We transfected HEK293T cells to express DNER. A, Immunostaining with patient cerebrospinal fluid (CSF) (1:20). B, Immunostaining with a goat DNER antibody. C, Merged images show the patient CSF in red, the goat antibody in green, and nuclei stained blue with 4′,6-diamidino-2-phenylindole (DAPI). The patient’s CSF reacts strongly with DNER-expressing cells.
Cultured Hela cells were transiently transfected to express DNER coupled to an enhanced green fluorescent protein tag (EGFP) (green). A, Immunostaining of fixed cells. B, Immunostaining of live cells. Immunostaining consisted of human serum (red) from patients with paraneoplastic cerebellar degeneration, anti-Tr, and Hodgkin lymphoma (serum samples 1-5) and from a patient with neuropathy and anti-Tr (serum sample 6). In merged images, nuclei are stained blue with 4′,6-diamidino-2-phenylindole (DAPI). Staining of live cells produced improved signal-to-noise compared with fixed cells, showing that serum samples 1 through 5 label surface epitopes of DNER, whereas serum 6 showed negative findings. A panel of control serum samples (not shown) likewise showed negative results. DNER appeared to cluster on the surface of cells.
A, Cultured embryonic rat hippocampal neuron immunostained with patient cerebrospinal fluid (CSF) (1:20). B, The hippocampal neuron immunostained with a goat DNER antibody (1:200). C, Merged images show the patient CSF staining in green and the goat DNER antibody in red to demonstrate colocalization.
Greene M, Lai Y, Baella N, Dalmau J, Lancaster E. Antibodies to Delta/Notch-like Epidermal Growth Factor–Related Receptor in Patients With Anti-Tr, Paraneoplastic Cerebellar Degeneration, and Hodgkin Lymphoma. JAMA Neurol. 2014;71(8):1003-1008. doi:10.1001/jamaneurol.2014.999
The anti-Tr immune response is associated with paraneoplastic cerebellar degeneration and Hodgkin lymphoma (HL). One case series has reported that the Delta/notch-like epidermal growth factor–related receptor (DNER) is the actual target for anti-Tr antibodies, but this result has not been replicated.
To describe a patient with anti-Tr and confirm that DNER is the autoantigen for a series of patients with anti-Tr.
Design, Setting, and Participants
Observational study and analysis of biological samples for antibodies to DNER at the hospital of the University of Pennsylvania. We examined a cerebrospinal fluid sample from 1 patient with anti-Tr and serum and/or cerebrospinal fluid samples from 5 other patients with anti-Tr.
Transfection of HEK293T and Hela cells to express DNER coupled to an enhanced green fluorescent protein tag using a plasmid previously used to detect human DNER antibodies.
A man in his 30s with paraneoplastic cerebellar degeneration and anti-Tr underwent treatment with corticosteroids and intravenous immunoglobulin, resulting in clinical improvement before chemotherapy. Despite close oncologic follow-up, a biopsy, positron emission tomography, and computed tomography, he was not diagnosed as having HL until 6 months after symptom onset. The cerebrospinal fluid sample from this patient reacted with cells transfected to express DNER, as did cerebrospinal fluid and/or serum samples from 5 other patients with paraneoplastic cerebellar degeneration, HL, and anti-Tr. Only 4 of the 5 serum samples reacted to permeabilized cells enough to be distinguished from background, but all 5 serum samples convincingly labeled live cells, which had considerably less background. All 6 control serum samples and 1 serum sample from a patient previously diagnosed as having anti-Tr (but without HL or cerebellitis) did not recognize DNER.
Conclusions and Relevance
This case demonstrates the importance of testing for the anti-Tr immune response in patients with cerebellar degeneration. The strong association of anti-Tr with HL requires careful surveillance for this tumor. We also confirm that DNER is the target antigen of the anti-Tr immune response. Screening for DNER antibodies against living transfected cells may offer an improved signal-to-noise characteristic compared with immunostaining of fixed, permeabilized cells.
In 1976, Trotter et al1 described a patient with Hodgkin lymphoma (HL), subacute cerebellar degeneration, and antibodies that stained cerebellar Purkinje neurons in a characteristic pattern. These findings were further described in a case series in 1992 by Hammack et al.2 This characteristic staining pattern was termed anti-Tr after the lead investigator on the initial report3 and was subsequently identified in other patients with paraneoplastic cerebellar degeneration, about 90% of whom had HL.4,5 Patients with paraneoplastic cerebellar degeneration typically have progressive nystagmus, limb ataxia, dysarthria, and gait ataxia. Magnetic resonance imaging of the brain may show signs of cerebellar inflammation, and cerebrospinal fluid (CSF) samples may show mild pleocytosis and/or elevated protein levels.6 Patients may improve with immunotherapy and/or therapy directed against the tumor, but even treated patients typically have permanent cerebellar dysfunction. Postmortem studies show a loss of cerebellar Purkinje neurons.4
Recently, the Delta/notch-like epidermal growth factor–related receptor (DNER)7 was identified as the target of anti-Tr.8 Serum samples from 12 patients with anti-Tr (but only 1 of 246 control individuals) bound to cells expressing DNER. Further, immunoabsorption of patient serum samples with DNER abolished cerebellar neuron reactivity, and knockdown of DNER in neurons prevented recognition by patient serum samples. These experiments provided compelling evidence that DNER is the true target of the anti-Tr response, although this result has not been replicated until now.
Studies with human specimens were approved by the institutional review board at the University of Pennsylvania under protocol 819113. Written informed consent was obtained from the participants.
We separately grew HEK293T and Hela cells to near confluence on 12-mm glass coverslips. Cells were transiently transfected to express DNER coupled to an enhanced green fluorescent protein tag using a plasmid previously used to detect human DNER antibodies.8 After allowing 24 hours for expression, cells were fixed with 4% paraformaldehyde for 5 minutes, washed 3 times with phosphate-buffered saline solution (PBS), permeabilized with 0.3% Triton X-100 (Sigma-Aldrich Corp) for 5 minutes in PBS, washed 3 times with PBS, and blocked for 1 hour in 5% normal goat serum in PBS. Patient serum (diluted 1:200 in blocking solution) or CSF (diluted 1:20 in blocking solution) samples were applied for 1 hour at room temperature. Coverslips were washed 3 times with PBS, then stained with an appropriate secondary antibody (tetramethylrhodamine [TRITC]-conjugated antihuman antibody; Molecular Probes) for 1 hour at room temperature. Coverslips were stained with 4′,6-diamidino-2-phenylindole (DAPI) to visualize cell nuclei and mounted, then imaged using a fluorescent microscope. For immunostaining of live, unfixed cells, these methods were modified to apply the human serum (1:200) or CSF (1:20) samples to the cultured cells in culture media in the incubator at 37°C before fixation.
Cultured rat embryonic neurons were generated as described previously9 and grown in culture for 10 to 21 days. Cells were fixed with 4% paraformaldehyde for 5 minutes, washed 3 times with PBS, permeabilized with the 0.3% detergent solution for 5 minutes, washed 3 times with PBS, and then placed in blocking solution (4% bovine serum albumin in PBS). Human CSF (1:20) and a goat antibody to DNER (R&D Systems) in blocking solution were applied for 1 hour at room temperature. Coverslips were washed 3 times with PBS and then treated with appropriate fluorescent secondary antibodies (as described above) for 1 hour at room temperature. Coverslips were washed 3 times with water and mounted on slides.
A 36-year-old man with no medical history presented with worsening dysarthria and ataxia for 4 days. He initially woke up with the sensations of dizziness, the room spinning, and unsteadiness associated with nausea. In retrospect he also noted double vision for the 3 weeks before admission, which improved when he closed one eye. About 1 week before admission, he had fever and chills at home, with a temperature of 37.3°C. Three days before admission he noticed that his speech was slurred, and he felt progressively more tired. His dizziness progressed to feeling severely ill with any movement, and his dysarthria worsened.
On the day of admission he stumbled down some stairs and hit his head with minor injury and no loss of consciousness. He denied numbness, weakness, dysphagia, photophobia, or phonophobia. Anisocoria was noted on admission, but the patient was not previously aware of this.
During his admission, magnetic resonance imaging initially showed subtle enhancement of the leptomeninges of the cerebellum, but repeated imaging did not demonstrate this. No parenchymal lesions were apparent on magnetic resonance imaging. However, magnetic resonance imaging of the cervical spine demonstrated enlarged cervical lymph nodes on the right side only. He underwent fine-needle aspiration biopsy of a right cervical lymph node, results of which demonstrated a few large atypical mononuclear cells present in a background of mixed lymphocytes and plasma cells, and a lymphoproliferative disorder was suspected. An excisional biopsy of a right cervical lymph node was performed, results of which were nonspecific. However, rare scattered cells were found that raised a concern for classic HL, without cytologically atypical or frequent enough findings to justify that diagnosis. Results of his initial lumbar puncture were markedly inflammatory, with white blood cell counts of 38/μL and 110/μL in tubes 1 and 4 (to convert to ×109 per liter, multiply by 0.001), with a protein level of 69 mg/dL. This finding subsequently resolved with lowering of the protein level to 36 mg/dL and a white blood cell count of 10/μL. Results of CSF flow cytometry and cytologic studies were negative for malignant disease.
During the first 2 weeks of his hospitalization, ataxia worsened to such a degree that he could not sit up or use his limbs meaningfully. His dysarthria worsened to the point where he could not generate any comprehensible speech. He lost the ability to tolerate oral intake, and he received total parenteral nutrition. He developed lethargy and encephalopathy, and only inconsistently followed simple commands. Two weeks after admission he was treated for presumed autoimmune/paraneoplastic encephalitis with 5 days of intravenous (IV) methylprednisolone sodium succinate (Solu-Medrol) at 1 g/d and then with 5 days of IV immunoglobulin (0.4 g/kg per day). He continued to receive IV methylprednisolone sodium succinate, 40 mg/d, and switched to prednisone, 60 mg/d, when he could resume oral intake. His paraneoplastic antibody studies returned results positive at 1:256 for anti-Tr and negative for antinuclear antibody types 1, 2, and 3; antiglial/neuronal nuclear antibody type 1; Purkinje cell cytoplasmic antibody types 1 and 2; amphiphysin; and collapsin response mediator protein-5 antibodies (Mayo Clinic Laboratories, Rochester, Minnesota). Owing to the strong association of anti-Tr with HL,4,5 the hematology/oncology department was consulted. He continued to receive prednisone, 60 mg/d, and monthly IV immunoglobulin, 2 g/kg. Approximately 1 week after the initial IV immunoglobulin treatment, his limb ataxia and dysarthria began to improve. He was discharged to acute rehabilitation, where he resumed oral intake and total parenteral nutrition was discontinued. At discharge he was confined to a wheelchair, but improved during the course of months to ambulate with a walker and make himself food.
He underwent 2 positron emission tomography and computed tomography scans after discharge that showed only mildly increased uptake in right cervical lymph nodes, which is less than is typical for HL. He underwent a second biopsy of a right cervical lymph node 6 months after symptom onset. This biopsy finding was diagnostic of classic HL of the lymphocyte-rich subtype. He was treated with 2 cycles of chemotherapy consisting of doxorubicin hydrochloride, bleomycin sulfate, vinblastine sulfate, and dacarbazine, followed by 10 fractions of radiotherapy (2000 cGy [to convert to rads, multiply by 1]) applied to the right side of his neck with sparing of the cerebellum), and is still in the process of weaning from corticosteroid therapy (he is currently receiving prednisone, 10 mg/d) and planning to discontinue IV immunoglobulin therapy eventually. At last follow-up, 11 months after symptoms began, he had dysarthria but was intelligible, could use his hands for most common tasks, including sending text messages and preparing food, and used a walker.
The CSF sample from the patient at diagnosis reacted with cells transfected to express DNER. A subsequent CSF sample from this patient taken at 6 months after diagnosis did not recognize DNER, suggesting that the immune response was transient. Results of repeated commercial testing for anti-Tr (Mayo Clinic Laboratories) were also negative for this second CSF sample. Furthermore, his serum sample had a negative finding on commercial testing for Purkinje cell or neuronal nuclear antibodies at 4 months after symptom onset (ARUP Laboratories).
We also acquired tissue-banked serum (n = 5) and CSF (n = 1) samples from 5 other patients with HL and cerebellitis that were previously described as having anti-Tr reactivity. The clinical characteristics of these patients are shown in the Table. The CSF sample bound to cells transfected to express DNER (Figure 1). The serum samples produced considerable background staining of fixed, permeabilized cells; under our conditions, only 4 of the serum samples tested in this fashion had clearly positive results. Testing serum samples against live cells transfected to express DNER produced far less background staining and a superior signal-to-noise characteristic. Under these conditions, all 5 serum samples selectively labeled DNER-expressing cells; DNER appears to cluster into small puncta on the cell membrane (Figure 2). Serum samples from 6 controls did not react with live or fixed/permeabilized DNER-expressing cells. We also analyzed a serum sample from a patient who reportedly had anti-Tr (based on the pattern of neuronal reactivity), but who had peripheral neuropathy instead of a cerebellar syndrome. This patient had no known tumor. Serum samples from this patient did not react with fixed/permeabilized or live cells transfected to express DNER. Following the methods of de Graaff et al,8 we also demonstrated that the 2 CSF samples reacted with cultured rat embryonic hippocampal neurons in a punctate fashion; this reactivity colocalized very strongly with the reactivity of a commercial goat antibody to DNER (Figure 3).
We describe a patient with paraneoplastic cerebellar degeneration and anti-Tr. His neurologic syndrome preceded the diagnosis of HL by 6 months. This case highlights the following important aspects of this disorder: (1) the strong associations among anti-Tr/anti-DNER, paraneoplastic cerebellar degeneration, and HL; (2) the importance of thorough and ongoing evaluation for HL in these patients; (3) the potential for significant clinical improvement with immunotherapy; and (4) the transient nature of the immune response in some patients.
We also confirm that DNER is the target antigen of the anti-Tr immune response. Our experience in screening a series of patients who were reported to have anti-Tr also suggests that reactivity to DNER may be more specific than analysis for the pattern of reactivity to cerebellar neurons. Testing the reactivity of serum samples against live cells also produced less background staining than testing against fixed cells and may be useful for diagnosis.
Testing of our patient’s serum samples yielded negative findings for the Tr antibody (ARUP Laboratories) 4 months after symptom onset, as did testing of his CSF sample (Mayo Clinic Laboratories) 6 months after symptom onset but before any chemotherapy or radiotherapy was given for his HL. The anti-DNER antibody response might be monophasic in nature, and intrathecal synthesis of his paraneoplastic antibodies might have ceased, or his therapy with corticosteroids and IV immunoglobulin may have been sufficient to suppress intrathecal synthesis of the anti-DNER antibody. Some patients with anti-Tr have been reported to relapse, but whether a negative CSF antibody response affects the risk for relapse or further neurologic deterioration remains unknown.
Delta/notch-like epidermal growth factor–related receptor is a central nervous system–specific ligand for the Notch receptors.10 The 4 different Notch receptors in mammals (1-4) mediate numerous cell-signaling events and are important for cellular differentiation, growth, and cancer.11 On binding to its ligand, Notch is cleaved and the intracellular domain recruits coactivators to initiate transcription of Notch target genes.12 Delta/notch-like epidermal growth factor–related receptor is expressed by cerebellar Purkinje neurons and is essential for the normal development of Purkinje neurons and Bergmann glia, which express 1 or more Notch receptors.13,14 Bergmann glia are intimately associated with Purkinje neurons throughout development,15 so it is plausible that autoantibodies disrupting signaling between Bergmann glia and Purkinje neurons could have profound effects on cerebellar function. The first 2 epidermal growth factor domains of DNER are sufficient and necessary to bind Notch10; this region does not contain the immunodominant epitope or the glycosylation sites.8
Anti-DNER represents, to our knowledge, the only autoimmune disorder targeting a Notch ligand. Antibodies to other cell-surface receptors, such as the N-methyl-d-aspartate (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamate receptors, cause direct effects on the target proteins.16,17 We find it plausible that DNER antibodies act by disrupting DNER-Notch signaling between Purkinje neurons and Bergmann glia. Mice with genetic disruption of DNER have cerebellar abnormalities on a gross and microscopic level, including abnormal Purkinje neuron morphologic characteristics and innervation, but they still have Purkinje neurons.13 In contrast, autopsy of humans with anti-Tr may show profound loss of Purkinje neurons.4 Whether this finding represents a species difference or a difference between a congenital and an acquired disruption of DNER or whether the antibodies cause more than a simple loss of DNER function remains unclear. The exact pathogenic mechanisms of this disorder remain to be demonstrated.
Accepted for Publication: April 4, 2014.
Corresponding Author: Eric Lancaster, MD, PhD, Department of Neurology, University of Pennsylvania School of Medicine, 3 W Gates, 3400 Spruce St, Philadelphia, PA 19104 (firstname.lastname@example.org).
Published Online: June 16, 2014. doi:10.1001/jamaneurol.2014.999.
Author Contributions: Dr Lancaster had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Greene, Lai, Dalmau, Lancaster.
Acquisition, analysis, or interpretation of data: Greene, Lai, Baella, Lancaster.
Drafting of the manuscript: Greene, Lancaster.
Critical revision of the manuscript for important intellectual content: All authors.
Obtained funding: Lancaster.
Administrative, technical, or material support: Greene, Lai, Baella, Lancaster.
Study supervision: Dalmau, Lancaster.
Conflict of Interest Disclosures: Dr Dalmau has a research grant from Euroimmun and receives royalties from patents for the use of Ma2 and NMDAR as autoantibody tests. No other disclosures were reported.
Funding/Support: This study was supported by a Dana Foundation Neuroimmunology award (Dr Lancaster) and by grant K08 NS075142 from the National Institutes of Health.
Role of the Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank the patients who donated their medical information and CSF and serum samples for these studies. Steven S. Scherer, MD, PhD (University of Pennsylvania), provided many useful discussions and advice regarding this project. Esther de Graaff, PhD, Peter Sillevis-Smitt, MD, PhD, and colleagues (Erasmus Medical Center, Rotterdam, the Netherlands) shared their plasmids for expressing DNER, which were originally from Mineko Kengaku, PhD, and her group (Institute for Integrated Cell-Material Sciences, Kyoto University, Japan). No financial compensation was given for these services.