyperglycemia is measured by A1C, and its value tends to be consistent in an individual over long periods of time (unpublished data). However, once ESRD or proteinuria has developed, A1C values become profoundly affected by kidney disease and its treatment and no longer reflect past exposure. Thus, this analysis cannot be stratified by intensity of exposure. However, variation in the intensity of exposures by DRB1 genotype in the control probands was examined for possible evidence that would suggest a mechanism of action for the DRB1*04 protective effect. A1C value increased with each additional DRB1*04 allele from 7.2 to 7.3 to 7.7% for individuals with 0, 1, and 2 copies, respectively (P = 0.013).
DISCUSSION
Enrichment of the DRB1*04 alleles in long-duration type 1 diabetes with normoalbuminuria despite long duration of diabetes suggests that these alleles are associated with reduced nephropathy. The mechanisms underlying this protective association are unknown. However, involvement of DRB1 is consistent with the known immunological processes in the pathogenesis of diabetic nephropathy.
Diabetic nephropathy was considered to be a chronic degenerative disease until Bohle et al. (17) described the of presence of monocytes, macrophages, T-cells, and fibroblasts associated with tubulo-interstitial changes, suggesting an immunological role in diabetic nephropathy. A study of renal biopsies from patients with diabetic nephropathy found increased steady-state mRNA levels of inflammatory genes associated with interstitial fibrosis and progressive nephropathy (18). Additionally, many studies have shown that changes in hormone levels, vaso-active peptides, cytokine, and growth factors are associated with diabetic nephropathy development and progression. Specifically, rat studies have suggested a functional role for the proinflammatory cytokine tumor necrosis factor (TNF)-α in the pathogenesis of diabetic nephropathy (19,20). Supporting this finding are intervention studies in patients with either diabetic or membranous nephropathy that found that the nonspecific TNF-α inhibitor pentoxifylline reduced both TNF-α and urinary proteinura (21,22).
While genetic susceptibility factors are important for the development of diabetic nephropathy, the presence of hyperglycemia is required. Since A1C values are not meaningful in nephropathy patients with proteinuria or ESRD, we examined the control probands who have not developed nephropathy despite many years of diabetes. The control probands who are heterozygous and homozygous for DRB1*04 have a higher A1C on average than those with other alleles. While this difference is small, it is consistent with protection against the toxic effects of hyperglycemia.
The major strengths of this study are its size and well-documented participant phenotypes. Power calculations indicated a >99% ability to detect the observed DRB1*04 effect regardless of the mode of inheritance (6). Genetic association studies with renal disease carry the possibility of a false-positive finding due to either population stratification or mortality effects. Arguing against population stratification is the consistent pattern of association with DRB1*04 when the data were adjusted for recruitment center, as well as the increased significance when duration of diabetes was taken into account as predicted in simulation studies (13). Stratification would also not explain the association of A1C with the number of DRB1*04 alleles. Furthermore, there was no significant association seen with nearby HLA genes or the insulin gene. Arguing against mortality effects as the cause of these results, it is important to consider that mortality is higher in patients with ESRD than with proteinuria. The allele and genotype distributions were similar in both of these populations, and the DRB1*04 frequency was similar regardless of ESRD duration. Overall, these data suggest that HLA-DRB1*04 has a protective effect against diabetic nephropathy; however, further investigation of its functional role is needed.
Acknowledgments
The GoKinD Coordinating Center was funded by the Juvenile Diabetes Research Foundation, and CDC’s contributions were funded by grants PL-105-33, PL-106-554, and PL-107-360 from the National Institutes of Health.
The GoKinD collaborators acknowledge the 28 recruitment centers’ contributions to recruitment (6). Additionally, the collaborators thank Dr. Andrzej Krolewski, Dr. Christopher Greene, and Dr. Janelle Noble for their critical comments and intellectual input.
Footnotes
Additional information for this article can be found in an online appendix at http://dx.doi.org./10.2337/db07-0826.
CDC, Centers for Disease Control; ESRD, end-stage renal disease; GoKinD, Genetics of Kidneys in Diabetes; GWU, George Washington University Biostatics Center; JDC, Joslin Diabetes Center; TNF, tumor necrosis factor.
References
- 1.Onkamo P, Vaananen S, Karvonen M, Tuomilehto J. Worldwide increase in incidence of type I diabetes: the analysis of the data on published incidence trends. Diabetologia. 1999;42:1395–1403. doi: 10.1007/s001250051309. [DOI] [PubMed] [Google Scholar]
- 2.Jones CA, Krolewski AS, Rogus J, Xue JL, Collins A, Warram JH. Epidemic of end-stage renal disease in people with diabetes in the United States population: do we know the cause? Kidney Int. 2005;67:1684–1691. doi: 10.1111/j.1523-1755.2005.00265.x. [DOI] [PubMed] [Google Scholar]
- 3.the Diabetes Control and Complications Trial Research Group. Clustering of long-term complications in families with diabetes in the diabetes control and complications trial. Diabetes. 1997;46:1829–1839. [PubMed] [Google Scholar]
- 4.Quinn M, Angelico MC, Warram JH, Krolewski AS. Familial factors determine the development of diabetic nephropathy in patients with IDDM. Diabetologia. 1996;39:940–945. doi: 10.1007/BF00403913. [DOI] [PubMed] [Google Scholar]
- 5.Seaquist ER, Goetz FC, Rich S, Barbosa J. Familial clustering of diabetic kidney disease: evidence for genetic susceptibility to diabetic nephropathy. N Engl J Med. 1989;320:1161–1165. doi: 10.1056/NEJM198905043201801. [DOI] [PubMed] [Google Scholar]
- 6.Mueller PW, Rogus JJ, Cleary PA, Zhao Y, Smiles AM, Steffes MW, Bucksa J, Gibson TB, Cordovado SK, Krolewski AS, Nierras CR, Warram JH. Genetics of Kidneys in Diabetes (GoKinD) study: a genetics collection available for identifying genetic susceptibility factors for diabetic nephropathy in type 1 diabetes. J Am Soc Nephrol. 2006;17:1782–1790. doi: 10.1681/ASN.2005080822. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Noble JA, Valdes AM, Cook M, Klitz W, Thomson G, Erlich HA. The role of HLA class II genes in insulin-dependent diabetes mellitus: molecular analysis of 180 Caucasian, multiplex families. Am J Hum Genet. 1996;59:1134–1148. [PMC free article] [PubMed] [Google Scholar]
- 8.Bennett ST, Lucassen AM, Gough SC, Powell EE, Undlien DE, Pritchard LE, Merriman ME, Kawaguchi Y, Dronsfield MJ, Pociot F, et al. Susceptibility to human type 1 diabetes at IDDM2 is determined by tandem repeat variation at the insulin gene minisatellite locus. Nat Genet. 1995;9:284–292. doi: 10.1038/ng0395-284. [DOI] [PubMed] [Google Scholar]
- 9.Barbosa J, Saner B. Do genetic factors play a role in the pathogenesis of diabetic microangiopathy? Diabetologia. 1984;27:487–492. doi: 10.1007/BF00290381. [DOI] [PubMed] [Google Scholar]
- 10.Chowdhury TA, Dyer PH, Mijovic CH, Dunger DB, Barnett AH, Bain SC. Human leucocyte antigen and insulin gene regions and nephropathy in type I diabetes. Diabetologia. 1999;42:1017–1020. doi: 10.1007/s001250051262. [DOI] [PubMed] [Google Scholar]
- 11.Cordovado SK, Hancock LN, Simone AE, Hendrix M, Mueller PW. High-resolution genotyping of HLA-DQA1 in the GoKinD study and identification of novel alleles HLA-DQA1*040102, HLA-DQA1*0402 and HLA-DQA1*0404. Tissue Antigens. 2005;65:448–458. doi: 10.1111/j.1399-0039.2005.00389.x. [DOI] [PubMed] [Google Scholar]
- 12.Cordovado SK, Simone AE, Mueller PW. High-resolution sequence-based typing strategy for HLA-DQA1 using SSP-PCR and subsequent genotyping analysis with novel spreadsheet program. Tissue Antigens. 2001;58:308–314. doi: 10.1034/j.1399-0039.2001.580504.x. [DOI] [PubMed] [Google Scholar]
- 13.Rogus JJ, Warram JH, Krolewski AS. Genetic studies of late diabetic complications. the overlooked importance of diabetes duration before complication onset. Diabetes. 2002;51:1655–1662. doi: 10.2337/diabetes.51.6.1655. [DOI] [PubMed] [Google Scholar]
- 14.Chacko VJ. Testing homogeneity against ordered alternatives. Ann Math Stat. 1963;34:945–956. [Google Scholar]
- 15.Mimura T, Funatsu H, Uchigata Y, Kitano S, Shimizu E, Amano S, Yamagami S, Noma H, Araie M, Hori S. Glutamic acid decarboxylase autoantibody prevalence and association with HLA genotype in patients with younger-onset type 1 diabetes and proliferative diabetic retinopathy. Ophthalmology. 2005;112:1904–1909. doi: 10.1016/j.ophtha.2005.05.016. [DOI] [PubMed] [Google Scholar]
- 16.Wong TY, Cruickshank KJ, Klein R, Klein BE, Moss SE, Palta M, Riley WJ, Maclaren NK, Vadheim CM, Rotter JI. HLA-DR3 and DR4 and their relation to the incidence and progression of diabetic retinopathy. Ophthalmology. 2002;109:275–281. doi: 10.1016/s0161-6420(01)00925-3. [DOI] [PubMed] [Google Scholar]
- 17.Bohle A, Wehrmann M, Bogenschutz O, Batz C, Muller CA, Muller GA. The pathogenesis of chronic renal failure in diabetic nephropathy. Investigation of 488 cases of diabetic glomerulosclerosis. Pathol Res Pract. 1991;187:251–259. doi: 10.1016/s0344-0338(11)80780-6. [DOI] [PubMed] [Google Scholar]
- 18.Schmid H, Boucherot A, Yasuda Y, Henger A, Brunner B, Eichinger F, Nitsche A, Kiss E, Bleich M, Grone HJ, Nelson PJ, Schlondorff D, Cohen CD, Kretzler M. Modular activation of nuclear factor-kappaB transcriptional programs in human diabetic nephropathy. Diabetes. 2006;55:2993–3003. doi: 10.2337/db06-0477. [DOI] [PubMed] [Google Scholar]
- 19.DiPetrillo K, Coutermarsh B, Soucy N, Hwa J, Gesek F. Tumor necrosis factor induces sodium retention in diabetic rats through sequential effects on distal tubule cells. Kidney Int. 2004;65:1676–1683. doi: 10.1111/j.1523-1755.2004.00606.x. [DOI] [PubMed] [Google Scholar]
- 20.Kalantarinia K, Awad AS, Siragy HM. Urinary and renal interstitial concentrations of TNF-alpha increase prior to the rise in albuminuria in diabetic rats. Kidney Int. 2003;64:1208–1213. doi: 10.1046/j.1523-1755.2003.00237.x. [DOI] [PubMed] [Google Scholar]
- 21.Ducloux D, Bresson-Vautrin C, Chalopin J. Use of pentoxifylline in membranous nephropathy. Lancet. 2001;357:1672–1673. doi: 10.1016/s0140-6736(00)04830-3. [DOI] [PubMed] [Google Scholar]
- 22.Navarro JF, Mora C, Rivero A, Gallego E, Chahin J, Macia M, Mendez ML, Garcia J. Urinary protein excretion and serum tumor necrosis factor in diabetic patients with advanced renal failure: effects of pentoxifylline administration. Am J Kidney Dis. 1999;33:458–463. doi: 10.1016/s0272-6386(99)70182-4. [DOI] [PubMed] [Google Scholar]