Introduction

Alemtuzumab is an anti-CD52 humanised monoclonal antibody recently approved in the European Union and by the US Food and Drug Administration as a therapeutic agent in relapsing–remitting multiple sclerosis (RRMS). Alemtuzumab induces a profound lymphopenia; reconstitution of the various lymphocyte subsets leads to prolonged modulation of the immune repertoire [1–3]. This leads to significant reductions in both relapse frequency and sustained accumulation of disability, compared to interferon beta-1a, in RRMS [4–6].

Neuromyelitis optica (NMO) spectrum disorders are a group of syndromes caused by antibodies, discovered in 2004 [7], targeting the aquaporin-4 water channel on astrocytic foot processes [8]. Alemtuzumab has occasionally been used to treat NMO, both before and after NMO was recognised as a distinct clinical entity. Here we report out experience.

Methods

We conducted a retrospective case review, from 1991 to the present day, of patients treated with alemtuzumab at the Department of Neurology in Addenbrooke’s Hospital, Cambridge, UK, to identify those patients who fulfil the criteria for NMO spectrum disorder under the revised Wingerchuk criteria [9] and the consensus criteria drawn up by the International Panel for NMO Diagnosis convened in 2011 [10]. The main outcome measures considered were relapse rate, accumulation of disability and long-term outcome after alemtuzumab treatment.

Results

In retrospect, three cases could be classified as neuromyelitis optica. One case (case 1) was from the 1990s; the remaining two were from the early 2000s.

Case 1

A 9-year-old girl presented in 1998 with protracted vomiting followed by transverse myelitis, two relapses of bilateral optic neuritis and two brainstem episodes (nystagmus, ataxia and facial sensory disturbance) within a year. Imaging showed one lower medullary and several longitudinally extensive cord lesions (images no longer available). Despite high-dose steroids and intensive immunosuppression (Table 1), she remained significantly visually impaired (acuities down to 6/60), as well as having a spastic paraparesis. She received alemtuzumab with no relapses for 12 months. At month 12, she relapsed with myelitis with longitudinally extensive thoracic cord demyelination and was retreated with alemtuzumab; only weeks later, a further relapse resulted in dense paraparesis, rendering her wheelchair bound. Her vision deteriorated further and she was registered blind by age 12. In 2005, her illness was recognised as antibody-mediated NMO [7] and she received plasma exchange (PLEX). She died at age 18, whilst awaiting rituximab treatment. Assays for aquaporin-4 immunoglobulin-G (AQP4-IgG) were not available at the time.

Table 1 Demographic and clinical data
Full size table

Case 2

A 41-year-old female presented in 2004 with bilateral optic neuritis and a brainstem deficit (left-sided horizontal nystagmus and ataxia) and two episodes of myelitis within 6 months; imaging showed enhancing brain and longitudinally extensive cervical cord lesions (Fig. 1a–c). She received pulsed high-dose steroids and remained with residual spastic paraparesis. Three cycles of alemtuzumab were administered 12 months apart. At months 3, 4 and 11 after the first cycle, she had discrete relapses attributable to new thoracic cord lesions (Fig. 1d), whilst cerebral lesions improved (not shown). At months 2 and 9 after second cycle, she relapsed with brainstem symptoms (vertigo and incoordination) and bilateral optic neuritis, with residual visual impairment (6/12). After the third cycle, she deteriorated with disabling cognitive deficits. She received rituximab, to no effect, later developing seizures and tumefactive white matter lesions (Fig. 1e, f). A brain biopsy performed to confirm active demyelination showed extensive loss of myelin with numerous CD68+ macrophages throughout, but the specific features of neuromyelitis optica were not sought. She went on to be treated with mitoxantrone with poor response. Late in the course of her illness, as the assay became available, she was tested for AQP4-IgG antibodies, which were absent. She died aged 51.

Fig. 1
figure 1

MRI images. Case 2: a pre-treatment T2-weighted axial views show high signal in posterior part of corpus callosum and left middle cerebellar peduncle. b T2-weighted sagittal view of cervical cord shows expansion with an extensive lesion from medulla to T1 level, and c gadolinium enhancement on T1-weighted image. d Post-treatment T2-weighted sagittal view shows new expansion of thoracic cord between T4 and T7 with increased T2 signal but no gadolinium enhancement. e T2-weighted axial images of brain (left) and FLAIR images (right) at month 20 after third dose of alemtuzumab show a large lesion in right parietal and posterior frontal white matter that extends to posterior corpus callosum and a smaller lesion in left parietal white matter. f Both lesions show an incomplete peripheral rim of enhancement on T1-weighted post-gadolinium axial image (left) and coronal image (right). Case 3: g pre-treatment T2-weighted sagittal MRI sequence (left) shows subtle expansion of cervical cord from C2 to C6 inclusive with diffuse increase in signal, more localised at level of C3/4 as demonstrated on axial view (right). h T1-weighed sagittal and axial sequences with gadolinium contrast show enhancement involving the left side of the cord. Worsening of disease at month 5 after alemtuzumab is shown by i T2-weighted sagittal views of spine showing progression of T2 signal change and swelling in cervical cord (left) and new thoracic hyperintensity extending from T2 to T7 (right); with gadolinium enhancement (j). k T2-weighted sagittal view of cord shows patchy increased intensity and evidence of cord atrophy, 7 years after disease onset

Full size image

Case 3

A 31-year-old female presented in 2007 with two episodes of transverse myelitis with longitudinally extensive cervical cord lesions (Fig. 1g, h) and tested seropositive for AQP4-IgG antibody. Two years prior, she had a prolonged episode of nausea, vomiting and diplopia that had resolved spontaneously. Response to high-dose steroids and PLEX was moderate, with residual paraesthesia, unsteadiness and tonic spasms. She received a single course of alemtuzumab (12 mg daily for 5 consecutive days) in an attempt to prevent further relapses but at months 3 and 5 after alemtuzumab she represented with spastic paraparesis; cord imaging showed disease progression with new enhancing longitudinally extensive spinal cord lesions (Fig. 1i, j). At months 10 and 18, she had further relapses of myelitis without cord enhancement, losing dominant hand function and ability to walk unassisted despite high-dose steroids and PLEX. She became relapse free for 4 years after two cycles of rituximab in combination with mycophenolate mofetil (MMF), though disability persists with cord atrophy (Fig. 1k). Serum AQP4-IgG antibody remained positive after alemtuzumab treatment.

Discussion

We report three cases which demonstrate that alemtuzumab failed to prevent disabling relapses and poor outcome in neuromyelitis optica. This converges from evidence from other drugs effective in multiple sclerosis which also fail to suppress—and indeed may exacerbate—the illness, like interferon beta and natalizumab [11, 12]. We are aware of two other cases of neuromyelitis optica treated using alemtuzumab. In one, there was no effect on the disease which was subsequently effectively treated using an IL-6 antibody [13]. In another, there was a cessation of relapses and reduction in lymphocytes detected in brain lesions pathologically, but the patient deteriorated and died with significant monocyte invasion of her brain lesions [14].

All three cases fulfil the criteria for NMO spectrum disorder under the revised Wingerchuk criteria [9] and the International Panel for NMO Diagnosis consensus [10]. Both cases 1 and 2 presented prior to the clinical availability of an AQP4-IgG antibody test; stored serum samples from case 2 were negative; case 3 was positive in life.

In one of our cases (case 3), the patient subsequently had a good outcome with rituximab and mycophenolate. We hypothesise that the important mechanistic difference between alemtuzumab and rituximab is the latter’s prolonged depletion of B lymphocytes, which are only depleted for 1–2 months after alemtuzumab. The use of alemtuzumab is associated with secondary autoimmunity, directed mainly against the thyroid, but we do not propose this mechanism for continued disease activity of NMO after alemtuzumab.

From this case series, and a review of the literature, we therefore caution against the use of alemtuzumab in cases of neuromyelitis optica.