Peroxidasin—a Novel Autoantigen in Anti-GBM Disease?

Basement membranes are highly specialized extracellular matrices, the evolution of which coincided with the emergence of multicellularity in animals. They associate with all epithelial and endothelial surfaces in the body, and they are essential for tissue morphogenesis both through provision of structural support and via their interaction with cell surface receptors to modulate cell behaviors, such as adhesion, migration, proliferation, and differentiation.

The lamellar structure of basement membranes arises from two independent polymeric networks: one of laminin and one of type IV collagen; they are linked by additional extracellular matrix proteins including nidogen (entactin) and perlecan, a heparan sulfate proteoglycan. The collagen IV network consists of a latticed arrangement of triple-helical protomers, which in humans are made up of a combination of six genetically distinct α-chains (α1–α6). These chains trimerize with each other to form specific triple-helical protomers: α1α1α2 (the most abundantly expressed), α3α4α5 (expression of which is almost exclusively restricted to glomerular and alveolar basement membranes), and α5α5α6. These triple helices polymerize extracellularly to form a two-dimensional mesh-like network via end-to-end association of their N-terminal 7S domains and by their C-terminal noncollagenous (NC1) domains, with two trimeric NC1 “caps” coming together to form hexameric NC1 complexes.

The quaternary structure of these hexamers is secured by the formation of specialized sulphilimine bonds that crosslink methionine and hydroxylysine residues on opposing NC1 domains.¹ These crosslinks provide critical structural reinforcement to collagen IV networks and, of note, they also confer immune privilege to pathogenic epitopes in the NC1 domains of α3 and α5 chains that are the target of the autoimmune response in antiglomerular basement membrane (anti-GBM) disease.²

It is known that the formation of these sulphilimine bonds is dependent on peroxidasin,³ the most recently identified member of the haem peroxidase family. This enzyme is embedded in the basement membrane, having a multidomain structure containing motifs found in other extracellular matrix proteins (leucine rich repeats, Ig domains, and a vWf domain) along with a catalytic peroxidase domain that acts locally to produce hydrobromous acid intermediates that oxidize methionine, forming the sulphilimine crosslink.⁴ Peroxidasin, like the basement membrane itself, is phylogenetically conserved throughout the animal kingdom and seems to be a critical component in the evolution of basement membranes and tissue development.⁵

In this issue of the Journal of the American Society of Nephrology, McCall et al.⁶ report the novel identification of antiperoxidasin antibodies in a significant proportion of patients with pulmonary renal syndromes. In anti-GBM disease, these antibodies were detected in almost 50% of patients. Using a unique resource of prediagnostic samples available via the US Department of Defense Serum Repository, in a small cohort, they show that antiperoxidasin antibodies could be detected several years before the onset of clinical disease in some patients. In vitro, the antibodies can inhibit hydrobromous acid formation by peroxidasin, one of its essential in vivo functions in this context.

Interestingly, they show that antiperoxidasin antibodies crossreact to myeloperoxidase (MPO), one of the recognized targets in ANCA-associated vasculitis (AAV). MPO is another member of the haem peroxidase family with which peroxidasin shares a degree of homology, and in a series of competitive ELISA studies, the authors find that antiperoxidasin positivity may account for a proportion of patients with “double-positive” anti-GBM and anti-MPO disease (with the remaining proportion showing “true” anti-MPO reactivity).

They further analyzed a group of patients with AAV without circulating anti-GBM antibodies. In this cohort, 14% of patients with MPO-AAV had reactivity to peroxidasin, which was not observed in patients with proteinase 3–AAV or drug-induced vasculitis. In this group, the presence of both MPO–specific and peroxidasin-specific antibodies was shown, confirming that their concurrence was not due to crossreactivity in patients with MPO-AAV. These observations suggest that a range of epitope specificities is present within the observed antiperoxidasin antibodies, with some recognizing peroxidasin but not MPO and some recognizing both antigens. In the studied AAV cohort, antiperoxidasin antibodies were associated with more severe vasculitis activity as determined by Birmingham Vasculitis Activity Score (BVAS), although the inclusion of a high proportion of patients in apparent disease remission in the anti-MPO alone group (with BVAS=0) in this comparison might suggest that they are a feature of active disease rather than a marker of disease severity specifically.

It is suggested that antiperoxidasin antibodies may contribute to disease pathogenesis in anti-GBM disease via inhibition of sulphilimine crosslinking in type IV collagen, which may then render epitopes in the NC1 domains of α3 and α5 chains susceptible to immune detection or binding of pathogenic anti-GBM antibodies. In both anti-GBM disease and MPO-AAV, they may contribute to vascular injury or impede vascular repair via inhibition of peroxidasin activity, a suggestion that is supported by the finding of upregulated peroxidasin expression in experimental models of kidney fibrosis.⁷

The origin and pathogenic significance of these autoantibodies, however, are currently unclear. It is possible that they arise as an “epiphenomenon” during AAV or anti-GBM disease, perhaps as a result of intra- and/or intermolecular epitope spreading after a primary response to MPO or GBM antigens, and thus they may not have a directly pathogenic role. Autoantibodies to other noncollagen constituents of the GBM, such as entactin for example, have been reported in patients with anti-GBM disease.⁸ In addition, it has been shown that, in vitro, chemical inhibition of peroxidasin does not break crosslinks after bond formation but specifically prevents bond generation,³ and therefore, it is not clear that the presence of inhibitory antiperoxidasin antibodies could disrupt established collagen IV networks to initiate anti-GBM disease, although they may contribute to impaired basement membrane turnover or repair.

Pathogenic autoreactivity to peroxidasin might also be expected to manifest systemic features outside of the renal and pulmonary systems given the expression of peroxidasin in other basement membranes. Severe loss-of-function mutations in peroxidasin in Drosophila (the species in which peroxidasin was first identified⁹) result in nonviable larvae, with widespread disruption of basement membranes in the midgut visceral muscles.³ In mice, peroxidasin mutations may result in perinatal lethality and developmental ocular defects.¹⁰ These defects recapitulate the features of anterior segment dysgenesis in humans, which has been associated with mutations in peroxidasin,¹¹ although such mutations have not, to our knowledge, been identified in association with human renal disease.

The findings reported by McCall et al.⁶ are, however, highly novel and identify a new autoantibody target to add to the repertoire seen in pulmonary renal syndromes. Peroxidasin is expressed within the GBM, and antiperoxidasin antibodies with inhibitory activity can be identified before the onset of clinical disease, suggesting that they may contribute to disease pathogenesis. In addition, they have clinical implications given that antiperoxidasin reactivity may define a subset of patients with anti-GBM disease who are not truly anti-MPO positive and thus do not require maintenance immunosuppression.¹² Further characterization of these antiperoxidasin antibodies and their associations with disease phenotype, particularly in patients found to be “double positive” for ANCA and anti-GBM autoantibodies, are now required to fully understand both their pathogenic and clinical significance.

Disclosures

None.