Characterization of a Subtype of Autoimmune Encephalitis With Anti–Contactin-Associated Protein-like 2 Antibodies in the Cerebrospinal Fluid, Prominent Limbic Symptoms, and Seizures | Epilepsy and Seizures | JAMA Neurology | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Fluid-Attenuated Inversion Recovery Imaging of a Patient With Epilepsy, Partial Temporal Seizures, and Anterograde Memory Impairments
Fluid-Attenuated Inversion Recovery Imaging of a Patient With Epilepsy, Partial Temporal Seizures, and Anterograde Memory Impairments

In these T2-weighted magnetic resonance images, note the asymmetrical hyperintensities involving both temporomesial regions (arrowheads).

Figure 2.  Reactivity of Anti–Contactin-Associated Protein-like 2 Antibodies from Patients’ Cerebrospinal Fluid (CSF) With Rat Brain
Reactivity of Anti–Contactin-Associated Protein-like 2 Antibodies from Patients’ Cerebrospinal Fluid (CSF) With Rat Brain

The CSF of 2 patients was used for immunostaining; fluorescent antihuman IgG mouse antibodies were used as secondary antibodies. Sagittal section of rat brain immunostained with a patient’s CSF (A, dilution 1:10; original magnification × 7). There was diffuse staining of the neuropil that was not observed when rat brain sections were incubated with a control patient’s CSF (B, dilution 1:10; original magnification × 7). Immunoreactivity was particularly strong in the molecular and the granular cell layers of the cerebellum (C; original magnification × 70) and the hippocampus (D; original magnification × 35); in addition, there was diffuse staining of the molecular layer of the dentate gyrus (E; original magnification × 70).

Figure 3.  CASPR2 Autoantibody Characterization
CASPR2 Autoantibody Characterization

A, Human embryonic kidney 293 cells were transfected to express the full-length contactin-associated protein-like 2 (CASPR2) protein- or domain-deletion constructs tagged with hemagglutinin (HA), visualized in the first column. The transfected cells were incubated with patients’ cerebrospinal fluid (CSF; dilution 1:10), visualized in the second column. Top left: CASPR2 cells were not recognized in the CSF of any control patients. Bottom left: Constructs only including the discoidin domain (Disc), laminin G1 domain (Lam), or both domains (not shown) were recognized in all 18 samples for patients with CASPR2 cells in the CSF, but the Del1 construct, lacking Disc and Lam, was not recognized in 8 of 18 patients. Top right: Example of 1 of 10 patients in whom the Del1 construct was recognized. Bottom right: Illustration of the different CASPR2 domains deleted in our constructs. B, Human embryonic kidney 293 cells were transfected with the HA-tagged full-length CASPR2 protein and incubated with patients’ CSF (dilution 1:10) to determine the IgG isotypes of CASPR2 autoantibodies. We observed 10 of 17 patients with both IgG4 and IgG1 antibody subtypes (left panel), but only IgG4 antibodies were observed in the remaining 7 patients (right panel).

Table 1.  Clinical Findings, Brain MRI, and Evolution of Patients With Anti-CASPR2 Antibodies in the CSF
Clinical Findings, Brain MRI, and Evolution of Patients With Anti-CASPR2 Antibodies in the CSF
Table 2.  Baseline Characteristics and Immunological Findings of Patients With Anti-CASPR2 Antibodies in the CSF Compared With Patients With NMT or MoS
Baseline Characteristics and Immunological Findings of Patients With Anti-CASPR2 Antibodies in the CSF Compared With Patients With NMT or MoS
1.
Irani  SR, Alexander  S, Waters  P,  et al.  Antibodies to Kv1 potassium channel-complex proteins leucine-rich, glioma inactivated 1 protein and contactin-associated protein-2 in limbic encephalitis, Morvan’s syndrome and acquired neuromyotonia.  Brain. 2010;133(9):2734-2748.PubMedGoogle ScholarCrossref
2.
Malter  MP, Frisch  C, Schoene-Bake  JC,  et al.  Outcome of limbic encephalitis with VGKC-complex antibodies: relation to antigenic specificity.  J Neurol. 2014;261(9):1695-1705.PubMedGoogle ScholarCrossref
3.
Irani  SR, Pettingill  P, Kleopa  KA,  et al.  Morvan syndrome: clinical and serological observations in 29 cases.  Ann Neurol. 2012;72(2):241-255.PubMedGoogle ScholarCrossref
4.
Laurencin  C, André-Obadia  N, Camdessanché  JP,  et al.  Peripheral small fiber dysfunction and neuropathic pain in patients with Morvan syndrome.  Neurology. 2015;85(23):2076-2078.PubMedGoogle ScholarCrossref
5.
Poliak  S, Salomon  D, Elhanany  H,  et al.  Juxtaparanodal clustering of Shaker-like K+ channels in myelinated axons depends on Caspr2 and TAG-1.  J Cell Biol. 2003;162(6):1149-1160.PubMedGoogle ScholarCrossref
6.
Bel  C, Oguievetskaia  K, Pitaval  C, Goutebroze  L, Faivre-Sarrailh  C.  Axonal targeting of Caspr2 in hippocampal neurons via selective somatodendritic endocytosis.  J Cell Sci. 2009;122(pt 18):3403-3413.PubMedGoogle ScholarCrossref
7.
Pinatel  D, Hivert  B, Boucraut  J,  et al.  Inhibitory axons are targeted in hippocampal cell culture by anti-Caspr2 autoantibodies associated with limbic encephalitis.  Front Cell Neurosci. 2015;9:265.PubMedGoogle ScholarCrossref
8.
Olsen  AL, Lai  Y, Dalmau  J, Scherer  SS, Lancaster  E.  Caspr2 autoantibodies target multiple epitopes.  Neurol Neuroimmunol Neuroinflamm. 2015;2(4):e127.PubMedGoogle ScholarCrossref
9.
Reiber  H, Lange  P.  Quantification of virus-specific antibodies in cerebrospinal fluid and serum: sensitive and specific detection of antibody synthesis in brain.  Clin Chem. 1991;37(7):1153-1160.PubMedGoogle Scholar
10.
Navarro  V, Kas  A, Apartis  E,  et al.  Motor cortex and hippocampus are the two main cortical targets in LGI1-antibody encephalitis.  Brain. 2016;139(pt 4):1079-1093.PubMedGoogle ScholarCrossref
11.
Lancaster  E, Huijbers  MG, Bar  V,  et al.  Investigations of caspr2, an autoantigen of encephalitis and neuromyotonia.  Ann Neurol. 2011;69(2):303-311.PubMedGoogle ScholarCrossref
12.
Becker  EBE, Zuliani  L, Pettingill  R,  et al.  Contactin-associated protein-2 antibodies in non-paraneoplastic cerebellar ataxia.  J Neurol Neurosurg Psychiatry. 2012;83(4):437-440.PubMedGoogle ScholarCrossref
13.
Sunwoo  JS, Lee  ST, Byun  JI,  et al.  Clinical manifestations of patients with CASPR2 antibodies.  J Neuroimmunol. 2015;281:17-22.PubMedGoogle ScholarCrossref
14.
Bindon  CI, Hale  G, Brüggemann  M, Waldmann  H.  Human monoclonal IgG isotypes differ in complement activating function at the level of C4 as well as C1q.  J Exp Med. 1988;168(1):127-142.PubMedGoogle ScholarCrossref
15.
Bruhns  P, Iannascoli  B, England  P,  et al.  Specificity and affinity of human Fcgamma receptors and their polymorphic variants for human IgG subclasses.  Blood. 2009;113(16):3716-3725.PubMedGoogle ScholarCrossref
16.
Schuurman  J, Van Ree  R, Perdok  GJ, Van Doorn  HR, Tan  KY, Aalberse  RC.  Normal human immunoglobulin G4 is bispecific: it has two different antigen-combining sites.  Immunology. 1999;97(4):693-698.PubMedGoogle ScholarCrossref
17.
van der Neut Kolfschoten  M, Schuurman  J, Losen  M,  et al.  Anti-inflammatory activity of human IgG4 antibodies by dynamic Fab arm exchange.  Science. 2007;317(5844):1554-1557.PubMedGoogle ScholarCrossref
18.
Huijbers  MG, Zhang  W, Klooster  R,  et al.  MuSK IgG4 autoantibodies cause myasthenia gravis by inhibiting binding between MuSK and Lrp4.  Proc Natl Acad Sci U S A. 2013;110(51):20783-20788.PubMedGoogle ScholarCrossref
19.
Körtvelyessy  P, Bauer  J, Stoppel  CM,  et al.  Complement-associated neuronal loss in a patient with CASPR2 antibody-associated encephalitis.  Neurol Neuroimmunol Neuroinflamm. 2015;2(2):e75.PubMedGoogle ScholarCrossref
Original Investigation
September 2016

Characterization of a Subtype of Autoimmune Encephalitis With Anti–Contactin-Associated Protein-like 2 Antibodies in the Cerebrospinal Fluid, Prominent Limbic Symptoms, and Seizures

Author Affiliations
  • 1Centre National de Référence pour les Syndromes Neurologiques Paranéoplasiques, Hôpital Neurologique, Hospices Civils de Lyon, Lyon, France.
  • 2Institut NeuroMyoGene, INSERM 1217/CNRS 5310, Université de Lyon, Lyon, France.
  • 3Université Claude-Bernard Lyon 1, Université de Lyon, Lyon, France.
  • 4Département de Neurologie, Groupe Hospitalier Pitié-Salpêtrière, Paris, France
  • 5Service de Neurologie, Hôpital Bellevue, Centre Hospitalier Universitaire de Saint-Étienne, Saint-Étienne, France
JAMA Neurol. 2016;73(9):1115-1124. doi:10.1001/jamaneurol.2016.1585
Abstract

Importance  Autoantibodies against contactin-associated protein-like 2 (CASPR2) are observed in several neurological syndromes, including neuromyotonia (NMT), Morvan syndrome (MoS), and limbic encephalitis.

Objective  To characterize the clinical and biological presentations of patients with anti-CASPR2 antibodies in the cerebrospinal fluid (CSF).

Design, Setting, and Participants  We conducted a retrospective cohort analysis of 18 patients who had anti-CASPR2 antibodies in their CSF between March 2009 and November 2015 at the Centre National de Référence pour les Syndromes Neurologiques Paranéoplasiques in Lyon, France. Additionally, we analyzed 15 patients who were diagnosed as having NMT or MoS as a comparative group.

Main Outcomes and Measures  Clinical presentations, anti-CASPR2 antibodies specificities, brain magnetic resonance imaging, and CSF analyses, cancer prevalence, and evolution.

Results  In this cohort of 18 patients with anti-CASPR2 antibodies in their CSF, 17 (94.4%) were male and had a median (range) age of 64.5 (53-75) years; in the second group, 9 of 15 patients (60.0%) with NMT or MoS were male and had a median (range) age of 51 years (1 month to 75 years). Only 3 patients (16.7%) in this cohort had a previous or concomitant history of cancer (prostate, hematological, or thyroid), whereas 9 patients (60.0%) in the second group had a malignant thymoma. Symptoms of limbic encephalitis were observed in all patients, including temporal lobe seizures in 16 patients (88.9%) and memory disorders in 17 patients (94.4%) from the cohort. Extralimbic signs were also evident in 12 of 18 patients (66.7%), including cerebellar ataxia in 6 patients (33.3%). Only 2 patients (11.1%) from the cohort were diagnosed as having NMT. Brain magnetic resonance imaging displayed T2-weighted temporolimbic abnormalities in 14 of 15 patients (93.3%) in the second group. Cerebrospinal fluid analysis was abnormal in 9 of 12 patients (75.0%). For 16 of 18 patients (88.9%), follow-up was performed for at least a 6-month period (median [range], 34 [11-114] months). Of these, 15 (93.8%) improved and 6 (37.5%) relapsed. In all patients in this cohort, IgG4 autoantibodies were detected in the CSF. Anti-CASPR2 antibodies in the CSF targeted the laminin G1 and discoidin domains of CASPR2 in all patients. Importantly, anti-CASPR2 antibodies were detected in the serum but not in the CSF of all patients with NMT or MoS.

Conclusions and Relevance  In this cohort study, anti-CASPR2 antibodies in the CSF are associated with a subtype of autoimmune encephalitis with prominent limbic involvement and seizures that is rarely associated with cancer. Conversely, patients with NMT and MoS have anti-CASPR2 antibodies only in the serum but not in the CSF and frequently present with a malignant thymoma. The anti-CASPR2 antibodies found in these patients targeted the discoidin and laminin G1 domains of CASPR2 and always included IgG4 autoantibodies.

Introduction

Antibodies directed against contactin-associated protein-like 2 (CASPR2) have been described in the sera of patients with peripheral and central neurological syndromes, including neuromyotonia (NMT), Morvan syndrome (MoS), and autoimmune limbic encephalitis.1 Anti-CASPR2 antibodies belong to the anti–voltage-gated potassium channel antibody complex, a biomarker that has been shown to include autoantibodies targeting CASPR2, leucin-rich glioma 1 protein, and other unknown antigens.2 In addition, CASPR2 antibody–related syndromes have been frequently described in patients with malignant thymoma, a potential source of impaired T-cell maturation that may lead to autoantibody production.3,4 Although CASPR2 is involved in the organization of the juxtaparanodal regions of myelinated nerves, it is also expressed in the axons of hippocampal cells.5,6 A recent study7 has suggested that anti-CASPR2 antibodies from the cerebrospinal fluid (CSF) of patients with limbic encephalitis bind to hippocampal inhibitory interneurons at the presynaptic level and have a disruptive effect on inhibitory synapses. Moreover, anti-CASPR2 antibodies in the serum and CSF have predominantly been shown to be IgG4 and to target multiple extracellular epitopes of CASPR2, including its discoidin and laminin G1 domains.7,8 Therefore, the diversity of CASPR2 antibody–associated diseases may rely on the implication of various factors, including the site of autoantibody synthesis, the epitope specificities of anti-CASPR2 antibodies, or the IgG subtypes involved. To further characterize patients with anti-CASPR2 antibodies, we analyzed clinical and biological data from patients who were positive for anti-CASPR2 antibodies in the CSF.

Box Section Ref ID

Key Points

  • Question What are the clinical features associated with anti–contactin-associated protein-like 2 (CASPR2) antibodies in the cerebrospinal fluid?

  • Findings In this cohort study of 18 patients with anti-CASPR2 antibodies in the cerebrospinal fluid, all patients displayed limbic symptoms, including temporal lobe seizures in 16 patients and memory disorders in 17 patients. Extralimbic signs were seen in 12 patients, including cerebellar ataxia in 6 patients.

  • Meaning The presence of anti-CASPR2 antibodies in the cerebrospinal fluid is associated with a subtype of autoimmune encephalitis with predominant temporal lobe epilepsy and memory disorders.

Methods

In this study, we retrospectively included all patients with CSF samples positive for anti-CASPR2 antibodies at the Centre National de Référence pour les Syndromes Neurologiques Paranéoplasiques in Lyon, France, between March 2009 and November 2015. During this period, we tested 6650 CSF samples from French hospitals sent for routine diagnostic workup in patients with suspected autoimmune encephalitis or paraneoplastic neurological disorders. Cerebrospinal fluid samples from the patients were screened by immunohistofluorescence on rat brain sections, and anti-CASPR2 antibodies were subsequently identified by a cell-based binding assay (CBA). In addition, serum samples from patients with confirmed anti-CASPR2 antibodies in the CSF were screened for anti-CASPR2 antibodies using a CBA. Approval for this study was granted by the institutional review board of the Hospices Civils de Lyon (Comité de Protection des Personnes SUD-EST IV). Written informed consent was obtained from all patients.

Serum and CSF samples were deposited in the NeuroBioTec biobank (Hospices Civils de Lyon). Clinical data were collected from the referring physicians by telephone and email. We focused on the modes of onset, clinical symptoms, ancillary results, therapy regimens, cancer prevalence, and outcomes. In addition, patients diagnosed in our center as having NMT or MoS were tested for anti-CASPR2 antibodies by CBA in both the serum and CSF and were used as a comparative group. Age differences between groups were compared by Mann-Whitney U test. Sex ratio, anti-CASPR2 antibody levels in the serum and CSF, and comorbidities between groups were compared using Fisher exact test. Statistical significance was set at P < .05.

For immunohistofluorescence assays, sagittal slices obtained from adult rat brains were incubated with patient CSF samples (diluted 1:10) overnight at 4°C. Brain slices were fixed by immersion into 4% paraformaldehyde for 1 hour, followed by immersion in a 30% sucrose solution for 24 hours. Histological slices were produced using the Leica CM1950 cryostat microtome (Leica Biosystems Nussloch GmbH). Tissues were subsequently washed and incubated with 488 Alexa Fluor–conjugated goat antihuman IgG (A-21433; Thermofisher). For the CBA, cultured human embryonic kidney 293 cells were transfected using the Lipofectamine LTX kit (Invitrogen), with plasmids coding for full-length or deleted CASPR2 fused to an influenza virus hemagglutinin tag. Cells were incubated for 48 hours after transfection with CSF (diluted 1:10) or sera (diluted 1:20) and for 2 hours with mouse anti-hemagglutinin antibody (diluted 1:1000; H3663; Sigma-Aldrich) in Dulbecco modified eagle medium, 25 mmol/L 4-(2-hydroxyetyl)-1-piperazineethanesulfonic acid, 3% bovine serum albumin, and 5% normal goat serum. Cells were subsequently washed in Dulbecco modified eagle medium and 4-(2-hydroxyetyl)-1-piperazineethanesulfonic acid for 15 minutes and incubated with Alexa Fluor–conjugated secondary antibodies (diluted 1:1000), including goat antihuman total IgG-555 (A-21433; Thermofisher), -488 (A-11013; Thermofisher), goat anti-mouse IgG-555 (A-21424; Thermofisher), -488 (A-11029; Thermofisher), mouse antihuman IgG1-488 (F9890; Sigma-Aldrich) and mouse antihuman IgG4-488 (F0767; Sigma-Aldrich). Bound antibodies were visualized using the Zeiss Axiophot Imager.Z1. Anti-CASPR2 antibodies titers in patient sera or CSF were assessed by serial 2-fold dilutions in CBA. Titers were determined as the lowest dilution that gave a positive signal, according to 2 blinded investigators (V.R., M.S.M.). The epitopes targeted by patient autoantibodies were examined on CBA using CASPR2 with hemagglutinin-deleted constructs as previously described7 with the same dilutions (CSF, 1:10; serum, 1:20).

Two blinded investigators (V.R., M.S.M.) assessed epitope and isotype specificities of anti-CASPR2 antibodies. The IgG and albumin concentrations in the CSF and serum were evaluated using nephelometry (IMMAGE Immunochemistry Systems; Beckman-Coulter). The CSF/serum albumin quotient was used to evaluate the integrity of the CSF-blood barrier. To characterize intrathecal immunoglobulin synthesis, the immunoglobulin quotient (immunoglobulin levels in the CSF divided by immunoglobulin levels in the serum) was calculated and adjusted to the corresponding maximum immunoglobulin quotient that can be expected at a given albumin quotient in the absence of intrathecal immunoglobulin synthesis9 to quantify the CSF/serum anti-CASPR2 antibodies index. An antibody index greater than 4.0 was considered to reflect intrathecal synthesis.9

Results
Clinical Findings

Among the 6650 CSF samples tested by immunohistochemistry, we identified 18 patients with anti-CASPR2 antibodies in the CSF. During the same period, we identified 15 patients with NMT or MoS and anti-CASPR2 antibodies in the serum. Samples of CSF from all patients with NMT or MoS were tested for anti-CASPR2 antibodies and found to be normal. The clinical findings, ancillary results, and outcomes of the patients with anti-CASPR2 antibodies in the CSF are shown in Table 1. Table 2 summarizes the baseline characteristics and immunological findings of patients with anti-CASPR2 antibodies in the CSF compared with patients with NMT or MoS. Compared with the group of patients with NMT or MoS, the cohort had a significantly higher number of males (94.4% vs 66.7%; P = .03)and median age (64.5 vs 51 years; P < .01). Only 1 patient (5.6%) in our cohort was diagnosed as having another autoimmune disorder (thyroiditis), whereas autoimmune comorbidities were seen in 8 of 15 patients (53.3%) with NMT or MoS (P < .01; 7 of these 8 had myasthenia gravis). Only 3 patients (16.7%) in our cohort had a previous or concomitant history of cancer (prostate adenocarcinoma and chronic lymphoid leukemia in 1 patient and papillary thyroid cancer in 2 patients). Although a malignant thymoma had been found prior to or following the neurological disorder in 9 of 15 patients (60.0%) with NMT or MoS, none of the patients with anti-CASPR2 antibodies in the CSF had such a tumor (P < .01).

For most patients with anti-CASPR2 antibodies in the CSF (13 of 18 [72.2%]), admission to the referring hospital occurred because of partial temporal seizures. In these patients, other symptoms installed acutely over the course of a few days to 1 week, simultaneously to epilepsy in 9 patients and only a few months after the onset of epilepsy in the 3 remaining patients (median [range] delay, 6 [5-18] months). In the remaining 5 patients (27.8%), admission was requested because of memory disorders, which had developed progressively over several months in 4 patients and acutely in 1. All patients in our cohort had symptoms of limbic structure impairment, including 17 (94.4%) with anterograde and episodic memory disorders, 16 (88.9%) with temporal lobe seizures, 10 (55.6%) with symptoms of frontal lobe dysfunction, and 4 (22.2%) with psychiatric symptoms (eg, depressed mood or persecutory thoughts). None of the patients presented with the acute confusion or behavioral disorders classically reported in autoimmune limbic encephalitis. Extralimbic symptoms were observed in 12 patients (66.7%), including 6 (33.3%) with cerebellar ataxia, 4 (22.2%) with sleep disorders, and 3 (16.7%) with peripheral neuropathy. Signs of NMT were present in only 2 patients (11.1%).

Ancillary

Hyponatremia was observed in only 1 of 18 patients (5.6%). Anti–SOX1 antibodies were detected in 1 patient (5.6%), although no evidence of cancer was demonstrated. None of the other patients had additional antibodies against neuronal antigens. Brain magnetic resonance imaging (MRI) data were available for 15 patients (83.3%). Of these 15 patients, brain MRI results remained normal in only 1 patient (6.7%). Twelve patients (80.0%) had temporolimbic hyperintensities on T2-weighted images (Figure 1), which appeared subsequently in 3 patients. Bilateral hippocampal atrophy was observed in 2 of 15 patients (13.3%) on the first imaging analysis. In 1 patient, temporomesial hyperintensities were observed in the first MRI and secondarily evolved into bilateral hippocampal atrophy. Cytochemical data analysis results in the CSF were available for 12 patients and were normal in only 3 patients (25.0%). Increased white blood cell counts were observed in 8 of 12 patients (66.7%; median [range], 4000 [3000-32 000] cells/μL [to convert to 109/L, multiply by .001]), and elevated protein levels were observed in 1 patient (8.3%; 0.11 g/dL). Of note, brain MRI and CSF analyses were normal in all 11 patients with NMT and all 9 with MoS for whom data were available. Interictal electroencephalograms displayed focal abnormalities in 4 of 18 patients (22.2%). Brain positron emission tomography scans were performed in 5 patients (27.8%) and were pathological in 4. Positron emission tomography scan findings were not reproducible from one patient to another; 1 patient had diffuse hypometabolism, another had temporomesial hypometabolism, and 2 patients had frontotemporal hypermetabolism. Therefore, we were unable to determine a pattern specifically associated with CASPR2 antibody–related encephalitis. Nonetheless, it has to be noted that in contrast with patients with anti–leucin-rich glioma 1 proteinencephalitis,10 increased metabolism in the basal ganglia was not observed in any of the 5 patients.

Treatments

Oral steroids were given to a total of 4 patients. One patient was treated only with oral steroids, while 3 patients were treated with oral steroids combined with other treatments. Eleven of 18 patients (61.1%) were treated with intravenous immunoglobulins (0.4 g/kg/d for 3 to 5 days as part of 1 to 9 monthly courses), 8 (44.4%) with high-dose corticosteroid boluses (0.5-1 g/d for 3 to 5 days as part of 1 to 6 monthly courses), 1 (5.6%) with plasmapheresis (6 monthly courses), 3 (16.7%) with intravenous cyclophosphamide (1-g infusion every month for 2 to 8 months), and 2 (11%) with rituximab (375 mg/m2 administered intravenously once a week for 1 month). Three patients (16.7%) were given mycophenolate mofetil (1 g/24 h administered orally) on a long-term basis. Four patients (22.2%) were given 2 lines of treatment, and 2 patients (11.1%) needed 3 different lines of treatment.

Follow-up and Outcomes

The median (interquartile range) follow-up time was 27.4 (19-51) months. All patients survived. Follow-up was performed for at least 6 months in 16 of 18 patients (88.9%). Of these, 15 (93.8%) progressively improved, with 11 improving after immunomodulatory treatment and 4 spontaneously (Table 1). One patient remained stable with ataxia and severe cognitive defects. Six of 16 patients (37.5%) had 1 or 2 relapses. Five of these patients had relapses presenting as an increase in seizure frequency. In 2 patients, epileptic relapses occurred when steroid medications were decreased and were easily controlled by increasing steroid dosages. The relapse of 1 patient consisted of the development of NMT 8 years after the onset of encephalitis, without concomitant worsening of the encephalitic symptoms. The median (interquartile range) modified Rankin score for all patients was 2 (2-3) at onset and 1 (1-2) at the end of follow-up. At the end of follow-up, 4 of 16 patients (25.0%) still had symptoms that significantly altered their quality of life (modified Rankin score ≥2), including NMT (1 patient), cerebellar ataxia (3 patients), and amnesia (3 patients). Two of 16 patients (12.5%) recovered completely. Brain MRI follow-up data were available for 9 patients. The MRI results remained stable in 6 of 9 patients (66.7%), while temporomesial hyperintensities decreased partially or completely in 2 patients (22.2%) and evolved from unilateral to bilateral in 1 patient (11.1%). One patient (11.1%) secondarily developed bilateral hippocampal atrophy.

Anti-CASPR2 Antibody Characteristics

As previously reported,11 studies on immunohistochemistry in the CSF revealed a similar staining pattern in all patients with strong fixation of the granular and molecular layers of the cerebellar cortex and hippocampus along with a diffuse staining of the molecular layer of the dentate gyrus (Figure 2). The median anti-CASPR2 antibodies end point dilution in the CSF was 1/1280 (range, 1/80-1/10 240). All serum samples from the patients were also positive for anti-CASPR2 antibodies. The median anti-CASPR2 antibodies end point dilution in the serum were significantly higher in patients in our cohort than in patients with NMT or MoS (1/15 360 vs 1/800; P < .001). Albumin levels in the CSF and serum were available for 2 patients, allowing for the assessment of intrathecal synthesis. In both of the patients, the immunoglobulin quotient values (0.03 and 0.06) were larger than the maximum immunoglobulin serum/CSF quotient values, calculated using the Reiber hyperbolic function9 (0.004 and 0.011). The CASPR2 antibody indices were 7.4 and 5.6, suggesting intrathecal synthesis of anti-CASPR2 antibodies in both patients. In all patients, anti-CASPR2 antibodies levels in the serum and CSF were directed against the discoidin and laminin G1 domains of CASPR2, as illustrated in Figure 3A. IgG isotypes were determined in 17 patients (Figure 3B). Importantly, IgG4 antibodies were detected in the CSF of all patients along with IgG1 antibodies in 10 of 17 patients (58.8%). Interestingly, IgG4 anti-CASPR2 antibodies were found in the serum of 11 of 12 patients (91.7%) with anti-CASPR2 antibodies in the CSF and in only 5 patients (41.7%) with NMT or MoS (P = .03). We observed 3 patients with anti-CASPR2 antibodies in the CSF that were of the IgG4 but not IgG1 isotype, which targeted the discoidin and laminin G1 domains only. All 3 of these patients had anterograde and episodic amnesia and temporal seizures. One patient also reported persecutory thoughts, and another patient exhibited hypersomnia. Notably, all 3 patients with hippocampal atrophy on brain MRI had anti-CASPR2 antibodies in the CSF of both the IgG4 and IgG1 isotypes.

Discussion
Clinical Patterns

It has been previously reported4,11-13 that patients with anti-CASPR2 antibodies present with symptoms of MNT, MoS, and various patterns of autoimmune encephalitis. However, our study indicates that the presence of anti-CASPR2 antibodies in the CSF is associated with a homogeneous clinical pattern of autoimmune encephalitis with prevalent limbic involvement and seizures. As reported before,1,11 we observed an overrepresentation of male patients in our cohort, which was significantly higher than in patients with NMT or MoS. Epilepsy was the main symptom that led the patients to seek medical attention. Neither confusion nor behavioral disorders were observed in any of the patients. All patients had anterograde or episodic memory disorders and/or temporal seizures, which suggests a constant involvement of the hippocampus. Brain MRIs were in accordance with this result, as most patients displayed temporomesial hyperintensities or hippocampal atrophy on T2-weighted images (Figure 1). Nonetheless, extralimbic symptoms were observed as well, the most frequent being cerebellar ataxia, in accordance with previous results.12 Neuromyotonia was rare in our cohort. Relapses were observed in 6 patients, most often in the form of an increase in seizure activity. In some cases, this increased seizure activity was steroid-dependent and steroid-responsive, indicating the possible direct responsibility of anti-CASPR2 autoimmunity for the epileptogenic process. In contrast with patients with NMT or MoS, none of the patients in our cohort had a malignant thymoma.

Our data suggest that the presence of anti-CASPR2 antibodies in the CSF is associated with a subtype of autoimmune encephalitis that is mostly nonparaneoplastic. By comparison, patients with NMT or MoS do not have detectable anti-CASPR2 antibodies in the CSF and are frequently affected by a malignant thymoma. The absence of detectable anti-CASPR2 antibodies in the CSF of patients with NMT or MoS could reflect an autoimmune reaction strictly developed outside of the central nervous system. Alternatively, it might be because of extremely low titers of anti-CASPR2 antibodies in the CSF, as serum titers were much lower in patients with NMT or MoS. Therefore, variations of the autoantibodies production within the intrathecal compartment are likely to underline the observed differences in clinical presentation between the 2 groups.

Epitope Specificities

Interestingly, the anti-CASPR2 antibodies in the CSF from all patients targeted the discoidin and laminin G1 domains of CASPR2, suggesting that recognition of these domains is involved in the genesis of symptoms. To our knowledge, the precise roles of the discoidin and laminin G1 domains have not yet been clearly determined.

IgG Subtypes

Importantly, IgG4 anti-CASPR2 antibodies were detected in the CSF of all patients. IgG4 antibodies cannot bind to complement or form immune complexes and have a low affinity for Fc receptors.14,15 Moreover, IgG4 antibodies have the ability to exchange fragment antigen-binding arms, and a fraction of them are therefore bispecific, suggesting that some of the anti-CASPR2 antibodies of our patients may have bispecificity for the laminin G1 and discoidin domains.16,17 Interestingly, fragment antigen-binding fragments of anti-Musk IgG4 autoantibodies from patients with myasthenia gravis have been shown to be pathogenic in vitro, suggesting the involvement of mechanisms other than the classical cross-linking and internalization that was observed with other autoantibodies directed against synaptic proteins.18 It remains to be seen whether IgG4 anti-CASPR2 antibodies are bispecific for the discoidin and laminin G1 domain and whether they are able to cross-link CASPR2 and induce its internalization or alter its function by other mechanisms. Additionally, IgG1 autoantibodies were found in all the patients with hippocampal atrophy, implying the possible involvement of complement or antibody-dependent cell death. This view is supported by 1 biopsied case of CASPR2 antibody encephalitis in which immunoglobulin and complement depositions were reported in neurons along with cell degeneration in the hippocampus.19 Therefore, distinct pathological processes can occur during the course of CASPR2 antibody encephalitis, ranging from functional alterations involving IgG4 autoantibodies to complement-mediated cell toxicity induced by IgG1 autoantibodies.

This study has several limitations. Because of the rarity of the disease, we only analyzed retrospective data from a small series of patients. We also cannot exclude low titers of anti-CASPR2 antibodies in the CSF of patients with NMT or MoS not detected by our methods. Our results need to be confirmed in futures studies.

Conclusions

In conclusion, anti-CASPR2 antibodies are found in the CSF only in patients with CASPR2 antibody–associated autoimmune encephalitis, whereas anti-CASPR2 antibodies are detected only in the serum in patients with NMT or MoS. Contactin-associated protein-like 2 antibody–associated encephalitis is observed predominantly in males and is characterized by prominent limbic symptoms and temporal lobe epilepsy. Malignant thymoma was found only in patients with NMT or MoS. Recognition of the discoidin and laminin G1 domains by the autoantibodies is constant and may be important in the genesis of limbic symptoms. Additionally, anti-CASPR2 antibodies of the IgG4 isotype may play a fundamental role in the pathogenic processes through a functional effect on CASPR2, although complement-mediated cell death induced by IgG1 autoantibodies may also occur in some patients. These hypotheses must be further supported by future functional studies focused on antibody specificities.

Back to top
Article Information

Corresponding Author: Jérôme Honnorat, MD, PhD, Centre de Référence National pour les Syndromes Neurologiques Paranéoplasiques, Hôpital Neurologique, 59 Boulevard Pinel, 69677 Bron Cedex, France (jerome.honnorat@chu-lyon.fr).

Accepted for Publication: April 8, 2016.

Published Online: July 18, 2016. doi:10.1001/jamaneurol.2016.1585.

Author Contributions: Drs Honnorat and Joubert had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Joubert, Honnorat.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Joubert, Honnorat.

Critical revision of the manuscript for important intellectual content: All authors.

Obtained funding: Honnorat.

Administrative, technical, or material support: Rogemond, Delattre.

Study supervision: Honnorat.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study is supported by research grant ANR-14-CE15-0001-MECANO from the L’Agence Nationale de la Recherche and the 2014 Abnormal Behavior and Immune Dysfunction grant from the Fondation pour la Recherche sur le Cerveau.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank the referral physicians who kindly transmitted the exhaustive clinical and ancillary information that allowed us to perform this study: Lejla Koric, MD (Assistance Publique–Hôpitaux de Marseille, Marseille, France); Isabelle Lambert, MD (Assistance Publique–Hôpitaux de Marseille, Marseille, France); Cécile Marchal, MD (Centre Hospitalier Universitaire de Bordeaux, Bordeaux, France); Philippe Kassiotis, MD (Centre Hospitalier Bretagne Atlantique, Vannes, France); Fabienne Ory-Magne, MD (Centre Hospitalier Universitaire de Toulouse, Toulouse, France); Jérémie Pariente, MD (Centre Hospitalier Universitaire de Toulouse, Toulouse, France); Jonathan Curot, MD (Centre Hospitalier Universitaire de Toulouse, Toulouse, France); Pierric Giraud, MD (Centre Hospitalier Annecy Genevois, Metz-Tessy, France); Laurent Martinez-Almoyna, MD (Centre Hospitalier d’Aix en Provence, Aix-en-Provence, France); Anne-Céline Zeghoudi, MD (Centre Hospitalier de Versailles, Versailles, France); Sylvain Rheims, MD, PhD (Hospices Civils de Lyon, Lyon, France); Ana Gales, MD (Assistance Publique–Hôpitaux de Paris, Paris, France); Charles Behr, MD (Centre Hospitalier Régional Universitaire de Strasbourg, Strasbourg, France); Eric Thouvenot, MD (Centre Hospitalier Universitaire de Nîmes, Nimes, France); Gil Petitnicolas, MD (Centre Hospitalier Intercommunal de Toulon, Toulon, France); and Claire Boutoleau-Bretonniere, MD (Centre Hospitalier Universitaire de Nantes, Nantes, France). None of these contributors were compensated for their work.

References
1.
Irani  SR, Alexander  S, Waters  P,  et al.  Antibodies to Kv1 potassium channel-complex proteins leucine-rich, glioma inactivated 1 protein and contactin-associated protein-2 in limbic encephalitis, Morvan’s syndrome and acquired neuromyotonia.  Brain. 2010;133(9):2734-2748.PubMedGoogle ScholarCrossref
2.
Malter  MP, Frisch  C, Schoene-Bake  JC,  et al.  Outcome of limbic encephalitis with VGKC-complex antibodies: relation to antigenic specificity.  J Neurol. 2014;261(9):1695-1705.PubMedGoogle ScholarCrossref
3.
Irani  SR, Pettingill  P, Kleopa  KA,  et al.  Morvan syndrome: clinical and serological observations in 29 cases.  Ann Neurol. 2012;72(2):241-255.PubMedGoogle ScholarCrossref
4.
Laurencin  C, André-Obadia  N, Camdessanché  JP,  et al.  Peripheral small fiber dysfunction and neuropathic pain in patients with Morvan syndrome.  Neurology. 2015;85(23):2076-2078.PubMedGoogle ScholarCrossref
5.
Poliak  S, Salomon  D, Elhanany  H,  et al.  Juxtaparanodal clustering of Shaker-like K+ channels in myelinated axons depends on Caspr2 and TAG-1.  J Cell Biol. 2003;162(6):1149-1160.PubMedGoogle ScholarCrossref
6.
Bel  C, Oguievetskaia  K, Pitaval  C, Goutebroze  L, Faivre-Sarrailh  C.  Axonal targeting of Caspr2 in hippocampal neurons via selective somatodendritic endocytosis.  J Cell Sci. 2009;122(pt 18):3403-3413.PubMedGoogle ScholarCrossref
7.
Pinatel  D, Hivert  B, Boucraut  J,  et al.  Inhibitory axons are targeted in hippocampal cell culture by anti-Caspr2 autoantibodies associated with limbic encephalitis.  Front Cell Neurosci. 2015;9:265.PubMedGoogle ScholarCrossref
8.
Olsen  AL, Lai  Y, Dalmau  J, Scherer  SS, Lancaster  E.  Caspr2 autoantibodies target multiple epitopes.  Neurol Neuroimmunol Neuroinflamm. 2015;2(4):e127.PubMedGoogle ScholarCrossref
9.
Reiber  H, Lange  P.  Quantification of virus-specific antibodies in cerebrospinal fluid and serum: sensitive and specific detection of antibody synthesis in brain.  Clin Chem. 1991;37(7):1153-1160.PubMedGoogle Scholar
10.
Navarro  V, Kas  A, Apartis  E,  et al.  Motor cortex and hippocampus are the two main cortical targets in LGI1-antibody encephalitis.  Brain. 2016;139(pt 4):1079-1093.PubMedGoogle ScholarCrossref
11.
Lancaster  E, Huijbers  MG, Bar  V,  et al.  Investigations of caspr2, an autoantigen of encephalitis and neuromyotonia.  Ann Neurol. 2011;69(2):303-311.PubMedGoogle ScholarCrossref
12.
Becker  EBE, Zuliani  L, Pettingill  R,  et al.  Contactin-associated protein-2 antibodies in non-paraneoplastic cerebellar ataxia.  J Neurol Neurosurg Psychiatry. 2012;83(4):437-440.PubMedGoogle ScholarCrossref
13.
Sunwoo  JS, Lee  ST, Byun  JI,  et al.  Clinical manifestations of patients with CASPR2 antibodies.  J Neuroimmunol. 2015;281:17-22.PubMedGoogle ScholarCrossref
14.
Bindon  CI, Hale  G, Brüggemann  M, Waldmann  H.  Human monoclonal IgG isotypes differ in complement activating function at the level of C4 as well as C1q.  J Exp Med. 1988;168(1):127-142.PubMedGoogle ScholarCrossref
15.
Bruhns  P, Iannascoli  B, England  P,  et al.  Specificity and affinity of human Fcgamma receptors and their polymorphic variants for human IgG subclasses.  Blood. 2009;113(16):3716-3725.PubMedGoogle ScholarCrossref
16.
Schuurman  J, Van Ree  R, Perdok  GJ, Van Doorn  HR, Tan  KY, Aalberse  RC.  Normal human immunoglobulin G4 is bispecific: it has two different antigen-combining sites.  Immunology. 1999;97(4):693-698.PubMedGoogle ScholarCrossref
17.
van der Neut Kolfschoten  M, Schuurman  J, Losen  M,  et al.  Anti-inflammatory activity of human IgG4 antibodies by dynamic Fab arm exchange.  Science. 2007;317(5844):1554-1557.PubMedGoogle ScholarCrossref
18.
Huijbers  MG, Zhang  W, Klooster  R,  et al.  MuSK IgG4 autoantibodies cause myasthenia gravis by inhibiting binding between MuSK and Lrp4.  Proc Natl Acad Sci U S A. 2013;110(51):20783-20788.PubMedGoogle ScholarCrossref
19.
Körtvelyessy  P, Bauer  J, Stoppel  CM,  et al.  Complement-associated neuronal loss in a patient with CASPR2 antibody-associated encephalitis.  Neurol Neuroimmunol Neuroinflamm. 2015;2(2):e75.PubMedGoogle ScholarCrossref
×