This article is commented on by Gorman on page 483 of this issue.
Dr Russell C Dale at Clinical School, The Children’s Hospital, Westmead, Locked Bag 4001, Sydney, NSW 2145, Australia. E-mail: firstname.lastname@example.org
Aim Systemic lupus erythematosus is a multi-organ autoimmune disorder associated with autoantibodies of complex diversity. Antiphospholipid antibodies (aPL), which are commonly associated with lupus, create a pro-thrombotic tendency, but are also associated with non-thrombotic neurological features. Movement disorders are rare neuropsychiatric complications of lupus and antiphospholipid syndrome, and autoimmune and thromboembolic disease mechanisms have been proposed.
Method We describe the clinical features, investigation findings, treatment, and outcome of six paediatric participants with movement disorders associated with lupus and/or aPL (six females, median age 13y, range 8–15). To examine the autoantibody hypothesis, we used a neuronal cell line with dopaminergic characteristics and measured serum antibody binding to neuronal cell-surface antigens using flow cytometry. For comparison with the six participants, we used serum from healthy individuals (n=12, six females, median age 11y, range 9–13) and children with other neurological diseases (n=13, seven females, median age 7y, range 2–15).
Results Of the six participants, two had lupus only, two had lupus with aPL, and two had aPL only. The movement disorder was chorea in four and parkinsonism in two. All four participants with chorea had aPL and movement disorder relapses. The two participants with parkinsonism did not have aPL, but had a progressive course until rituximab or plasma exchange resulted in neuropsychiatric remission. All six participants demonstrated elevated serum antibody binding to neuronal cell-surface antigens compared with healthy individuals and those with other neurological diseases.
Interpretation This report supports the association of chorea with aPL, but suggests a different autoimmune mechanism operates in lupus parkinsonism. The presence of antibody binding to neuronal cell-surface antigens supports a possible direct action of autoantibodies on neurons in patients with movement disorders associated with lupus and aPL.
• The classic movement disorder associated with antiphospholipid antibodies is chorea.
• Therapeutic benefits of rituximab and plasma exchange support a humoural autoimmune pathogenesis.
• The presence of IgG binding to the cell surface of neuronal cells supports the possible presence of pathogenic autoantibodies in these patients.
Systemic lupus erythematosus (SLE) is a multisystem autoimmune disease associated with a complex array of autoantibodies. Approximately 20% of patients with SLE will present during the first two decades of life, and neuropsychiatric disease occurs in 25% of paediatric patients with SLE, often in the first year of disease.1,2 Movement disorders are rare but recognized neurological complications of SLE:1,3 chorea is described in approximately 5% of paediatric patients with SLE,1 and parkinsonism has been reported occasionally.3
Antiphospholipid syndrome (APS) is a systemic autoimmune condition characterized by hypercoagulability, arterial and venous thromboses, pregnancy morbidities, and the presence of antiphospholipid antibodies (aPL), which include anticardiolipin antibodies, lupus anticoagulant, and anti-β2-glycoprotein I antibodies.4 APS can be primary or secondary to systemic autoimmune disorders, particularly SLE.5,6 It is rare in the paediatric population as associated thrombotic risk factors such as atherosclerosis, smoking, hypertension, contraceptive hormonal treatment, and pregnancy are rare.5 Non-thrombotic neurological complications are also described in patients with aPL, including migraine, seizures, and chorea.5 Non-thrombotic complications, and specifically chorea, have been described in 16% and 4% of paediatric patients with APS respectively.5
The proposed pathology of movement disorders associated with SLE and aPL is heterogenous and includes direct effects of autoantibodies on neurons, or thromboembolic disease of the lenticulostriate arteries.4,7,8 These heterogenous proposed disease mechanisms have sometimes made therapeutic decision-making difficult in neuropsychiatric complications of SLE and APS.5 We describe the clinical features, investigation, treatment, and outcome of six paediatric patients with movement disorders associated with SLE or aPL. We also demonstrate the presence of immunoglobulin-G (IgG) binding to the cell surface of live dopaminergic neurons in vitro in these patients, providing support for possible direct effects of autoantibodies on neurons.
The six participants were all sequential patients with movement disorders associated with SLE or aPL seen by RCD at Great Ormond Street Hospital for Children NHS Trust, London, UK or the Children’s Hospital Westmead, Sydney, between 2000 and 2009. The movement disorders exhibited by these patients satisfied the definitions and classifications of movement disorders according to the Movement Disorder Society (http://www.movementdisorders.org).
Participants diagnosed with SLE satisfied the American College of Rheumatology criteria.9 Four participants had evidence of persistently elevated aPL, with a persistently high titre of anticardiolipin IgG antibodies, and persistently detected lupus anticoagulant on multiple occasions for at least 1 year, therefore ruling out transient elevation of aPL.5 These participants, therefore, had laboratory evidence of APS and demonstrated a non-thrombotic manifestation of APS, but did not fulfil Sapporo criteria for APS, as previously described and discussed.4,5,7,10 Serum samples were taken from the acute movement disorder phase of the disease, and IgG levels were within normal limits. The study was approved by the ethics committee at the Children’s Hospital at Westmead, and all the patients’ families gave informed written consent.
IgG binding to neuronal cell-surface antigens
This approach has been used to demonstrate IgG binding to cell-surface antigens in immune-mediated disorders of the central nervous system (CNS).11,12 In brief, SH-SY5Y cells were used because of their neuronal characteristics: they have been shown to express dopaminergic characteristics after incubation with retinoic acid.13 Fifty thousand SH-SY5Y cells were incubated with serum at 1:100 dilution for 1 hour, washed with phosphate buffered saline containing fetal bovine serum, and incubated with Alexa Fluor (Invitrogen, Carlsbad, CA, USA) 488-conjugated goat antihuman IgG secondary antibody for 1 hour. After further washing, and re-suspending in phosphate buffered saline containing fetal bovine serum, a viability dye (7-ADD) was added to exclude dead cells. Ten thousand events per well were recorded on a BD LSR II fluorescence-activated cell-sorting instrument (BD Biosciences, Sparks, MD, USA) with a high-throughput sampler. Data were analysed using FlowJo software, (Tree Star, Ashland, OR, USA). Binding was expressed as mean fluorescence intensity. All experiments were performed three times. For comparison with the six participants, we used serum from healthy individuals (n=12, six females, mean age 11y, median age 11, range 9–13), and children with other neurological diseases (n=13, seven females, mean age 6y, median age 7, range 2–15). A Mann–Whitney U test was used to compare antibody titre between patients and the comparison groups.
Table I provides an overview of the six participants, their presentations, investigations, management, and progress. A diagnosis of SLE only was made in two participants, SLE with associated aPL in two, and aPL only in two. All six participants were female, with a mean age of 12 years 10 months, median age 13 years, range 8 to 15 years. The movement disorders exhibited by the participants were chorea (n=4) and parkinsonism (n=2). The movement disorders presented acutely, with maximal severity of the movement disorder occurring within 1 to 4 weeks of onset.
Table I. Summary of six patients with movement disorders associated with systemic lupus erythematosus (SLE) and/or antiphospholipid antibodies (aPL)
SLE Full neuropsychiatric recovery Cutaneous lupus (2y)
Participants with chorea (n=4)
The four youngest participants had generalized chorea, all as their first presentation. Two participants had SLE with aPL, and two had aPL only. Two had associated psychiatric symptoms (Table I). All four participants had persistently abnormal aPL, with high titre of anticardiolipin antibodies and lupus anticoagulant detected on multiple occasions for at least 1 year (Table I). All four participants with chorea had testing for streptococcal serology (two of the four had low positive antistreptolysin-O titres [365 and 600 international units/mL]). Cerebrospinal fluid (CSF) was for the most part unremarkable: only one patient had a mild pleocytosis, and two had mirrored oligoclonal bands. Magnetic resonance imaging (MRI) of the brain was normal in two, and showed multiple small white-matter lesions in two.
Symptomatic management of the chorea included haloperidol, tetrabenazine, and carbamazepine. Methylprednisolone for 3 days was given to all participants except for one patient with aPL only (Patient 2). All four participants made a complete recovery within 2 months, but all have had movement disorder relapses (chorea in three, ataxia in one), again with subsequent recovery (Table I). Three participants have chronic headache or migraine. Ongoing immune therapy includes prednisolone, azathioprine, mycophenylate, and hydroxychloroquine; and all four participants are taking aspirin prophylaxis.
Case example (patient 2)
A 13-year-old Australian Caucasian female presented with a 4-day history of abnormal movements of the face, trunk, and limbs. Her father had APS and was on lifelong warfarin.
On examination, she had generalized chorea and dysphasia. Functionally, she was unable to hold a pencil, or to feed herself with a spoon. Her gait was very abnormal, with lurching, random movements of the limbs (Video S1, supporting material published online).
She was initially suspected to have Sydenham chorea; however, her streptococcal serology, throat swab, and echocardiogram were normal. Her aPL were positive, with anticardiolipin IgG of 72 GPL units (1 GPL unit is equivalent to 1μL/mL of IgG anticardiolipin immunoreactivity; normal <15) and anticardiolipin immunoglobulin-M of 14 GPL units (normal <12). Lupus anticoagulant was also detected in her blood. Furthermore, she had persistently elevated antinuclear antibodies (1:640 to 1:2560; speckled pattern), elevated anti-double-stranded DNA antibodies, as well as persistently raised ESR. An acute lumbar puncture during the chorea illness was normal. T2-weighted MRI of the brain showed a few non-specific small white-matter lesions (Fig. S1, supporting material published online). A provisional diagnosis of SLE with associated aPL was made.
She was treated with methylprednisolone for 3 days, which resulted in significant improvement in her movement disorder and gait over the first week. She subsequently received 4 weeks of oral prednisolone and her chorea completely resolved within 6 weeks.
Four months after her initial presentation, she developed weekly unilateral, migraine-like headache. Repeat MRI, MR angiogram, MR venogram, and diffusion-weighted imaging were normal.
Over the next 3.5 years she experienced three relapses of chorea, one generalized and two localized to the right hand. So far, she has not had any thrombotic events and no other organ disease typical of SLE. However, her IgG anticardiolipin antibodies, lupus anticoagulant, antinuclear antibody and anti-double-stranded DNA antibodies have remained persistently elevated or abnormal for over 3 years. Each chorea relapse was treated with oral prednisolone, which resulted in remission of the chorea over 4 weeks. She has also been treated with aspirin and mycophenylate (1000mg twice daily). Her baseline neurological examination between relapses has been normal.
Participants with parkinsonism (n=2)
The participants with parkinsonism had previously had other organs affected by SLE, and presented with parkinsonism with bradykinesia (n=2), rigidity (n=2), tremor (n=1), and loss of postural reflexes (n=1). Both females had significant psychiatric symptoms, with psychosis and emotional lability (Table I). Both were negative for aPL, but both had typical SLE autoantibodies. CSF was unremarkable in both, apart from mildly elevated neopterin in the CSF in Patient 6 (the only participant in whom neopterin in the CSF was measured). Both participants had mild volume loss on MRI, but no localizing features.
Despite administration of methylprednisolone and cyclophosphamide, both participants deteriorated, with progression of their movement and psychiatric disorders. The participants only improved after plasma exchange (Patient 5) or rituximab (Patient 6). Patient 5 demonstrated clear improvements within the first week of plasma exchange, and was fully recovered within 2 months. Despite prolonged admissions, both patients have made a complete neuropsychiatric recovery, but have ongoing cutaneous lupus.
Case example (patient 6)
A month before her neuropsychiatric syndrome, a 15-year-old Australian female of Philippine origin had night sweats, arthralgia, myalgia, alopecia, and malaise. She had a malar rash and distal skin vasculitis. She had a positive antinuclear antibody (1:2560), positive ribosomal P antibodies and Smith antibodies, as well as a positive direct antiglobulin antibody.
A diagnosis of SLE was made, and she was started on 30mg oral prednisolone daily. Two weeks later, she developed headache, confusion, reduced concentration, and altered behaviour. She had persecutory auditory hallucinations and heard voices saying ‘look at her, she is bad and mean’, to which she responded by continuously apologizing ‘I’m sorry’, ‘there’s a misunderstanding’, and ‘I need to be strong, I’ve hurt many people’. She denied any insertion or deletion of thoughts. She also demonstrated ideas of reference, where she believed that the television was talking to her. Her psychiatric symptoms were treated with risperidone and benztropine.
Electroencephalography, MRI, and MR angiography were all normal. CSF was negative for microscopy, glucose, protein, and oligoclonal bands, but neopterin in it was marginally elevated at 40nmol/L (normal <28), consistent with inflammation of the CNS.14
Over the next week, her symptoms progressed to include hypersalivation, increasing anxiety, and intermittent fixed upward gaze, which were interpreted as oculogyric crises. She became mute, was increasingly agitated, and paced back and forth in her room. She was treated with 750mg/m2 induction and monthly maintenance of cyclophosphamide (guided by lymphopenia), while continuing prednisolone 60mg/day.
After 2 months and despite cyclophosphamide, she became bradykinetic, had reduced blinking, positive glabellar tap, abnormal postural reflexes, and rigidity of her right arm with cogwheeling (Video S1). She exhibited self-injurious behaviour and tried to slash her wrists with a bread knife. Repeat MRI showed mild reduction in brain volume.
In the third month, she exhibited progression of her parkinsonian symptoms. A trial of l-3,4-dihydroxyphenylalanine (l-DOPA) was started, and a marginal improvement in bradykinesia and gait was noted. In view of the progression of her disease despite cyclophosphamide and high-dose oral prednisolone, she was started on rituximab 375mg/m2 weekly for 1 month. Within 2 weeks of starting rituximab, there were significant improvements in her agitation, cognition, and spontaneity. She continued to improve over the next month and was discharged from hospital after a 6-months admission, with some residual parkinsonism, but normal cognition and behaviour.
She has been followed for 2 years and her SLE is currently well controlled on a combination of low dose steroid treatment, mycophenylate, and hydroxychloroquine. She has no parkinsonian features, no cognitive or behavioural problems, and her l-DOPA has been stopped. Her MRI shows persistent mild cerebral atrophy.
IgG binding to neuronal cell-surface antigens
All six patients demonstrated elevated IgG binding to the cell surface of live neurons (n=6, mean fluorescence intensity 275, range 189–382) compared with healthy individuals (n=12, mean 42, range 26–56) and those with other neurological disease (n=13, mean 59, range 26–90; Mann–Whitney U test p<0.001; Fig. 1). There was no difference in the IgG cell-surface binding between the chorea and participants with parkinsonism. Participants with SLE, but no neuropsychiatric features (n=6), also had elevated IgG binding to the cell surface of live neurons, at levels similar to the patients with movement disorders (data not shown).
Movement disorders are rare neuropsychiatric features of SLE or aPL.1,5 Chorea is the most common association with both, but parkinsonism and dystonia are also described.1,3,15 Chorea is often an early symptom in SLE, and often a first symptom of APS before the occurrence of thrombosis.15 It is common for psychiatric features and migraine to accompany the movement disorders, as in the cases in this report.1,3 Only one of our patients had seizures, suggesting that the pathological process was predominantly directed towards the subcortical structures rather than the cerebral cortex. Relapses of chorea are considered unusual; however, all of the participants with chorea in this report had neuropsychiatric relapses (three with chorea), demonstrating a persistent predisposition to chorea in these participants.1,15
Several reports have described the association of chorea with aPL.1,6,7,15 In this report, all four individuals with chorea had aPL, further strengthening this association. In this study, no patient had documented thromboses, and therefore no patient fulfilled the Sapporo diagnostic criteria of APS,10 a situation described frequently in chorea associated with aPL.7,15 However, all four patients with chorea had persistently elevated high titres of anticardiolipin antibodies and lupus anticoagulant for over 12 months (ongoing), increasing our diagnostic confidence that these patients have non-thrombotic complications of APS. A limitation of our study is that no patient had anti-β2-glycoprotein I antibodies measured, although the patients still fulfilled category 1 laboratory criteria of APS.10 Interestingly, two of the four patients with aPL chorea also had positive streptococcal serology at onset. Although it is possible that this is an incidental finding, the similarities between Sydenham chorea and aPL chorea are notable, and Streptococcus may play a pathogenic role in these patients.
The treatment of neurological complications associated with aPL is contentious. If there is a definite thrombosis associated with aPL, a diagnosis of definite APS is made, and anticoagulation is recommended.4,10 However, there is no clear recommendation for the treatment of non-thrombotic complications associated with aPL.4,5 Although aspirin is often given, its efficacy is unclear.4 Anticoagulation with warfarin is not recommended in chorea associated with aPL, as anticoagulation has been associated with fatal cerebral haemorrhagic complications.7 All of our patients with aPL have been told to manage or minimize thrombotic risk factors such as immobilization and surgery,5 and are aware that oestrogen can precipitate relapses.15
The evidence that cerebrovascular disease caused the chorea in our four patients was lacking although two had non-specific white-matter lesions, none had abnormalities of the basal ganglia, and no patient had acute evidence of stroke using MR angiogram or diffusion-weighted imaging. Previous studies have shown that 62% of patients with chorea associated with aPL have white-matter lesions, but basal ganglia lesions are rarely reported.15 It is conceivable that these participants had subtle small-vessel thromboses involving the lenticulo-striate arteries, which was beyond the resolution of the imaging techniques used. The alternative hypothesis is that the chorea is mediated by aPL binding directly to neurons.4,6,8
In contrast to the participants with chorea, those with parkinsonism in this report did not have aPL, and case reports of juvenile SLE parkinsonism have found variable aPL.3,16,17 Therefore, lupus parkinsonism appears to be mediated by a different autoimmune mechanism to aPL, possibly by antibodies directed against dopaminergic cells.18 MRI in lupus parkinsonism is often normal,3,17 although it may occasionally reveal basal ganglia lesions.16 The two patients with parkinsonism in this report both deteriorated with conventional immune suppression (steroids and cyclophosphamide) until therapies targeting humoural immunity were used (plasma exchange or rituximab). Rituximab is an anti-CD20 monoclonal antibody that depletes B-cells, and has been used in refractory paediatric SLE.19 Therefore, the therapeutic responses in both patients with lupus parkinsonism supports the hypothesis that autoantibodies are the probable mediators.
The role of autoantibodies in neuropsychiatric complications of SLE and aPL is likely to be complex, and many antibodies have been described.20,21 Many of the antibodies associated with neuropsychiatric SLE bind to intracellular antigens, and their direct pathogenic role in neuropsychiatric syndromes is therefore uncertain.20 Many of these antibodies have been defined using western blotting, a method that alters protein conformation and releases intracellular antigens. These antibodies have questionable pathogenic significance. By contrast, antibodies that bind to antigens on the cell surface of neurons are likely to be pathogenic.22–25 As most pathogenic autoantibodies involved in autoimmunity of the CNS bind to the cell surface of neurons, we used an assay to measure serum IgG binding to the cell surface of neurons in our patients. We demonstrated that the patients with movement disorders associated with SLE or aPL all had elevated IgG neuronal cell-surface binding compared with comparison groups. We did not find any difference between the patients with chorea and those with parkinsonism. This assay does not define the exact antigens involved in IgG binding, nor does it demonstrate whether the antibodies are pathogenic. Therefore this technique is useful at demonstrating a ‘proof of principle’ that these patients have increased antibodies that could mediate autoimmunity of the CNS. This finding is not specific to SLE or aPL, as we have demonstrated elevated IgG cell-surface binding in other autoimmune movement disorders including Sydenham chorea (data not shown). A previous report using similar methods found higher antineuronal antibodies in patients with SLE with neuropsychiatric complications (mainly headache), than those with SLE with no neuropsychiatric complications.26 By contrast, we found that patients with SLE without obvious neuropsychiatric complications had similar elevated IgG binding to the cell surface of neurons (data not shown). It has been previously proposed that the integrity of the blood–brain barrier may be important in protecting patients with SLE from peripheral circulating antibody and other immune factors.27 Examination of CSF IgG binding to the cell surface of neurons, and examination of the integrity of the blood–brain barrier including using gadolinium in MRI studies, may be able to test this hypothesis further.20,27
In conclusion, we have described our experience of movement disorders associated with SLE and aPL, and demonstrated increased IgG binding to the cell surface of neurons. Defining the exact antigens involved in IgG cell-surface binding would further our understanding of the neuropsychiatric complications of SLE. The investigations, responses to immunosuppressive therapy, and demonstration of IgG cell-surface binding strongly suggest an underlying autoimmune pathophysiology in these patients, rather than a cerebrovascular process. These findings lend support to the use of immune-suppressant therapies rather than anticoagulants in movement disorders associated with SLE and aPL.
RCD and FB have support from the University of Sydney postdoctoral research scheme. They also received grants from Multiple Sclerosis Research Australia and the Tourette Syndrome Association.