Chapter 22C Antineuronal Antibodies

INTRODUCTION

Autoantibody production is a cardinal feature in the pathogenesis of systemic lupus erythematosus (SLE) and is associated with most of the clinical manifestations of the disease. Antineuronal antibodies have been the subject of extensive research since they were first described in SLE patients in 1978,¹ although their exact role in the pathogenesis and diagnosis of neuropsychiatric (NP) lupus is still unclear. Here we review the methodology for the detection of antineuronal antibodies, and explore their prevalence and association with NP-SLE and their potential role in the pathogenesis of the disease.

DETECTION OF ANTINEURONAL ANTIBODIES

Antineuronal antibodies have been measured by a variety of assays using different substrates as the source of neuronal antigens (Table 22C.1). These have included cultured viable and fixed neuroblastoma cell lines¹^–⁷ and brain sections of human origin.⁸ Complement-dependent cytotoxicity,¹ indirect immunofluorescence,^2–4,⁷ mixed hemadsorption,⁵ radioimmunoassay,^7,⁹ and whole-cell ELISA⁶ assays have been some of the techniques used. Due in part to the wide heterogeneity of assay techniques, the prevalence of antineuronal antibodies in SLE patients has varied considerably (Table 22C.1). In general, IgG antibodies have been less frequent than the IgM isotype and the prevalence has varied between 13 and 82% and 26 and 75%, respectively. Virtually all studies included healthy and disease control groups, and it is noteworthy that antineuronal antibodies have also been reported in other neurologic disease states of both an inflammatory and degenerative nature¹⁰^–¹⁵ and in rheumatoid arthritis.⁸

TABLE 22C.1 MEASUREMENT AND PREVALENCE OF ANTINEURONAL ANTIBODIES IN SLE

CLINICAL ASSOCIATIONS WITH ANTINEURONAL ANTIBODIES

Some groups have reported an association between circulating antineuronal antibodies and clinically overt NP manifestations of SLE (Table 22C.2).^4,^5,⁸ The correlation is strongest between antibodies of the IgG isotype and diffuse NP manifestations such as psychosis. For example, Wilson and colleagues⁴ demonstrated serum IgG reactivity to surface antigens on the human neuroblastoma cell line SK-N-SH in 45% of SLE patients with NP disease compared to 5% of SLE patients without NP manifestations. Similarly, 63% of patients with diffuse NP disease demonstrated antineuronal reactivity compared to none of the patients with focal disease.⁴ A similar but less striking correlation was reported by How and colleagues.⁵ However, others have not confirmed this clinical-serologic correlation.^2,^6,^7,^9,¹⁵ In particular, the association between circulating antineuronal antibodies and cognitive impairment has been inconsistent in cross-sectional studies. There is one report of a positive correlation between the presence of circulating antineuronal antibodies and cognitive impairment in SLE patients,¹⁶ but this was not confirmed in subsequent independent studies.^2,⁷ Intuitively, the presence of antineuronal antibodies in a location where they would have ready access to brain tissue would support a pathogenic role. Thus, it is of interest that Bluestein and colleagues reported a striking association between the presence of antibodies in CSF of SLE patients with NP disease,⁹ an observation supported by two subsequent studies.^6,¹⁵

TABLE 22C.2 ASSOCIATION BETWEEN IgG ANTINEURONAL ANTIBODY AND NP-SLE

Information is limited on the association between serial changes in antineuronal antibody reactivity with NP disease. Wilson and colleagues ⁴ (8 patients) and Bresnihan and colleagues⁸ (2 patients) reported a close temporal association between serum IgG reactivity and clinical disease. Hanly and colleagues¹⁷ demonstrated a similar although less striking correlation between cognitive function and serum IgG antineuronal antibodies in a group of 20 SLE patients evaluated retrospectively over a 2-year period. In this study, changes in antineuronal antibody levels were frequently associated with concurrent changes in anti-DNA antibodies and overall lupus disease activity. When neuropsychiatric disease or cognitive dysfunction was present, their course showed a close correlation with changes in antineuronal antibody levels. However, the concordance with anti-DNA antibodies limits the clinical relevance of these findings.

ANTIGENIC SPECIFICITY OF ANTINEURONAL ANTIBODIES

Further insight into the pathogenic role and the diagnostic value of antineuronal antibodies in SLE patients might be derived from characterization of their antigenic specificities (Table 22C.3). Such information could permit the distinction between pathogenic and nonpathogenic antibodies and correlation between specific autoantibodies and subsets of nervous system lupus. Reactivity to a variety of surface neuronal antigens by SLE sera was first suggested by Bluestein,¹⁸ who demonstrated at least six different antibody specificities in immunoabsorption experiments using a panel of human neuroblastoma cell lines. Subsequent studies demonstrated cross-reactivity between antineuronal antibodies with surface proteins on lymphocytes,¹⁹^–²¹ red blood cells,⁸ and glial cells²² and with mycobacterial antigens.²³

TABLE 22C.3 ANTIGENIC SPECIFICITY OF ANTINEURONAL ANTIBODIES

Antigenic Specificity	Source of Antigen	Reference
Gangliosides	Neuronal and myelin membranes	Hirano (1980) (24)
97-Kd plasma membrane protein	Human neuroblastoma cell line IMR-6	Hanly (1988) (31)
50-Kd synaptic terminal membrane protein	Adult bovine and fetal human brain	Hanson (1992) (32)
Brain synaptosomes	Rat brain	Hanly (1993) (33)
Brain integral membrane proteins	Human and rat brain	Hanly (1993) (34)
NR2 glutamate receptor	Mouse brain	DeGiorgio (2001) (42)

Gangliosides are members of a family of glycolipids predominantly located on neuronal and myelin membranes in the central and peripheral nervous systems. Antiganglioside antibodies have been found in both the serum ²⁴^–²⁷ and CSF²⁸ of SLE patients and other autoimmune inflammatory neurological disorders such as multiple sclerosis.^29,³⁰ Hanly and colleagues³¹ identified reactivity to a 97-Kd surface protein on two human neuroblastoma cell lines, IMR-6 and NMB-7. This protein was not identified on non-neuronal cell lines and was not precipitated by sera from control subjects. Hanson and colleagues ³² described antibodies to a 50-Kd neuronal membrane protein isolated from bovine brain synaptic plasma membrane. Affinity purified antibodies from SLE sera bound to the surface of cultured rat neuroblastoma cells, and upon Western blotting identified a protein of comparable size in human fetal brain. Additional antibody specificities have been described against proteins of different size derived from synaptosomes³³ and integral membrane proteins from homogenized brain tissue.³⁴ Although these studies clearly indicate that SLE patients have autoantibodies that target multiple neuronal and brain antigens, a reproducible and convincing association with clinical or subclinical nervous system disease has not so far been demonstrated.

Most recently, attention has been focused on anti-NR2 glutamate receptor antibodies as a potentially novel system that could explain some of the complexities of NP-SLE and provide a useful diagnostic tool. The NMDA (N-methyl-D-aspartate) receptors NR2a and NR2b bind the neurotransmitter glutamate and are present on neurons throughout the forebrain.³⁵^–³⁷ The hippocampus, which is the anatomical structure closely linked to learning and memory, has the highest density of brain NMDA receptors.³⁶ In addition to their putative role in learning and memory,³⁸ these receptors display altered expression in major psychoses³⁹ and if engaged by receptor antagonists cause hallucinations and paranoia.⁴⁰

Studies have shown that a subset of anti-DNA antibodies, derived from both murine models of SLE and from a limited number of human subjects with the disease, cross-react with a pentapeptide consensus sequence ^41,⁴² that is present in the extracellular ligand binding domain of NR2 receptors. Moreover, these antibodies were present in the CSF of one SLE patient with progressive cognitive decline. Although of considerable interest, these findings are largely derived from animal studies and require confirmation in human subjects with NP-SLE. To date, the studies in human lupus examining the association between this subset of antineuronal antibodies and cognitive impairment have yielded conflicting results.⁴³^–⁴⁵

OTHER AUTOANTIBODIES WITH BRAIN/NEURONAL CROSS-REACTIVITY

There are a number of lupus autoantibodies that cross-react with brain/neuronal antigens and thus may also be involved in the pathogenesis of nervous system disease Table 22C.4.

TABLE 22C.4 AUTOANTIBODIES THAT CROSS-REACT WITH NEURONAL CELLS

Autoantibodies	Neuronal Cross-Reactivity	Reference
Lymphocytotoxic antibodies	Human brain homogenates	Bluestein (1976) (21)
Antiribosomal P antibodies	38 Kd surface human neuroblastoma cell line	Koren (1992) (65)
Antiphospholipid antibodies	Rat brain	Kent (2000) (69)

Lymphocytotoxic Antibodies

Two classes of such antibodies have been demonstrated in SLE patients.⁴⁶ IgM antibodies demonstrate maximal binding at 4°C and are cytotoxic for resting peripheral blood lymphocytes (hence the term cold-reactive lymphocytotoxic antibodies). In contrast, IgG antibodies demonstrate maximum binding at 37°C, preferentially react with activated lymphocytes, and result in target cell death by antibody-dependent cell-mediated cytotoxicity in the absence of complement. Lymphocytotoxic antibodies have been demonstrated in serum and CSF of SLE patients and have been associated with nervous system manifestations, including cognitive dysfunction in some ^1,^8,^21,⁴⁷ but not all studies.^2,⁴⁸ In patients with NP-SLE, lymphocytotoxic reactivity was frequently removed by preabsorbing with brain homogenate,²¹ suggesting the presence of cross-reactivity between lymphocytes and neuronal cells. Eluates with lymphocytotoxicity from brain homogenates were of both IgM and IgG isotypes.⁴⁹^–⁵¹

The antigenic specificity of IgM and IgG lymphocytotoxic antibodies has been studied. Cold-reactive IgM lymphocytotoxic antibodies demonstrate preferential binding to CD4+ T lymphocytes and the CD4+ 2H4+ subset.^52,⁵³ Using normal human peripheral blood lymphocytes and lymphocyte cell lines, reactivity to proteins of 55-, 70-, and 105- to 110-Kd proteins have been identified by Western blotting.⁵⁴ Another study demonstrated IgM reactivity to 46- and 200-Kd glycoproteins isolated by lectin affinity chromatography from normal lymphocytes and lymphocyte cell lines.⁵⁵ IgG anti-lymphocyte reactivity is directed predominantly against two proteins of 90- and 55-Kd molecular weight.⁵⁶ Reactivity to a shared 52-Kd protein present on both lymphocytes (CD4+ and HUT-78 cell line) and neruoblastoma cells (SK-N-SH and IMR-6) has also been described,⁵⁷ and the same group have reported an association with cognitive impairment in SLE.⁴⁹