Antibodies to an Epstein Barr Virus protein that cross-react with dsDNA have pathogenic potential

Highlights

• Monoclonal Antibodies to EBNA-1, that cross-react with dsDNA.

• Can bind to isolated glomeruli and can deposit in the kidney when injected into mice.

• Can induce proteinuria and histopathologic alterations in the kidney consistent with glomerulonephritis.

• Can bind extracellular matrix antigens found in the GBM and mesangial matrix.

• Have pathogenic potential and may play a significant role in triggering SLE and lupus nephritis.

Abstract

Pathogens such as the Epstein Barr virus (EBV) have long been implicated in the etiology of systemic lupus erythematosus (SLE). The Epstein Barr virus nuclear antigen I (EBNA-1) has been shown to play a role in the development of anti-nuclear antibodies characteristic of SLE. One mechanism by which EBV may play a role in SLE is molecular mimicry. We previously generated two monoclonal antibodies (mAbs) to EBNA-1 and demonstrated that they cross-react with double-stranded DNA (dsDNA). In the present study, we demonstrate that these mAbs have pathogenic potential. We show that they can bind to isolated rat glomeruli and that binding can be greatly diminished by pretreatment of glomeruli with DNase I, suggesting that these mAbs bind dsDNA in the kidney. We also demonstrate that these antibodies can deposit in the kidney when injected into mice and can induce proteinuria and elicit histopathological alterations consistent with glomerulonephritis. Finally, we show that these antibodies can cross-react with laminin and collagen IV in the extracellular matrix suggesting that direct binding to the glomerular basement membrane or mesangial matrix may also contribute to the antibody deposition in the kidney. In summary, our results indicate that EBNA-1 can elicit antibodies that cross-react with dsDNA, that can deposit in the kidney, and induce kidney damage. These results are significant because they support the role of a viral protein in SLE and lupus nephritis.

Introduction

Pathogenic IgG antibodies to double-stranded DNA (dsDNA) are the hallmark of the autoimmune disease systemic lupus erythematosus (SLE) and are present in greater than 70 % of SLE patients (Fabrizio et al., 2015; Isenberg et al., 2007). In some patients, these antibodies can deposit in the kidney and induce kidney damage leading to lupus nephritis. It is unclear, how these antibodies arise, since dsDNA is not very immunogenic. One mechanism to explain the origin of anti-dsDNA antibodies is molecular mimicry by a viral or bacterial pathogen. It is thought that certain pathogens may express an epitope that structurally mimics dsDNA, thereby eliciting an antibody response that cross-reacts with dsDNA (Rojas et al., 2018). Environmental pathogens have long been thought to play a role in eliciting autoantibodies and triggering autoimmunity. Several studies have reported that bacterial pathogens can mimic dsDNA and elicit anti-dsDNA antibodies [4,5,6]. Mouse monoclonal antibodies to phosphorylcholine, a major component of the pneumococcal cell wall of Streptococcus pneumoniae, have been shown to cross-react with dsDNA. In human studies, anti-dsDNA antibodies in the sera of lupus patients have been found to cross-react with phosphorylcholine, suggesting that these antibodies may arise following a pneumococcal infection (Ray et al., 1996; Sharma et al., 2001). Antibodies to dsDNA obtained from human SLE sera have also been found to cross-react with a bacterial protein from Burkholderia fungorum, an organism that is both a plant and human pathogen (Zhang and Reichlin, 2008).

Several viruses have also been associated with lupus including Human Herpes Viruses HHV 4 (Epstein Barr Virus (EBV)), HHV 5 (Cytomegalovirus (CMV)), HHV-6, HHV-7, and HHV-8 as well as parvovirus B19, Human Papilloma Virus (HPV), and the exogenous human retroviruses human immunodeficiency virus (HIV) and human endogenous retroviruses (HERVs) (Nelson et al., 2014). HPV infection is prevalent among SLE patients and HPV peptides have been observed to share homology with the ribonucleoproteins, La, Sm B/B’ and Sm D (Segal et al., 2017). Antibodies to Sm B/B’ and Ro are observed in 30–50 % of lupus patients respectively. Cross-reactivity of antibodies to HPV with Sm proteins has led investigators to hypothesize that HPV could elicit autoantibodies that predispose to SLE. Hsieh et al. have demonstrated that peptide epitopes in the CMV protein, pp65, can molecularly mimic dsDNA and elicit an anti-dsDNA response when injected into mice (A.-H. Hsieh et al., 2017; A. H. Hsieh et al., 2011). In addition, they have observed that SLE patients have elevated titers of antibodies to pp65 compared to healthy individuals. Bonsignori et al. found that broadly neutralizing antibodies to HIV-1 gp120, present in the sera of SLE/HIV-1 patients, cross-react with dsDNA (Bonsignori et al., 2014).

EBV is a virus that has been extensively studied for its role in eliciting autoantibodies and initiating autoimmunity via molecular mimicry and it has long been linked to SLE. EBV infection is more prevalent among adolescent and adult lupus patients than age-matched healthy individuals and the levels of EBV dsDNA, antibodies to EBV proteins, and EBV transcripts are also higher among lupus patients than healthy individuals (Draborg et al., 2012; James et al., 1997a,b; James et al., 2001; James and Robertson, 2012; Moon et al., 2004; Sculley et al., 1986; Yokochi et al., 1989). In addition, antibodies to the EBV major nuclear protein 1, EBNA-1, have been shown to cross-react with several RNA binding proteins (Poole et al., 2006; Sabbatini et al., 1993). Previous studies revealed that EBNA-1 shares a proline-rich amino acid sequence with SmB/B’and nRNP and immunization of mice and rabbits with this EBNA-1 epitope elicited the production of antibodies to Sm B/B’ and nRNP (James et al., 1997b; Poole et al., 2008). Antibodies to an epitope in Ro60 have also been found to cross-react with an epitope in EBNA-1, although both epitopes do not have any primary sequence homology (McClain et al., 2005). The basis for this cross-reactivity is unknown, however, both cross-reactive epitopes have similar isoelectric points, so charge may play a role.

Our laboratory previously demonstrated that some monoclonal antibodies (mAbs) generated in response to EBNA-1, cross-react with dsDNA (Sundar et al., 2004; Yadav et al., 2016, 2011). Following immunization of BALB/c mice with the EBNA-1 protein, we generated anti-EBNA-1 mAbs that bind strongly to dsDNA, as demonstrated by ELISA and by an immunofluorescent Crithidia luciliae assay (Yadav et al., 2016, 2011). We characterized two unrelated mAbs to EBNA-1 that cross-react with dsDNA; an IgM antibody designated, 16D2, and an IgG antibody designated 3D4. Both mAbs were previously shown to utilize different V_H and V_L germline genes but the same J_H2 and J_k1 gene segments in their heavy and light chain genes, respectively. 3D4 utilizes the V_H7183 gene, which is frequently used by anti-dsDNA antibodies and the bb1 V_k light chain gene, while 16D2 utilizes the J558 V_H gene and the 8–30 V_k light chain gene (Yadav et al., 2016). Neither antibody is heavily mutated in their CDR regions, but both contain at least one arginine, one phenylalanine, and one tyrosine residue in their CDR3 V_H regions (Yadav et al., 2016). Basic and aromatic amino acids are frequently observed in the binding sites of anti-dsDNA antibodies (Ono et al., 1995; Radic et al., 1993; Sathyapriya and Vishveshwara, 2004). Both antibodies were also found to bind strongly to a 148 amino acid domain in the carboxyl region of EBNA-1, designated LS9C, which contains the EBV viral DNA binding site (VBS). Epitope mapping using overlapping peptides derived from LS9C, identified a 15 amino acid peptide designated, PFM-15 (aa_549−563) with the following sequence; PQPGPLRESIVCYFM, that is recognized by 3D4 and 16D2 (Yadav et al., 2016). This peptide was shown to partially inhibit 3D4 and 16D2 from binding to EBNA-1 and dsDNA by ELISA (Yadav et al., 2016). However, binding to this amino acid sequence was not as strong when tested by ELISA, as binding to the entire LS9C region, suggesting that PFM-15 is only a partial epitope and that the optimum epitope is conformational.

In the present study, we sought to further characterize the mAbs, 3D4 and 16D2, and examine whether they could deposit in the kidney of mice and induce nephritis. Not all anti-dsDNA antibodies are pathogenic. Our goal was to determine whether antibodies that arise following exposure to a viral protein such as EBNA-1, that cross-react with dsDNA, could have pathogenic potential and play a role in SLE.