Anti-LAMP-2 autoantibodies in ANCA-associated pauci-immune glomerulonephritis.

The association between anti–lysosome-associatedmembrane protein 2 (LAMP-2) autoantibodies and vasculitis is not new; it dates back to work by Kain et al.1 in 1995 that identified LAMP-2 as a target of antineutrophil cytoplasmic antibodies (ANCA) in 14 of 16 patients with a pauci-immune necrotizing crescentic glomerulonephritis. LAMP-2 makes an intriguing target for ANCA; it is a major constituent of the lysosomal membrane,2 plays a role in autophagy,3 and is involved inHLA class II antigen presentation. It is a highly glycosylated protein that is found not only lining neutrophil granules but also on the cell surface, where it is a ligand for E-selectin. Taken together, these diverse roles link LAMP-2 to the clearance of intracellular pathogens and the immune response, thereby raising the possibility that anti–LAMP-2 antibodies contribute significantly to the pathogenicity of ANCA. The work by Kain et al. addressed this directly in a follow-up 2008 study, showing that the immunization of rats with anti–LAMP-2 IgG was sufficient to precipitate a crescentic glomerulonephritis. They also showed that anti–LAMP-2 IgG activates human neutrophils as efficiently as antiproteinase 3 (anti-PR3) autoantibodies and—unlike anti-PR3 antibodies—anti–LAMP-2 antibodies are sufficient to induce the apoptosis of endothelial cells in vitro. Perhaps the most-discussed finding of that study, however, was evidence for molecular mimicry as a potential initiating factor in anti–LAMP-2 autoantibody formation. Specifically, an antibody-binding continuous epitope, P41–49 of LAMP-2, was found to share significant homology with the bacterial adhesion protein FimH, expressed on a subset of gramnegative bacteria. The fact that the homology should have been with gram-negative bacteria was surprising given the accumulated evidence for Staphylococcus aureus involvement in the pathogenesis of ANCA vasculitis. Nonetheless, 9 of 10 rats immunized with recombinant FimH proceeded to develop cross-reactive anti–LAMP-2 antibodies and a pauciimmune glomerulonephritis. This finding is important, because although long hypothesized, it represented the first time that molecular mimicry had been directly shown in ANCA; neither of the two other major ANCA classes, antiPR3 or antimyeloperoxidase, shares homology with known bacterial or viral antigens. Anti–LAMP-2 autoantibodies have the potential, therefore, to signal a major advance in our understanding of ANCAassociated glomerulonephritis: they are widely prevalent in active disease (14 of 16 patients in ref. 1 and 78 of 84 patients in ref. 6), precipitate pauci-immune glomerulonephritis in a rat model, activate human neutrophils, induce endothelial cell apoptosis in vitro, and bind an epitope on LAMP-2 shared with common gram-negative bacteria. However, these findings represent the work of one group of investigators and have not been yet been replicated in an independent laboratory; in JASN, the work by Roth et al. reports their attempts to replicate these results. In an initial cross-sectional cohort of 103 University of North Carolina patients with ANCA glomerulonephritis, they found an anti–LAMP-2 antibody prevalence of 21.1%, substantially lower than the prevalence reported by Kain et al. and similar to the prevalence of healthy volunteers and two disease control cohorts, including patients with active urinary tract infection and a variety of non-ANCA–associated glomerulonephritides. In a second cohort of 226 patients fromMassachusetts General Hospital with ANCA-associated vasculitis, anti–LAMP-2 antibody prevalence was 21.2%. The work by Roth et al. then determines autoantibody reactivity against the LAMP-2 P41–49 epitope possessing FimH homology, again finding a similar prevalence to their control cohort. Finally, they injected WKY rats with antibodies shown to be reactive against LAMP-2 and FimH. None of these rats went on to develop clinical or histologic glomerulonephritis. In contrast, JASN also contains a report by Kain et al. in which their initial findings of high anti–LAMP-2 antibody prevalence in ANCA-associated glomerulonephritis are recapitulated in additional cohorts drawn from three European centers: Vienna, Groningen, and Cambridge. Using three independent assays (ELISA, Western blot, and immunofluorescence), they report an overall frequency of 81% in patients presenting with an ANCA-associated glomerulonephritis, significantly higher than healthy controls and two sets of disease controls, one with a range of non-ANCA–associated renal diseases and another with active systemic lupus erythematosus. Together, these findings represent data from an additional 74 patients with untreated, new-onset ANCA-associated disease and 52 patients with treated ANCA-associated disease either during a flare or in remission. 9 The discordance between these two studies is concerning, and reconciling themwill be critical to determining the role of anti–LAMP-2 antibodies in the pathogenesis and clinical management of ANCA-associated vasculitis. The two most obvious sources of variation are the patient populations and the assay design, especially the choice of LAMP-2 antigen. A consistent feature of the studies by Kain Published online ahead of print. Publication date available at www.jasn.org.