1 Department of Medicine and Immunology, University of Colorado Health Sciences Center and National Jewish Medical and Research Center Denver, Colorado, USA
2 Department of Psychiatry and Medicine, University of Colorado Health Sciences Center and National Jewish Medical and Research Center Denver, Colorado, USA brian.kotzin@uchsc.edu
Systemic lupus erythematosus can cause various forms of central nervous system disorders, ranging from subtle cognitive dysfunction to life-threatening coma. How lupus autoantibodies target neurons and cause brain injury remains a mystery. New research suggests that a subset of autoantibodies to double-stranded DNA in lupus patients cross-reacts with the NMDA glutamate receptor, and produces neuronal injury and death. (pages 1189−1193)
Systemic lupus erythematosus (SLE) is a systemic autoimmune disease that primarily affects women of childbearing age. Its manifestations are extremely diverse and variable1, ranging from skin rash and arthritis to disorders of the kidneys (lupus nephritis), heart, lungs, blood cells or brain in varying degrees. The hallmark of autoimmunity in SLE is production of autoantibodies to nuclear constituents1. Elevated serum levels of antinuclear antibodies are present in more than 98% of SLE patients, and high levels of antibodies to double-stranded DNA (dsDNA) are rarely seen in any other disease. SLE patients can make different types of autoantibodies, both to intracellular and extracellular antigens, and these autoantibodies cause the clinical manifestations of the disease. Although we partially understand how certain autoantibodies may cause disease (such as how antibodies to dsDNA mediate glomerulonephritis), we know little about the types of autoantibodies that target neurons and the mechanisms of brain injury in patients with central nervous system (CNS) involvement. How could lupus autoantibodies cause such diverse complications as psychosis, acute confusional states or slowly progressive intellectual dysfunction? In this issue, DeGiorgio et al.2 provide a clue to understanding this by showing that a subset of anti-DNA antibodies cross-reacts with the NMDA (N-methyl-d-aspartate) subtype of glutamate receptors and induces neuronal cell injury.
CNS involvement in SLE encompasses a wide range of neuropsychiatric syndromes1,
3,
4, such as generalized seizures, acute confusional states, cognitive dysfunction and major psychopathology (such as depression or psychosis), which suggest diffuse brain effects. Other complications such as cerebrovascular stroke syndromes reflect focal brain involvement. Manifestations also vary in severity, ranging from mild headaches, subtle mood swings or subclinical learning deficits to life-threatening coma. Because none of the syndromes are specific for SLE and because there is no reliable diagnostic test, the diagnosis of neuropsychiatric lupus is difficult1. Reports on the prevalence of CNS disease in SLE vary widely from 15 to 75%.
Although immune complexes are important for some types of tissue injury in SLE, immune-complex−mediated inflammation is not the central mechanism for CNS lupus. Autopsy studies have demonstrated little evidence of vasculitis and little inflammation in sites of injury. One explanation for the diversity of CNS problems in neuropsychiatric lupus is that there are different autoantibodies with different CNS targets. For example, antiphospholipid antibodies (associated with focal CNS manifestations) have been implicated in vasculopathy and blood vessel occlusion1,
3. Although controversial, elevated serum levels of antibodies to ribosomal P proteins may be another example5. P proteins are primarily associated with the 60S ribosomal subunit in cytoplasm, and antibodies to these proteins have been associated with psychosis and severe depression in SLE (ref. 5). Antibodies against P proteins have not been implicated in other CNS lupus syndromes, such as cognitive dysfunction or confusional states.
A separate autoantibody group, and the type relevant to the study by DeGiorgio et al.2, are those that bind to neuronal cells in vitro6. The specific membrane antigens targeted by antineuronal antibodies have not been well characterized. Clinical studies have suggested that active CNS disease is best correlated with elevated levels of antineuronal antibodies in the cerebrospinal fluid (CSF) versus blood. Both crossing through the blood−brain barrier and in situ synthesis from B cells within the CNS may contribute to the elevated CSF levels. Standardizing the assays for these antineuronal antibodies has been difficult; in various studies, 30−90% of patients with active diffuse CNS disease show elevated antineuronal antibodies in CSF.
DeGiorgio et al.2 used a different approach to identify the specificity of antineuronal antibodies possibly relevant to CNS disease in SLE. Using a phage display peptide library, these investigators had previously noted that anti-dsDNA antibodies cross-react with a particular pentapeptide (Asp/Glu-Trp-Asp/Glu-Tyr-Ser/Gly). The consensus sequence is in the extracellular domain of mouse and human NMDA receptors NR2a and NR2b. A prototype monoclonal antibody to this pentapeptide caused neuronal injury and loss after injection into the hippocampal region of mice and also caused apoptosis of primary neuronal cells in culture2. Serum anti-DNA antibodies from four SLE patients, after enrichment for the pentapeptide reactivity, produced similar effects. Neuron injury seemed to be related to NMDA-receptor engagement and stimulation, similar to that seen in excitatory amino-acid toxicity7. The authors also found that antibodies with the same characteristics were present in the CSF of one SLE patient with cognitive deficits. Based on these findings, the authors postulate a novel pathway for neuron injury in SLE, which may explain some of the variability in symptoms and why certain brain areas are targeted in CNS lupus. If confirmed by additional studies, these findings may even suggest an adjunctive therapeutic strategy in some patients.
The NMDA subtype of glutamate receptors8 is critical in long-term potentiation and synaptic plasticity, processes that influence learning and memory. NMDA receptor antagonists or the excessive stimulation of these receptors (excitatory amino-acid toxicity), which results in excessive calcium entry into the cell and apoptosis7, can impair learning. SLE patients, even those without overt CNS symptomatology, frequently demonstrate deficits in these cognitive domains9. Altered expression and antagonism of NMDA receptors can also cause symptoms of psychosis8, another manifestation of CNS lupus. Interestingly, the hippocampus, implicated in learning and memory deficits, has the highest density of neuronal NMDA receptors8; it may be particularly vulnerable to excitatory amino-acid toxicity, and it might be selectively affected in some patients with SLE (ref. 9). Although studies have not found a correlation between learning deficits and global cerebral atrophy in SLE (ref. 3), newer imaging techniques that quantify regional atrophy and metabolic changes may be able to show selective hippocampal abnormalities in SLE. The ability of antibodies against NMDA to cause varying degrees of excitatory toxicity or even to block receptor function may also underlie differences in CNS symptomatology and unpredictable reversibility.
Although provocative, the report by DeGiorgio et al.2 is preliminary. They only studied sera from a few patients and CSF from only one patient. Also, there are no data yet to show that this subset of anti-DNA antibodies is associated with CNS disease in SLE, especially the relevant diffuse CNS syndromes. Standard assays for anti-dsDNA antibodies have shown correlations with systemic disease activity in SLE and particularly renal involvement, but not CNS disease1. Thus, future association studies will require assays that quantify the subset with anti-pentapeptide activity in both sera and CSF. In studies of CNS lupus, comparing the frequencies and levels of the relevant autoantibody in SLE patients with and without evidence of CNS disease is critical. This may require neuropsychological testing and newer imaging techniques to identify and properly group individuals with subclinical disease. Follow-up serial studies in individual patients may help determine whether such antibodies correlate with or even predict CNS events.
Animal studies also offer opportunities to further examine this subset of anti-DNA antibodies. Are there any functional consequences of these autoantibodies after binding to neuronal cells? Can the cross-reactive autoantibodies produced in mice after immunization with the pentapeptide10 localize to neuronal cells in vivo? Although the production of circulating autoantibodies with anti-NMDA activity may be insufficient to mediate pathology in the CNS, there should be ways to test potential pathogenicity in a more physiologic manner than direct injection of affinity-purified material into the hippocampus. There are also some excellent spontaneous mouse models of SLE (ref. 1), and these mice may produce this particular anti-dsDNA antibody subset. Notably, learning deficits and behavioral abnormalities have also been reported in these spontaneous mouse models, and may relate to a specific type of autoantibody.
Finally, the cross-reactivity described by DeGeorgio et al.2 is a reminder that we know too little about the generation and pathogenicity of antinuclear antibodies, particularly anti-DNA antibodies, in SLE. Cross-reactivity of anti-DNA antibodies with different non−nucleic-acid molecules have been described (Fig. 1). In lupus nephritis, the evidence supports a model wherein anti-DNA antibodies bind to extracellular DNA (or chromatin) in glomeruli, which leads to immune-complex formation in situ1. However, studies show that only a subset of anti-DNA antibodies are nephritogenic. The characteristics that separate pathogenic from non-pathogenic anti-DNA antibodies are unclear. Even in lupus nephritis, cross-reactivity to certain glomerular antigens may be involved1,
2. Unexpected cross-reactivities, such as the one described by DeGeorgio et al.2, may eventually explain different types of tissue injury in SLE.
Figure 1. Production of anti-dsDNA antibodies and possible mechanisms of organ damage in SLE.
In lupus nephritis, a subset of antibodies bind to 'planted' DNA or chromatin in glomeruli, forming immune complexes and causing complement activation and glomerulonephritis. Anti-dsDNA antibodies could also cross-react with non−nucleic-acid substrates. A subset of these antibodies cross-react with the NMDA receptor and may be involved in CNS lupus. Not shown are the other multiple (not anti-DNA) autoantibodies that participate in renal, CNS and other forms of SLE injury.
Kotzin, B.L. & West, S.G. Systemic lupus erythematosus. in Clinical Immunology: Principles and Practice, 2nd edn. (eds. Rich, R.R. et al.) 60.1–60.24 (Mosby, London, 2001).
DeGiorgio, L.A. et al. A subset of lupus anti-DNA antibodies cross-reacts with the NR2 glutamate receptor in systemic lupus erythematosus. Nature Med.7, 1189–1193 (2001). | Article | PubMed | ISI | ChemPort |
Moore, P.M. & Lahita, R.G. (eds.) Neuropsychiatric manifestations of systemic lupus erythematosus. Proceedings of a conference, New York, New York, September 27–30, 1997. Ann. NY Acad. Sci.823, 1–328 (1997). | PubMed | ChemPort |
ACR Ad Hoc Committee on Neuropsychiatric Lupus Nomenclature. The American College of Rheumatology nomenclature and case definitions for neuropsychiatric lupus syndromes. Arthritis Rheum.42, 599–608 (1999). | PubMed | ISI |
Teh, L-S. & Isenberg, D.A. Antiribosomal P protein antibodies in systemic lupus erythematosus: A reappraisal. Arthritis Rheum.37, 307–315 (1994). | PubMed | ISI | ChemPort |
Lipton, S.A. & Rosenberg, P.A. Excitatory amino acids as a final common pathway for neurologic disorders. N. Engl. J. Med.330, 613–622 (1994). | Article | PubMed | ISI | ChemPort |
Ozawa, S., Kamiya, H. & Tsuzuki, K. Glutamate receptors in the mammalian central nervous system. Prog. Neurobiol.54, 581–618 (1998). | Article | PubMed | ISI | ChemPort |
Kozora, L. et al. Analysis of cognitive and psychological deficits in systemic lupus erythematosus patients without overt central nervous system disease. Arthritis Rheum.39, 2035–2045 (1996). | PubMed | ISI |
Putterman, C & Diamond, B. Immunization with a peptide surrogate for double-stranded DNA (dsDNA) induces autoantibody production and renal immunoglobulin deposition. J. Exp Med.188, 29–38 (1998). | Article | PubMed | ISI | ChemPort |
