|Home | About | Journals | Submit | Contact Us | Français|
The recent appreciation that a subset of anti-DNA antibodies cross-reacts with the N-methyl-D-aspartate receptor (NMDAR) encourages a renewed examination of anti-brain reactivity in SLE autoantibodies. Moreover, investigations of their autospecificity present a paradigm for studies of anti-brain reactivity and demonstrate that 1) serum antibodies access brain tissue only after a compromise of blood-brain barrier integrity; 2) the same antibodies have differential effects on brain function depending on the region of brain exposed to the antibodies; and 3) insults to the blood-brain barrier are regional rather than diffuse. Finally, these studies suggest that an anatomic classification scheme for neuropsychiatric SLE may facilitate research on etiopathogenesis and the design of clinical trials.
Over the past few decades, as patients with systemic lupus erythematosus (SLE) are experiencing increased longevity, we have become more aware of the late sequelae of this disease [1, 2]. It is clear that most SLE patients develop some manifestation of neuropsychiatric disease (NPSLE) and that the incidence of NPSLE is greater in those with longer duration of disease. It is also clear that many of the most common manifestations of NPSLE do not associate with other metrics of disease, such as flare or severity. Thus, there is a need for exploring new paradigms for pathophysiologic mechanisms to explain this paradoxical and increasingly vexing problem in NPSLE. In this chapter we discuss the impact of the classification scheme for NPSLE and new thoughts regarding the role of anti-N-methyl-D-aspartate receptor (NMDAR) antibodies in the pathogenesis of some of the diffuse CNS manifestations of NPSLE.
Prior to 1999, characterization of CNS events in lupus was hampered by confusing terminology and differences among studies in attribution and methods of ascertainment. A consensus conference convened by the American College of Rheumatology (ACR) in 1999 to facilitate clinical and basic research of NPSLE resulted in the elucidation of nineteen different neuropsychiatric syndromes attributable to SLE (Box 1) . Case definitions, reporting standards and diagnostic criteria were provided by the group. Identification of these 19 syndromes has allowed the rheumatology community to classify more precisely and universally individual clinical presentations thereby paving the way for translational research investigating mechanisms of disease.
|Acute Confusional State|
|Acute Inflammatory Demyelinating Polyradiculoneuropathy (Guillain–Barré Syndrome)|
|Movement Disorder (Chorea)|
Effective use of the NPSLE classification scheme relies on correct attribution of the NP event. Approximately two thirds of NP events occurring in lupus patients are attributable to other causes; it is critically important that all other possible entities have been investigated and excluded for each syndrome [4, 5]. Three conditions, in particular, must be excluded as they may mimic central nervous system (CNS) disease resulting from active SLE. First, infections are a major confounding condition. Immunosuppressive therapies and inherent immune abnormalities in lupus patients contribute to the increased infectious risk in SLE. In North America and Western Europe, most infections are bacterial while in other parts of the world, fungal and mycobacterial infections are common. If unrecognized and untreated, these conditions can be fatal. Reports of PML (Progressive Multifocal Leukoencephalopathy) in SLE patients treated with rituximab or other immunosuppressive therapies highlight the need for increased vigilance in detecting infection in immunosuppressed patients with altered NP status [6, 7]. Another condition, thrombotic thrombocytopenic purpura (TTP), presents with mental status changes as well as thrombocytopenia, microanigopathic hemolytic anemia, renal disease and fever. Appropriate treatment is mandatory; untreated, TTP is 100% fatal. The pathologic lesion is platelet microthrombi, often due to a failure to cleave von Willebrand factor and ensuing platelet activation. Finally, treatment of hypertension in lupus patients is crucial. Posterior reversible encephalopathy syndrome (PRES) occurs in hypertensive lupus patients, frequently in the setting of acute renal failure, recent cyclophosphamide treatment, TTP or pre-eclampsia, and leads to increased cerebral vascular permeability and brain edema. Thus, three potentially fatal conditions, infection, TTP and PRES may be confused with SLE disease activity as they can all mimic an acute, diffuse presentation of CNS NPSLE.
The 1999 classification scheme has been useful to the clinician considering diagnostic and therapeutic options in an individual patient, but is perhaps less useful in probing disease pathogenesis. Of the multiple symptoms encompassed by NPSLE, CNS symptoms occur much more frequently than peripheral nervous system symptoms . Moreover, diffuse CNS symptoms, such as cognitive dysfunction, psychosis, acute confusional state, anxiety and mood disorders, occur more commonly than focal CNS symptoms in most studies. The focal CNS symptoms, including stroke, demyelinating syndromes, movement disorders and transverse myelitis are most frequently secondary to vascular events caused by antiphospholipid antibodies [8, 9]. The diffuse CNS symptoms have a less certain pathogenic mechanism. Cognitive impairment and mood disturbance are among the most frequently reported diffuse CNS syndromes. The cognitive impairment derives most often from memory impairment involving verbal memory as well as executive function and attention. These symptoms are insidious and usually develop slowly over time, independent of disease activity. Their presence is also independent of current or previous medication use and cannot be explained solely on the basis of co-existing antiphospholipid antibodies that are known to cause chronic cerebrovascular disease that may result in cognitive difficulty.
Multiple studies conducted in lupus cohorts worldwide have consistently demonstrated that cognitive impairment occurs with a high frequency [4, 9–12]. Comparisons among studies are, however, difficult. Variability in reported results is attributable to the different instruments used for cognitive assessment, differences in definition of impairment as well as potential inherent differences in selected populations . Traditionally, cognitive ability has been assessed in a one on one setting by a neuropsychologist who administers a battery of tests. Assessment of cognitive ability recommended by the ACR consensus panel is comprised of ten tests administered over one hour that evaluate 8 cognitive domains (intelligence, reasoning, attention, learning, recall, fluency, language, perception). However, despite these recommendations, there has been little uniformity in the selection of tests used. More recently, a computer-based neuropsychiatric assessment (ANAM) has been used to assess cognitive function . This is, in general, less time consuming, less dependent on strong language skills and less dependent on the establishment of a rapport between the tester and the subject. The ANAM has the additional advantage that the practice effect, improvement over time from repeated performances of the test, is less pronounced in longitudinal studies of cognitive function. The individual tests chosen by investigative groups remain, however, variable. More significantly, the performance criteria to identify impairment vary among investigators. The lower frequency of cognitive impairment in some cohorts reflects a more stringent definition of impairment. Patients with memory deficits only are not identified as cognitively impaired in those cohorts, although they would be considered impaired by investigators reporting on other lupus cohorts .
The etiopathogeneis of cognitive impairment and mood disorder remain a mystery. Studies of serum antibodies and cytokines have failed to show a reproducible signal that predicts the development of diffuse NPSLE symptoms in the CNS or that correlates with the presence of these symptoms. For example, serum antiphospholipid antibodies have been shown to correlate with cognitive decline in some studies but not in others [10, 15–18]. Numerous studies of serum anti-neuronal and anti-ribosomal p antibodies demonstrate inconclusive results [18–20]. Further complexity is introduced by the fact that these symptoms can wax and wane, or, can be irreversible. Thus, it is not clear if one mechanism or multiple mechanisms are responsible for these symptom complexes. Many, if not most, neuroimaging studies of NPSLE have sought to associate active NPSLE with reproducible neuroimaging abnormalities. Given that “active NPSLE” comprises a fairly large group of disparate syndromes, it is not surprising that this has proven to be extremely difficult, despite employment of numerous imaging modalities including CT, MRI, functional MRI (fMRI), MRS, diffusion tensor imaging (DTI) and positron emission tomography (PET) scans [21–25]. Differences in patient populations, instrumentation, technique, and metrics for interpretation all prevent comparisons among studies. Nonetheless, magnetic resonance spectroscopy (MRS) identified regional increased choline/creatine ratios in grey and white matter in patients with cognitive dysfunction [26, 27]. Functional MRI also distinguished differences in global and regional brain activation patterns between lupus patients, healthy controls and disease controls (RA) in response to specific memory tasks [28, 29].
Studies of cerebrospinal fluid (CSF) in patients with active NPSLE have proven to be more revealing and are furthering our understanding of the pathophysiology of NPSLE. Significantly elevated IL-6 levels are a consistent finding in patients with active CNS disease, particularly psychosis, and the concentration of this cytokine correlates with symptom severity [30–32]. Additionally, patients with active NP disease (defined as ≥ 2 of the following: psychosis, aseptic meningitis, transverse myelitis, seizures, pathological brain MRI, severely abnormal NP tests, oligoclonal bands in CSF) have 20 and 200 fold increases in intrathecal levels of APRIL and BAFF respectively compared to non-NPSLE lupus patients . Immune complexes formed by autoantibodies in CSF from patients with active diffuse CNS symptoms (psychosis, acute confusional state, seizure disorders, mood and anxiety disorders) and added apoptotic debris has been shown to significantly induce production of IFNα by IFN-producing cells although CSF alone does not do so . Other inflammatory mediators identified in CSF of active NPSLE patients include chemokines (MCP-1, RANTES, MIG, IP-10), IL-8, and MMP-9 [35, 36]. There are no studies of CSF abnormalities specific for insidious manifestations of NPSLE such as cognitive impairment as CSF is not routinely obtained from these individuals. Markers for cognitive impairment and mood disorder that can be reliably measured in an easy-to-use assay are clearly needed. The inability to identify such a marker for these manifestations of NP disease has hampered our understanding of its pathogenesis and has also made design of clinical trials in NPSLE extremely problematic. Clinical investigation of the course of NPSLE is also made difficult as there is no reliable assessment for measurement of improvement in symptomatology or to determine if progression of symptomatology has been retarded.
In SLE, tissue injury is initiated by antibodies. This is true in the kidneys, the skin, blood vessels and in all organs for which we have an appreciation of pathogenesis and inflammatory pathways. For decades it has been know that the serum of many SLE patients contain brain-reactive antibodies. The specific antigens that are recognized by these antibodies were not identified, nor was their functionality known. Additionally, no correlations were found between the presence of these antibodies in serum and aspects of NPSLE. Anti-ribosomal p antibody has been extensively studied with respect to NPSLE. Several clinical studies examining whether serum anti-ribosomal p correlates with psychosis have yielded conflicting results and a recent meta-analysis of 14 published studies concluded that serum anti- ribosomal p measurements were not sensitive in diagnosing NPSLE and did not distinguish between NPSLE subsets [18, 37–40]. Interest in the antibody diminished because it was also not clear how an antibody directed against an intracellular protein could mediate brain dysfunction. Recently, a team of investigators from Chile have demonstrated that the anti-ribosomal p antibody cross-reacts with a membrane protein on neurons and that binding of the antibody to neurons can initiate an apoptotic cascade . Thus, there is now a plausible mechanism for brain pathology resulting from anti-ribosomal p antibodies.
Our own interest is in a subset of anti-DNA antibodies that cross-reacts with a consensus pentapeptide present in the NR2A and NR2B subunits of the N-methyl-D-aspartate (NMDA) receptor (NMDAR).
Many anti-DNA antibodies derived from patients with lupus and from some spontaneous mouse models of SLE are of the IgG isotype and display extensive somatic mutation in variable region sequences . These are characteristics of the molecular signature of a T cell dependent, germinal center matured B cell response. Generally, protein antigens induce a germinal center B cell response; we therefore, asked whether an anti-DNA antibody can bind to a peptide sequence. The anti-DNA antibody that we used in these studies, R4A, deposits in glomeruli, causes proteinuria, and therefore has features of a pathogenic lupus anti-DNA antibody. R4A binds a consensus pentapeptide sequence D/E W D/E Y S/G comprised of either L or D amino acids . This sequence is contained within the NR2A and NR2B subunits of both rodent and human NMDARs. Indeed, the antibody binds each subunit on ELISA and Western blot, and can immunoprecipitate the subunits from a mouse brain lysate. 
NMDARs are receptors for the neurotransmitter glutamate, the major excitatory neurotransmitter in the brain and critically important for many brain functions. Most neurons in the brain contain high levels of glutamate stored inside synaptic vesicles that is released, in a carefully controlled fashion, to convey sensory information, respond to motor commands and to form thoughts and memories that translate to cognitive and emotional abilities. Excessive exposure to glutamate results in increased excitotoxic cell death  and disturbances of glutamate or NMDAR activity have been implicated in several neurologic syndromes including traumatic brain and cord injuries, stroke, Alzheimer’s disease, Huntington’s disease, Parkinson’s disease, seizures, multiple sclerosis, HIV-associated dementia, schizophrenia and ALS .
NMDARs are present throughout the brain and subunits are differentially expressed both regionally in the brain and temporally during development. They are composed of two NRI subunits which have a binding site for glycine, a co-agonist, and two of any four NR2 subunits (A–D) (Figure 1). Receptors containing NR2A and NR2B are most dense on neurons in the CA1 region of the hippocampus, as well as in the amygdala [48–50]. Pertinent to our concerns in SLE, hippocampal NMDARs subserve learning and memory and in the amygdala, NMDARs are critical in the fear conditioning response. These receptors function as voltage-gated calcium channels; following electrical stimulation to the nerve, glutamate and glycine bind an NR2 or NR1 subunit respectively and allow calcium to flux into the cell. Activation of the receptor requires that magnesium exit from the pore of the receptor, at which time calcium is free to enter . The magnitude of the calcium influx is proportional to the time the pore remains in the open position. The change in intracellular calcium is crucial for cellular function. An excessive flux of calcium into neurons causes mitochondrial stress and activates caspase cascades leading to neuronal death [52–54]. Proper regulation of NMDAR activation is, therefore, essential for both cognitive performance and appropriate emotional responses. Consequences of alterations in NMDAR function can be severe; MK-801 is an NMDAR antagonist that effectively blocks excitotoxicity but can produce seizures and coma. Memantine is another NMDAR antagonist that successfully blocks the open channel with few of the sedating side effects . Other NMDAR antagonists with different kinetics produce hallucinations (phencyclidine/PCP/”angel dust”) or excessive drowsiness (ketamine) . The observation that PCP produces hallucinations suggests the possibility that NMDAR abnormalities may contribute to schizophrenia .
The murine monoclonal antibody R4A which binds both DNA and NMDAR functions to enhance NMDAR activation. Studies of hippocampal slices from mice demonstrate that binding of cross-reactive, anti-DNA antibodies to NMDARs on neurons results in apoptotic neuronal death . Further study has demonstrated that the antibody modulates NMDAR activation, synergizing with the natural agonist glutamate to increase excitatory post synaptic potentials. At higher concentrations, it synergizes with glutamate to cause mitochondrial stress and caspase activation. The death inducing function is mediated through NMDAR binding as NMDAR antagonists block caspase activation. The antibody’s effects are dependent only on antibody binding and do not require complement activation or antibody-dependent cell-mediated cytotoxicity as Fab’2 fragments of antibody provoke cell death .
Mechanistic studies of the interaction of R4A with NMDARs show that R4A preferentially binds the NMDAR when the pore is open; thus, the antibody can be presumed to augment the time of opening of the pore and enhances the calcium influx. R4A decreases the concentration of glutamate needed to trigger excitatory post synaptic potentials and to induce apoptosis.
In order to study the potential effects of anti-NMDAR antibody on cognition we have immunized mice with a multimeric form of the DWEYS peptide. These mice develop anti-DNA/anti-NMDAR cross-reactive antibodies. Although these antibodies are present in the circulation, there is no evidence of brain pathology and no alteration of learning or memory, presumably because the endothelial cells in the brain microvasculature form a blood-brain barrier (BBB) that is impenetrable to antibody . There are, however, conditions known to compromise the integrity of the BBB. Infection has long been recognized as a threat to barrier integrity. Bacterial lipolysaccharide (LPS) induces production of IL-1 and TNF, both of which alter permeability of the BBB. When mice immunized with the multimeric peptide are subsequently given LPS, antibody penetrates the BBB and preferentially targets the hippocampus. Anti-NMDAR antibodies bind hippocampal neurons with an ensuing death of hippocampal neurons that is immediate. As the BBB integrity is reconstituted quickly following LPS administration, there is no accumulation of damage following the initial event. Most notably, there is no inflammation in regions of neuronal loss nor activation of resident inflammatory cells in the brain and no influx of blood borne inflammatory cells. Mice with antibody-mediated neuronal loss perform poorly on several tests of memory function, but are unimpaired in tasks that measure other cognitive domains .
Similar findings are observed when normal mice are given human lupus serum or CSF containing anti DNA/anti-NMDAR antibody followed by administration of LPS. Human lupus anti-NMDAR antibody binds to hippocampal neurons in the mouse causing apoptotic neuronal loss . These mice also perform poorly in tests of memory function.
Another agent recognized to compromise the integrity of the BBB is epinephrine. When mice immunized with multimeric peptide and harboring high titers of anti-DNA/anti-NMDAR antibodies are given epinephrine systemically, the antibodies transit from the vasculature into the amygdala. The hippocampi of these mice are histologically normal and there are no cognitive deficits detected in testing of memory and learning functions . However, these mice do have impaired performance in a fear conditioning paradigm. In this paradigm, normal mice exposed to a neutral stimulus followed by a noxious stimulus learn to associate the noxious stimulus with the neutral one. Therefore, conditioned mice will freeze as soon as the neutral stimulus is delivered in anticipation of the noxious stimulus. Mice with antibody-mediated amygdala damage fail to freeze appropriately. Since there is no impairment of memory, the failure to freeze is a behavioral impairment. This study was highly informative for several reasons. It showed that anti-NMDAR antibodies could cause behavioral changes. It also showed that the same anti-NMDAR antibodies could result in two distinct manifestations of NPSLE. Finally, it showed that agents that breach the BBB do so with regional specificity. Thus antibody-related brain symptomatology will depend on the nature of the agent that permits antibody penetration into brain as well as the specific antibody.
These models demonstrate permanent loss of function either in the hippocampus or the amygdala. It is clear, however, in patients that some cognitive or behavioral changes are transient. We postulate that this reflects the observation that lower concentrations of antibodies are needed to affect synaptic plasticity than to cause apoptosis. Thus, low titers of antibody may lead to transient dysfunction and high titers lead to permanent impairment (Figure 1).
Multiple studies performed on cohorts in Asia, Europe and North America demonstrate that approximately 40–50% of patients have antibody reactivity to the DWEYS peptide. This reactivity is only observed in patients with anti-DNA antibody and, when the peptide-reactive antibodies are purified, they display cross-reactivity to DNA . Many studies have attempted to correlate the presence of these antibodies in serum with aspects of NPSLE. Two cross-sectional studies demonstrated correlations with cognitive impairment and depression [27, 61] and another study demonstrated a weak association between decreased amygdala size and serum anti-NMDAR antibody  but several other studies, including a prospective, longitudinal study, have found no correlations [63–65]. In general, there is no reproducible data to correlate serum titers of anti-NMDAR antibodies with any aspect of NPSLE.
More recently, some studies have explored CSF titers of these antibodies. Anti-NMDAR antibody present in CSF of lupus patients cross-reacts with DNA, is neurotoxic and is functionally indistinguishable from the serum antibody. In contrast to serum titers, the presence of these antibodies in the CSF correlates with acute, diffuse CNS manifestations of NPSLE that include seizure disorders, acute confusional state, psychosis, severe, refractory headache and cerebrovascular disease. Moreover, titers correlate with symptom severity [66, 67]. A six month follow-up of CSF from patients studied at the time of an acute episode showed a decrease but not a total absence of CSF anti-NMDAR antibody . Thus, antibody may be present in the CSF of patients without clinically apparent CNS disease. Further studies may determine whether antibody, anti-NMDAR antibody or others, can account for the insidious manifestations of disease, the changes in cognitive function and mood that occur in individuals who never have manifested acute clinical CNS disease.
It is interesting to note that the presence of anti-NMDAR antibody in serum or in CSF fails to correlate with peripheral nervous system manifestations of SLE. Undoubtedly, this reflects the absence of NMDARs on peripheral nerves. Compared to other autoantibodies studied; anti-DNA, ANA, anti-ribosomal p, and anticardiolipin antibodies, anti-NMDAR antibodies have been shown to clearly distinguish between patients with central, diffuse CNS manifestations and those with peripheral nervous system involvement and those with no NP. It also underscores the need when studying “NPSLE” to consider central and peripheral manifestations separately as the pathogenic agents and mechanisms are likely to be quite distinct.
How anti-NMDAR antibodies access the brain in human SLE remains unclear. Raised intrathecal levels imply two possible mechanisms; intrathecal production and increased permeability of the BBB. Grossly elevated intrathecal levels of all autoantibodies, including anti-NMDAR antibodies, occurs during a clear state of BBB disruption such as septic meningitis. While intrathecal production may occur in some patients, it would seem an unlikely event in all patients.
In addition to serum and CSF, anti-NMDAR antibody may be demonstrated in brain tissue. We have eluted anti-NMDAR antibody from post-mortem brain tissue of an SLE patient who died with severe cognitive deficits . The eluted antibody displays neurotoxic effects when injected into a mouse brain. In several other brains, it has been possible to discern antibody bound to neurons and co-localizing with NMDARs . It is clear that under some circumstances brain parenchyma is exposed to anti-NMDAR antibody.
It has been reported in several small studies that the children of women with SLE have an increased frequency of learning disorders [70–74]. These disabilities are not related to birth weight or prematurity. Only one study has asked if the offspring of male SLE patients are similarly affected and found no evidence for this . Together these observations suggest that the in utero environment might be a major contributor to abnormal fetal brain development. It is know that after the first trimester of pregnancy maternal antibody crosses the placenta and enters fetal circulation. Since there is no BBB during fetal development, the fetal brain is exposed to circulating maternal antibody. We therefore studied whether anti-NMDAR antibody can impair normal brain development in pregnant mice.
To address the possibility that maternal anti-NMDAR antibody impaired normal brain development, female mice were immunized with multimeric peptide and demonstrated high titers of anti-DNA anti-NMDAR antibodies. They were then mated with male mice. On day 15 of embryonic development (E15), fetal brains exposed to anti-NMDAR antibody display a thin cortical plate with an increase in the number of apoptotic neurons . When mice exposed to anti-NMDAR antibody during gestation are born, they exhibit a delay in acquisition of neonatal reflexes. As adults, they display normal function of the hippocampus and amygdala, but are impaired in a few isolated tasks requiring intact cortical function . While these impairments cannot be termed “learning disabilities” (as learning disabilities refer specifically to difficulties with reading or mathematical skills), they are isolated impairments of cortical function and in that respect, they are similar to learning disabilities. Thus, anti-NMDAR antibodies have the potential in mice to affect fetal brain development. To date there are no studies in humans, but intravenous administration of human monoclonal anti-DNA/anti-NMDAR antibodies to pregnant mice causes the same histologic insult to the brains of their offspring as that seen in the brains of immunized mice with polyclonal anti-NMDAR antibody. Administration of soluble DWEYS peptide during pregnancy in immunized mice blocks antibody-mediated damage; presumably by formation of peptide-antibody immune complexes and prevention of antibody binding to fetal tissue. This may be a useful therapeutic strategy if it becomes clear that the offspring of women with lupus can be damaged by this antibody during pregnancy.
It is important and interesting to note that anti-NMDAR antibodies and their toxic effects are not limited to SLE. Anti-NMDAR encephalitis is a recently described syndrome characterized by diffuse cerebral dysfunction with acute organic psychiatric disturbances progressing to seizures, dyskinesias, autonomic instability, abnormal cardiac conduction, decreased level of consciousness and central hypoventilation . This syndrome is highly associated with ovarian teratomas but can occur in the absence of tumor . It has also been implicated in the pathogenesis of new-onset seizure disorder in young women . All patients with anti-NMDAR associated encephalitis or seizure disorder demonstrate the antibody in serum and CSF and measurements of BBB integrity suggest intrathecal synthesis of the antibody. The target antigen for this anti-NMDAR antibody differs from than that seen in the lupus patients. Whereas the lupus-associated anti-DNA/anti-NMDAR antibodies target the NR2A and NR2B subunits of the NMDAR, the non-lupus associated anti-NMDAR antibodies bind the NR1 subunit and the two antibodies are not cross-reactive. In contrast, perhaps, to the apoptotic neuronal death induced by the lupus-associated antibody, the non-lupus associated antibody produces a selective and reversible concentration-dependent decrease in the cell-surface density of NMDA receptors . Generally, these patients do well with removal of the tumor (if present) and/or immunosuppressive therapy; clinical symptoms and abnormal MRI findings are reversed as the antibody titer declines. Other anti-NMDAR antibodies have been reported. An anti-NMDAR antibody directed against epitopes on the NR2A subunit was seen in 18% of a cohort of pediatric patients with epilepsy  suggesting the possibility of autoimmune epilepsy or alternatively suggesting that these antibodies can develop after neuronal damage.
Studies to date demonstrate that cross-reactive anti-DNA/anti-NMDAR antibodies are present in serum, CSF and brain tissue of a significant number of patients with SLE. Using a mouse model, it has been possible to demonstrate that these antibodies can cause either cognitive or behavioral impairments. The nature of the brain dysfunction depends on the region of the brain exposed to antibody. This in turn, is influenced by the agent breaching the BBB. These studies explain why serum titers of a pathologic antineuronal antibody will not correlate with CNS symptoms; antibodies must first gain access to brain tissue before causing neuronal dysfunction. In contrast, antibodies present within the CSF will be closely related to disease manifestation.
The demonstration that anti-NMDAR antibody alters synaptic activity at concentrations below those needed to cause neuronal death explains how some episodes of CNS disease are characterized by transient symptomalogy and recovery, while others lead to a fixed impairment. Moreover, the fact that antibody synergizes with ligand, but does not bind a quiescent receptor suggests that regions of brain experiencing synaptic activity will be preferentially targeted.
Whether these antibodies can affect the offspring of mothers with lupus is a fascinating area that needs further study.
In summary, three important points need to be made. First, it is highly likely that there are multiple lupus autoantibodies that can bind targets in the brain and mediate aspects of CNS NPSLE. Of course, there will be non-antibody mediated mechanisms of brain injury as well. Second, to study CNS aspects of NPSLE further, we will need non-invasive assessments of BBB integrity and reliable markers for changes in brain function. Third, if brain dysfunction in lupus patients is indeed mediated by anti-NMDAR antibody, it may be possible to develop the D-DWEYS peptide as a therapeutic to prevent neuronal damage from antibody. Similarly, it may be possible to use D-DWEYS to protect fetal brain. Thus, this mechanism of tissue injury may be treatable, as there is no evidence that antibody triggers any inflammatory cascades in either the adult or fetal brain.
Finally, the paradigm that has been developed suggests that it may be useful to reclassify NPSLE according to region of damage: vascular, CNS or peripheral nerve. Antibodies or other toxic agents that differentially affect brain endothelial cells, brain cells (neurons, astrocytes and microglia) or peripheral nerves, will result in distinct symptomatologies. This classification scheme (vascular, central, and peripheral) may facilitate mechanistic studies of NPSLE, and ultimately clinical trials as well.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Cynthia Aranow, Associate Investigator, Feinstein Institute for Medical Research, Manhasset, NY, 350 Community Drive, Manhasset, NY 11030, 516 562-3837, fax: 516 562-2537.
Betty Diamond, Investigator, Feinstein Institute for Medical Research, Manhasset, NY, 350 Community Drive, Manhasset, NY 11030, 516 562-3830, fax: 516 562-2921.
Meggan Mackay, Assistant Investigator, Feinstein Institute for Medical Research, Manhasset, NY, 350 Community Drive, Manhasset, NY 11030, 516 562-3838, fax: 516 562-2537.