Genes that link nephritis to autoantibodies and innate immunity

The rapid evolution of array techniques and the completion of the map of the human genome have led to a host of genomewide association studies over the past few years. Yet many of these studies have been disappointing; they identify genes that confer only a moderate increase in risk and explain only a small fraction of the known hereditary characteristics of the diseases.1 However, the results ofgenomewide association studies for idiopathic membranous nephropathy conducted by Stanescu and coworkers, reported in this issue of the Journal, are more striking than usual.2 The authors report highly significant associations between membranous nephropathy and single-nucleotide polymorphisms (SNPs) in genes that encode HLA-DQA1 and the M-type phospholipase A2 receptor (PLA2R).

The relativerisks for the two implicated alleles are remarkably similar in three independent French, Dutch, and British cohorts; for the HLA-DQA1 SNP (rs2187668), the odds ratios were calculated as 4.48, 3.76, and 5.33, respectively, and the corresponding odds ratios for the receptor SNP (rs4664308) were 1.87, 2.27, and 2.10. Even more striking is the multiplicative increase in the odds ratio for each riskallele added. Among persons carrying one risk allele at the rs4664308 locus, the odds ratio increases from 2.22 to 8.49 with the addition of one risk allele and to 31.03 with the addition of two risk alleles at the rs2187668 locus. Furthermore, the odds ratio for persons who are doubly homozygous is 78.46. Even though membranous nephropathy is rare, such an increase in risk has substantialimplications on the individual level.

The annual incidence of membranous nephropathy is approximately 1 case per 100,000 population — representing a lifelong risk of about 1 per 1000 for a person in a Western country who reaches average life expectancy. This implies a lifelong increased risk for persons with this specific genotype — at a level where screening and preventive measures would be meaningful,if effective interventions become available.

The results reported by Stanescu et al. confirm the findings of two recent studies from Asia (Taiwan and Korea) that used a candidate gene approach.3,4 Both those studies examined only SNPs in the PLA2R gene that resulted in amino acid changes, and a significant association was found for one SNP (rs35771982). Stanescu et al. do not report onrs35771982, but for a more C-terminally oriented SNP, rs3828323, both the European and Korean studies show weak but significant associations with the C allele. Taken together, these findings suggest that similar genetic mechanisms confer a predisposition to membranous nephropathy in different human populations.

Although a relation between the HLA system and membranous nephropathy has been recognized forsome time,5 the knowledge of an association between this disease and PLA2R is recent. In 2009, Beck and colleagues reported autoantibodies with specificity for PLA2R that were present in 70% of their patients with idiopathic membranous nephropathy but not in controls or patients with secondary membranous nephropathy.6 These results are now confirmed and extended in a letter by Debiec and Ronco inthis issue of the Journal.

The results of the genomewide association study by Stanescu et al. strongly suggest a direct interaction between HLA-DQ and PLA2R in the pathogenesis of membranous nephropathy. How would this occur? The most straightforward interaction would be that certain genetic variants of PLA2R yielded peptides with strong affinity for specific HLA-DQA1 variants thatsubsequently would confer a predisposition to autoantibody generation. However, the SNP in the PLA2R gene with the strongest association does not alter the amino acid sequence. The authors provide two possible explanations for this observation. First, the true association might not be with rs4664308 but rather with rs3749117, which is a nonsynonymous SNP reported to be in linkage disequilibrium with...