Portal de Periódicos da CAPES

TNFA2 and d2 alleles of the tumor necrosis factor alpha gene polymorphism are associated with onset/occurrence of idiopathic membranous nephropathy

2007; Elsevier BV; Volume: 71; Issue: 5 Linguagem: Inglês

10.1038/sj.ki.5002054

1523-1755

Damien Thibaudin, L. Thibaudin, P Berthoux, Christophe Mariat, Jean Pierre de Filippis, Laurent Busé, É. Alamartine, F. Berthoux,

Pregnancy and Medication Impact

Idiopathic membranous nephropathy (IMN) has a strong association with the major histocompatibility complex HLA B8DR3(17)DQ2 haplotype. The tumor necrosis factor (TNF)A gene is located within the major histocompatibility complex region on chromosome 6. We have studied the influence of two functional polymorphisms; the -308 (promoter region) and the TNFd microsatellites on initiation and/or progression of IMN. This was a case–control study comparing data from 100 Caucasians patients (67 male subjects; 67%) with IMN to 232 Caucasians local controls (171 male subjects; 74%). We have analyzed genotypes and alleles distributions and the role of these polymorphisms in disease progression towards end-stage renal failure or patient death. For -308 TNFA polymorphism, distribution of genotypes was significantly different between IMN and controls (χ2=16.25; P=0.0003): the A2 allele frequency was 28.0% in IMN vs 15.3% in controls (χ2=14.57; P=0.0001). For TNFd polymorphism, alleles distribution (from d1 to d7) was also significantly different between IMN and controls (χ2=56.74; P<0.0001) with both diminished d3 allele frequency (χ2=27.30; P<0.0001; Pc=0.001) and increased d2 allele frequency (χ2=29.95; P<0.0001; Pc=0.001) in IMN. We could not isolate any significant and independent influence of these different genotypes on IMN disease progression. The TNFA2 and TNFd2 alleles were strongly associated with occurrence/initiation of IMN and should be considered as susceptibility genes for this disease. Idiopathic membranous nephropathy (IMN) has a strong association with the major histocompatibility complex HLA B8DR3(17)DQ2 haplotype. The tumor necrosis factor (TNF)A gene is located within the major histocompatibility complex region on chromosome 6. We have studied the influence of two functional polymorphisms; the -308 (promoter region) and the TNFd microsatellites on initiation and/or progression of IMN. This was a case–control study comparing data from 100 Caucasians patients (67 male subjects; 67%) with IMN to 232 Caucasians local controls (171 male subjects; 74%). We have analyzed genotypes and alleles distributions and the role of these polymorphisms in disease progression towards end-stage renal failure or patient death. For -308 TNFA polymorphism, distribution of genotypes was significantly different between IMN and controls (χ2=16.25; P=0.0003): the A2 allele frequency was 28.0% in IMN vs 15.3% in controls (χ2=14.57; P=0.0001). For TNFd polymorphism, alleles distribution (from d1 to d7) was also significantly different between IMN and controls (χ2=56.74; P<0.0001) with both diminished d3 allele frequency (χ2=27.30; P<0.0001; Pc=0.001) and increased d2 allele frequency (χ2=29.95; P<0.0001; Pc=0.001) in IMN. We could not isolate any significant and independent influence of these different genotypes on IMN disease progression. The TNFA2 and TNFd2 alleles were strongly associated with occurrence/initiation of IMN and should be considered as susceptibility genes for this disease. Membranous nephropathy (MN) is a rare disease and is classically the prototype of immune complexes glomerulonephritis. The rat model of MN is the heymann glomerulonephritis (active and passive forms).1.Heymann W. Hackel D.B. Harwood S. et al.Production of nephrotic syndrome in rats by Freund's adjuvants and rat kidney suspensions.Proc Exp Biol Med. 1959; 100: 660-664Crossref PubMed Scopus (397) Google Scholar Autoantibodies directed against glycoprotein 330 (gp330) or megalin-bound complexes were found in rats but not in humans.2.kerjaschki D. Pathogenic concepts of membranous glomerulopathy (MGN).J Nephrol. 2000; 13: S96-S100PubMed Google Scholar Hypothesis on antibodies deposition mechanisms are: (i) deposition of circulating immune complexes, (ii) deposition of antibodies on kidney constitutive antigen and, (iii) deposition of antibodies on kidney cell-fixed antigen so called planted antigens.3.Shankland S.J. New insights into pathogenesis of membranous nephropathy, editorial.Kidney Int. 2000; 57: 1204-1205Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar The pivotal cell of the disease is the glomerular visceral epithelial cell also called podocyte. The complement membrane attack complex, C5b-9, plays a crucial role and acts as a sublytic pattern on podocyte with phospholipase A2 activation and DNA damage rather than lytic injury.4.Montinaro V. Lopez A. Monno R. et al.Renal C3 synthesis in idiopathic membranous nephropathy: correlation to urinary C5b-9 excretion.Kidney Int. 2000; 57: 137-146Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar, 5.Cybulsky A.V. Takano T. Papillon J. et al.Complement-induced phospholipase A2 activation in experimental membranous nephropathy.Kidney Int. 2000; 57: 1052-1062Abstract Full Text Full Text PDF PubMed Scopus (37) Google Scholar, 6.Pippin J.W. Durvasula R. Petermann A. et al.DNA damage is a novel response to sublytic complement C5b-9-induced injury in podocytes.J Clin Invest. 2003; 111: 877-885Crossref PubMed Scopus (113) Google Scholar In humans, a large variety of antigens have been identified and deposited in the kidney, more precisely on the epithelial side of glomerular basement membrane together with immunoglobulins (as antibodies) and complement. These antigenic epitopes may belong to infectious agent (such as hepatitis B virus as an example), tumors or toxic agent (such as gold or mercury). When a specific cause is found, the disease is called secondary MN. By contrast, in the majority of human cases, no cause could be identified and the disease is then called idiopathic membranous nephropathy (IMN). Tumor necrosis factor (TNF)α is a proinflammatory, antitumoral, and activator of immune system cytokine.7.Rink L. Kircner H. Recent progress in tumor necrosis factor-alpha field.Int Arch Allergy Immunol. 1996; 111: 199-209Crossref PubMed Scopus (133) Google Scholar,8.Ruddle N.H. Tumor necrosis factor (TNFα) and lymphotoxin (TNFβ).Curr opin Immunol. 1992; 4: 327-332Crossref PubMed Scopus (138) Google Scholar The TNFA gene lies within the major histocompatibility complex genes on short arm of chromosome 6, in the human leukocyte antigen (HLA) class III region, in a position defined as 250 kb centromeric to HLA-B locus and about 850 kb telomeric to HLA-DR locus.9.Wilson A.G. de Vries N. Pociot F. et al.An allelic polymorphism within the human tumor necrosis factor alpha promoter region is strongly associated with HLA, A1, B8 and DR3 alleles.J Exp Med. 1993; 177: 557-560Crossref PubMed Scopus (670) Google Scholar Different polymorphisms have been described for TNFα cytokine gene(s), but we focused on two functional variants associated with increased TNFα serum level:•The TNFA single-nucleotide polymorphism at position -308 in the promoter region9.Wilson A.G. de Vries N. Pociot F. et al.An allelic polymorphism within the human tumor necrosis factor alpha promoter region is strongly associated with HLA, A1, B8 and DR3 alleles.J Exp Med. 1993; 177: 557-560Crossref PubMed Scopus (670) Google Scholar with a guanine (TNFA1 allele) to adenine (TNFA2 allele) mutation (G/A). This TNFA2 allele is associated with the HLA A1/B8/DR17/DQ2 extended haplotype, which was also reported as a marker of autoimmunity.9.Wilson A.G. de Vries N. Pociot F. et al.An allelic polymorphism within the human tumor necrosis factor alpha promoter region is strongly associated with HLA, A1, B8 and DR3 alleles.J Exp Med. 1993; 177: 557-560Crossref PubMed Scopus (670) Google Scholar•The TNFd microsatellite polymorphism10.Udalova I.A. Nedospasov S.A. Webb G.C. et al.Highly informative typing of human TNF locus using six adjacent polymorphic markers.Genomics. 1993; 16: 180-186Crossref PubMed Scopus (291) Google Scholar located 8 kb downstream to TNFA gene. It consist of various dinucleotide repeats (GA) termed TNFd1–d7 with a greater TNFα cytokine production associated with TNFd3 allele. IMN was also found to be an HLA-associated disease and more precisely with B8 and/or DR17(3) in the caucasian population.11.Glotz D. Bariety J. Druet P.H. Glomérulonéphrites extramembraneuses.Rev Prat. 1991; 41: 2424-2429PubMed Google Scholar, 12.Berthoux F.C. Berthoux P. Hassan A.A. et al.Immunogénétique des glomérulonéphrites extramembraneuses primitives.Presse Med. 1990; 19: 990-993PubMed Google Scholar, 13.Berthoux F.C. Laurent B. Le Petit J.C. et al.Immunogenetics and immunopathology of human primary membranous glomerulonephritis: HLA-A B DR antigens; functional activity of splenic macrophage Fc-receptors and peripheral blood T-lymphocyte subpopulations.Clin Nephrol. 1984; 22: 15-20PubMed Google Scholar The relative risk to develop a MN with a gold treatment was increased 32-fold if you carry the DR3 antigen.11.Glotz D. Bariety J. Druet P.H. Glomérulonéphrites extramembraneuses.Rev Prat. 1991; 41: 2424-2429PubMed Google Scholar Moreover, B8/DR3 haplotype was associated with high production of TNFα in vivo and in vitro.14.Lio D. Candore G. Colombo A. A genetically determined high setting of TNFα influences immunologic parameters of HLA-B8, DR3 positives subjects: implications for autoimmunity.Hum Immunol. 2001; 62: 705-713Crossref PubMed Scopus (84) Google Scholar A direct role of TNFα in IMN is possible: Wu et al.15.Wu T.H. Tsai C.Y. Yang W.C. Excessive expression of the tumor necrosis factor-alpha gene in the kidneys of patients with membranous glomerulonephritis.Chug Hua I Hsueh Tsa Chih. 1998; 61: 524-530Google Scholar found an excessive expression of the TNFA gene in the kidneys of patients with membranous glomerulonephritis. TNFα was expressed by glomerular visceral epithelial cells in human MN.16.Neale T.J. Ruger B.M. Macaulay H. et al.Tumor necrosis factor-alpha is expressed by glomerular visceral epithelial cells in human membranous nephropathy.Am J Pathol. 1995; 146: 1444-1455PubMed Google Scholar Urinary excretion of TNFα was increased in IMN vs controls; this was correlated with overexpression of leukocyte function antigen 1 in kidney interstitial tissue.17.Honkanen E. von Willebrand E. Teppo A.M. et al.Adhesion molecules and urinary tumor necrosis factor-alpha in idiopathic membranous glomerulonephritis.Kidney Int. 1998; 53: 909-917Abstract Full Text Full Text PDF PubMed Google Scholar Moreover, TNFα was able to induce dyscohesion at the slit membrane level with disassembly of actin filaments;18.Tabibzadeh S. Kong K.F. Kapur S. et al.TNF-alpha induces dyscohesion of epithelial cells – association with disassembly of actin filaments.Endocr J. 1995; 3: 549-556Crossref Scopus (9) Google Scholar indeed, Yuan et al.19.Yuan H. Takeuchi E. Taylor G.A. et al.Nephrin dissociates from actin, and its expression is reduced in early experimental membranous nephropathy.J Am Soc Nephrol. 2002; 13: 946-956Crossref PubMed Google Scholar found that nephrin dissociates from actin with a reduced expression in early experimental MN. Taken together, these data led us to investigate whether TNFα gene polymorphisms could play a significant role in patients with IMN. The aim of this study was to detect any influence of these two TNFα polymorphisms: – on the onset/occurrence of the disease by comparing alleles and genotypes distributions between IMN and appropriate controls and – on the progression of the disease by comparing clinical/pathological parameters according to the different genotypes either single or in combination. The results are given in Table 1. The genotypes distribution between IMN patients and controls was significantly different (χ2=16.25; P=0.0003). More precisely, the A2 allele was significantly more frequent in IMN patients vs controls (28 vs 15.3%; χ2=14.57; P=0.0001). Patients carrying the A2 allele (A2/A2+A1/A2) were 51 vs 28.0% in controls (χ2=16.24; P<0.0001). IMN, idiopathic membranous nephropathy; TNF, tumor necrosis factor. The results are given in Table 2. The six alleles distribution was significantly different in IMN vs controls (χ2=56.74; P<0.0001, Pc=0.001). Subanalysis demonstrated in IMN patients a significant decreased frequency of the d3 allele vs controls (χ2=27.30; P<0.0001, Pc=0.001) and a contrario a significant increase of d2 allele as compared to controls (χ2=29.95; P<0.0001, Pc=0.001). IMN, idiopathic membranous nephropathy; TNF, tumor necrosis factor. The results are given in Table 3. The increased frequencies of A2 and/or d2 alleles were pooled (52.0% in IMN patients vs 30.2% in controls, χ2=14.32; P=0.0002). Note that there were only four A2/d2 individuals in the control group vs 24 in the IMN group. IMN, idiopathic membranous nephropathy; TNF, tumor necrosis factor. The number of observed genotypes was not significantly different by the χ2 test from the theoretical numbers calculated from the Hardy–Weinberg equilibrium using the allelic frequencies (see Appendix). The results are given in Table 4. We found the B8/DR3/DQ2 haplotype in 54% of IMN patients. All A2A2 patients were B8/DR3/DQ2. There was a strong association in patients between A2 allele and extended haplotype (P<0.001) on one hand and between d2 allele and extended haplotype (P<0.003). We observed the same finding for combined genotype A2 a/o d2 and extended haplotype (75.9 vs 23.9%, P<0.0001). In addition, d2 and d3 alleles were mutually exclusives. It is important to note that in the IMN subgroup without HLA-B8/DR3/DQ2 haplotype, the genotypes and alleles concerning the different TNF polymorphisms were not different from controls. TNF, tumor necrosis factor. The HLA-B8 and HLA-DR3 antigens distribution in IMN patients compared with controls confirmed the already well-known and significant association between HLA-B8 and/or HLA-DR17(3) in IMN (data not shown). We compared clinical and pathological data between patients bearing the A1A1 genotype vs those with A1A2/A2A2 genotypes, that is, with the A2 allele. There was no significant difference both at diagnosis and at last follow-up. We compared patients with d2/non-d2 vs non-d2/non-d2 genotypes. There were no patients homozygous for d2. Again, we could not find any difference. We compared A2 a/o d2 with non-A2/non-d2 patients. Once again, we found no difference. All these TNFα genotypes played no significant role if any in the progression of the disease. Fifty-four patients (54.0%) had the B8/DR3/DQ2 haplotype. We found a significant difference for follow-up time from onset: 104.8 (±88.4) for patients with B8/DR3/DQ2 haplotype vs 70.4 (±62.4) months (P=0.03) but these patients were younger at the onset of their disease (46.3 vs 49.6 years, NS). But again, this haplotype had no role in the progression of the disease. Results are given in Figures 1, 2 and 3: All these TNFα polymorphisms have no influence on survival without renal/patient death. We have used the Cox monovariate analysis in the search of significant factors influencing survival without renal/patient death (Table 5). We have found only two factors: gender and amount of proteinuria. Multivariate Cox regression demonstrated that only gender was an independent risk factor for progression to renal/patient death: male gender was associated with a relative risk of about 11 compared to 1 for female. CI, confidence interval; CRF, chronic renal failure; d, day; ESRF, end-stage renal failure; IMN, idiopathic membranous nephropathy; RB1, renal biopsy 1; RR, relative risk; s.e., standard error; y, year. In this report, TNFA2 allele (-308 promoter) and TNFd2 allele (microsatellite TNFd) were strongly associated with the onset/occurrence of IMN; they are susceptibility genes for IMN and a contrario, d3 allele could be considered as a protective gene. D2 and d3 alleles are mutually exclusive in our patients. Bantis et al.20.Bantis C. Heering P.J. Aker S. et al.Tumor necrosis factor-α gene G-308A polymorphism is a risk factor for the development of membranous glomerulonephritis.Am J Nephrol. 2006; 26: 12-15Crossref PubMed Scopus (27) Google Scholar had performed a similar study but limited to TNFA polymorphism in 53 IMN patients and 100 controls. A2A2 was found in 9.4% in IMN patients and 2.0% in controls, and A2 allele in 34% in IMN patients and 15% in controls. Our present study found similar results (A2 allele: 15.3% in controls vs 28.0% in IMN). This finding is different from a population of immunoglobulin A nephropathy patients where d3 allele is the most frequent.21.Tuglular S. Berthoux P. Berthoux F. Polymorphisms of the tumour necrosis factor α gene at position -308 and TNFd microsatellite in primary IgA nephropathy.Nephrol Dial Transplant. 2003; 18: 724-731Crossref PubMed Scopus (37) Google Scholar In addition, d2 and A2 alleles were strongly associated with the highly conserved HLA-B8/DR3/DQ2 extended haplotype. These different markers are in strong disequilibrium. This HLA haplotype is classically associated with IMN.11.Glotz D. Bariety J. Druet P.H. Glomérulonéphrites extramembraneuses.Rev Prat. 1991; 41: 2424-2429PubMed Google Scholar, 12.Berthoux F.C. Berthoux P. Hassan A.A. et al.Immunogénétique des glomérulonéphrites extramembraneuses primitives.Presse Med. 1990; 19: 990-993PubMed Google Scholar, 13.Berthoux F.C. Laurent B. Le Petit J.C. et al.Immunogenetics and immunopathology of human primary membranous glomerulonephritis: HLA-A B DR antigens; functional activity of splenic macrophage Fc-receptors and peripheral blood T-lymphocyte subpopulations.Clin Nephrol. 1984; 22: 15-20PubMed Google Scholar By contrast, we found that A2 and/or d2 alleles played no significant role if any in the progression of IMN disease. We have observed a global proportion of end-stage renal failure (ESRF)/dialysis death around 20%; a contrario the proportion of patients without is about 80%. The small number of patients could permit only to detect a large difference of outcome (very unlikely): proportion reduces to 0.55 with the power of 80% (54 × 2 patients needed from the Table).22.Arkin C.F. Wachtel M.S. How many patients are necessary to assess test performance?.JAMA. 1990; 263: 275-278Crossref PubMed Scopus (113) Google Scholar So clearly we were under powered to detect a smaller difference. Our negative findings concerning the association between the different alleles and IMN outcome could be due to this underpower (number of patients too limited). Actually it seems important to study polymorphisms that have been shown to have a direct effect on gene function.23.Colhoun H.M. McKeigue P.M. Smith G.D. Problems of reporting genetic associations with complex outcomes.Lancet. 2003; 361: 865-872Abstract Full Text Full Text PDF PubMed Scopus (974) Google Scholar This is the case for the two studied TNFA polymorphisms: in hematopoietic stem cell transplantation, the presence of an adenine at position -308 (TNFA2 allele) has been reported to result in a twofold greater level of transcription than the common A1 allele;24.Kroeger K.M. Carville K.S. Abraham L.J. The -308 tumor necrosis factor-alpha promoter polymorphism effects transcription.Mol Immunol. 1997; 34: 391-399Crossref PubMed Scopus (806) Google Scholar – in Crohn's patients with fistulizing disease, those carrying A1A2 as compared to A1A1 patients had a significant increase of serum level of TNFα (58+/-79 vs 8+/-19 pg/ml, P 140/90 mmHg on at least two occasions and/or the regular use of antihypertensive medications, the presence of CRF, defined as calculated creatinine clearance according to Cockcroft and Gault formula below 60 ml/min for at least 3 months39.National Kidney Foundation K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: Evaluation, Classification and Stratification.Am J Kidney Dis. 2002; 39: S46-S75Google Scholar and the presence of ESRF, defined as the need for renal replacement therapy (Cockcroft was always below 15 ml/min, but usually between 8 and 10). The scoring was performed by our renal pathologist as a routine procedure without knowledge of the clinical data. We use classification in four stages by optical microscopy: I–IV according to the description by Ehrenreich.40.Ehrenreich T. Churg J. Pathology of membranous nephropathy.Pathol Annu. 1968; 3: 145-186Google Scholar The primary combined end point was occurrence of either ESRF (renal death) or patient death. The secondary end points were occurrence of hypertension and occurrence of CRF, as already defined. Our control group was matched for gender according to the male predominance of IMN and consisted of 232 local Caucasian healthy adults (171 male subjects, 74%) and will represent the local general population. Age at sampling was different in controls (41.6 years). vs IMN patients (47.8 years) (χ2=3.41, P<0.001). In controls, weight was 69.96 kg, height 1.72 m and body mass index 23.58. Controls are 77 local healthy volunteers and 155 blood donors, all living in the Saint-Etienne area. Genomic DNA was obtained from peripheral blood mononuclear cells with a classical phenol–chloroform extraction method. A polymerase chain reaction coupled with restriction fragment length polymorphism (PCR) was used for amplification and accurate genotyping. The sequence of the primers was chosen for the amplification of a 289 bp DNA fragment, including the polymorphic nucleotide ((G-308A), were the following: 5′-AGGCAATAGGTTTTGAGGGCCAT-3′ (sense) and 5′-CAGCGGAAAACTTCCTTGGT-3′ (antisense). The PCR technique was performed on 250 ng of extracted DNA: DNA amplification with 1.25 U of Taq DNA polymerase (Gibco-BRL-Invitrogen, Cergy Pontoise, France) in 20 mM Tris-HCl containing 50 mM KCl, 1.5 mM MgCl2, 200 μmol/l dNTP, and 0.5 μM of each primer. After an initial denaturing time of 5 min at 90°C, PCR reactions were run for 35 cycles including 30 s at 94°C, 30 s at 60°C, and 1 min at 72°C. The PCR product was digested by 5 U of the enzyme NcoI at 37°C for 3 h and then subjected to electrophoresis in 2% agarose gel (Metaphor agarose; FMC-BioProducts, Rocklands, Maine, USA) stained with ethidium bromide. The TNFA-1 allele gave two bands at 244 and 20 bp respectively and the TNFA-2 allele gave only one band at 264 bp. As the very light band of 20 bp migrates too fast to be detected on the gel, the TNFA-1 allele was in practice characterized by only one band at 244 bp. Finally, the genotypes were defined as 1/1 (homozygous) in case of a unique band at 244 bp, as 2/2 (homozygous) in case of a unique band at 264 bp and as 1/2 (heterozygous) in case of two bands – one at 244 bp and one at 264 bp. The polymorphism is based on the number of GA repeats. It was detected by PCR followed by polyacrylamide gel electrophoresis to separate the different DNA fragments of 126, 128, 130, 132, 134, and 138 bp. The primers used were: TNFd12 sense, 5′-CATAGTGGGACTCTGTCTCCAAAG-3′; TNFd11 antisense, 5′-AGATCCTTCCCTGTGAGTTCTGCT-3′. For the PCR, we used 250 ng of extracted DNA, amplified with 1.25 U of Taq polymerase (Gibco BRL) in 20 mM Tris-HCl containing 50 mM KCl, 1.5 mM MgCl2, 200 μm of each dNTP, and 0.4 μM of each primer. After an initial denaturation at 94°C for 2 min, PCR reactions were run for 30 cycles including 15 s at 94°C, 15 s at 55°C, 30 s at 72°C and one cycle with 2 min at 72°C and 1 min at 30°C. The PCR products obtained were submitted to polyacrylamide gel electrophoresis (12% acrylamide/bisacrylamide 19:1) and the bands obtained permitted the characterization of the different alleles d1, d2, d3, d4, d5, and d7; d6 was never found. For both polymorphisms, we have verified that the genotype distributions observed were in accordance with the Hardy–Weinberg equilibrium. These determinations were carried out on a routine basis by the HLA typing laboratory of the Etablissement Français du Sang in Saint-Etienne (Dr JC Le Petit), involved in our transplant program. All 100 patients were HLA types using molecular techniques for class II antigens (except for few patients) and serological techniques for class I antigens. The HLA-B8 and HLA-DR17 antigens distributions in IMN patients were compared to the local controls established by the HLA typing laboratory. The different allelic frequencies and genotype distributions were compared using the χ2 contingency table for qualitative variables. When applicable, the P-value obtained was corrected (Pc) for the number of alleles tested (six for TNFd). With the results of A1 allele distribution in controls approximately 0.85 and in patients 0.70 in a total group of 664 alleles (332 individuals but imbalance of numbers between controls and IMN) the power is over 95% with risk α=0.05 as shown in specialized table of Arkim and Wachted.22.Arkin C.F. Wachtel M.S. How many patients are necessary to assess test performance?.JAMA. 1990; 263: 275-278Crossref PubMed Scopus (113) Google Scholar The qualitative variables, such as pathological stages (I+II vs III+IV), hypertension, CRF, and ESRF, were compared using the χ2 contingency table. The quantitative variables, such as proteinuria, follow-up duration and serum creatinine, were compared using either the unpaired t-test or non-parametric test (Mann–Whitney U-test) when applicable. Survival curves were analyzed according to the Kaplan–Meier method: the events were ESRF/death or CRF; time zero was the date of RB1 (onset of the disease) and the final date was either the event date or the last follow-up date (censored patients). The curves obtained were compared according to the log-rank test. In case of significant results, additional multivariate Cox regression analysis was performed with the following covariates at RB1: age, gender, amount of proteinuria, and class of RB1 stages. Example for TNFA (Table 1) in IMN patients: A1 allele frequency=0.720; A2 allele frequency=0.280; A1A1 genotype=(0.720)2 × 100=52; A2A2 genotype=(0.280)2 × 100=8; A1A2 genotype=2 (0.720 × 0.280) × 100=40. The observed genotype distribution was 49, 5, and 46, respectively, but not different from the theoretical distribution: 52, 8, and 40 (χ2=1.20, P=NS).