This article is commented on by Pranzatelli on page 782 of this issue.
Cerebrospinal fluid B-cell expansion in longitudinally extensive transverse myelitis associated with neuromyelitis optica immunoglobulin G
Version of Record online: 17 JUN 2011
© The Authors. Developmental Medicine & Child Neurology © 2011 Mac Keith Press
Developmental Medicine & Child Neurology
Volume 53, Issue 9, pages 856–860, September 2011
How to Cite
DALE, R. C., TANTSIS, E., MERHEB, V. and BRILOT, F. (2011), Cerebrospinal fluid B-cell expansion in longitudinally extensive transverse myelitis associated with neuromyelitis optica immunoglobulin G. Developmental Medicine & Child Neurology, 53: 856–860. doi: 10.1111/j.1469-8749.2011.03975.x
- Issue online: 2 AUG 2011
- Version of Record online: 17 JUN 2011
- PUBLICATION DATA Accepted for publication 28th February 2011. Published online 17th June 2011.
A first episode of central nervous system (CNS) demyelination may represent heterogeneous entities such as acute disseminated encephalomyelitis, clinically isolated syndrome, neuromyelitis optica (NMO), or multiple sclerosis. As new immune therapies become available, it is increasingly important to make an early diagnosis. Autoantibodies such as NMO immunoglobulin G (IgG) and myelin oligodendrocyte glycoprotein IgG are increasingly being employed to define subgroups of CNS demyelination or guide treatment. Similarly, cerebrospinal fluid (CSF) immunophenotyping can demonstrate B-lymphocyte subpopulation expansion, which has been used to guide therapy in other autoimmune CNS disorders. We present a report on a 15-year-old male with longitudinally extensive transverse myelitis with magnetic resonance imaging findings of oedema, cavitation, and gadolinium enhancement. NMO-IgG and aquaporin 4 IgG were positive; thus, we diagnosed a limited form of NMO. Acute CSF immunophenotyping revealed a 3.6% expansion of CD19 B-cell populations, whereas a comparison group of five children (4 males, age range 2–15y; mean age 7y) with other neurological disorders showed only a 0.51% expansion (SD 0.25%). In view of the diagnosis of a ‘limited form of neuromyelitis optica’, we therefore elected to treat him aggressively from the outset with a prolonged steroid regimen and mycophenylate mofetil. This case demonstrates a correlation between autoantibody production and CSF B lymphocyte expansion in an individual with CNS demyelination. These approaches could be used in individuals with a first episode of CNS demyelination to help delineate immunological subgroups and guide treatment.
Acute disseminated encephalomyelitis
Longitudinally extensive transverse myelitis
Other neurological disorders
What this paper adds
- •Transverse myelitis associated with neuromyelitis optica immunoglobulin G can support a diagnosis of a ‘limited form of neuromyelitis optica’ in children.
- •CSF immunophenotyping demonstrated B-cell expansion in our patient, who had neuromyelitis optica.
- •This approach may be useful in defining immunological subgroups of CNS demyelination in children.
A first episode of central nervous system (CNS) demyelination in children can represent a monophasic disorder such as acute disseminated encephalomyelitis (ADEM) or, alternatively, represent a relapsing disorder such as multiple sclerosis or neuromyelitis optica (NMO).1 NMO is a rare inflammatory autoimmune demyelinating disorder of the spinal cord and optic nerves.2 Untreated, NMO will have a relapsing course in 80 to 90% of individuals, and those with NMO experience significant morbidity and mortality.2 The inflammatory demyelinating process is primarily directed against astrocytes with secondary oligodendrocyte involvement, and may be associated with profound oedema, inflammatory infiltrate, and necrosis with cavitation.2 NMO is often associated with a defining pathogenic autoantibody (NMO immunoglobulin G [IgG]), which binds to the water channel aquaporin 4 (AQP4),3,4 a protein that is mainly expressed on astroglial foot processes.3,4 Revised diagnostic criteria for NMO were created in 2006 that include the presence of NMO-IgG.5
The presence of NMO-IgG in adults with a first episode of longitudinally extensive transverse myelitis (LETM) is termed a ‘limited’ form of NMO and has a high risk of relapse.6 The majority of children with LETM have a monophasic course and are negative for NMO-IgG.1,7 The few children described with LETM who are associated with positive NMO-IgG had recurrent disease and significant disability.7,8
Encouragingly, NMO is often responsive to acute immunotherapy with steroids, intravenous immunoglobulin, or plasma exchange.2 In view of the high risk of relapse in NMO, there is an increasing move to use immune-suppressive therapies that substantially reduce the rates of relapse.9,10
Another useful measure of CNS inflammation is cerebrospinal fluid (CSF) immunophenotyping, which can define expansion of different lymphocyte subtypes.11 This approach has been utilized in opsoclonus–myoclonus syndrome to define B-cell expansion and support the use of B-cell-depleting therapies such as rituximab.12,13
We report the case of an individual with atypical LETM who had positive NMO-IgG and an expansion of CD19 B-cell population in his acute CSF. We elected to treat this individual aggressively with chronic immune suppression from the outset.9 This case demonstrates the potential value of autoantibodies and CSF immunophenotyping in individuals with a first episode of CNS demyelination.11
Acute CSF was taken from our index patient before treatment, and routine CSF microscopy revealed three mononuclear cells per mm3, no polymorphs per mm3, and no red blood cells per mm3. For comparison, we present five participants (mean age 7y; range 2–15y) with other neurological disorders: two had hydrocephalus, one had Tourette syndrome, one had epilepsy, and one had dystonic cerebral palsy. The routine microscopy in the comparison group revealed a mean mononuclear count of 1/mm3 (range 0–2/mm3) and a mean red blood cell count of 1/mm3 (range 0–4/mm3). The total volumes of CSF varied according to CSF availability and ranged from 10 ml to 500μl. It is necessary to minimize contamination of CSF by blood, although in our experience CSF immunophenotyping is possible if the red blood cell count is lower than 100 red blood cells per mm3. We used a staining method adapted from Kivisakk et al.14 A 900μl sample of CSF was collected and immediately placed on ice, and then centrifuged at 600g for 7 minutes at 4°C. Non-specific Fc receptor binding was blocked with 0.2mg/ml normal mouse IgG (Invitrogen, Carlsbad, CA, USA) for 15 minutes at room temperature. Directly labelled anti-CD3 (clone UCHT1; BD Biosciences, Sparks, MD USA), anti-CD19 (clone SJ25-C1; Invitrogen), anti-CD4 (clone RPA-T4; BD Biosciences), anti-CD8 (clone SK1; BD Biosciences), anti-CD27 (clone L128; BD Biosciences), anti-CD14 (clone TUK4; Invitrogen), anti-CD138 (clone MI15; BD Biosciences), and anti-CD38 (clone HIT2; eBioscience, San Diego, CA, USA) were added to the cells without washing the mouse IgG. The samples were incubated for 15 minutes at room temperature in the dark. Cells were washed once with 1ml of ice-cold phosphate-buffered saline supplemented with 2% fetal bovine saline and centrifuged at room temperature for 30 seconds at 5000g in a table-top microcentrifuge. The supernatant was carefully removed by pipetting. Cells were then re-suspended in 100μl of phosphate-buffered saline and fetal bovine saline and 1% paraformaldehyde. All cells in the sample were then acquired on a BD LSRII instrument (BD Biosciences) and compensation was performed using BD CompBeads (BD Biosciences) compensation particles. Data were analysed with FlowJo software (TreeStar, Ashland, OR, USA). The absolute number of events acquired by flow cytometry ranged from 995 to 11643. This experimental approach received ethical approval from the Children’s Hospital at Westmead Ethics Committee and the family of the 15-year-old male gave written consent to publish this case report.
A previously healthy 15-year-old white Australian male presented with a progressive subacute spinal cord syndrome. Ten weeks before presentation he complained of pins and needles in his feet, which evolved to affect the whole of each leg. During the 4 weeks before admission, his walking became increasingly altered with unsteadiness and reduced exercise capacity. In addition, in the weeks preceding admission, his upper limbs became involved and he had urinary frequency. He had no alteration in thinking, behaviour, or confusion and his vision was unchanged.
He had a normal cranial nerve and visual examination. His upper limbs showed loss of light touch and proprioception, plus impaired repetitive finger movements, but normal power and reflexes. His lower limb examination revealed mild pyramidal weakness, pathologically brisk reflexes, upgoing plantars, and loss of light touch and proprioception. He had a sensory level at T2, absent abdominal reflexes, altered anal sphincter tone, and evidence of urinary retention. He had a spastic gait and was able to walk for only 50m.
Magnetic resonance imaging (MRI) of the spine using T2 sequences showed severe oedema and enhancement in a patchy distribution from C1 to T7. In addition, MRI T1 sequences revealed cavitation and patchy gadolinium enhancement (Fig. 1a). MRI of the brain demonstrated asymptomatic white matter lesions (Fig. 1c). Tests on the CSF revealed three mononuclear cells, no polymorphs, no red blood cells, normal protein levels (0.3g/dl), normal glucose levels, neopterin, cytospin, and negative oligoclonal bands. CSF polymerase chain reaction was negative for enterovirus and herpes simplex virus. Serology for Epstein–Barr virus was positive, compatible with previous infection. Serology for antinuclear antibody and anticardiolipin antibody was negative, and vitamin D titre was normal. Serology for NMO-IgG from the National Reference Laboratory was positive, with a titre of 1:640. Testing for autoantibodies against AQP4 was positive (score 3+).15
A diagnosis of a ‘limited form of NMO’ was made, and he was given intravenous methylprednisolone at 1g per day for 5 days and 2g/kg intravenous immunoglobulin. He improved within a week of starting therapy, and was discharged after a 2-week admission with residual sensory ataxia and pyramidal signs. He was given a prolonged tapering course of prednisolone starting at 60mg per day. He remains on 20mg of prednisolone on alternate days at 6 months. In addition, during the first month of treatment, he was started on mycophenylate at 1000mg twice a day, which he is tolerating. At 6 months, his functional motor performance is normal, and he is running and surfing at his previous ability. However, his examination still shows brisk reflexes and upgoing plantars in his lower limbs, as well as absent abdominal reflexes. His follow-up spine MRI revealed residual damage with cavitation and his brain MRI showed partial improvement in the asymptomatic white matter lesions (Fig. 1b,d). His NMO-IgG remains positive, with a titre of 1:160.
CSF phenotyping by flow cytometry
The CSF of the participant with NMO (NMO-CSF) and the CSF of the five comparison children with other neurological disorders (OND-CSF) were stained immediately after collection and were phenotyped for the presence of B and T lymphocytes. The number and percentage of B and T cells in the CSF was determined in each specimen by flow cytometry. An increased percentage of B cells (CD19) was found in the NMO CSF (3.62%) compared with the OND-CSF (0.51%, SD 0.25%; Fig. 2a,b; Table I). T-cell percentages were similar in the NMO-CSF and the OND-CSF CD4 and CD8 T-cell percentages in the NMO-CSF were within the range observed in the comparison group (59.2% and 29.8% in the NMO-CSF vs 64.16% [SD 5.5%] and 25.2% [SD 6.6%] in the OND-CSF respectively; Fig. 2). Memory CD4 T cells constituted the majority of T cells in both NMO- and OND-CSF (Fig. 2). These results suggest that the individual with NMO has B-cell expansion in the CSF. Absolute numbers of B cells reflected the same trend. Low absolute numbers of CD19 B cells were found in the OND-CSF (3–4 B cells), whereas 33 B cells were detected in the NMO-CSF (Table I). We did not see a correlation between the cells counted by routine microscopy and the absolute B-cell number or B-cell percentage analysed by flow cytometry. Peripheral lymphocyte immunophenotyping performed at the same time in the individual with NMO revealed a typical number of CD19 B cells (18%, typical 8–24%), CD4 T cells (38%, typical 25–48%), and CD8 T cells (34%, typical 9–35%).
|Specimen||Number of mononuclear cells per mm3 as per routine CSF microscopy||Gated subpopulation of lymphocytes and monocytes (%)a||Absolute number of lymphocytes and monocytesb||Gated CD19+ B cells (%)c||Absolute number of CD19+ B cells|
|OND CSF 1||0||16.7||1667||0.18||3|
|OND CSF 2||0||10.2||563||0.71||4|
|OND CSF 3||0||12.3||1111||0.36||4|
|OND CSF 4||2||53.4||500||0.8||4|
|OND CSF 5||0||23.5||588||0.51||3|
NMO is a rare inflammatory autoimmune demyelinating disorder, and there are standardized criteria for diagnosis, which now include the use of NMO-IgG as a diagnostic marker.5 NMO is rare in adults and very rare in children. The three largest studies of NMO in children reported a total of 38 children and were mainly multicentre or international collaborative studies.7,8,16 Although a diagnosis of NMO previously required the presence of both optic neuritis and myelitis, it is now clear that some individuals with isolated optic neuritis or isolated LETM harbour NMO-IgG and are part of the NMO spectrum; these individuals are termed as having ‘limited forms of NMO’ and have a high risk of relapse or disability.2,6,8,17
We have presented the case of an individual with LETM associated with serum NMO-IgG whose presentation was subacute and progressive, and whose MRI findings were typical of NMO with oedema, cavitation, and patchy gadolinium enhancement on T1 MRI sequences.2 Brain MRI lesions are described in 60% of adults with NMO and are often asymptomatic, as in our case.2
NMO, if untreated, can have a poor outcome with significant motor or visual morbidity. After 5 years of relapsing NMO, more than 50% of individuals are blind in one or both eyes, or require ambulatory help.2 Although monophasic NMO can occur, relapsing courses are more common in NMO, and the presence of NMO-IgG appears to be strongly associated with a relapsing course in children.2,7 As in our patient, NMO is often responsive to immune therapies, and most individuals respond well to steroids, although some may become steroid dependent. In view of the relapsing course in many individuals, chronic immune suppression is often employed, and mycophenylate, cyclophosphamide, and rituximab have all been shown to reduce substantially the relapse rate in adults with NMO.9,10 In view of the severity of the LETM in this case, the residual cavitating spinal cord findings, and the presence of NMO-IgG, we elected to use chronic immune suppression from the outset. The duration of treatment will be partly guided by longitudinal NMO-IgG monitoring. Rituximab will be considered if this individual has a severe relapse despite receiving mycophenylate.
There is now persuasive evidence that NMO-IgG is a pathogenic autoantibody. NMO-IgG was originally described as IgG that binds to perivascular regions of the CNS, and has been shown to be very specific to NMO.4,18 Not all adults and children with clinical NMO have NMO-IgG, and the sensitivity is approximately 60 to 90% in adults.2,18 In children, NMO-IgG has 47% sensitivity in NMO, and 78% sensitivity in relapsing NMO.7 The specific target of NMO-IgG is AQP4, a water channel that is highly expressed on astrocyte foot processes.3 The titre of AQP4 antibody correlates strongly with the clinical course and risk of relapse.19,20 Pathological and in vitro studies have shown that NMO-IgG has pathogenic effects and mediates AQP4 depletion, complement-mediated astrocyte cytotoxicity, and secondary oligodendrocyte damage.21–23 A recent paper examining the CSF of one adult with NMO demonstrated clonally expanded plasma cell populations with a targeted response against AQP4.24
Other than NMO-IgG, autoantibodies against native myelin oligodendrocyte glycoprotein have been described in demyelination in children (ADEM, clinically isolated syndrome, and multiple sclerosis), and may have a pathogenic function, suggesting that there may be multiple subgroups of CNS demyelination in children associated with defining autoantibodies.25,26
A further demonstration of the role of autoantibodies, and humoral autoimmunity in general, is to measure the relative presence of B cells.19 CD19 B cells usually represent less than 0.5% of the total CSF lymphocyte population in the CSF of healthy children or children with OND.12 The proportion of B cells has been shown to be increased in individuals with the putative autoimmune disorder opsoclonus–myoclonus syndrome, and the degree of B-cell expansion in individuals with opsoclonus–myoclonus syndrome has been correlated with disease severity.12 This finding in opsoclonus–myoclonus syndrome has supported the rationale of B-cell-depleting therapies such as rituximab.13 In our participant, CD19 B-cell percentage and absolute number in the CSF were significantly increased. It is possible that this increased number is associated with B-cell differentiation. However, there were inadequate numbers of B cells in the CSF of our patient to confidently count subpopulations such as plasmablast cells (CD38) and plasma cells (CD138; result not shown). B-cell expansion is also reported in other autoimmune CNS disorders, including multiple sclerosis.27 Previous studies have presented B-cell expansion in percentages rather than as an absolute number of cells.12,27 There was no correlation between routine CSF microscopy and flow cytometry immunophenotyping, suggesting that only flow cytometry can provide information on B-cell expansion. We also confirmed that memory CD4 T cells (CD3+ CD4+ CD27+) constituted the majority of T cells in both the NMO CSF and OND-CSF, as previously reported in adults.28
This report must be considered provisional as only one individual with NMO is presented. Further individuals with a first episode of CNS demyelination should be enrolled to determine the usefulness of CSF immunophenotyping in defining subgroups and guiding therapy. However, it is hoped that CSF immunophenotyping in association with autoantibody measurement may allow a ‘tailoring’ of therapy according to the individual’s dominant immune response.11
RCD and FB have funding from Star Scientific Foundation, the Australian National Health and Medical Research Council, the University of Sydney, Multiple Sclerosis Research Australia, and the Trish Multiple Sclerosis Research Foundation. Flow cytometry was performed in the Flow Cytometry Core Facility that is supported by Westmead Millennium Institute, National Health and Medical Research Council, and Cancer Institute New South Wales. We are grateful to the family of the patient, who gave permission for this study.