Abstract

Neurofascin (NF) is a cell-adhesion molecule that is found at the nodes of Ranvier. The 186 kDa isoform of neurofascin (NF186) is expressed on the axon in the exposed node, and the 155 kDa isoform (NF155) is expressed on myelinating glia at the paranode. NF186 is essential for clustering of sodium channels to the nodes while NF155 is needed for close paranodal interactions between myelinating glia and axons. The neurofascins are found in both the peripheral and central nervous system (PNS and CNS). NF-specific autoantibodies were identified in serum of multiple sclerosis (MS) patients using a proteomics approach with two-dimensional Western blotting of human myelin glycoproteins. A monoclonal antibody (mAb) specific for NF was shown to induce axonal injury in an animal model of MS, experimental autoimmune encephalomyelitis. This indicated that NF is a relevant autoantibody target in patients with inflammatory diseases of the nervous system (central and peripheral), but actual abundance of anti-NF autoantibodies is unknown. The objectives of the thesis were the following: 1) Develop assays to detect autoantibodies against human NF. 2) Determine the prevalence in patients with MS and with inflammatory diseases of the PNS. 3) Characterize the reactivity by immunoglobulin isotyping, serial dilution, epitope mapping, and staining of nodal structures in tissue sections. 4) Affinity purify anti-NF antibodies from plasma exchange material. 5) Determine possible in vivo effects of anti-NF antibodies in the PNS using a neuritis animal model. First, we expressed the complete human NF155 and NF186 on the surface of stable human cell lines, produced the complete extracellular portion of the NFs in HEK293 cells, and expressed truncated variants of the NFs in E. coli. With these reagents, we set up three antibody detection assays: cell-based assay by flow cytometry, ELISA, and Western blot. These assays were validated using NF-specific monoclonal and polyclonal antibodies, and optimized with a test cohort of serum samples. We screened 687 serum and 48 plasma exchange samples from patients with MS (n = 233), inflammatory diseases in the PNS (n = 294), and controls (n = 208). From serum analysis, we observed low prevalence of anti-NF reactivity (3%) by flow cytometry and/or ELISA despite broad reactivity in almost half of the serum samples analyzed by Western blot. Reactivity observed by flow cytometry and by ELISA were congruent only in the patients with the highest reactivities. The anti-NF antibodies were NF-isoform specific, mainly IgG subclasses, and at high titres in some cases. Using truncated variants of NF fused to super green fluorescence protein (sGFP), we showed that reactivity of anti-NF Abs was largely directed towards the membrane proximal extracellular domains that are unique to each isoform, while the membrane distal immunoglobulin-like domains and fibronectin domains were not recognized. A small proportion (3%; 8/254) of patients with GBS and CIDP showed reactivity to human NF by ELISA. A few showed a particularly high reactivity (up to 1:10 000 dilution) to NF. Two CIDP patients showed a particularly high (up to 1:10 000 dilution) anti-NF155 reactivity by FACS and ELISA, recognized paranodes in tissue sections, and exhibited dominant IgG4 subclass usage. Another CIDP patient who benefited from plasma exchange had a persistent anti-NF155 reactivity by ELISA in serum, and after affinity purification, anti-NF186 and -NF155 reactivity by FACS and ELISA were detected in addition. These antibodies were mainly IgG3, with minor contribution of IgM and IgA. To investigate possible functions of anti-NF antibodies in inflammatory PNS diseases, we injected two different monoclonal antibodies (mAbs) into a P2-peptide induced experimental autoimmune neuritis (EAN) animal model at disease onset. We found that while the anti-NF mAbs prolonged and exacerbated clinical disease in these animals, they could not induce disease on their own. We detected NF-reactivity in a small proportion of MS samples (3%; 7/225) by ELISA and flow cytometry. We obtained follow-up material from two NF-reactive patients and saw a persistent NF reactivity in one of them. To increase detection sensitivity, we affinity purified anti-NF antibodies from plasma exchange material of patients with MS (n = 8). IgG, IgM, and IgA were isolated from most of the samples; they were found to recognize NF155 and to a lower extent NF186 by ELISA and in a few also by flow cytometry. This indicates that low levels of anti-NF antibodies exist in a proportion of MS patients. In conclusion, 3% of serum samples from patients with PNS inflammatory neuropathies (GBS and CIDP) showed reactivity by ELISA and none of the controls. In an animal model of autoimmune peripheral nerve inflammation, we showed, using two anti-NF mAbs, that antibody targeting of NF can enhance and prolong disease course. This suggests that antibodies to NF may be relevant for a small group of patients with peripheral inflammatory neuropathies. In MS patients, 3% showed anti-NF reactivity by flow cytometry and ELISA. Furthermore, low levels of anti-NF antibodies that could be detected by ELISA and flow cytometry after affinity purification were additionally found in some MS patient samples that were unreactive by serum screening. This raises the possibility that low levels of antibodies to NF are present in some MS patients and may contribute to the pathogenesis of this chronic disease.