Article Text
PDF
Abstract
Chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids (CLIPPERS) is a recently described inflammatory disease of the CNS with a predilection for the hindbrain and responsive to immunotherapy. Five further cases are described with detailed pathology and long term evaluation. CLIPPERS does not represent a benign condition, and without chronic immunosuppression the disease may relapse. The radiological distribution is focused not only in the pons but also in the brachium ponti and cerebellum. Pontocerebellar atrophy occurred early, even in cases treated promptly. Significant cognitive impairment was seen in some cases and was associated with additional cerebral atrophy. The pathology included distinctive histiocytic as well as lymphocytic components and evidence of neuro-axonal injury. Additional subclinical systemic findings on investigation were identified. Relapse was associated with withdrawal of corticosteroids, and disability was least marked in cases where both the presentation and relapses were treated promptly. We propose that the title of the syndrome be amended to chronic lymphocytic inflammation with pontocerebellar perivascular enhancement responsive to steroids to more accurately reflect the distribution of the radiological findings.
Statistics from Altmetric.com
Introduction
A series of cases with corticosteroid responsive perivascular brainstem and cerebellar inflammation have recently been reported and given the name chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids (CLIPPERS).1 This condition shows a typical pattern of nodular or punctate contrast enhancement in the brachium pontis, pons and cerebellum on MRI, and a predominantly T lymphocytic, perivascular infiltrate with parenchymal extension on brain biopsy. Five biopsy proven cases are described here, which expand on the clinical, radiological and pathological features of the disease. The presence of cognitive impairment, new characteristic features of MRI lesions, significant axonal injury in biopsy specimens and subsequent development of profound local and, in some cases, additional diffuse cerebral atrophy, are novel findings.
Methods
Patients
The study was approved by the Sydney University Human Research Ethics Committee. Five patients with symptoms referable to the brainstem or cerebellum were assessed by a neurologist (SWR, MHB, JDP, MWH) and followed clinically for a mean of 57 months (range 12–100 months). The modified Rankin Scale2 was applied retrospectively based on review of the clinical records from each clinical visit. All patients underwent brain biopsy early in the course of their disease and additional biopsies (parotid (n=2), conjunctiva (n=1), bone marrow (n=2) and skin (n=2)) were performed in some patients. Extensive blood and CSF investigations were performed in all cases to exclude infection, vasculitis, paraneoplastic inflammation, lymphoma, sarcoidosis, other autoimmune diseases and Langerhans cell histiocytosis (LCH) as possible causes of the radiological abnormalities identified. MRI was obtained before and soon after treatment and up to 84 months after commencement of treatment (range 9–84 months, mean 44 months). Additional imaging investigations were performed including chest x-ray (n=5), MRI of the spinal cord (n=2), cerebral angiography (n=3), CT scan of the chest, abdomen and pelvis (n=3), positron emission tomography (PET) scan of the brain (n=3), whole body PET scan (n=2) and whole body gallium scan (n=3).
Pathological methods
Brain biopsies targeted MRI lesions in the posterior fossa in four cases and in the basal ganglia in one case. Biopsy was performed prior to any corticosteroid or other immunosuppressive treatment in four cases. Case No 5 had received two isolated pulses of intravenous methylprednisolone (IVMP) but no treatment had been given for 9 months prior to biopsy. Biopsy specimens were evaluated using routine formalin fixed, paraffin embedded sections after staining with haematoxylin–eosin. Additional stains were performed including Luxol fast blue, Garvey modified Bielschowsky silver stain, periodic acid Schiff reagent with and without diastase digestion, Warthin Starry stain and Ziehl Neelson stains for acid fast bacilli. Immunohistochemical staining for glial fibrillar acidic protein, CD3, CD4, CD8, CD5, CD7, CD20, CD10, CD23, CD68, CD138, κ and λ light chains, CD1a, SV40 large T antigen (polyoma virus) and antibody for toxoplasma were performed. In situ hybridisation studies were also performed for Epstein–Barr virus early RNA-1.
Results
Clinical characteristics and demographics
The clinical characteristics of the cases are shown in table 1. Mean age at the time of onset was 43.4 years (range 20–65); four of the five patients were men. Subacute gait ataxia marked the clinical onset of the disease in three of the five cases. Case No 4 first presented with a right facial nerve palsy, which improved over the course of several weeks with ataxia developing after its resolution. Case No 5 initially developed headaches and fatigue followed by ataxia. At the point of maximal disability, the ataxia was accompanied by appendicular cerebellar signs (n=4), dysarthria (n=3), pyramidal weakness (n=3), eye movement abnormalities (n=3), tremor (n=1), cognitive impairment (n=4) and pseudobulbar affect (n=3). Cognitive features included dysexecutive syndromes, language disturbance, perseveration and frontal release reflexes. Symptoms and disability progressed after onset in all patients. Figure 1 demonstrates the change in disability status over time, which paralleled the accumulation of clinical signs.
Clinical characteristics and investigation findings of the five cases
Graphical representation of the change in disability, as assessed by the modified Rankin Scale (mRS), over the course of the illness. Treatment is documented above the graph and the nature of the treatment is indicated in the boxed legend. AZA, azathioprine; IV MP, intravenous methylprednisolone; MMF, mycophenolate mofetil; MTX, methotrexate.
Case No 1 had a history of left parietal stroke 9 years prior to onset, complicated by a single post-stroke seizure. There was a mild residual speech disturbance as the only permanent sequela of the stroke. She had a colonic malignancy resected 6 years prior to the onset of ataxia. This was treated with adjuvant chemotherapy and no recurrence had been detected on subsequent screening. Case No 4 had a background of type 2 diabetes mellitus complicated by peripheral neuropathy. Case No 3 had a significant drug use history, consuming more than 10 g cannabis per day. The remaining cases had no prior significant medical history. In particular, no patient had sicca symptoms, rash, arthritis, oral or genital ulcers, or other symptoms or signs of systemic immune mediated diseases. In no case was there evidence of keratoconjunctivits sicca, uveitis or retinal vasculitis on formal ophthalmological examination, including Rose–Bengal staining and the Schirmer test. No case demonstrated evidence of pathergy after needle induced cutaneous injury.
Neuroimaging results
MRI of the brain showed punctate gadolinium enhancement with a perivascular pattern centred on the brachium pontis in all cases, with variable involvement of the pons and cerebellar white matter. While the lesion burden on imaging was always most dense in the brachium pontis, pons and cerebellum, typical lesions were also seen in the medulla, midbrain, cervical spinal cord, basal ganglia, thalamus and cerebral white matter in a number of cases (table 1). Contrast enhancing lesions were associated with mild to moderate hyperintensity on T2 weighted MRI, including fluid attenuated inversion recovery sequences. However, T2 weighted hyperintensity did not significantly extend beyond the boundaries of the contrast enhancement of individual lesions, and this striking deficiency of vasogenic oedema is an unusual feature of this disorder (figure 2). PET-CT of the brain showed no change or minor hypermetabolism of contrast enhancing areas, considerably less than seen in lymphoma.
Coronal MRI brain in case No 1 shows punctate contrast enhancement affecting the brachium pontis, pons and cerebellar white matter on contrast enhanced T1 weighted images (1a) with relatively modest hyperintensity on T2 weighted images (1b). Contrast enhanced T1 weighted MRI of case Nos 2–5 show perivascular contrast enhancement in the brachium pontis, pons and cerebellar white matter.
Laboratory investigations
All patients had normal or negative serology for syphilis, HIV, cryptococcus, toxoplasma, Lyme, brucella, Q fever, Chlamydia psittaci, mycoplasma, Epstein–Barr virus, cytomegalovirus, herpes simplex virus, varicella zoster virus, Barmah–Forest virus, human T lymphocyte virus 1 and 2, flavivirus panel, Murray Valley encephalitis virus and Ross River Fever virus serologies. One had positive IgM toxoplasma serology but no other evidence of toxoplasmosis on laboratory testing, including toxoplasma PCR on CSF and brain samples. Autoantibodies were tested in all patients and found in two cases (table 1). None of the five cases had an elevated serum angiotensin converting enzyme level, antineuronal antibodies, serum paraprotein or elevated tumour markers.
CSF examination was performed on all cases (table 1). Abnormal findings were mild lymphocytic pleocytosis and/or elevated protein in some but not all cases. Oligoclonal bands were detected in one of five cases (CSF analysis performed 3 months after symptom onset). Antibody and PCR tests for Whipple's disease, syphilis and viral panels were negative.
Evidence of systemic involvement
There was evidence of subtle systemic involvement on investigations without symptoms in three cases. Whole body gallium scans were abnormal in two out of three cases. Case No 1 had abnormal uptake in the left parotid and case No 4 had abnormal uptake in the right parotid. Parotid biopsy of the abnormal side was performed in both cases prior to the initiation of treatment. The biopsies, prior to treatment, showed chronic lymphocytic sialadenitis without granulomas in case No 1, and case No 4 was normal. Case No 4 had additional bilateral conjunctival biopsies that showed chronic lymphocytic inflammation without granuloma formation. Case No 3 had mildly elevated serum corrected calcium (2.73 mmol/l) and elevated 1,25-dihydroxycholecalciferol prior to any treatment, which normalised with corticosteroid therapy, and had normal thoracic CT and whole body gallium scans. Case No 3 had a normal thoracic CT and gallium scan, and neither case No 1 nor case No 4 had evidence of hilar lymphadenopathy on thoracic CT or other features of sarcoidosis on additional imaging.
Neuropathology
Five biopsies were analysed (four cerebellar, one basal ganglia). The histopathological characteristics are summarised in table 2. All five cases showed a predominantly white matter based perivascular and parenchymal lymphohistiocytic inflammatory infiltrate (figure 3A,B) comprised of small lymphocytes, macrophages and activated microglia. Accompanying reactive gliosis was noted in all cases. The severity of inflammation was variable among the five cases and ranged from large nodules and sheets of inflammation to scattered foci of inflammatory cells. Lymphocytes and histiocytes were observed in similar numbers in four of the cases while lymphocytes were predominant in one case. Plasma cells were rare and found in only two cases. Neutrophils and eosinophils were absent in all cases.
Neuropathological characteristics of the five cases
Representative neuropathology from case Nos 3, 4 and 5. Haematoxylin and eosin stain demonstrates a perivascular lymphocytic infiltrate (A: 100×, case No 3) with extension of the infiltrate into the interstitium (B: 100×, case No 3). Loose histiocytic aggregates and occasional granulomas (arrowhead) accompanied perivascular inflammation in case Nos 1, 3 and 5 (C: 200×, case No 3). The majority of the inflammatory cells are CD3 positive (D: 200×, case No 3), of which the majority are CD4 positive (E: 200×, case No 3) and the remainder CD8 positive (F: 200×, case No 3). CD68 positive histiocytes are numerous, comprising 50% or less of the inflammatory cells (G: 200×, case No 3). CD20 positive B cells comprise a minority of lymphocytes (H: 200×, case No 3). Reactive astrocytosis is seen in all cases and is clearly demonstrated with glial fibrillar acidic protein stain (I: 200×, case No 4). Neuronophagia is seen with lymphocytes and histiocytes surrounding a shrunken neuron (J: 400×, solid arrow; case No 5). Silver stains demonstrate axonal injury (K: 630×; case No 3), including axon retraction balls (K: inset). Other cases showed a similar pattern of neuropathology.
The most striking histopathological feature in all five cases was perivascular lymphohistiocytic inflammatory infiltrate involving both small arteries and veins. Occasional aggregates of histiocytes resembling loose granulomas, without significant peripheral lymphocytic cuffing or necrosis were observed, centred on small blood vessels in three cases (case Nos 1, 3 and 5; figure 3C, arrowhead). Focal transmural lymphocytic infiltration of the vessel wall was seen in three cases. Characteristic histological features of vasculitis (destruction of the vessel wall with fibrinoid necrosis, leukocytoclasia, fibrin thrombi) were not identified in any case. In addition to perivascular inflammation, inflammation of the white matter was also seen to a variable extent in all cases, ranging from scattered cells to large and relatively well circumscribed sheets of lymphohistiocytic inflammation. Axonal injury with evidence of axonal spheroids and/or torpedoes was observed in all cases assessed by modified Bielschowsky's stain (figure 3K, inset). Varying degrees of myelin loss accompanied the perivascular or interstitial inflammatory infiltrate in three cases but focal demyelination in isolation was not seen. Leptomeningeal tissue was included in the biopsy in three cases, and only scanty perivascular lymphocytes were identified in each case. Focal neuronophagia, characterised by shrunken neurons with cytoplasmic retraction and hyperchromatic, pyknotic nuclei engulfed by lymphocytes and histiocytes (figure 3J) was observed in the basal ganglia and one cerebellar biopsy, while Purkinje cell loss was seen in one of the four cerebellar biopsies.
Immunohistochemical stains demonstrated that the cellular infiltrate was predominantly composed of CD4 positive T cells and histiocytes (table 2, figure 3D–H). CD20 positive B cells were polyclonal on light chain immunohistochemistry in the two cases assessed. Immunostains for glial fibrillary acidic protein confirmed reactive astrocytosis (figure 3I) except in areas of intense perivascular inflammation where fragmentation and beading of astrocyte processes occurred. There was no evidence of atypical lymphocytes, multinucleate giant cells, viral inclusions, fungi, mycobacteria or Whipple's disease. Immunohistochemical analysis did not reveal neoplastic lymphoid proliferation, LCH, Epstein–Barr virus infection or toxoplasmosis.
Treatment and outcome
All patients received immunosuppressive therapy as treatment for their condition. An outline of the timing and nature of treatment and the clinical response, as represented by the modified Rankin Scale grading, is shown in figure 1. All patients were treated initially with pulse IVMP. In all cases, initial corticosteroid therapy induced a prompt and significant improvement in the level of disability, clinical symptoms and neurological signs, but complete resolution did not occur. Pulse IVMP was given at other times and was associated with improvement in clinical features and improvement or stabilisation of disability. With the intention of achieving more sustained clinical improvement, four of five cases received pulse intravenous cyclophosphamide. Cyclophosphamide or azathioprine given without sustained corticosteroid in case No 4 did not result in clinical improvement. All patients received chronic oral immunosuppression at some time with prednisolone, methotrexate, azathioprine, mycophenolate or combinations thereof. In general, following the local custom for immunotherapy, the cyclophosphamide or antiproliferative agent was withdrawn first, followed by a tapered withdrawal of prednisone, and the symptoms and signs of relapse occurred soon after final withdrawal of prednisone. The clinical relapses correlated with radiological reappearance of enhancing lesions.
Contrast enhancing lesions on MRI responded promptly to treatment. There was a complete or significant resolution of contrast enhancement in all cases on follow-up imaging shortly after pulse IVMP therapy. Follow-up scans later in the course of the disease demonstrated significant atrophy of the cerebellum and brachium pontis in all cases (figure 4). Cerebellar atrophy was evident despite good clinical response to treatment. Generalised cerebral atrophy was also evident on follow-up scans of case Nos 1, 2 and 5 (figure 5), and cognitive dysfunction was a component of the illness in the two most severely atrophic cases (Nos 1 and 5).
Axial T1 weighted MRI of the cerebellum without contrast early in the course of the disease (row A) and at the last radiological follow-up (row B), showing cerebellar atrophy in all cases.
MRI brain (axial T1 weighted images without gadolinium) at the level of the Sylvian fissures early in the course of the disease (row A) and at the last radiological follow-up (row B), showing generalised cortical atrophy in three cases.
Discussion
The case series presented here concurs with the recent report by Pittock et al1 that CLIPPERS represents a discrete CNS inflammatory disorder with lesions that are characteristic in distribution, morphology and histology; the clinical findings can be predicted by the lesion locations and the disease responds to corticosteroid based immunosuppression. This series presents a further five biopsy proven cases that exhibited extensive histiocytic as well as lymphocytic involvement and also demonstrated neuro-axonal injury that may explain the atrophy and disability that develops over time, particularly if untreated. The lesions have characteristic enhancement without marked surrounding T2 weighted MRI hyperintensity, and are most marked in the brachium pontis, pons or cerebellum rather than being always predominant in the pons. A review of the radiological images in the original description1 also suggests that the most marked involvement is not always in the pons. We suggest that the name of the syndrome should be adjusted to chronic lymphocytic inflammation with pontocerebellar perivascular enhancement responsive to steroids to more accurately reflect the distribution of the radiological findings.
Histopathology from all subjects was thoroughly assessed to exclude other possible diagnoses which we have seen resemble the radiological appearance of CLIPPERS. Furthermore, it has been observed that the radiological appearance, while suggestive, is not specific enough to obviate biopsy.3 Isolated CNS vasculitis, and in particular the lymphocytic variant, shares some similarities with the pathology seen in CLIPPERS. However, typical pathology includes vessel wall destruction with fibrinoid necrosis and leukocytoclasia,4 5 which were not present in any biopsy specimen. Primary CNS lymphoma or lymphomatoid granulomatosis generally has perivascular large atypical B cells and an associated reactive T cell infiltrate.6 7 When present in the biopsies in this series, B cells did not show atypical morphology, evidence of Epstein–Barr infection on Epstein–Barr virus early RNA-1 in situ hybridisation or evidence of light chain restriction on immunohistochemistry. There were small areas of demyelination present in three of the cases. Demyelination was associated with the most intense inflammatory infiltrates, was not restricted to perivascular areas and was considered to be a secondary phenomenon. Similarly, in areas of intense perivascular inflammation, astrocytes demonstrated fragmentation and beading of processes, again most likely a result of secondary cellular injury. The absence of Langerhans cells in the pathology specimens distinguishes this from LCH, which can also involve the pons, cerebellum and brachium pontis, and is associated with cerebellar atrophy.8
Neurosarcoidosis must be carefully considered, and none of the biopsies demonstrated neuropathological changes diagnostic of this disease, in particular a predominance of discrete, lymphocyte rimmed granulomas with giant cells.9 None of the five cases met the previously published criteria for definite or probable neurosarcoidosis.10 Sjögren's syndrome was also considered but over the course of their illness, none of the patients demonstrated specific clinical features suggesting this diagnosis, and the radiological features were not consistent with previously reported cases of encephalitis associated with the syndrome.11 12 However, in three of the five cases there was subtle paraclinical evidence of systemic involvement. Two of the cases showed a lymphocytic infiltrate involving other organs but without specific histopathological features of sarcoidosis or Sjögren's syndrome. Two cases had positive autoantibodies, and there was abnormal calcium homeostasis that responded to immunosuppressive treatment in a third case, and it is noted that serum autoantibodies were also detected in the original CLIPPERS series.1 This suggests that, while clinically CLIPPERS appears to be a disease isolated to the CNS, there are subtle systemic manifestations of an autoimmune nature.
All of the patients described in the original series,1 this series and additional reported cases13–16 developed a subacute and progressive neurological syndrome dominated by gait ataxia and dysarthria, but with varying additional deficits, including cranial nerve lesions and long tract signs. Despite this partial clinical heterogeneity, the characteristic imaging and histopathological assessments appear to confirm the same neuroinflammatory process.1 Due to the varied clinical presentation of CLIPPERS and the potential for diagnostic confusion,17 18 we have highlighted the core clinical, radiological and histopathological findings of the syndrome in box 1.
Core features of CLIPPERS*
Clinical
Subacute progressive ataxia and dysarthria.†
Radiological
(a) Numerous punctate or nodular enhancing lesions bilaterally in at least two of the three following anatomical locations: pons, brachium pontis, cerebellum.
(b) Individual radiological lesions are small but may coalesce to form larger lesions (mass effect has not been seen and to date suggests an alternative diagnosis).
(c) Lesions may occur in the spinal cord, basal ganglia or cerebral white matter but should be of decreasing density with increasing distance from the hindbrain.
(d) Absence of the following radiological features
(i) Restricted diffusion on diffusion weighted imaging
(ii) Marked hyperintensity on T2 weighted images
(iii) Abnormal cerebral angiography
Corticosteroid responsiveness
Prompt and significant clinical and radiological response to corticosteroids.
Histopathological
(a) White matter perivascular lymphohistiocytic infiltrate with or without parenchymal extension.
(b) Infiltrate contains predominantly CD3+ and CD4+ lymphocytes.
(c) Absence of the following histopathological characteristics:
(i) Monoclonal or atypical lymphocyte population
(ii) Necrotising granulomas or giant cells
(iii) Histological features of vasculitis (destruction of the vessel wall, fibrinoid necrosis, leukocytoclasia, fibrin thrombi).
*Differential diagnoses, including neurosarcoidosis, Sjögren's syndrome, neuro-Behçet's disease, vasculitis and lymphoma should be excluded.
†A range of other clinical features referable to brainstem pathology plus cognitive and spinal features occur in some patients.
Functional deterioration occurred in all cases, and at the end of the follow-up period persisted to varying degrees, despite treatment (figure 1). Even with improvement of the inflammatory process, as documented by resolution of contrast enhancement on follow-up MRI, significant cerebellar atrophy developed. Four cases were documented to have cognitive deficits, over and above a pseudobulbar affect, that were not readily explained by involvement of the cerebellum and brainstem alone, although one (case No 3) had a significant drug abuse history. Two of the cases with cognitive dysfunction developed cerebral atrophy over time (case Nos 1 and 5; figure 5). Neuro-axonal injury was observed in a number of biopsy specimens, and this is likely to have contributed to the pontocerebellar and cerebral atrophy seen. The most intense inflammatory activity documented in this syndrome has been evident in the brainstem as described, but it is likely that the disease is more diffuse with widespread microscopic involvement of other areas of the brain. This is reinforced by the recent description of cerebral lesions in a case of CLIPPERS.15
The cases in this series were typical of CLIPPERS in that they demonstrated a rapid clinical response to corticosteroids. In addition, there was prompt improvement of the imaging correlates of the disease. However, recurrence of disease occurred following withdrawal of corticosteroid, or when a pulse of IVMP was given in isolation. In the cases that progressed off therapy, intravenous cyclophosphamide was generally effective at improving the clinical syndrome although that did not protect against subsequent relapse off treatment. Various additional immunsuppressive agents combined with prednisone appeared effective at maintaining clinical improvements after pulse IVMP or cyclophosphamide therapy. We cannot, however, establish an independent benefit of steroid sparing therapies without concurrent corticosteroid. From this series, it appeared that early and vigorous immunosuppressive treatment with continuing maintenance immunosuppression resulted in the best long term functional outcome. This is particularly important given the development of atrophy in CLIPPERS and the potential that fixed neurological deficits are the clinical correlate of this atrophy. We suggest that treatment be initiated as soon as practical after the diagnosis has been made, and that it is initiated with high dose IVMP. Subsequent chronic immunosuppression with oral corticosteroids is required to maintain remission, and the addition of a steroid sparing immunosuppressive agent is suggested given the chronicity of the disease. Given the histiocytic pathology, the use of tumour necrosis factor inhibiting drugs such as infliximab should also be explored in CLIPPERS.
Conclusion
This report contributes to the clinical understanding of the newly described neuroinflammatory disease CLIPPERS. The typical presentation of a subacute cerebellar and brainstem syndrome and suggestive neuroimaging pattern has been confirmed in this report. Further detail about the histopathological characteristics of the disease has been provided. Additionally, more widespread cerebral involvement with cognitive dysfunction and cerebral atrophy has been added to the clinical phenotype. Despite its steroid responsiveness, CLIPPERS does not represent a benign condition. No patient fully recovered, and without chronic immunosuppression the disease may relapse. Marked cerebellar atrophy accompanies or follows disease activity, as measured by clinical and MRI parameters, and this may contribute to long term neurological morbidity. While the exact pathophysiological mechanisms of the disease are unclear, there appears to be clinical overlap with other autoimmune brainstem predominant encephalitides. The neurologist should be aware of the condition and the typical manifestations as prompt and appropriate treatment may limit permanent neurological sequelae.
References
Footnotes
See Editorial commentary, p 2
Linked article 301626.
Competing interests None.
Patient consent Detail has been removed from these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.
Ethics approval This study was approved by University of Sydney Human Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Editorial commentary