Proteinase 3 associated with Wegener's granulomatosis

Wegener's granulomatosis (WG) is a form of systemic vasculitis characterized by granulomatous inflammation of the upper and lower airways, vasculitis, and necrotizing glomerulonephritis. It is strongly associated with anti‐neutrophil cytoplasmic antibodies against proteinase 3 (PR3‐ANCAs). Various in vitro observations provided strong evidence that autoimmune PR3‐ANCAs are directly involved in glomerular and vascular inflammation. However, little is known about the pathogenic significance of PR3‐ANCAs in vivo. Therefore, the generation of animal models helped to validate the suggested autoimmune origin and pathophysiology in WG. To characterize and improve the models, numerous studies were carried out to elucidate the effect of mouse/rat PR3‐ANCAs on neutrophil function as well as the role of CD4/CD8 in T and B cells and antibodies in the pathogenesis of the disease. Understanding the pathogenesis is therefore critical to relate these models to human studies hoping that they will be useful for better insight of WG and the development of specific therapies for the disease.


| INTRODUCTION
Wegener's granulomatosis (WG) is a debilitating and life-threatening autoimmune disease of unknown etiology and a major cause of pauciimmune necrotizing and crescentic glomerulonephritis. The classical triad of WG consists of (i) necrotizing granuloma of the upper and lower respiratory tract typically with mucosal inflammation and ulceration, (ii) necrotizing vasculitis involving both arteries and veins, and (iii) nephritis, which is a focal necrotizing glomerulitis with thrombosis of capillary loops. It has a strong and specific association with autoantibodies directed against proteinase 3 (PR3) ( Van der Geld et al., 2001). These circulating antineutrophil cytoplasmatic autoantibodies (ANCAs) are directed against conformational epitopes of PR3 and are highly sensitive and specific markers for the disease. The involvement of ANCA in the inflammatory tissue injury is supported by several observations and ANCA titers are well correlated with disease activity in several clinical studies (Franssen et al., 2000;Specks, 2000). In vitro, the interaction of ANCA with neutrophils resulted in the activation of polymorphonuclear neutrophil degranulation, superoxide secretion, release of lipid mediators, stimulation of neutrophil-related endothelial cytotoxicity, and secretion of cytokines (interleukin-1β) (Daouk et al., 1995;Han et al., 2003;Harper & Savage, 2000;Müller-Kobold et al., 1998).
PR3 is a neutral serine protease found in neutrophils and monocytes (Pankhurst & Savage, 2006). It is a highly folded protein with four disulfide bridges keeping its three-dimensional (3D) structure intact (Goldmann et al., 1999). One of its biological roles is the microbicidal activity independent of the proteolytic activity (Brooks et al., 1996). PR3 has an elastase-like enzymatic activity and can degrade extracellular matrix and basement membrane proteins leading to the migration of neutrophils through the basement membranes (Fujinaga et al., 1996;Pankhurst & Savage, 2006). In azurophilic granules as well as secretory vesicles, PR3 expression is also observed on the plasma membrane of resting neutrophils (Wiesner et al., 2005). The PR3 expression on neutrophils is increased in patients with active WG, and the expression level is correlated with disease activity (Niles et al., 1991).
2 | PR3-BINDING TO LIPID MEMBRANES, SIMULATIONS, AND MODELING Goldmann et al. (1999) showed that PR3 inserts into the hydrophobic region of liposomes and that the viability of this enzyme in the lipid environment is accessible to the natural and pathologic inhibitors such as α1-PI and ANCA. Although the attachment of PR3 to liposomes decreased its esterolytic activity by approximately 50% compared to control, that is, in the absence of lipids, the enzyme/ substrate complex did not induce a conformational change in PR3 when associating with liposomes. The insertion of PR3 into liposomes impaired either the "proper positioning" of PR3 for the substrate or the active site. However, the binding of the natural inhibitor α1-PI to PR3 showed a higher reduction of the enzyme activity in the presence compared to the absence of liposomes, whilst the autoantibody (ANCA) bound to PR3 in the presence of lipids was slightly more effective. These researchers also showed that the 3D structure of PR3 is like other proteases of the chymotrypsin family proteases and that its quaternary arrangement allows the formation of a hydrophobic "pore-like" structure, with Phe166, Ile217, Trp218, Leu223, and Phe224 from each monomer contributing to this hydrophobic patch and to the formation of the central cavity of the crystallographic tetramer (Goldmann et al., 1999). As the catalytic triad (Ser195, His57, and Asp102) and the putative substrate-binding site lies within a shallow depression on the "front" surface of PR3 and the antigenic loop for ANCA on the "back," that is, relatively remote from the surface of PR3 and from the catalytic and substrate-binding pocket, Goldmann et al. (1999) speculated that the hydrophobic patch may be involved in the insertion of PR3 into the lipid membrane. Whether additional surface receptor(s) can contribute to this interaction was also addressed in a study by Witko-Sarat et al. (1999). Broemstrup and Reuter (2010) carried out molecular dynamics simulations of PR bound to pure dimyristoylphosphatidylcholine (DMPC) and dimyristoyl-phosphatidylglycerol lipid bilayer and mixed bilayer of equimolar ratio. In this comprehensive study, these researchers described the interaction of the enzyme PR3 with the lipid membranes and proposed three types of interaction: (i) that five basic amino acids of the enzyme associate with the phosphate head groups via hydrogen bounds; (ii) that six hydrophobic amino acids V163, F165, F166, I217, L223, and F224 insert into the hydrophobic core below the carbonyl groups of the bilayers which confirmed experimental data from Goldmann et al. (1999); and (iii) that those six amino acids of the aromatic side-chain interact via cation-pi interaction with the choline groups of DMPC. They further found that amino acids of PR3 (believed to be important for membrane binding) are not conserved in the homolog of human neutrophil elastase (HNE) and therefore proposed that the amino acids R186, K187, R222, and W218 make major contributions to the stabilization and orientation of PR3 in the interfacial region which are not conserved in HNE.
These observations are in accordance with reports from (Goldmann et al., 1999;Witko-Sarat et al., 1999) that PR3 has only one specific membrane binding site compared to HNE. The esterolytic activity of PR3 is not impaired according to Campbell et al. (2000); however, Goldmann et al. (1999) reported an approximately 50% reduction of its catalytic activity when inserted in membrane liposomes. antibodies against human PR3 did not react with rat PR3 and rat myeloid cells, that is, there is probably no strong homology between rodents and human PR3. Jenne et al. (1997) found that human ANCA antibodies from 40 different patients with WG did not bind to the murine homolog of PR3. Other studies also showed that the epitopes of human PR3 recognized by the human autoantibody are not preserved on mouse PR3 (Shochet et al., 2020). In fact, the cloning of mouse PR3 demonstrated that the nucleotide sequence of murine PR3 is only 73% and the protein sequence only 69% identical to human PR3 sequences within the coding region of the mature PR3 enzyme (Jenne et al., 1997). Additionally, significant species' differences in physiochemical properties, that is, substrate specificities, enzyme kinetics towards synthetic peptide substrates, oxidized insulin B chains, and F I G U R E 1 PR3-ANCA pathogenicity. Two types of interaction between PR3-ANCA and neutrophils are presented by Granel et al. (2021): "one includes a link between the PR3-ANCA Fab and mbPR3 exposed at the surface of neutrophils and the other involves a bond between Fc of PR3-ANCA and FcgR. Pathogenicity of PR3-ANCA depends on many factors such as the expression of membrane PR3 on neutrophils, recognized epitopes, the presence/absence of Fcg-receptor polymorphisms, subclasses, and isotypes of PR3-ANCA, and finally the Fc glycosylation of PR3-ANCA" (taken from Granel et al. (2021) in Frontiers in Immunology with permission). ANCA, antineutrophil cytoplasmic antibody; PR3, proteinase 3

C Cell ell B Biology iology I International nternational
fibrinogen were detected between mouse and human PR3 (Wiesner et al., 2005).
Another way to induce PR3-ANCA in mice has been described by Rauova et al. (2002). Two weeks after the injection of human PR3-ANCA, mice developed PR3-specific anti-idiotypic antibodies and months later were positive for anti-idiotypic antibodies that reacted with PR3. Surprisingly, serum from these mice also reacted to human haplotype (H-2 g7 ) (Tisch & McDevitt, 1996). The mechanism of the association of major histocompatibility complex (MHC) class II molecules with autoimmune disease is not well understood, but a model by Ridgeway et al. (1999) suggests that NOD MHC class II molecules (binding self-peptides with low efficiency during thymic selection) permit self-reactive T cells to enter the periphery after escaping negative selection processes. In support of this hypothesis, reports by Latek et al. (2000) indicated that I-A g7 is unstable and a "poor peptide binder," and the immunization with multiple self-peptides induces an autoimmune response in NOD mice due to their lower threshold for breaking self-tolerance (Ridgeway et al., 1996). Therefore, Primo et al. (2010) hypothesized that it would be easier to induce a response to self-PR3 in NOD mice than in other mouse strains. Wiesner et al. (2005) reported a higher kinetic value (K m ) and a lower value for (k cat ) for murine PR3 compared to Goldmann et al. (1999). One reason for the difference between the kinetic mea- There is a strong association between the disease and the detection of high titers of c-ANCA and the presence of a granular cytoplasmatic staining pattern using an indirect immunofluorescence to ethanolfixed human neutrophils (Ewert et al., 1991;Franssen et al., 2000). In addition, activated granulocytes are present in the circulation of WG patients and the amount of surface-expressed PR3 correlates with disease activity (Müller-Kobold et al., 1998). Despite some similarities to the human disease, the presence of mPR3-ANCAs was not sufficient to produce a clinical manifestation of WG in these mice.

DATA AVAILABILITY STATEMENT
Data availability statement will be published.