“Autoimmune Epilepsy”: Encephalitis With Autoantibodies for Epileptologists

Abstract

Autoimmune encephalitides may account for epilepsies of so far unknown cause. These “autoimmune epilepsies” may respond well to immunotherapy. More than a dozen autoantibodies have been found with this constellation; therefore, broad autoantibody testing of serum-CSF pairs offers the best diagnostic yield. Several particular features raise the suspicion of an autoimmune cause in otherwise unexplained seizure disorders.

Commentary

One of the most important challenges of scientific and clinical epileptology is the elucidation of the causes of seizure disorders: in a general sense and on the individual patient level. In recent years, neuroimmunology has contributed to these efforts by identifying immunoglobulin G (IgG) antibodies against neural antigens in patients with so far unrecognized autoimmune encephalitides, which often lead to recurrent seizures. This review article does not try to give a comprehensive review of what has been achieved in this field. The authors of this review article, two neurologists with long-lasting expertise in clinical epileptology, one of whom has particular interest in the diagnosis of neural antibodies, try to give a clinically meaningful overview and useful management suggestions for autoimmune encephalitides with predominant epileptic features, that is, the “autoimmune epilepsies” (1).

Some Basic Facts

In the first decade of this century, IgG autoantibodies against proteins on the surfaces of neurons were identified as markers and pathogens in autoimmune encephalitides that in approximately 80% of cases are accompanied by repetitive focal seizures (1). Apart from those surface targets, intracellular antigens (most of them first identified in the 1980s and 1990s) account for a relevant, albeit smaller number of antibody associated encephalitides: the onconeural antibodies–those potentially occurring with seizures are directed against Hu, Ma, CV2, amphiphysin, Delta/Notch-like EGF-related receptor (previously: Tr), or Sox1 and those against glutamic acid decarboxylase (GAD). For an overview of autoantibodies with epileptologic relevance, see Figure 1.

Figure 1.

Autoantigens in autoimmune encephalitides with seizures.

From an epileptologic viewpoint, a highlight was the discovery of a new seizure type, faciobrachial dystonic seizures (FBDS), by a team of neuroimmunologists and epileptologists based on the demonstration of antibodies against the voltage-gated potassium channel (VGKC) complex; the antigenic target was later identified as Leucine-rich, glioma inactivated 1 (LGI1) (2, 3), an element of the VGKC complex.

The pathophysiology of antibody-defined autoimmune encephalitides, including their epilepsy-related aspects, has been studied by in vitro and in vivo experiments. The results suggest that the antibodies against surface antigens contribute directly to the disease processes. This issue has been reviewed by others recently (4).

Testing for autoantibodies has been incorporated into practical neurology and epilepsy because many tests have been shown to be reliable and specific for distinct phenotypes. The key technique for the demonstration of antibodies against the N-methyl-D-aspartate receptor (NMDAR), LGI1, contactin-associated protein-like 2 (CASPR2), the gamma-amino-butyric B receptor (GABA_BR), and other surface antigens is the incubation of patient's diluted serum or CSF (undiluted or mildly diluted) with human embryonic kidney (HEK) cells transfected with the antigens of interest. Potential antibody binding is then visualized with a “secondary” antibody against human IgG carrying a fluorochrome or other chemicals permitting the detection of a surface staining of the HEK cells under the microscope (5).

Other techniques include testing for autoantigen binding on rodent brain, immunoblotting, or radioimmunoprecipitation. All investigations follow the same principle: binding of autoantibodies to the matrix is detected by an anti-human IgG antibody. Formal requirements for making antibody diagnoses are given in Table 1.

TABLE 1.

Formal requirements for making antibody diagnoses

Antibody-Negative Autoimmune Epilepsy

Some patients may have phenotypes typical of autoimmune epilepsy (e.g., an acute, unexplained epilepsy onset with accompanying neuropsychiatric features; elevated cell count or intrathecal IgG synthesis demonstrated by CSF studies; or inflammatory brain lesions according to neuroimaging or histopathology) in the absence of a known autoantibody despite extensive testing. These patients may benefit from immunotherapy (8). In the absence of a reliable biomarker, it is difficult to gain systematic and general insights into these patients' diseases. It can be hoped that detection of novel antibodies or other encephalitis markers will decrease the number of patients in this group.

Which Epilepsy Patients Should be Tested, and How?

The answer to this “deontological” question is only in part a biological–medical one. It rather depends on how one weighs the costs of the diagnostic evaluations, the likelihood of having a positive result, and the risks of immunotherapies in antibody-positive cases. For some of these issues, only limited data are available. There are, however, certainly clinical circumstances in which a positive test result is more likely than in others. A general suggestion for the identification of autoimmune encephalitides prior to autoantibody diagnostics has recently been published (9). Features suggestive of autoimmune epilepsies are summarized in Table 2. In recent years, several authors have recommended to test CSF and serum at the same time (11–13). They have done so because in some instances, antibodies may not be detectable in both materials, for example, NMDAR antibodies not always in serum (14) and LGI1 antibodies not always in CSF (12, 15). Exceptions from the general suggestion to test CSF-serum pairs may be young females with an encephalopathic presentation suggestive of anti-NMDAR encephalitis (CSF sufficient) or faciobrachial dystonic seizures (serum for LGI1 antibody testing usually sufficient). If the results of those specific tests are negative, there may be need for subsequent more extensive testing.

TABLE 2.

Features of patients with seizures and increased likelihood of having autoimmune epilepsy (modified from [10])

Similarly, the selection of antibodies to be tested for should probably not be too small. The phenotypic features of autoimmune epilepsy rarely permit a targeted testing for a specific antibody, except for a typical presentation of anti-NMDAR encephalitis and the faciobrachial dystonic seizures seen with anti-LGI1 antibodies. In most other cases, a broader approach should be taken to avoid time consuming and expensive sequential testing for single antibodies. One strategy is the use of rodent brain as a “general” assay to select for samples containing any antibody before subtyping it by cell-based assays (taking into account the staining pattern on rodent brain) as suggested by the Philadelphia and Barcelona groups (16). Another approach is the use of multivalent panels of specific tests in form of commercially available mosaics of cell-based assays presenting surface antigens plus multivalent immunoblots with intracellular antigens (17).

What Does a “Positive Result” Mean?

The clinician is most interested in antibodies with two characteristics: 1) confirmed specificity for distinct syndromes (two criteria fulfilled: the antibodies do not or only exceptionally occur in persons with other or without neurological disease; and, the syndromic specificity has been found in more than one series), and 2) predictive potency for responsiveness to immunologic therapy.

Using these criteria, not all neural autoantibodies have the same “value.” Here, we propose a simple dichotomy of “high-rank antibodies” (reproducibly specific and treatment responsive) and “grey cases” (not fulfilling both criteria) (Table 3). In some cases, the site of antibody detection (blood or CSF) makes the difference between the two categories. In the absence of comparative studies across all antibodies, this classification has a certain degree of arbitrariness, and may require refinement with future data.

TABLE 3.

“High-rank antibodies” versus “grey cases” according to the published case series (N ≥ 10 where available)

TABLE 3. Continued.

“High-rank antibodies” versus “grey cases” according to the published case series (N ≥ 10 where available)

Applying these categories, the percentage of epilepsy patients with “high rank antibodies” in series of patients without encephalitic features is less than 5 percent, with a larger proportion of “grey cases” of less clear specificity (18–21). The proportion of positive antibodies goes up with inclusion of patients clinically suggestive of having an encephalitis as a cause of their recurrent seizures, but again, most series contain “grey cases” of unknown significance (22, 23). To conclude, autoimmune epilepsy is rare in the whole epilepsy population unless features of Table 2 are considered. In anti-NMDAR encephalitis, monosymptomatic cases account for only 1 percent in the largest available series (“seizures only” being just a part of this fraction) (24). Even if one concedes some sampling bias in favor of more severely (polysymptomatically) affected patients, this seems to be characteristic for many antibodies. In contrast, there may be prolonged disease periods with “seizures only” with LGI1 antibodies (18% [25]). One or more “seizure(s) only” as a presenting symptom are frequent with CASPR2 antibodies (24% [26] to 72% [27]) and in male patients with anti-NMDAR encephalitis (27% [28] to 62% [29]); in these patients, other CNS symptoms evolved within 1 to 17 days (29, 27) emphasizing again the importance of antibody testing especially in patients with complex phenotypes.

Which Treatment to Choose?

The conditions associated with antibodies against surface antigens often have a favorable prognosis with immunotherapy. From an epilepsy standpoint, this is particularly striking with FBDS. They hardly respond to antiepileptic drugs, and these medications surprisingly often cause allergic reactions in this condition. On the other hand, FBDS respond well to steroids (55). In the absence of prospective, let alone randomized trials, therapeutic recommendations have relied on retrospective evaluations. The most influential one comes from the experience in anti-NMDAR encephalitis (56), which suggested the sequence of first line therapies (steroids, intravenous immunoglobulin, or apheresis therapy–or combinations of these interventions) and second line therapies (cyclophosphamide and/or rituximab). There are no evidence-based dosing recommendations. Recently, rituximab (57, 58) and apheresis techniques (59, 60) have received increasing attention.

It is to be expected that local preferences, individual patient features (i.e., special risk situations), reimbursement issues, and personal experience of treating physicians will influence the therapeutic sequence. A suggested sequence of interventions in cases with “high-rank antibodies” is given in Figure 2. In addition to prednisolone, an apheresis technique may be used to reduce the antibody burden rapidly. Prior to the oral prednisolone, a course of three to five pulses of 500- to 1000-mg methylprednisolone may be given on consecutive days. For the “grey cases,” a similar approach has been suggested recently (8).

Figure 2.

Suggestion for a treatment pathway in patients with “high-rank antibodies.” This treatment pathway is used in the author's (CGB) institution. It is based on previous recommendations (56).

Summary, Final Remarks

Autoimmunity by antibodies is causative for a small proportion of epilepsies. The identification of such cases is, nevertheless, of importance because many of the affected patients have a favorable prognosis with immunotherapy. Neural autoantibodies can be specifically diagnosed by established techniques. Clinical challenges include identifying which patients should undergo antibody testing and delineating the most appropriate therapeutic interventions. Preliminary, readily applicable suggestions exist, and will continue to be refined with increased knowledge over time.

Conflicts of interest

CGB gave scientific advice to Eisai (Frankfurt, Germany) and UCB (Monheim, Germany), undertook industry-funded travel with support of Eisai (Frankfurt, Germany), UCB (Monheim, Germany), Desitin (Hamburg, Germany), and Grifols (Frankfurt, Germany), obtained honoraria for speaking engagements from Eisai (Frankfurt, Germany), UCB (Monheim, Germany), Desitin (Hamburg, Germany), diamed (Köln, Germany), Fresenius Medical Care (Bad Homburg, Germany), and Biogen (Ismaning, Germany). He received research support from diamed (Köln, Germany) and Fresenius Medical Care (Bad Homburg, Germany). He is a consultant to the Laboratory Krone, Bad Salzuflen, Germany, regarding neural antibodies and therapeutic drug monitoring for antiepileptic drugs.

MH received speaker's honoraria and/or consultancy fees from Cyberonics, Desitin, Eisai, GlaxoSmithKline, Janssen-Cilag, Novartis, UCB and Viropharma.