Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Intracellular and non-neuronal targets of voltage-gated potassium channel complex antibodies

Intracellular and non-neuronal targets of voltage-gated potassium channel complex antibodies Neuro-inflammation RESEARCH PAPER Intracellular and non-neuronal targets of voltage-gated potassium channel complex antibodies 1 1 1 2 Bethan Lang, Mateusz Makuch, Teresa Moloney, Inga Dettmann, 2 2 2 1 Swantje Mindorf, Christian Probst, Winfried Stoecker, Camilla Buckley, 3 1 4 2 Charles R Newton, M Isabel Leite, Paul Maddison, Lars Komorowski, 1 1 1 1 Jane Adcock, Angela Vincent, Patrick Waters, Sarosh R Irani ► Additional material is ABSTRACT Kv1.6), and was used to label these channels in the published online only. To view Objectives Autoantibodies against the extracellular radioimmunoassay which first identified VGKC please visit the journal online domains of the voltage-gated potassium channel (VGKC) complex autoantibodies in patients with neuromyo- (http:// dx. doi. org/ 10. 1136/ complex proteins, leucine-rich glioma-inactivated 1 (LGI1) tonia (NMT), and then in Morvan’s syndrome jnnp- 2016- 314758). 3 45 and contactin-associated protein-2 (CASPR2), are found (MoS), limbic encephalitis (LE) and faciobra- 1 67 Nuffield Department of Clinical in patients with limbic encephalitis, faciobrachial dystonic chial dystonic seizures (FBDS). These autoanti- Neurosciences, University of seizures, Morvan’s syndrome and neuromyotonia. bodies were assumed to be directed against the Kv1 Oxford, Oxford, UK However, in routine testing, VGKC complex antibodies subunits themselves, but subsequent studies showed Institute for Experimental without LGI1 or CASPR2 reactivities (double-negative) are that the Kv1 subunits were part of a multiprotein Immunology, Lubeck, Germany Department of Psychiatry, more common than LGI1 or CASPR2 specificities. neuronal complex that includes leucine-rich glioma University of Oxford, Oxford, UK Therefore, the target(s) and clinical associations of inactivated 1 (LGI1), contactin-associated protein 2 Department of Neurology, double-negative antibodies need to be determined. (CASPR2) and contactin-2. Almost all of the anti- Queen’s Medical Centre, Methods Sera (n=1131) from several clinically defined bodies from patients with LE, FBDS or MoS, and Nottingham, UK cohorts were tested for IgG radioimmunoprecipitation of some with NMT, are directed against the extracellu- radioiodinated α-dendrotoxin ( I-αDTX)-labelled VGKC lar domains of LGI1 or CASPR2, and the anti- Correspondence to Professor Sarosh R Irani, West complexes from mammalian brain extracts. Positive bodies often co-immunoprecipitate the Wing, Level 6, John Radcliffe samples were systematically tested for live hippocampal I-αDTX-labelled Kv1 subunits from brain Hospital, Oxford OX3 9DU, UK; 7–10 neuron reactivity, IgG precipitation of I-αDTX and extracts. Patients with LGI1 or CASPR2 anti- sarosh. irani@ ndcn. ox. ac. uk I-αDTX-labelled Kv1 subunits, and by cell-based assays bodies often respond very well to immunotherapies, which expressed Kv1 subunits, LGI1 and CASPR2. and their antibody levels broadly correlate with Received 25 August 2016 711 12 Revised 3 November 2016 Results VGKC complex antibodies were found in 162 of clinical status. Accepted 30 November 2016 1131 (14%) sera. 90 of these (56%) had antibodies Therefore, there has been an increasing interest in Published Online First targeting the extracellular domains of LGI1 or CASPR2. Of diagnosing these autoimmune neurological dis- 23 January 2017 13 14 the remaining 72 double-negative sera, 10 (14%) eases, leading to large numbers of requests for immunoprecipitated I-αDTX itself, and 27 (38%) bound testing in patients who are unlikely to have well- to solubilised co-expressed Kv1.1/1.2/1.6 subunits and/or defined autoimmune syndromes. This has generated Kv1.2 subunits alone, at levels proportionate to VGKC an increase in the number of patient serum IgGs complex antibody levels (r=0.57, p=0.0017). The sera with which precipitate the VGKC complex but lack LGI1 LGI1 and CASPR2 antibodies immunoprecipitated neither or CASPR2 reactivity (‘double-negative’ samples). preparation. None of the 27 Kv1-precipitating samples bound These double-negatives can account for up to 80% live hippocampal neurons or Kv1 extracellular domains, but of samples with positive VGKC complex antibodies 16 (59%) bound to permeabilised Kv1-expressing human in studies which most closely recapitulate clinical 15 16 embryonic kidney 293T cells. These intracellular Kv1 practice. Moreover, the clinical syndromes in antibodies mainly associated with non-immune disease the double-negative patients are diverse and include 17 18 aetiologies, poor longitudinal clinical–serological correlations patients with pain syndromes, status epilepticus, 19 20 and a limited immunotherapy response. acute and chronic epilepsies, inflammatory poly- Conclusions Double-negative VGKC complex antibodies radiculopathies, children with a variety of neuroin- are often directed against cytosolic epitopes of Kv1 subunits flammatory diseases, systemic and central nervous and occasionally against non-mammalian αDTX. These system (CNS)-directed infections, a few patients antibodies should no longer be classified as neuronal-surface with Creutzfeldt-Jakob disease, and up to 5% of antibodies. They consequently lack pathogenic potential and elderly clinic controls. This clinical heterogeneity do not in themselves support the use of immunotherapies. has questioned both the pathological relevance of the antibodies and the justification for immunother- 13 14 16 25 apies in these patients. Some studies have suggested that higher titres of INTRODUCTION To cite: Lang B, Makuch M, the double-negative VGKC complex antibodies Moloney T, et al. J Neurol Radioiodinated α-dendrotoxin ( I-αDTX) binds Neurosurg Psychiatry help to increase the likelihood of pathogen- to neuronal voltage-gated potassium channels 22 26 2017;88:353–361. icity. However, until now, the few available (VGKC) of the Shaker-family (Kv1.1, Kv1.2 and Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 Neuro-inflammation studies of double-negative patients have classified these patients formalin followed by pure acetone, using a 1:10 dilution of patient serum from coded vials, with unblinding after study by their clinical features and relied on arbitrary non-validated completion. diagnostic criteria, and the subjective retrospective response to 16 25 26 Commercial antibodies against the extracellular domain of immunotherapies. Here, to definitively determine Kv1.1 (Neuromab, 75/105), and intracellular domains of Kv1.1 whether double-negative VGKC complex antibodies have pathogenic potential, we explored the epitopes they bound, (Chemicon, AB9782), Kv1.2 (Millipore, AB5924 and Neuromab their titres and clinical associations across a large variety of 75/008) and Kv1.6 (Chemicon, AB5184 and Neuromab 75/012) clinical syndromes. were used for immunoprecipitation and CBA studies. Commercial antibodies against the extracellular domains of METHODS Kv1.2 or Kv1.6 were not available. Statistics were performed Patients studied using GraphPad Prism V.6, and individual tests are stated below. To assess the frequencies of VGKC complex antibodies in a large number of varied patient phenotypes, and include syn- RESULTS dromes reported to associate with double-negative samples, Antibodies against the VGKC complex 1131 sera were tested from nine groups, including those with: Overall, across the varied cohorts, 162 of 1131 (14%) patients (1) known LE, FBDS, MoS or NMT, LGI1 or CASPR2 anti- had VGKC complex antibodies, mainly from the groups with bodies and VGKC complex antibody levels >400 pM known positivity (figure 1A). Live CBAs showed that 90 of 162 (n=84); (2) a consecutive clinic cohort known to have VGKC (56%) patients had LGI1 or CASPR2 antibodies (4 with coexist- complex antibodies without LGI1/CASPR2 reactivities (n=27; ent contactin-2 antibodies). Eighty-four of these 90 (90%) detailed in online supplementary table S1, which included patients had LE, FBDS, MoS and NMT, and 80 of the 90 sera patients with encephalopathies (n=10), NMT (n=2), stiff (87%), those with higher titres, also had IgG antibodies that person syndrome (n=2), psychiatric conditions (n=6), isolated bound the surface of hippocampal neurons. Of the remaining amnesia (n=2), Parkinson’s disease dementia (n=1), 72 (44%) double-negative VGKC complex antibody-positive Guillain-Barre syndrome (n=1) and neuropathic pain (n=3)); samples, only one—from a patient with LE in the clinic cohort (3) adult-onset epilepsies (n=582); (4) infectious diseases (see online supplementary table S1)—bound to live hippocam- (n=107: herpes simplex virus encephalitis (n=29), varicella pal neurons, suggesting a possible novel surface antigen. Among zoster virus encephalitis (VZVE, n=20), measles encephalitis the cohorts (3–9) without known VGKC complex antibodies, (n=30) and malaria (n=28, 12 with cerebral involvement)); (5) the percentage of positives ranged between 0% and 4%, with dysautonomia (n=95); (6) Lambert-Eaton myasthenic syndrome the exception of the infectious group (19%) some of which had (n=45); (7) Hu-antibodies (n=78); (8) healthy smokers (n=38) very high titres (figure 1A). and (9) healthy laboratory controls (n=75). Approval for anti- body studies was from the Oxfordshire Regional Ethical Antibodies against I-αDTX Committee A (07/Q1604/28). One possibility was that the double-negative sera bound to I-αDTX which is used to radiolabel the VGKC complex. Laboratory techniques Indeed, 10 of the 72 (14%) double-negative samples immuno- VGKC complex antibodies were detected by a radioimmunoassay 125 precipitated very high levels of the I-αDTX itself (figure 1B), which uses I-αDTX (Perkin Elmer, USA) to label VGKC com- correlating broadly with the corresponding VGKC complex plexes from 2% digitonin-solubilised rabbit whole brain mem- 24 antibody titres (figure 1C, r=0.54; p=0.02, Spearman’s rank branes. In order to closely mimic these conditions, but detect correlation). All these 10 were found in patients with infectious antibodies exclusively against the αDTX-sensitive Kv1 subunits, diseases (malaria (n=4), cerebral malaria (n=4), VZVE (n=1) Kv1.1-tranfected, Kv1.2-tranfected and Kv1.6-transfected and measles encephalitis (n=1)). Similarly, three sera from a human embryonic kidney 293T (HEK) cells were used in place snake handler, with vocational exposure to αDTX, immunopre- of brain tissue to prepare the extracts. In other respects, the cipitated I-αDTX (figure 1B). The remaining 62 radioimmunoassays were identical. To see if results were con- 125 double-negative samples without I-αDTX reactivity were founded by antibodies binding the I-αDTX itself, the tissue/ carried forward to the next experiments. cell extracts were replaced by solubilisation buffer. In each case, 5 μL of patient serum was incubated with 50 μL brain extract, HEK cell extract or buffer overnight and precipitated with 50 μL Antibodies against the αDTX binding VGKC subunits antihuman immunoglobulin G (IgG; Binding Site). The cut-off detected in solution for positivity based on the mean plus three SDs of results from To test selectively for binding to the Kv1 subunits themselves, 20 healthy control sera was 100 pM for VGKC complex anti- under conditions of the VGKC complex antibody assay, bodies, 80 pM for Kv1 subunit antibodies and 137 pM for anti- I-αDTX-labelled digitonin extracts of Kv1-transfected HEK bodies against I-αDTX alone. cells were examined. Strikingly, none of the LGI1 or CASPR2 Culture and staining procedures for live hippocampal antibody-positive sera precipitated the Kv1 subunits (figure 2A). neurons, and for live cell-based assays (CBAs) to detect anti- In contrast, 27 of 62 (44%) double-negative samples bound in bodies against LGI1, CASPR2, contactin-2, Kv1.1, Kv1.2 and solution to I-αDTX-labelled Kv1.1/1.2/1.6 heteromers only 8 125 Kv1.6 were performed as described previously. To validate (n=6), I-αDTX-labelled Kv1.2 homomers only (n=6) or to CBA results, flow cytometry was performed with live both (n=15; figure 2A and see online supplementary figure Kv1-transfected HEK cells incubated with patient serum (1:20), S1A). Furthermore, their binding to Kv1s correlated well with and bound-IgG detected using a phycoerythrin-conjugated anti- their corresponding VGKC complex antibody levels (figure 2A, human IgG secondary antibody. Samples were analysed on a r=0.57, p=0.0017, Spearman’s rank correlation). Ten of these LSRII flow cytometer with FlowJo V.10.0.8 software. Fixed 27 samples (37%) had VGKC complex antibody levels over CBAs were performed (at Euroimmun AG, Lübeck, Germany) 400 pM. No increase in binding was observed with additional with Kv1-transfected HEK cells, after fixation with 1.8% co-transfection of postsynaptic density protein 95 (PSD95, see Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 354 Neuro-inflammation Figure 1 Detection of VGKC complex antibodies and antibodies to dendrotoxin. (A) VGKC complex antibodies were determined from 1131 samples, including those with known VGKC complex antibodies (both with (n=84) and without (n=27) LGI1 or CASPR2 reactivities), and unselected patients with adult-onset epilepsies, infectious diseases, autonomic syndromes, LEMS, Hu, healthy smokers and HC. Samples with LGI1 antibodies (n=69), CASPR2 antibodies (n=21) and all available samples with VGKC complex antibody levels above 100 pM and unknown antigenic targets (red; n=72) were carried forward to other assays. Dotted lines represent this cut-off and the 400 pM cut-off from a previous study; (B) 10 of the 72 samples with unknown antigens immunoprecipitated substantial quantities of of I-αDTX alone (dotted line at 137 pM represents the mean plus 125 125 three standard deviations from 20 HCs). Three serum samples from a snake handler (grey dots) also had antibodies to I-αDTX; (C) I-αDTX antibody levels correlated with their corresponding VGKC complex antibody levels (r=0.54, p=0.015, Spearman’s rank correlation). I-αDTX, radioiodinated α-dendrotoxin; CASPR2, contactin-associated protein-2; HC, healthy controls; Hu, Hu antibodies; LEMS, Lambert-Eaton myasthenic syndrome; LGI1, leucine-rich glioma-inactivated 1; VGKC, voltage-gated potassium channel. online supplementary figure S1B), a putative Kv1-clustering Antibodies bind the intracellular domains molecule. of αDTX-sensitive VGKCs These results prove the existence of double-negative VGKC complex antibodies which bind Kv1 subunits in solution, but not Kv1 extracellular epitopes. Therefore, HEK cells expressing Antibodies do not bind the extracellular domains of Kv1.1, Kv1.2 and Kv1.6 were fixed and permeabilised so that αDTX-sensitive VGKCs antibodies against intracellular epitopes could be detected. The Kv1 reactivities available in solution could include intracel- Commercial antibodies raised against intracellular sequences of lular or extracellular epitopes. To restrict detection to extracellu- all the αDTX-sensitive Kv1 subunits bound specifically to the lar epitopes, live Kv1-transfected CBAs were tested. I-αDTX appropriate fixed HEK cells (example in figure 2C and see surface-binding studies on live Kv1-transfected HEK cells (see online supplementary figure S2A,B), and their binding was abro- online supplementary figure S1C) and commercial antibodies to gated after absorption of the commercial antibody with the the extracellular domain of Kv1.1 (figure 2B) confirmed immunising cytosolic Kv1 subunit peptide (shown for Kv1.2, adequate surface expression of Kv1.1, Kv1.2 and Kv1.6. see online supplementary figure S2A). This confirmed accessibil- However, none of the double-negative sera bound to live HEK ity of antibodies to intracellular Kv1-subunit epitopes. cells transfected with individual Kv1.1, Kv1.2 or Kv1.6, or all Subsequently, 175 coded sera were tested for binding to the three Kv1 subunits (representative example in figure 2B). To fixed permeabilised Kv1-expressing cells. These included first ensure maximal sensitivity, these negative results were con- samples of all double-negative patients without αDTX reactivity firmed using flow cytometry on the live HEK cells (n=62), 6 with αDTX antibodies, 57 sequential samples from co-transfected with Kv1.1, Kv1.2 and Kv1.6: there was no evi- the 27 patients with Kv1 antibodies demonstrated in solution dence of surface binding in the 16 sera from figure 2A with the (from figure 2A), patients with known LGI1 and CASPR2 anti- highest Kv1 antibody radioimmunoassay values (see online bodies (n=20), and disease and healthy controls without VGKC supplementary figures S1D–F). Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 355 Neuro-inflammation Figure 2 Kv1 antibodies target intracellular epitopes. (A) Twenty-seven of the remaining 62 patients with unknown VGKC complex antigenic 125 125 targets precipitated either I-αDTX-labelled Kv1.1/Kv1.2/Kv1.6 co-transfected HEK cell extracts (red circles) or I-αDTX-labelled Kv1.2-transfected HEK cell extracts (red circles with black outline). No sera with LGI1 or CASPR2 antibodies showed positive results. HEK cells transfected with Kv1.1 alone or Kv1.6 alone did not bind I-αDTX in solution; (B) a commercial antibody to the extracellular domain of Kv1.1 (anti-Kv1.1e) labelled the cell surface of live HEK cells co-transfected with Kv1.1, Kv1.2 and Kv1.6 (and enhanced green fluorescent protein (EGFP)). No patient antibodies (n=175, including the 62 double-negative samples without αDTX reactivity) showed similar binding to these live cells or live cells expressing only one of these subunits; (C) binding to fixed Kv1-transfected HEK cells was seen using serum samples which precipitated Kv1s from solution. This co-localised with binding of commercial antibodies against the intracellular domain of Kv1.2 (anti-Kv1.2). Examples for Kv1.2 and Kv1.6 are shown in online supplementary figure S1C. Scale bar=10 microns. I-αDTX, radioiodinated α-dendrotoxin; CASPR2, contactin-associated protein-2; HEK, human embryonic kidney 293T; LGI1, leucine-rich glioma-inactivated 1; VGKC, voltage-gated potassium channel. complex antibodies (n=30). Binding was observed in 41 of 175 no peak age at onset (figure 4B). Only 5 (19%) had classical (23%) samples (examples in figure 2C and online supplemen- autoimmune syndromes (paraneoplastic, NMT or LE; tary figure S2A, B): all of these were from the group of 27 figure 4C). The other patients had symptomatic or idiopathic patients whose first sample immunoprecipitated Kv1.1/Kv1.2/ generalised epilepsies (n=5), neurodegenerative diseases (n=2, Kv1.6 subunits (from figure 2A, Mann-Whitney test, Parkinson’s disease dementia and Alzheimer’s disease), and 14 p<0.0001). The first available sample from 16 of these 27 presented with conditions of unknown aetiology including patients bound to Kv1.2 (n=9), Kv1.6 (n=3), Kv1.1 (n=1), cryptogenic epilepsies (n=9), neuropathic pain (n=2), chronic Kv1.1, Kv1.2 and Kv1.6 (n=2), or both Kv1.2 and Kv1.6 encephalopathy (n=1), dysautonomia (n=1) or spontaneously (n=1). These patient IgGs showed consistent co-localisation resolving amnesia (n=1). In addition, one healthy smoker had with the Kv1 commercial antibodies (see online supplementary intracellular Kv1 antibodies. figure S2B). Overall, of the 27 samples which precipitated Despite this serological and clinical heterogeneity, 6 of the 9 I-αDTX-labelled Kv1 subunits in solution, those which did (67%) patients with cryptogenic epilepsies had Kv1.2-specific not show fixed CBA positivity tended towards lower VGKC antibodies, both patients with neuropathic pain had Kv1.6 reac- complex levels (Mann-Whitney test, p=0.001; see online tivities, and both patients with small cell lung carcinoma had supplementary figure S3A). The overall flow of samples and antibodies directed against all three subunits (Kv1.1, Kv1.2 and results by cohort is described in figure 3, and the extracellular Kv1.6). and intracellular molecular reactivities of the VGKC complex Eleven of the 27 patients were administered immunotherapies antibodies are summarised in figure 4A. and only 3 (27%) showed a sustained clinical benefit. In add- ition, only 4 of the 19 (21%) patients with serial serum samples demonstrated a relationship between intracellular antibody Correlations between Kv1 antibodies, clinical features and levels and clinical outcome (LE, NMT and 2 with cryptogenic treatment responses epilepsies), while the remaining 15 demonstrated poor correla- Serological and clinical details of the 27 patients with intracellu- tions. This contrasts with the majority of patients with LGI1 or lar Kv1 antibodies are summarised in table 1. There were 16 CASPR2 antibodies (see online supplementary figure S3B–E). men and 11 women, with ages ranging from 18 to 85 years and Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 356 Neuro-inflammation Figure 3 Summary of the sequential flow of assays through the study. As shown in figure 1A, 1131 samples were initially tested for VGKC complex antibodies by RIA (VGKC complex RIA, A) and subsequently using LGI1 and CASPR2 antibody live CBAs (B), live neuronal cultures (C) and precipitation of αDTX (αDTX RIA, D). As detailed in figure 2, double-negative samples were then tested for binding to the extracellular domains of live Kv1-tranfected HEK cells (Kv1-live CBA), for immunoprecipitations of I-αDTX-labelled Kv1-transfected HEK cells (Kv1-HEK RIA, E) and for binding to fixed permeabilised Kv1-transfected HEK cells (Kv1-fixed CBA). Cohorts are defined in more detail in the Methods section and online supplementary table S1. I-αDTX, radioiodinated α-dendrotoxin; CASPR2, contactin-associated protein-2; CBA, cell-based assay; HEK, human embryonic kidney 293T; Hu, Hu antibodies; LEMS, Lambert-Eaton myasthenic syndrome; LGI1, leucine-rich glioma-inactivated 1; RIA, radioimmunoassay; VGKC, voltage-gated potassium channel. Interestingly, in all individual patients, the targeted Kv1 subunit important because they use a systematic biochemical approach to (s) and their levels relative to VGKC complex antibodies demonstrate conclusively that a proportion of VGKC complex remained constant over time, strongly suggesting that these two antibodies binds to intracellular VGKC epitopes, or to the assays were measuring the same antibody populations (see non-mammalian-expressed αDTX. Discovery of these important online supplementary figures S3 C–E). antigenic targets should influence ongoing clinical practice, and prompt re-evaluation of several reports describing the clinical associations of patients with double-negative VGKC complex DISCUSSION 15 17–19 26 33 antibodies. In this study, double-negative antibodies Autoantibodies directed against the extracellular domains of were observed in several non-autoimmune conditions including LGI1 and CASPR2 usually associate with distinctive highly variable central and peripheral nervous system syndromes immunotherapy-responsive syndromes. Indeed, clinical and accu- with limited responses to immunotherapies, and poor correla- mulating paraclinical data strongly suggest that they are directly 891129–31 tions were noted between clinical data and serial antibody levels. pathogenic. In contrast, concerns have been raised Taken together, these double-negative autoantibodies often have about the clinical relevance of the double-negative VGKC targets which are inaccessible or non-existent in vivo, and they complex antibodies that do not bind either of these proteins, par- appear to be associated with limited clinical relevance. ticularly as they can be found in a proportion of patients with dis- Overall, LGI1 and CASPR2 antibody CBAs conferred greater eases which are unlikely to be of autoimmune 13 14 16 22 25 clinical utility and better sensitivity and specificity than VGKC aetiology. Studies which have interpreted the clinical complex antibody testing in providing a diagnosis and rationale relevance of such antibody results are limited by the inevitable for immunotherapy. This should prompt clinicians to use LGI1 difficulties in defining what constitutes an autoimmune 16 22 25 32 and CASPR2 antibodies over VGKC complex antibodies as disease. Ultimately, the definition of autoimmune first-line testing for pathogenic antibody-associated neurological diseases relies on the demonstration of a pathogenic immunotherapy-responsive syndromes, and limit immunotherapy immune factor. Therefore, the findings described here are Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 357 Neuro-inflammation Figure 4 Molecular and clinical features associated with double-negative VGKC complex antibodies. (A) The illustration of study results demonstrates that the antibodies with pathogenic potential (blue and green) target the extracellular domains of LGI1 and CASPR2, respectively, whereas likely non-pathogenic antibodies (red) target the intracellular domain of Kv1 channels, especially Kv1.2, and the α-dendrotoxin molecule itself (yellow), which is not present in mammalian tissue. Other intracellular targets may include the Kv-β2 subunit (pink). (B) The patients with intracellular Kv1 antibodies had no clear peak age of onset, and (C) 12 showed varied, known diagnoses (*), and 15 had conditions of unknown aetiology, unlikely to be autoimmune. CASPR2, contactin-associated protein-2; HS, hippocampal sclerosis; LGI1, leucine-rich glioma-inactivated 1; VGKC, voltage-gated potassium channel. administration to patients with double-negative antibodies and However, since a few disease-relevant LGI1 and CASPR2 atypical clinical syndromes. antibody-positive samples had no detectable live hippocampal A striking finding was the discovery of antibodies against neuron binding, the absence of live neuronal binding should not αDTX itself both in 10 patients with CNS infections and in a necessarily imply the presence of intracellular reactivities. snake handler. αDTX is a polypeptide toxin found in dendroas- Indeed, it is most likely that not all surface neuronal proteins pis snake venoms and was essential for labelling the VGKC are expressed in a hippocampal cell culture system. 2 4 5 15 34 complex in the earliest studies. Some of these studies It is curious that the double-negative VGKC complex epitopes included control tests to avoid detecting αDTX antibodies, appear especially immunogenic: they are found after both but commercial VGKC complex antibody assays do not provide human exposure to porcine brain aerosols and murine nasal this specific control. The αDTX antibodies were almost exclu- immunisations with brain extracts, suggesting that these anti- sively in sera with high titres of VGKC complex antibodies. In bodies are easily induced. Collectively, alongside their presence contrast, the intracellular epitopes identified on the Kv1 subu- in a variety of epilepsies, neurodegenerative diseases and idio- nits themselves, particularly on Kv1.2, were found both at high pathic syndromes, these observations imply that double-negative and low VGKC complex antibody levels. Therefore, this study antibodies arise secondary to diverse pathologies, where the shows that the antigenic target—rather than the VGKC complex primary pathological insult may not be immune. Indeed, antibody level—aids prediction of potential pathogenicity. perhaps non-immune causes of neuronal destruction are suffi- This is consistent with the observation that while VGKC cient to release immunogenic Kv1 antigens and provoke auto- 16 22 26 complex antibody levels associated with LGI1 and CASPR2 antibody production. Interestingly, the immunotherapy reactivities are often high, they can be low or even response observed in 27% of patients is very similar to the 8 152535 36 undetectable. placebo response rates observed in several other neurological Furthermore, this is the first clear demonstration of diseases, and much lower than those reported in LGI1 or 7–911152935 double-negative VGKC complex antibodies which target intra- CASPR2 antibody-associated syndromes. However, cellular epitopes. This positive demonstration is important as we conclusions regarding treatment outcomes are only definitive and others have previously hypothesised their existence, mainly after blinded interventional studies. Also, the demographics of 13 14 22 25 inferring from the absence of live neuronal binding. the patients with Kv1 antibodies did not show any age or sex Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 358 Neuro-inflammation Neuro-inflammation Table 1 Clinical–serological details of patients with antibodies against the intracellular aspects of Kv1.1, Kv1.2 and/or Kv1.6 Serological results Age/ VGKC complex Kv1-HEK Kv1-fixed Disease Sex antibody (pM) RIA CBA Clinical features aetiology Outcome and treatments Summary of clinical–serological correlations 59/M 1346 + 1.2 only Cryptogenic TLE with amnesia/ Unknown No sustained response to three AEDs and Poor. Persistent VGKC-c Abs despite markedly varied SZ frequencies. depression CS Online supplementary figure S3C 49/F 767 + 1.2 only NMT with EMG confirmation Autoimmune Limited response to AEDs and CS Poor. Highest VGKC-c Abs during clinical remission 55/M 533 + 1.2 only NMT with EMG confirmation Autoimmune Symptoms over 8 years despite AEDs/AZA/ Good. Patient symptomatic with persistent VGKC-c Abs CS 77/M 470 + 1.2 only Chronic myelopathy/ Unknown Transient response to CS and PLEX Poor. High VGKC-c Abs despite symptom fluctuations over 3 years encephalopathy 21/M 461 + 1.2 only Cryptogenic probable frontal lobe Unknown Ongoing SZs despite two AEDs Poor. Online supplementary figure S3E epilepsy 50/M 448 + 1.2 only Isolated amnesia Unknown Spontaneous resolution over a few days Poor. Abs reduced over 12 months. Symptoms resolved at 7 days 28/F 278 + 1.2 only Cryptogenic TLE Unknown SZ freedom after first AED Poor. Persistent VGKC-c Abs during SZ freedom 62/M 236 + 1.2 only Cryptogenic TLE with amnesia Unknown SZs and amnesia at 5 years with 3 AEDs Poor. VGKC-c Abs disappeared while SZs were ongoing 18/M 141 + 1.2 only Limbic encephalitis Autoimmune Good response to CS and PLEX Good. VGKC-c Abs reduced and clinical improvement at 12 months 25/F 529 + 1.6 only Diffuse neuropathic pain and Unknown No response to opioids, CS or IVIG Poor. High VGKC-c Abs despite marked fluctuations in symptoms depression 30/F 117 + 1.6 only Idiopathic generalised epilepsy Genetic Good response to AEDs NA 68/M 240 + 1.6 only Alzheimer’s disease; one SZ Degenerative Ongoing fall in memory despite CS/IVIG Poor. Slight fall in VGKC-c Abs but marked reduction in memory over 4 years 34/M 577 + 1.2 and 1.6 Widespread neuropathic pain Unknown No response to CS, IVIG and PLEX Poor. Constant pain despite highly varied VGKC-c Abs. Online supplementary figure S3D 71/F 2489 + 1.1, 1.2 and NMT plus SCLC Paraneoplastic Palliative care only NA. 1.6 67/F 2120 + 1.1, 1.2 and LEMS plus Hu antibody Paraneoplastic Good response to CS Poor. Persistently high VGKC-c Abs over 5 years despite clinical 1.6 neuropathy and SCLC improvements 84/F 282 + 1.1 only Parkinson’s disease dementia Degenerative No response to CS Poor. VGKC-c Abs reduced over 12 months despite clinical worsening 58/M 304 + Negative Dysautonomia Unknown NA NA 37/M 253 + Negative TLE related to left HS Structural NA NA 59/M 236 + Negative Healthy smoker Healthy Not relevant Not relevant 48/F 223 + Negative TLE related to left HS Structural NA NA 85/F 205 + Negative Cryptogenic focal motor SZs Unknown SZ free on 1 AED. Amnesia and anxiety Poor. VGKC-c Abs disappeared over 1 year; SZ freedom at 2 years. benefited from CS/IVIG VGKC-c Abs returned at 4 years without SZs 33/F 182 + Negative Cryptogenic TLE Unknown SZ freedom after second AED Moderate. VGKC-c Abs reduced over 6 months and SZ freedom at 1 year 76/M 181 + Negative Cryptogenic epilepsy Unknown SZ freedom with 1 AED Poor. SZ freedom at 6 months; VGKC-c Abs sampled after 15 years 54/F 139 + Negative Epilepsy after childhood Structural NA NA meningitis 22/M 137 + Negative Cryptogenic epilepsy Unknown SZ freedom after 1 AED NA 54/M 131 + Negative Cryptogenic epilepsy Unknown SZ free at 1 year with 1 AED Moderate. SZ free at 1 year; VGKC-c Abs absent at 3 months 77/M 123 + Negative TLE secondary to CVA Structural Ongoing SZs at 4 years NA Patients grouped by the VGKC-c antibody levels, precipitation from Kv1.1/Kv1.2/Kv1.6-cotransfected HEK cell and Kv1.2-transfected HEK cell radioimmunoassays (Kv1-HEK RIA; denoted as positive (+) or negative (−)) and the Kv1 subunit expressed fixed CBAs (Kv1-fixed CBA). All these patients had negative results in live CBAs for antibodies against Kv1s, LGI1, CASPR2, contactin-2 and for binding to live hippocampal neurons. Two patients had SCLC and tumours were not found in the remaining patients. Ab, antibody; AEDs, antiepileptic drugs; AZA, azathioprine; CASPR2, contactin-associated protein-2; CS, corticosteroids; CBA, cell-based assay; CVA, cerebrovascular accident; EMG, electromyography; F, female; HEK, human embryonic kidney 293T; HS, hippocampal sclerosis; IVIG, intravenous immunoglobulins; LEMS, Lambert-Eaton myasthenic syndrome; LGI1, leucine-rich glioma-inactivated 1; M, male; NA, not available (only single serum sample obtained); PLEX, plasma exchange; SCLC, small cell lung carcinomas; SZ, seizure; TLE, temporal lobe epilepsy; VGKC, voltage-gated potassium channel; VGKC-c, VGKC complex. Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 359 Neuro-inflammation 2 Shillito P, Molenaar PC, Vincent A, et al. Acquired neuromyotonia: evidence for bias, quite different from the occurrence of LGI1 antibody LE autoantibodies directed against K+ channels of peripheral nerves. Ann Neurol in later years or the male predominance of CASPR2 anti- 1995;38:714–22. 8 9 11 29 bodies. Alongside the 5% rate of double-negative VGKC 3 Liguori R, Vincent A, Clover L, et al. Morvan’s syndrome: peripheral and central complex antibodies in healthy individuals, these observations nervous system and cardiac involvement with antibodies to voltage-gated potassium suggest that double-negative reactivities have very limited clin- channels. Brain 2001;124:2417–26. 4 Vincent A, Buckley C, Schott JM, et al. Potassium channel antibody-associated ical significance. encephalopathy: a potentially immunotherapy-responsive form of limbic encephalitis. On the other hand, we noted relationships between specific Brain 2004;127:701–12. Kv1 subunit reactivities and clinical syndromes, particularly 5 Thieben MJ, Lennon VA, Boeve BF, et al. Potentially reversible autoimmune limbic Kv1.2 antibodies in cryptogenic epilepsies, Kv1.6 antibodies in encephalitis with neuronal potassium channel antibody. Neurology 2004;62:1177–82. the two patients with neuropathic pain, and Kv1.1, Kv1.2 and 6 Irani SR, Buckley C, Vincent A, et al. Immunotherapy-responsive seizure-like Kv1.6 antibodies in paraneoplastic syndromes. These prelimin- episodes with potassium channel antibodies. Neurology 2008;71:1647–8. ary findings suggest that specific patterns of Kv1 reactivities may 7 Irani SR, Michell AW, Lang B, et al. Faciobrachial dystonic seizures precede Lgi1 be generated in response to certain pathologies. Indeed, a few antibody limbic encephalitis. Ann Neurol 2011;69:892–900. 8 Irani SR, Alexander S, Waters P, et al. Antibodies to Kv1 potassium patients with classical VGKC complex antibody-associated syn- channel-complex proteins leucine-rich, glioma inactivated 1 protein and dromes, such as NMT and LE, did show good antibody-clinical contactin-associated protein-2 in limbic encephalitis, Morvan’s syndrome and correlations, and in these patients the Kv1 antibodies may be acquired neuromyotonia. Brain 2010;133:2734–48. 38 39 markers of a coexistent pathogenic antibody. However, 9 Lai M, Huijbers MG, Lancaster E, et al. Investigation of LGI1 as the antigen in while many of the clinic cohort had an underlying immune limbic encephalitis previously attributed to potassium channels: a case series. Lancet Neurol 2010;9:776–85. basis to their neurological syndrome, ascertainment of patients 10 Irani SR, Pettingill P, Kleopa KA, et al. Morvan syndrome: clinical and serological attending a specialist autoimmune clinic may well carry this observations in 29 cases. Ann Neurol 2012;72:241–55. intrinsic bias. 11 Lancaster E, Huijbers MGM, Bar V, et al. Investigations of caspr2, an autoantigen of In summary, our results show that many double-negative encephalitis and neuromyotonia. Ann Neurol 2011;69:303–11. 12 Shin YW, Lee ST, Shin JW, et al. VGKC-complex/LGI1-antibody encephalitis: clinical VGKC complex antibodies lack pathogenic potential, and that manifestations and response to immunotherapy. J Neuroimmunol 2013;265:75–81. direct examination of LGI1 and CASPR2 antibodies is more Irani SR, Gelfand JM, Al-Diwani A, et al. Cell-surface central nervous system informative than first-line VGKC complex antibody autoantibodies: clinical relevance and emerging paradigms. Ann Neurol 13 14 25 35 36 testing. Also, direct LGI1 and CASPR2 antibody 2014;76:168–84. CBAs appear more sensitive than VGKC complex antibody or 14 Leypoldt F, Armangue T, Dalmau J. Autoimmune encephalopathies. Ann N Y Acad 25 35 36 Sci 2015;1338:94–114. live neuronal antibody determination. Therefore, direct 15 Klein CJ, Lennon VA, Aston PA, et al. Insights from LGI1 and CASPR2 potassium LGI1 and CASPR2 antibody assays are the preferred tests for channel complex autoantibody subtyping. JAMA Neurol 2013;70:229–6. clinical purposes. This approach should limit the use of unneces- 16 Paterson RW, Zandi MS, Armstrong R, et al. Clinical relevance of positive sary immunotherapies in double-negative patients. voltage-gated potassium channel (VGKC)-complex antibodies: experience from a tertiary referral centre. J Neurol Neurosurg Psychiatry 2014;85:625–30. Until all centres update their practices, treatment of 17 Klein CJ, Lennon VA, Aston PA, et al. Chronic pain as a manifestation of potassium double-negative cases should be limited to those with a high channel-complex autoimmunity. Neurology 2012;79:1136–44. pretest probability of an underlying autoimmune condition. 18 Suleiman J, Brenner T, Gill D, et al. VGKC antibodies in pediatric encephalitis However, for research purposes, it may be useful for further presenting with status epilepticus. Neurology 2011;76:1252–5. studies to identify the remaining, likely intracellular, 19 Brenner T, Sills GJ, Hart Y, et al. Prevalence of neurologic autoantibodies in cohorts of patients with new and established epilepsy. Epilepsia 2013;54:1028–35. double-negative VGKC complex antigens such as the Kv-β2 20 Toledano M, Britton JW, McKeon A, et al. Utility of an immunotherapy trial in subunit and PSD95, and explore whether they are useful bio- evaluating patients with presumed autoimmune epilepsy. Neurology markers of traditionally non-immune or immune neurological 2014;82:1578–86. syndromes. 21 Meeusen JW, Klein CJ, Pirko I, et al. Potassium channel complex autoimmunity induced by inhaled brain tissue aerosol. Ann Neurol 2012;71:417–26. Funding BL is supported by Epilepsy Research UK (ERUK; P1201); SRI is supported 22 Hacohen Y, Singh R, Rossi M, et al. Clinical relevance of voltage-gated potassium by a Wellcome Trust Intermediate Fellowship (104079/Z/14/Z), BMA Research channel-complex antibodies in children. Neurology 2015;85:967–75. Grants—Vera Down grant (2013) , the Fulbright UK-US commission and the MS 23 Pillai SC, Hacohen Y, Tantsis E, et al. Infectious and autoantibody-associated society. Research in the Neuroimmunology Laboratory is supported by the Oxford encephalitis: clinical features and long-term outcome. Pediatrics 2015;135: NIHR Biomedical Research Centre. PW and MIL are supported by the NHS National e974–84. Specialized Commissioning Group for Neuromyelitis optica, UK. 24 Rossi M, Mead S, Collinge J, et al. Neuronal antibodies in patients with suspected or confirmed sporadic Creutzfeldt-Jakob disease: table 1. J Neurol Neurosurg Competing interests AV, SRI, BL and PW are coapplicants and receive royalties Psychiatry 2015;86:692–4. on patent application WO/2010/046716 entitled ‘Neurological Autoimmune 25 van Sonderen A, Schreurs MWJ, de Bruijn MAAM, et al. The relevance of VGKC Disorders’. The patent has been licensed to Euroimmun AG for the development of positivity in the absence of LGI1 and Caspr2 antibodies. Neurology assays for LGI1 and other VGKC-complex antibodies. BL and SRI had full access to 2016;86:1692–9. all of the data in the study and take responsibility for the integrity of the data and 26 Olberg H, Haugen M, Storstein A, et al. Neurological manifestations related to level the accuracy of the data analysis. of voltage-gated potassium channel antibodies. J Neurol Neurosurg Psychiatry Ethics approval Oxfordshire Regional Ethical Committee A (RECA). 2013;84:941–3. 27 Horresh I, Poliak S, Grant S, et al. Multiple molecular interactions determine the Provenance and peer review Not commissioned; externally peer reviewed. clustering of Caspr2 and Kv1 channels in myelinated axons. J Neurosci Open Access This is an Open Access article distributed in accordance with the 2008;28:14213–22. terms of the Creative Commons Attribution (CC BY 4.0) license, which permits 28 Waters PJ, McKeon A, Leite MI, et al. Serologic diagnosis of NMO: a multicenter others to distribute, remix, adapt and build upon this work, for commercial use, comparison of aquaporin-4-IgG assays. Neurology 2012;78:665–71–discussion669. provided the original work is properly cited. See: http://creativecommons.org/licenses/ 29 Flanagan EP, Kotsenas AL, Britton JW, et al. Basal ganglia T1 hyperintensity in by/4.0/ LGI1-autoantibody faciobrachial dystonic seizures. Neurol Neuroimmunol Neuroinflamm 2015;2:e161–1. 30 Ohkawa T, Fukata Y, Yamasaki M, et al. Autoantibodies to epilepsy-related LGI1 in limbic encephalitis neutralize LGI1-ADAM22 interaction and reduce synaptic AMPA receptors. J Neurosci 2013;33:18161–74. REFERENCES 31 Lalic T, Pettingill P, Vincent A, et al. Human limbic encephalitis serum enhances 1 Shamotienko OG, Parcej DN, Dolly JO. Subunit combinations defined for K+ hippocampal mossy fiber-CA3 pyramidal cell synaptic transmission. Epilepsia channel Kv1 subtypes in synaptic membranes from bovine brain. Biochemistry 2011;52:121–31. 1997;36:8195–201. Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 360 Neuro-inflammation 32 Graus F, Titulaer MJ, Balu R, et al. A clinical approach to diagnosis of autoimmune 36 Becker EBE, Zuliani L, Pettingill R, et al. Contactin-associated protein-2 antibodies encephalitis. Lancet Neurol 2016;15:391–404. in non-paraneoplastic cerebellar ataxia. J Neurol Neurosurg Psychiatry 33 McKnight K, Jiang Y, Hart Y, et al. Serum antibodies in epilepsy and 2012;83:437–40. seizure-associated disorders. Neurology 2005;65:1730–6. 37 Bittar C, Nascimento OJM. Placebo and nocebo effects in the neurological practice. 34 Hart IK, Waters C, Vincent A, et al. Autoantibodies detected to expressed K+ Arq Neuro Psiquiatr 2015;73:58–63. channels are implicated in neuromyotonia. Ann Neurol 1997;41:238–46. 38 Boronat A, Sabater L, Saiz A, et al. GABA(B) receptor antibodies in limbic encephalitis 35 Irani SR, Stagg CJ, Schott JM, et al. Faciobrachial dystonic seizures: the influence of and anti-GAD-associated neurologic disorders. Neurology 2011;76:795–800. immunotherapy on seizure control and prevention of cognitive impairment in a 39 Pettingill P, Kramer HB, Coebergh JA, et al. Antibodies to GABAA receptor α1 and broadening phenotype. Brain 2013;136:3151–62. γ2 subunits: clinical and serologic characterization. Neurology 2015;84:1233–41. Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Neurology Neurosurgery & Psychiatry British Medical Journal

Intracellular and non-neuronal targets of voltage-gated potassium channel complex antibodies

Download PDF
9 pages

Loading next page...
 
/lp/british-medical-journal/intracellular-and-non-neuronal-targets-of-voltage-gated-potassium-CjIK6hSP1H

References (48)

Publisher
British Medical Journal
Copyright
Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://www.bmj.com/company/products-services/rights-and-licensing/
ISSN
0022-3050
eISSN
1468-330X
DOI
10.1136/jnnp-2016-314758
Publisher site
See Article on Publisher Site

Abstract

Neuro-inflammation RESEARCH PAPER Intracellular and non-neuronal targets of voltage-gated potassium channel complex antibodies 1 1 1 2 Bethan Lang, Mateusz Makuch, Teresa Moloney, Inga Dettmann, 2 2 2 1 Swantje Mindorf, Christian Probst, Winfried Stoecker, Camilla Buckley, 3 1 4 2 Charles R Newton, M Isabel Leite, Paul Maddison, Lars Komorowski, 1 1 1 1 Jane Adcock, Angela Vincent, Patrick Waters, Sarosh R Irani ► Additional material is ABSTRACT Kv1.6), and was used to label these channels in the published online only. To view Objectives Autoantibodies against the extracellular radioimmunoassay which first identified VGKC please visit the journal online domains of the voltage-gated potassium channel (VGKC) complex autoantibodies in patients with neuromyo- (http:// dx. doi. org/ 10. 1136/ complex proteins, leucine-rich glioma-inactivated 1 (LGI1) tonia (NMT), and then in Morvan’s syndrome jnnp- 2016- 314758). 3 45 and contactin-associated protein-2 (CASPR2), are found (MoS), limbic encephalitis (LE) and faciobra- 1 67 Nuffield Department of Clinical in patients with limbic encephalitis, faciobrachial dystonic chial dystonic seizures (FBDS). These autoanti- Neurosciences, University of seizures, Morvan’s syndrome and neuromyotonia. bodies were assumed to be directed against the Kv1 Oxford, Oxford, UK However, in routine testing, VGKC complex antibodies subunits themselves, but subsequent studies showed Institute for Experimental without LGI1 or CASPR2 reactivities (double-negative) are that the Kv1 subunits were part of a multiprotein Immunology, Lubeck, Germany Department of Psychiatry, more common than LGI1 or CASPR2 specificities. neuronal complex that includes leucine-rich glioma University of Oxford, Oxford, UK Therefore, the target(s) and clinical associations of inactivated 1 (LGI1), contactin-associated protein 2 Department of Neurology, double-negative antibodies need to be determined. (CASPR2) and contactin-2. Almost all of the anti- Queen’s Medical Centre, Methods Sera (n=1131) from several clinically defined bodies from patients with LE, FBDS or MoS, and Nottingham, UK cohorts were tested for IgG radioimmunoprecipitation of some with NMT, are directed against the extracellu- radioiodinated α-dendrotoxin ( I-αDTX)-labelled VGKC lar domains of LGI1 or CASPR2, and the anti- Correspondence to Professor Sarosh R Irani, West complexes from mammalian brain extracts. Positive bodies often co-immunoprecipitate the Wing, Level 6, John Radcliffe samples were systematically tested for live hippocampal I-αDTX-labelled Kv1 subunits from brain Hospital, Oxford OX3 9DU, UK; 7–10 neuron reactivity, IgG precipitation of I-αDTX and extracts. Patients with LGI1 or CASPR2 anti- sarosh. irani@ ndcn. ox. ac. uk I-αDTX-labelled Kv1 subunits, and by cell-based assays bodies often respond very well to immunotherapies, which expressed Kv1 subunits, LGI1 and CASPR2. and their antibody levels broadly correlate with Received 25 August 2016 711 12 Revised 3 November 2016 Results VGKC complex antibodies were found in 162 of clinical status. Accepted 30 November 2016 1131 (14%) sera. 90 of these (56%) had antibodies Therefore, there has been an increasing interest in Published Online First targeting the extracellular domains of LGI1 or CASPR2. Of diagnosing these autoimmune neurological dis- 23 January 2017 13 14 the remaining 72 double-negative sera, 10 (14%) eases, leading to large numbers of requests for immunoprecipitated I-αDTX itself, and 27 (38%) bound testing in patients who are unlikely to have well- to solubilised co-expressed Kv1.1/1.2/1.6 subunits and/or defined autoimmune syndromes. This has generated Kv1.2 subunits alone, at levels proportionate to VGKC an increase in the number of patient serum IgGs complex antibody levels (r=0.57, p=0.0017). The sera with which precipitate the VGKC complex but lack LGI1 LGI1 and CASPR2 antibodies immunoprecipitated neither or CASPR2 reactivity (‘double-negative’ samples). preparation. None of the 27 Kv1-precipitating samples bound These double-negatives can account for up to 80% live hippocampal neurons or Kv1 extracellular domains, but of samples with positive VGKC complex antibodies 16 (59%) bound to permeabilised Kv1-expressing human in studies which most closely recapitulate clinical 15 16 embryonic kidney 293T cells. These intracellular Kv1 practice. Moreover, the clinical syndromes in antibodies mainly associated with non-immune disease the double-negative patients are diverse and include 17 18 aetiologies, poor longitudinal clinical–serological correlations patients with pain syndromes, status epilepticus, 19 20 and a limited immunotherapy response. acute and chronic epilepsies, inflammatory poly- Conclusions Double-negative VGKC complex antibodies radiculopathies, children with a variety of neuroin- are often directed against cytosolic epitopes of Kv1 subunits flammatory diseases, systemic and central nervous and occasionally against non-mammalian αDTX. These system (CNS)-directed infections, a few patients antibodies should no longer be classified as neuronal-surface with Creutzfeldt-Jakob disease, and up to 5% of antibodies. They consequently lack pathogenic potential and elderly clinic controls. This clinical heterogeneity do not in themselves support the use of immunotherapies. has questioned both the pathological relevance of the antibodies and the justification for immunother- 13 14 16 25 apies in these patients. Some studies have suggested that higher titres of INTRODUCTION To cite: Lang B, Makuch M, the double-negative VGKC complex antibodies Moloney T, et al. J Neurol Radioiodinated α-dendrotoxin ( I-αDTX) binds Neurosurg Psychiatry help to increase the likelihood of pathogen- to neuronal voltage-gated potassium channels 22 26 2017;88:353–361. icity. However, until now, the few available (VGKC) of the Shaker-family (Kv1.1, Kv1.2 and Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 Neuro-inflammation studies of double-negative patients have classified these patients formalin followed by pure acetone, using a 1:10 dilution of patient serum from coded vials, with unblinding after study by their clinical features and relied on arbitrary non-validated completion. diagnostic criteria, and the subjective retrospective response to 16 25 26 Commercial antibodies against the extracellular domain of immunotherapies. Here, to definitively determine Kv1.1 (Neuromab, 75/105), and intracellular domains of Kv1.1 whether double-negative VGKC complex antibodies have pathogenic potential, we explored the epitopes they bound, (Chemicon, AB9782), Kv1.2 (Millipore, AB5924 and Neuromab their titres and clinical associations across a large variety of 75/008) and Kv1.6 (Chemicon, AB5184 and Neuromab 75/012) clinical syndromes. were used for immunoprecipitation and CBA studies. Commercial antibodies against the extracellular domains of METHODS Kv1.2 or Kv1.6 were not available. Statistics were performed Patients studied using GraphPad Prism V.6, and individual tests are stated below. To assess the frequencies of VGKC complex antibodies in a large number of varied patient phenotypes, and include syn- RESULTS dromes reported to associate with double-negative samples, Antibodies against the VGKC complex 1131 sera were tested from nine groups, including those with: Overall, across the varied cohorts, 162 of 1131 (14%) patients (1) known LE, FBDS, MoS or NMT, LGI1 or CASPR2 anti- had VGKC complex antibodies, mainly from the groups with bodies and VGKC complex antibody levels >400 pM known positivity (figure 1A). Live CBAs showed that 90 of 162 (n=84); (2) a consecutive clinic cohort known to have VGKC (56%) patients had LGI1 or CASPR2 antibodies (4 with coexist- complex antibodies without LGI1/CASPR2 reactivities (n=27; ent contactin-2 antibodies). Eighty-four of these 90 (90%) detailed in online supplementary table S1, which included patients had LE, FBDS, MoS and NMT, and 80 of the 90 sera patients with encephalopathies (n=10), NMT (n=2), stiff (87%), those with higher titres, also had IgG antibodies that person syndrome (n=2), psychiatric conditions (n=6), isolated bound the surface of hippocampal neurons. Of the remaining amnesia (n=2), Parkinson’s disease dementia (n=1), 72 (44%) double-negative VGKC complex antibody-positive Guillain-Barre syndrome (n=1) and neuropathic pain (n=3)); samples, only one—from a patient with LE in the clinic cohort (3) adult-onset epilepsies (n=582); (4) infectious diseases (see online supplementary table S1)—bound to live hippocam- (n=107: herpes simplex virus encephalitis (n=29), varicella pal neurons, suggesting a possible novel surface antigen. Among zoster virus encephalitis (VZVE, n=20), measles encephalitis the cohorts (3–9) without known VGKC complex antibodies, (n=30) and malaria (n=28, 12 with cerebral involvement)); (5) the percentage of positives ranged between 0% and 4%, with dysautonomia (n=95); (6) Lambert-Eaton myasthenic syndrome the exception of the infectious group (19%) some of which had (n=45); (7) Hu-antibodies (n=78); (8) healthy smokers (n=38) very high titres (figure 1A). and (9) healthy laboratory controls (n=75). Approval for anti- body studies was from the Oxfordshire Regional Ethical Antibodies against I-αDTX Committee A (07/Q1604/28). One possibility was that the double-negative sera bound to I-αDTX which is used to radiolabel the VGKC complex. Laboratory techniques Indeed, 10 of the 72 (14%) double-negative samples immuno- VGKC complex antibodies were detected by a radioimmunoassay 125 precipitated very high levels of the I-αDTX itself (figure 1B), which uses I-αDTX (Perkin Elmer, USA) to label VGKC com- correlating broadly with the corresponding VGKC complex plexes from 2% digitonin-solubilised rabbit whole brain mem- 24 antibody titres (figure 1C, r=0.54; p=0.02, Spearman’s rank branes. In order to closely mimic these conditions, but detect correlation). All these 10 were found in patients with infectious antibodies exclusively against the αDTX-sensitive Kv1 subunits, diseases (malaria (n=4), cerebral malaria (n=4), VZVE (n=1) Kv1.1-tranfected, Kv1.2-tranfected and Kv1.6-transfected and measles encephalitis (n=1)). Similarly, three sera from a human embryonic kidney 293T (HEK) cells were used in place snake handler, with vocational exposure to αDTX, immunopre- of brain tissue to prepare the extracts. In other respects, the cipitated I-αDTX (figure 1B). The remaining 62 radioimmunoassays were identical. To see if results were con- 125 double-negative samples without I-αDTX reactivity were founded by antibodies binding the I-αDTX itself, the tissue/ carried forward to the next experiments. cell extracts were replaced by solubilisation buffer. In each case, 5 μL of patient serum was incubated with 50 μL brain extract, HEK cell extract or buffer overnight and precipitated with 50 μL Antibodies against the αDTX binding VGKC subunits antihuman immunoglobulin G (IgG; Binding Site). The cut-off detected in solution for positivity based on the mean plus three SDs of results from To test selectively for binding to the Kv1 subunits themselves, 20 healthy control sera was 100 pM for VGKC complex anti- under conditions of the VGKC complex antibody assay, bodies, 80 pM for Kv1 subunit antibodies and 137 pM for anti- I-αDTX-labelled digitonin extracts of Kv1-transfected HEK bodies against I-αDTX alone. cells were examined. Strikingly, none of the LGI1 or CASPR2 Culture and staining procedures for live hippocampal antibody-positive sera precipitated the Kv1 subunits (figure 2A). neurons, and for live cell-based assays (CBAs) to detect anti- In contrast, 27 of 62 (44%) double-negative samples bound in bodies against LGI1, CASPR2, contactin-2, Kv1.1, Kv1.2 and solution to I-αDTX-labelled Kv1.1/1.2/1.6 heteromers only 8 125 Kv1.6 were performed as described previously. To validate (n=6), I-αDTX-labelled Kv1.2 homomers only (n=6) or to CBA results, flow cytometry was performed with live both (n=15; figure 2A and see online supplementary figure Kv1-transfected HEK cells incubated with patient serum (1:20), S1A). Furthermore, their binding to Kv1s correlated well with and bound-IgG detected using a phycoerythrin-conjugated anti- their corresponding VGKC complex antibody levels (figure 2A, human IgG secondary antibody. Samples were analysed on a r=0.57, p=0.0017, Spearman’s rank correlation). Ten of these LSRII flow cytometer with FlowJo V.10.0.8 software. Fixed 27 samples (37%) had VGKC complex antibody levels over CBAs were performed (at Euroimmun AG, Lübeck, Germany) 400 pM. No increase in binding was observed with additional with Kv1-transfected HEK cells, after fixation with 1.8% co-transfection of postsynaptic density protein 95 (PSD95, see Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 354 Neuro-inflammation Figure 1 Detection of VGKC complex antibodies and antibodies to dendrotoxin. (A) VGKC complex antibodies were determined from 1131 samples, including those with known VGKC complex antibodies (both with (n=84) and without (n=27) LGI1 or CASPR2 reactivities), and unselected patients with adult-onset epilepsies, infectious diseases, autonomic syndromes, LEMS, Hu, healthy smokers and HC. Samples with LGI1 antibodies (n=69), CASPR2 antibodies (n=21) and all available samples with VGKC complex antibody levels above 100 pM and unknown antigenic targets (red; n=72) were carried forward to other assays. Dotted lines represent this cut-off and the 400 pM cut-off from a previous study; (B) 10 of the 72 samples with unknown antigens immunoprecipitated substantial quantities of of I-αDTX alone (dotted line at 137 pM represents the mean plus 125 125 three standard deviations from 20 HCs). Three serum samples from a snake handler (grey dots) also had antibodies to I-αDTX; (C) I-αDTX antibody levels correlated with their corresponding VGKC complex antibody levels (r=0.54, p=0.015, Spearman’s rank correlation). I-αDTX, radioiodinated α-dendrotoxin; CASPR2, contactin-associated protein-2; HC, healthy controls; Hu, Hu antibodies; LEMS, Lambert-Eaton myasthenic syndrome; LGI1, leucine-rich glioma-inactivated 1; VGKC, voltage-gated potassium channel. online supplementary figure S1B), a putative Kv1-clustering Antibodies bind the intracellular domains molecule. of αDTX-sensitive VGKCs These results prove the existence of double-negative VGKC complex antibodies which bind Kv1 subunits in solution, but not Kv1 extracellular epitopes. Therefore, HEK cells expressing Antibodies do not bind the extracellular domains of Kv1.1, Kv1.2 and Kv1.6 were fixed and permeabilised so that αDTX-sensitive VGKCs antibodies against intracellular epitopes could be detected. The Kv1 reactivities available in solution could include intracel- Commercial antibodies raised against intracellular sequences of lular or extracellular epitopes. To restrict detection to extracellu- all the αDTX-sensitive Kv1 subunits bound specifically to the lar epitopes, live Kv1-transfected CBAs were tested. I-αDTX appropriate fixed HEK cells (example in figure 2C and see surface-binding studies on live Kv1-transfected HEK cells (see online supplementary figure S2A,B), and their binding was abro- online supplementary figure S1C) and commercial antibodies to gated after absorption of the commercial antibody with the the extracellular domain of Kv1.1 (figure 2B) confirmed immunising cytosolic Kv1 subunit peptide (shown for Kv1.2, adequate surface expression of Kv1.1, Kv1.2 and Kv1.6. see online supplementary figure S2A). This confirmed accessibil- However, none of the double-negative sera bound to live HEK ity of antibodies to intracellular Kv1-subunit epitopes. cells transfected with individual Kv1.1, Kv1.2 or Kv1.6, or all Subsequently, 175 coded sera were tested for binding to the three Kv1 subunits (representative example in figure 2B). To fixed permeabilised Kv1-expressing cells. These included first ensure maximal sensitivity, these negative results were con- samples of all double-negative patients without αDTX reactivity firmed using flow cytometry on the live HEK cells (n=62), 6 with αDTX antibodies, 57 sequential samples from co-transfected with Kv1.1, Kv1.2 and Kv1.6: there was no evi- the 27 patients with Kv1 antibodies demonstrated in solution dence of surface binding in the 16 sera from figure 2A with the (from figure 2A), patients with known LGI1 and CASPR2 anti- highest Kv1 antibody radioimmunoassay values (see online bodies (n=20), and disease and healthy controls without VGKC supplementary figures S1D–F). Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 355 Neuro-inflammation Figure 2 Kv1 antibodies target intracellular epitopes. (A) Twenty-seven of the remaining 62 patients with unknown VGKC complex antigenic 125 125 targets precipitated either I-αDTX-labelled Kv1.1/Kv1.2/Kv1.6 co-transfected HEK cell extracts (red circles) or I-αDTX-labelled Kv1.2-transfected HEK cell extracts (red circles with black outline). No sera with LGI1 or CASPR2 antibodies showed positive results. HEK cells transfected with Kv1.1 alone or Kv1.6 alone did not bind I-αDTX in solution; (B) a commercial antibody to the extracellular domain of Kv1.1 (anti-Kv1.1e) labelled the cell surface of live HEK cells co-transfected with Kv1.1, Kv1.2 and Kv1.6 (and enhanced green fluorescent protein (EGFP)). No patient antibodies (n=175, including the 62 double-negative samples without αDTX reactivity) showed similar binding to these live cells or live cells expressing only one of these subunits; (C) binding to fixed Kv1-transfected HEK cells was seen using serum samples which precipitated Kv1s from solution. This co-localised with binding of commercial antibodies against the intracellular domain of Kv1.2 (anti-Kv1.2). Examples for Kv1.2 and Kv1.6 are shown in online supplementary figure S1C. Scale bar=10 microns. I-αDTX, radioiodinated α-dendrotoxin; CASPR2, contactin-associated protein-2; HEK, human embryonic kidney 293T; LGI1, leucine-rich glioma-inactivated 1; VGKC, voltage-gated potassium channel. complex antibodies (n=30). Binding was observed in 41 of 175 no peak age at onset (figure 4B). Only 5 (19%) had classical (23%) samples (examples in figure 2C and online supplemen- autoimmune syndromes (paraneoplastic, NMT or LE; tary figure S2A, B): all of these were from the group of 27 figure 4C). The other patients had symptomatic or idiopathic patients whose first sample immunoprecipitated Kv1.1/Kv1.2/ generalised epilepsies (n=5), neurodegenerative diseases (n=2, Kv1.6 subunits (from figure 2A, Mann-Whitney test, Parkinson’s disease dementia and Alzheimer’s disease), and 14 p<0.0001). The first available sample from 16 of these 27 presented with conditions of unknown aetiology including patients bound to Kv1.2 (n=9), Kv1.6 (n=3), Kv1.1 (n=1), cryptogenic epilepsies (n=9), neuropathic pain (n=2), chronic Kv1.1, Kv1.2 and Kv1.6 (n=2), or both Kv1.2 and Kv1.6 encephalopathy (n=1), dysautonomia (n=1) or spontaneously (n=1). These patient IgGs showed consistent co-localisation resolving amnesia (n=1). In addition, one healthy smoker had with the Kv1 commercial antibodies (see online supplementary intracellular Kv1 antibodies. figure S2B). Overall, of the 27 samples which precipitated Despite this serological and clinical heterogeneity, 6 of the 9 I-αDTX-labelled Kv1 subunits in solution, those which did (67%) patients with cryptogenic epilepsies had Kv1.2-specific not show fixed CBA positivity tended towards lower VGKC antibodies, both patients with neuropathic pain had Kv1.6 reac- complex levels (Mann-Whitney test, p=0.001; see online tivities, and both patients with small cell lung carcinoma had supplementary figure S3A). The overall flow of samples and antibodies directed against all three subunits (Kv1.1, Kv1.2 and results by cohort is described in figure 3, and the extracellular Kv1.6). and intracellular molecular reactivities of the VGKC complex Eleven of the 27 patients were administered immunotherapies antibodies are summarised in figure 4A. and only 3 (27%) showed a sustained clinical benefit. In add- ition, only 4 of the 19 (21%) patients with serial serum samples demonstrated a relationship between intracellular antibody Correlations between Kv1 antibodies, clinical features and levels and clinical outcome (LE, NMT and 2 with cryptogenic treatment responses epilepsies), while the remaining 15 demonstrated poor correla- Serological and clinical details of the 27 patients with intracellu- tions. This contrasts with the majority of patients with LGI1 or lar Kv1 antibodies are summarised in table 1. There were 16 CASPR2 antibodies (see online supplementary figure S3B–E). men and 11 women, with ages ranging from 18 to 85 years and Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 356 Neuro-inflammation Figure 3 Summary of the sequential flow of assays through the study. As shown in figure 1A, 1131 samples were initially tested for VGKC complex antibodies by RIA (VGKC complex RIA, A) and subsequently using LGI1 and CASPR2 antibody live CBAs (B), live neuronal cultures (C) and precipitation of αDTX (αDTX RIA, D). As detailed in figure 2, double-negative samples were then tested for binding to the extracellular domains of live Kv1-tranfected HEK cells (Kv1-live CBA), for immunoprecipitations of I-αDTX-labelled Kv1-transfected HEK cells (Kv1-HEK RIA, E) and for binding to fixed permeabilised Kv1-transfected HEK cells (Kv1-fixed CBA). Cohorts are defined in more detail in the Methods section and online supplementary table S1. I-αDTX, radioiodinated α-dendrotoxin; CASPR2, contactin-associated protein-2; CBA, cell-based assay; HEK, human embryonic kidney 293T; Hu, Hu antibodies; LEMS, Lambert-Eaton myasthenic syndrome; LGI1, leucine-rich glioma-inactivated 1; RIA, radioimmunoassay; VGKC, voltage-gated potassium channel. Interestingly, in all individual patients, the targeted Kv1 subunit important because they use a systematic biochemical approach to (s) and their levels relative to VGKC complex antibodies demonstrate conclusively that a proportion of VGKC complex remained constant over time, strongly suggesting that these two antibodies binds to intracellular VGKC epitopes, or to the assays were measuring the same antibody populations (see non-mammalian-expressed αDTX. Discovery of these important online supplementary figures S3 C–E). antigenic targets should influence ongoing clinical practice, and prompt re-evaluation of several reports describing the clinical associations of patients with double-negative VGKC complex DISCUSSION 15 17–19 26 33 antibodies. In this study, double-negative antibodies Autoantibodies directed against the extracellular domains of were observed in several non-autoimmune conditions including LGI1 and CASPR2 usually associate with distinctive highly variable central and peripheral nervous system syndromes immunotherapy-responsive syndromes. Indeed, clinical and accu- with limited responses to immunotherapies, and poor correla- mulating paraclinical data strongly suggest that they are directly 891129–31 tions were noted between clinical data and serial antibody levels. pathogenic. In contrast, concerns have been raised Taken together, these double-negative autoantibodies often have about the clinical relevance of the double-negative VGKC targets which are inaccessible or non-existent in vivo, and they complex antibodies that do not bind either of these proteins, par- appear to be associated with limited clinical relevance. ticularly as they can be found in a proportion of patients with dis- Overall, LGI1 and CASPR2 antibody CBAs conferred greater eases which are unlikely to be of autoimmune 13 14 16 22 25 clinical utility and better sensitivity and specificity than VGKC aetiology. Studies which have interpreted the clinical complex antibody testing in providing a diagnosis and rationale relevance of such antibody results are limited by the inevitable for immunotherapy. This should prompt clinicians to use LGI1 difficulties in defining what constitutes an autoimmune 16 22 25 32 and CASPR2 antibodies over VGKC complex antibodies as disease. Ultimately, the definition of autoimmune first-line testing for pathogenic antibody-associated neurological diseases relies on the demonstration of a pathogenic immunotherapy-responsive syndromes, and limit immunotherapy immune factor. Therefore, the findings described here are Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 357 Neuro-inflammation Figure 4 Molecular and clinical features associated with double-negative VGKC complex antibodies. (A) The illustration of study results demonstrates that the antibodies with pathogenic potential (blue and green) target the extracellular domains of LGI1 and CASPR2, respectively, whereas likely non-pathogenic antibodies (red) target the intracellular domain of Kv1 channels, especially Kv1.2, and the α-dendrotoxin molecule itself (yellow), which is not present in mammalian tissue. Other intracellular targets may include the Kv-β2 subunit (pink). (B) The patients with intracellular Kv1 antibodies had no clear peak age of onset, and (C) 12 showed varied, known diagnoses (*), and 15 had conditions of unknown aetiology, unlikely to be autoimmune. CASPR2, contactin-associated protein-2; HS, hippocampal sclerosis; LGI1, leucine-rich glioma-inactivated 1; VGKC, voltage-gated potassium channel. administration to patients with double-negative antibodies and However, since a few disease-relevant LGI1 and CASPR2 atypical clinical syndromes. antibody-positive samples had no detectable live hippocampal A striking finding was the discovery of antibodies against neuron binding, the absence of live neuronal binding should not αDTX itself both in 10 patients with CNS infections and in a necessarily imply the presence of intracellular reactivities. snake handler. αDTX is a polypeptide toxin found in dendroas- Indeed, it is most likely that not all surface neuronal proteins pis snake venoms and was essential for labelling the VGKC are expressed in a hippocampal cell culture system. 2 4 5 15 34 complex in the earliest studies. Some of these studies It is curious that the double-negative VGKC complex epitopes included control tests to avoid detecting αDTX antibodies, appear especially immunogenic: they are found after both but commercial VGKC complex antibody assays do not provide human exposure to porcine brain aerosols and murine nasal this specific control. The αDTX antibodies were almost exclu- immunisations with brain extracts, suggesting that these anti- sively in sera with high titres of VGKC complex antibodies. In bodies are easily induced. Collectively, alongside their presence contrast, the intracellular epitopes identified on the Kv1 subu- in a variety of epilepsies, neurodegenerative diseases and idio- nits themselves, particularly on Kv1.2, were found both at high pathic syndromes, these observations imply that double-negative and low VGKC complex antibody levels. Therefore, this study antibodies arise secondary to diverse pathologies, where the shows that the antigenic target—rather than the VGKC complex primary pathological insult may not be immune. Indeed, antibody level—aids prediction of potential pathogenicity. perhaps non-immune causes of neuronal destruction are suffi- This is consistent with the observation that while VGKC cient to release immunogenic Kv1 antigens and provoke auto- 16 22 26 complex antibody levels associated with LGI1 and CASPR2 antibody production. Interestingly, the immunotherapy reactivities are often high, they can be low or even response observed in 27% of patients is very similar to the 8 152535 36 undetectable. placebo response rates observed in several other neurological Furthermore, this is the first clear demonstration of diseases, and much lower than those reported in LGI1 or 7–911152935 double-negative VGKC complex antibodies which target intra- CASPR2 antibody-associated syndromes. However, cellular epitopes. This positive demonstration is important as we conclusions regarding treatment outcomes are only definitive and others have previously hypothesised their existence, mainly after blinded interventional studies. Also, the demographics of 13 14 22 25 inferring from the absence of live neuronal binding. the patients with Kv1 antibodies did not show any age or sex Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 358 Neuro-inflammation Neuro-inflammation Table 1 Clinical–serological details of patients with antibodies against the intracellular aspects of Kv1.1, Kv1.2 and/or Kv1.6 Serological results Age/ VGKC complex Kv1-HEK Kv1-fixed Disease Sex antibody (pM) RIA CBA Clinical features aetiology Outcome and treatments Summary of clinical–serological correlations 59/M 1346 + 1.2 only Cryptogenic TLE with amnesia/ Unknown No sustained response to three AEDs and Poor. Persistent VGKC-c Abs despite markedly varied SZ frequencies. depression CS Online supplementary figure S3C 49/F 767 + 1.2 only NMT with EMG confirmation Autoimmune Limited response to AEDs and CS Poor. Highest VGKC-c Abs during clinical remission 55/M 533 + 1.2 only NMT with EMG confirmation Autoimmune Symptoms over 8 years despite AEDs/AZA/ Good. Patient symptomatic with persistent VGKC-c Abs CS 77/M 470 + 1.2 only Chronic myelopathy/ Unknown Transient response to CS and PLEX Poor. High VGKC-c Abs despite symptom fluctuations over 3 years encephalopathy 21/M 461 + 1.2 only Cryptogenic probable frontal lobe Unknown Ongoing SZs despite two AEDs Poor. Online supplementary figure S3E epilepsy 50/M 448 + 1.2 only Isolated amnesia Unknown Spontaneous resolution over a few days Poor. Abs reduced over 12 months. Symptoms resolved at 7 days 28/F 278 + 1.2 only Cryptogenic TLE Unknown SZ freedom after first AED Poor. Persistent VGKC-c Abs during SZ freedom 62/M 236 + 1.2 only Cryptogenic TLE with amnesia Unknown SZs and amnesia at 5 years with 3 AEDs Poor. VGKC-c Abs disappeared while SZs were ongoing 18/M 141 + 1.2 only Limbic encephalitis Autoimmune Good response to CS and PLEX Good. VGKC-c Abs reduced and clinical improvement at 12 months 25/F 529 + 1.6 only Diffuse neuropathic pain and Unknown No response to opioids, CS or IVIG Poor. High VGKC-c Abs despite marked fluctuations in symptoms depression 30/F 117 + 1.6 only Idiopathic generalised epilepsy Genetic Good response to AEDs NA 68/M 240 + 1.6 only Alzheimer’s disease; one SZ Degenerative Ongoing fall in memory despite CS/IVIG Poor. Slight fall in VGKC-c Abs but marked reduction in memory over 4 years 34/M 577 + 1.2 and 1.6 Widespread neuropathic pain Unknown No response to CS, IVIG and PLEX Poor. Constant pain despite highly varied VGKC-c Abs. Online supplementary figure S3D 71/F 2489 + 1.1, 1.2 and NMT plus SCLC Paraneoplastic Palliative care only NA. 1.6 67/F 2120 + 1.1, 1.2 and LEMS plus Hu antibody Paraneoplastic Good response to CS Poor. Persistently high VGKC-c Abs over 5 years despite clinical 1.6 neuropathy and SCLC improvements 84/F 282 + 1.1 only Parkinson’s disease dementia Degenerative No response to CS Poor. VGKC-c Abs reduced over 12 months despite clinical worsening 58/M 304 + Negative Dysautonomia Unknown NA NA 37/M 253 + Negative TLE related to left HS Structural NA NA 59/M 236 + Negative Healthy smoker Healthy Not relevant Not relevant 48/F 223 + Negative TLE related to left HS Structural NA NA 85/F 205 + Negative Cryptogenic focal motor SZs Unknown SZ free on 1 AED. Amnesia and anxiety Poor. VGKC-c Abs disappeared over 1 year; SZ freedom at 2 years. benefited from CS/IVIG VGKC-c Abs returned at 4 years without SZs 33/F 182 + Negative Cryptogenic TLE Unknown SZ freedom after second AED Moderate. VGKC-c Abs reduced over 6 months and SZ freedom at 1 year 76/M 181 + Negative Cryptogenic epilepsy Unknown SZ freedom with 1 AED Poor. SZ freedom at 6 months; VGKC-c Abs sampled after 15 years 54/F 139 + Negative Epilepsy after childhood Structural NA NA meningitis 22/M 137 + Negative Cryptogenic epilepsy Unknown SZ freedom after 1 AED NA 54/M 131 + Negative Cryptogenic epilepsy Unknown SZ free at 1 year with 1 AED Moderate. SZ free at 1 year; VGKC-c Abs absent at 3 months 77/M 123 + Negative TLE secondary to CVA Structural Ongoing SZs at 4 years NA Patients grouped by the VGKC-c antibody levels, precipitation from Kv1.1/Kv1.2/Kv1.6-cotransfected HEK cell and Kv1.2-transfected HEK cell radioimmunoassays (Kv1-HEK RIA; denoted as positive (+) or negative (−)) and the Kv1 subunit expressed fixed CBAs (Kv1-fixed CBA). All these patients had negative results in live CBAs for antibodies against Kv1s, LGI1, CASPR2, contactin-2 and for binding to live hippocampal neurons. Two patients had SCLC and tumours were not found in the remaining patients. Ab, antibody; AEDs, antiepileptic drugs; AZA, azathioprine; CASPR2, contactin-associated protein-2; CS, corticosteroids; CBA, cell-based assay; CVA, cerebrovascular accident; EMG, electromyography; F, female; HEK, human embryonic kidney 293T; HS, hippocampal sclerosis; IVIG, intravenous immunoglobulins; LEMS, Lambert-Eaton myasthenic syndrome; LGI1, leucine-rich glioma-inactivated 1; M, male; NA, not available (only single serum sample obtained); PLEX, plasma exchange; SCLC, small cell lung carcinomas; SZ, seizure; TLE, temporal lobe epilepsy; VGKC, voltage-gated potassium channel; VGKC-c, VGKC complex. Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 359 Neuro-inflammation 2 Shillito P, Molenaar PC, Vincent A, et al. Acquired neuromyotonia: evidence for bias, quite different from the occurrence of LGI1 antibody LE autoantibodies directed against K+ channels of peripheral nerves. Ann Neurol in later years or the male predominance of CASPR2 anti- 1995;38:714–22. 8 9 11 29 bodies. Alongside the 5% rate of double-negative VGKC 3 Liguori R, Vincent A, Clover L, et al. Morvan’s syndrome: peripheral and central complex antibodies in healthy individuals, these observations nervous system and cardiac involvement with antibodies to voltage-gated potassium suggest that double-negative reactivities have very limited clin- channels. Brain 2001;124:2417–26. 4 Vincent A, Buckley C, Schott JM, et al. Potassium channel antibody-associated ical significance. encephalopathy: a potentially immunotherapy-responsive form of limbic encephalitis. On the other hand, we noted relationships between specific Brain 2004;127:701–12. Kv1 subunit reactivities and clinical syndromes, particularly 5 Thieben MJ, Lennon VA, Boeve BF, et al. Potentially reversible autoimmune limbic Kv1.2 antibodies in cryptogenic epilepsies, Kv1.6 antibodies in encephalitis with neuronal potassium channel antibody. Neurology 2004;62:1177–82. the two patients with neuropathic pain, and Kv1.1, Kv1.2 and 6 Irani SR, Buckley C, Vincent A, et al. Immunotherapy-responsive seizure-like Kv1.6 antibodies in paraneoplastic syndromes. These prelimin- episodes with potassium channel antibodies. Neurology 2008;71:1647–8. ary findings suggest that specific patterns of Kv1 reactivities may 7 Irani SR, Michell AW, Lang B, et al. Faciobrachial dystonic seizures precede Lgi1 be generated in response to certain pathologies. Indeed, a few antibody limbic encephalitis. Ann Neurol 2011;69:892–900. 8 Irani SR, Alexander S, Waters P, et al. Antibodies to Kv1 potassium patients with classical VGKC complex antibody-associated syn- channel-complex proteins leucine-rich, glioma inactivated 1 protein and dromes, such as NMT and LE, did show good antibody-clinical contactin-associated protein-2 in limbic encephalitis, Morvan’s syndrome and correlations, and in these patients the Kv1 antibodies may be acquired neuromyotonia. Brain 2010;133:2734–48. 38 39 markers of a coexistent pathogenic antibody. However, 9 Lai M, Huijbers MG, Lancaster E, et al. Investigation of LGI1 as the antigen in while many of the clinic cohort had an underlying immune limbic encephalitis previously attributed to potassium channels: a case series. Lancet Neurol 2010;9:776–85. basis to their neurological syndrome, ascertainment of patients 10 Irani SR, Pettingill P, Kleopa KA, et al. Morvan syndrome: clinical and serological attending a specialist autoimmune clinic may well carry this observations in 29 cases. Ann Neurol 2012;72:241–55. intrinsic bias. 11 Lancaster E, Huijbers MGM, Bar V, et al. Investigations of caspr2, an autoantigen of In summary, our results show that many double-negative encephalitis and neuromyotonia. Ann Neurol 2011;69:303–11. 12 Shin YW, Lee ST, Shin JW, et al. VGKC-complex/LGI1-antibody encephalitis: clinical VGKC complex antibodies lack pathogenic potential, and that manifestations and response to immunotherapy. J Neuroimmunol 2013;265:75–81. direct examination of LGI1 and CASPR2 antibodies is more Irani SR, Gelfand JM, Al-Diwani A, et al. Cell-surface central nervous system informative than first-line VGKC complex antibody autoantibodies: clinical relevance and emerging paradigms. Ann Neurol 13 14 25 35 36 testing. Also, direct LGI1 and CASPR2 antibody 2014;76:168–84. CBAs appear more sensitive than VGKC complex antibody or 14 Leypoldt F, Armangue T, Dalmau J. Autoimmune encephalopathies. Ann N Y Acad 25 35 36 Sci 2015;1338:94–114. live neuronal antibody determination. Therefore, direct 15 Klein CJ, Lennon VA, Aston PA, et al. Insights from LGI1 and CASPR2 potassium LGI1 and CASPR2 antibody assays are the preferred tests for channel complex autoantibody subtyping. JAMA Neurol 2013;70:229–6. clinical purposes. This approach should limit the use of unneces- 16 Paterson RW, Zandi MS, Armstrong R, et al. Clinical relevance of positive sary immunotherapies in double-negative patients. voltage-gated potassium channel (VGKC)-complex antibodies: experience from a tertiary referral centre. J Neurol Neurosurg Psychiatry 2014;85:625–30. Until all centres update their practices, treatment of 17 Klein CJ, Lennon VA, Aston PA, et al. Chronic pain as a manifestation of potassium double-negative cases should be limited to those with a high channel-complex autoimmunity. Neurology 2012;79:1136–44. pretest probability of an underlying autoimmune condition. 18 Suleiman J, Brenner T, Gill D, et al. VGKC antibodies in pediatric encephalitis However, for research purposes, it may be useful for further presenting with status epilepticus. Neurology 2011;76:1252–5. studies to identify the remaining, likely intracellular, 19 Brenner T, Sills GJ, Hart Y, et al. Prevalence of neurologic autoantibodies in cohorts of patients with new and established epilepsy. Epilepsia 2013;54:1028–35. double-negative VGKC complex antigens such as the Kv-β2 20 Toledano M, Britton JW, McKeon A, et al. Utility of an immunotherapy trial in subunit and PSD95, and explore whether they are useful bio- evaluating patients with presumed autoimmune epilepsy. Neurology markers of traditionally non-immune or immune neurological 2014;82:1578–86. syndromes. 21 Meeusen JW, Klein CJ, Pirko I, et al. Potassium channel complex autoimmunity induced by inhaled brain tissue aerosol. Ann Neurol 2012;71:417–26. Funding BL is supported by Epilepsy Research UK (ERUK; P1201); SRI is supported 22 Hacohen Y, Singh R, Rossi M, et al. Clinical relevance of voltage-gated potassium by a Wellcome Trust Intermediate Fellowship (104079/Z/14/Z), BMA Research channel-complex antibodies in children. Neurology 2015;85:967–75. Grants—Vera Down grant (2013) , the Fulbright UK-US commission and the MS 23 Pillai SC, Hacohen Y, Tantsis E, et al. Infectious and autoantibody-associated society. Research in the Neuroimmunology Laboratory is supported by the Oxford encephalitis: clinical features and long-term outcome. Pediatrics 2015;135: NIHR Biomedical Research Centre. PW and MIL are supported by the NHS National e974–84. Specialized Commissioning Group for Neuromyelitis optica, UK. 24 Rossi M, Mead S, Collinge J, et al. Neuronal antibodies in patients with suspected or confirmed sporadic Creutzfeldt-Jakob disease: table 1. J Neurol Neurosurg Competing interests AV, SRI, BL and PW are coapplicants and receive royalties Psychiatry 2015;86:692–4. on patent application WO/2010/046716 entitled ‘Neurological Autoimmune 25 van Sonderen A, Schreurs MWJ, de Bruijn MAAM, et al. The relevance of VGKC Disorders’. The patent has been licensed to Euroimmun AG for the development of positivity in the absence of LGI1 and Caspr2 antibodies. Neurology assays for LGI1 and other VGKC-complex antibodies. BL and SRI had full access to 2016;86:1692–9. all of the data in the study and take responsibility for the integrity of the data and 26 Olberg H, Haugen M, Storstein A, et al. Neurological manifestations related to level the accuracy of the data analysis. of voltage-gated potassium channel antibodies. J Neurol Neurosurg Psychiatry Ethics approval Oxfordshire Regional Ethical Committee A (RECA). 2013;84:941–3. 27 Horresh I, Poliak S, Grant S, et al. Multiple molecular interactions determine the Provenance and peer review Not commissioned; externally peer reviewed. clustering of Caspr2 and Kv1 channels in myelinated axons. J Neurosci Open Access This is an Open Access article distributed in accordance with the 2008;28:14213–22. terms of the Creative Commons Attribution (CC BY 4.0) license, which permits 28 Waters PJ, McKeon A, Leite MI, et al. Serologic diagnosis of NMO: a multicenter others to distribute, remix, adapt and build upon this work, for commercial use, comparison of aquaporin-4-IgG assays. Neurology 2012;78:665–71–discussion669. provided the original work is properly cited. See: http://creativecommons.org/licenses/ 29 Flanagan EP, Kotsenas AL, Britton JW, et al. Basal ganglia T1 hyperintensity in by/4.0/ LGI1-autoantibody faciobrachial dystonic seizures. Neurol Neuroimmunol Neuroinflamm 2015;2:e161–1. 30 Ohkawa T, Fukata Y, Yamasaki M, et al. Autoantibodies to epilepsy-related LGI1 in limbic encephalitis neutralize LGI1-ADAM22 interaction and reduce synaptic AMPA receptors. J Neurosci 2013;33:18161–74. REFERENCES 31 Lalic T, Pettingill P, Vincent A, et al. Human limbic encephalitis serum enhances 1 Shamotienko OG, Parcej DN, Dolly JO. Subunit combinations defined for K+ hippocampal mossy fiber-CA3 pyramidal cell synaptic transmission. Epilepsia channel Kv1 subtypes in synaptic membranes from bovine brain. Biochemistry 2011;52:121–31. 1997;36:8195–201. Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758 360 Neuro-inflammation 32 Graus F, Titulaer MJ, Balu R, et al. A clinical approach to diagnosis of autoimmune 36 Becker EBE, Zuliani L, Pettingill R, et al. Contactin-associated protein-2 antibodies encephalitis. Lancet Neurol 2016;15:391–404. in non-paraneoplastic cerebellar ataxia. J Neurol Neurosurg Psychiatry 33 McKnight K, Jiang Y, Hart Y, et al. Serum antibodies in epilepsy and 2012;83:437–40. seizure-associated disorders. Neurology 2005;65:1730–6. 37 Bittar C, Nascimento OJM. Placebo and nocebo effects in the neurological practice. 34 Hart IK, Waters C, Vincent A, et al. Autoantibodies detected to expressed K+ Arq Neuro Psiquiatr 2015;73:58–63. channels are implicated in neuromyotonia. Ann Neurol 1997;41:238–46. 38 Boronat A, Sabater L, Saiz A, et al. GABA(B) receptor antibodies in limbic encephalitis 35 Irani SR, Stagg CJ, Schott JM, et al. Faciobrachial dystonic seizures: the influence of and anti-GAD-associated neurologic disorders. Neurology 2011;76:795–800. immunotherapy on seizure control and prevention of cognitive impairment in a 39 Pettingill P, Kramer HB, Coebergh JA, et al. Antibodies to GABAA receptor α1 and broadening phenotype. Brain 2013;136:3151–62. γ2 subunits: clinical and serologic characterization. Neurology 2015;84:1233–41. Lang B, et al. J Neurol Neurosurg Psychiatry 2017;88:353–361. doi:10.1136/jnnp-2016-314758

Journal

Journal of Neurology Neurosurgery & Psychiatry – British Medical Journal

Published: Apr 23, 2017

There are no references for this article.