The aim and main findings of the present study

The current study was performed to try resolving the mystery of two intractable Epilepsy patients, and test for evidences of ‘Autoimmune Epilepsy’. These Epilepsy patients were enigmatic for many years before this study because of the negative results of numerous tests performed in the hospital prior to this study, specified in Table 2. Our study revealed several findings regarding the Epilepsy patients studied, which are listed in brief in this section, and schematically drawn in the Graphical Abstract (Fig. 13). Some of our discoveries and their conclusions, and citation of relevant studies, are discussed further in separate paragraphs, below the list of findings.

Fig. 13

Graphical summary of the current study on ‘autoimmune Epilepsy’, and of all its findings

1st finding: The Epilepsy patients we studied have elevated IgG antibodies in blood.

2nd finding: The Epilepsy patients have elevated levels of three types of Glutamate receptors antibodies, namely: GluR3B peptide antibodies, NMDA-NR1 peptide antibodies, and NMDA-NR2 peptide antibodies. Previous studies have shown that each of these GluR antibodies can bind and damage neural cells in tissue culture, and cause brain damage In vivo [reviewed in Levite and Goldberg (2021)]. We elaborate on the pathogenic activity of these GluR antibodies in a separate section below.

3rd finding: The Epilepsy patient’s IgG antibodies bind human neural cells In vitro.

4th finding: The Epilepsy patient’s IgG antibodies kill human neural cells In vitro.

5th finding: The Epilepsy patient’s IgG antibodies induce repeated GTCs in naïve rats, following their continuous infusion into naïve rat’s brain ventricular region.

6th finding: The Epilepsy patient’s IgG antibodies bind rat neurons and astrocytes in the cortex and hippocampus CA3 region of naïve rats.

7th finding: The Epilepsy patient’s IgG antibodies induce neuronal loss in the cortex and hippocampus CA3 region of naïve rats.

The patient’s IgG antibodies induced GTCs in naïve rats

We found that the type of seizures generated by the purified IgG antibodies of the Epilepsy patients IE-3 and IE-15 in naïve rats was typical of GTCs. This conclusion is based on the fact that the repeated seizures featured an Ictal EEG showing generalized polyspikes, building up to an increasing frequency in the tonic phase, followed by a generalized spike-wave pattern that gradually slows to a lower frequency (Smith 2005), and eventually a postictal suppression follows termination of the seizure.

We further found that the average duration of the seizures induced by the Epilepsy patient’s IgG antibodies was 68 s. The seizure duration ranged between 20 and 90 s, which is typical of the duration of GTCs that last up to 90 s or less, as described by Britton et al. (2016).

Wei et al. (2021) employed the use of EEG monitoring for the detection of spontaneous seizures in multiple experimental mouse models of Epilepsy. In that study, when the duration of the seizure event was greater than 10 s, it was defined as a seizure (Wei et al. 2021).

The type, duration and recurrent seizures induced by the Epilepsy patient’s IgG antibodies in naïve rats in the present study strongly suggest that the patient’s autoimmune IgG antibodies are inducing such seizures, i.e. recurrent GTCSs in the patient’s own body.

If this happens, these Epilepsy patients indeed suffer from ‘Autoimmune Epilepsy’.

EEG proves invaluable in assessing individuals suspected of seizures, Epilepsy or experiencing atypical episodes. In nearly all cases of Epilepsy, distinct EEG changes manifest during seizures (ictal recordings). Furthermore, many Epilepsy patients exhibit identifiable interictal epileptiform discharges (IEDs), such as spikes (lasting < 70 s), spike and wave, or sharp-wave discharges (lasting 70–200 s) (Paxinos and Watson 2014).

Our study and findings in naïve rats show that testing Epilepsy patients’ purified IgG antibodies (not serum!) In vivo, in similar animal models to the ones we developed, seems to be an essential and appropriate method to find out if a given Epilepsy patient has seizure-inducing autoimmune antibodies, and prove their pathogenicity, and as such to diagnose that the patient suffers from ‘Autoimmune Epilepsy’ and that he/she should be treated accordingly. The sophisticated video EEG tests we performed in naïve animals are expensive and complex and require profound unique knowledge, specific experience and appropriate very costly equipment. As such, they are not in hand in the majority of hospitals. Due to that, we recommend that whenever possible, small serum samples of patients with intractable enigmatic Epilepsy will be sent for In vitro and In vivo investigation, to external laboratories in the Academic Institutes or elsewhere, that have, or will be able to establish, animal models such as the video EEG setup we have established, and used successfully.

Brain damage in the cortex and the hippocampus CA3 region, as caused in this study in rats by IgG antibodies of Epilepsy patients, can induce seizures and other neurological problems in the patients themselves

Paroxysmal alteration of neurological function caused by an excessive hypersynchronous neuronal discharge in the brain is known as seizure. The hippocampus, amygdala, frontal cortex, temporal cortex, and olfactory cortex are the common areas involved in seizures (Chauhan et al. 2022). Brain lesions are the most common cause of adult‐onset Epilepsy (Nordberg et al. 2023).

The brain’s cortex. The cerebral cortex, also called the gray matter, is the brain’s outermost layer of nerve cell tissue. The cerebral cortex carries out essential brain functions, including memory, thinking, learning, reasoning, problem-solving, emotions, consciousness, and sensory functions. In general, lesions in the cerebral cortex are associated with a higher risk for Epilepsy than lesions in other parts of the brain (Pitkanen et al. 2016; Kerkhof and Vecht 2013).

The hippocampal CA3 brain region. The hippocampus CA3 region is important for rapid encoding of memory (Rebola et al. 2017). In recent years, the CA3 region has attracted major attention not only for its central role in memory, but also for its specific role in susceptibility to seizures and neurodegeneration [see for example a paper entitled “The CA3 region of the hippocampus: how is it? What is it for? How does it do it?”(Cherubini and Miles 2015)]. Internal connectivity in the CA3 subfield is richer than in other hippocampal regions (Cherubini and Miles 2015). Recurrent axon collaterals of CA3 pyramidal cells ramify extensively, making excitatory contacts with neighboring excitatory and inhibitory neurons. Evidence obtained in an animal model shows that the CA3 excitatory output is required for both the generation of epileptiform oscillatory activity and the progression of behavioral seizures (Yu et al. 2016).

Taken together, multiple direct and indirect evidences described in reliable papers show that damage in the cortex and the hippocampus CA3 region can lead to, or at least contribute to, outbursts of seizures and multiple other impaired brain functions.

Based on these evidences, we conclude that autoimmune antibodies—Glutamate receptor antibodies and/or others—of Epilepsy patients, that damage neural cells in the cortex and hippocampus CA3 region, as the purified IgG of the Epilepsy patients IE-3 and IE-15 induced in naïve rats in the present study, can induce or contribute to the patient’s own seizures.

In addition, we suspect that Epilepsy patient’s autoimmune antibodies that damage the patient’s cortex and the hippocampus CA3 region could not only induce seizures, but also: impair multiple other brain functions, among them thinking, learning, reasoning, understanding, memorizing, solving problems and others, and cause abnormal behavior, emotions, consciousness, and sensory functions, and even lead to psychiatric problems.

What can be studied in animal models cannot be studied in the patients themselves

A question that could not be answered in this study is whether the Epilepsy patient’s IgG antibodies indeed induce seizures and brain damage in the patient’s own body, like they did in naïve rats in the current study. This critical question will be left without a decisive and unequivocal answer, and only with a reasonable assumption that this is indeed the case, because it is impossible to carry out suitable In vivo research on the brain of a living Epilepsy patient, and isolate the seizures, brain damage and abnormal neural functions caused solely by the patient’s own autoimmune antibodies. Therefore, the development and usage of reliable and applicable animal models that make it possible to imitate what happens in the patients themselves, is so important.

The most important information for the Epilepsy patients themselves is whether they have pathological autoimmune antibodies, regardless of their antigenic specificity

The specific antigenic identity of the autoimmune antibodies of the Epilepsy patients IE-3 and IE-15 that induced all the pathological effects documented in this study by the patient’s IgG antibodies In vitro and In vivo was not determined in this study, since we studied and observed all the functional pathological effects, both in tissue culture and in the rats, with a purified preparation of all the IgG antibodies present in the patient’s sera.

Purified biological material that contains all the antibodies of IgG class surely contains IgG antibodies directed against many different non-self and self-antigens. That being said, let us argue that while testing a heterogeneous population of patient’s IgG antibodies, rather than only one type of autoimmune antibody to a known antigen, can be considered a scientific disadvantage, the most important information for the patients themselves, and to the clinicians that treat them, is whether or not the patient has any type of pathological autoimmune antibodies that induce seizures, neural death and impaired brain function. Their antigenic identity of the autoimmune antibodies is of value of course but of secondary importance, since available immunotherapies for most if not all autoimmune diseases are not tailored to autoimmune antibodies with a distinct antigenic specificity. Rather, they either silence, neutralize or purge many types of antibodies, with different specificities.

Glutamate receptor antibodies in intractable Epilepsy and Nodding syndrome

We suspect (but cannot prove) that in the case of the Epilepsy patients IE-3 and IE-15, it is their own Glutamate Receptor antibodies that bind neural cells, damage the brain and induce GTCSs. This suspicion is based on: (a) our finding that these Epilepsy patients have 3 types of GluR antibodies in their sera, and (b) these specific types of GluR antibodies, have already been shown to induce multiple pathological effects In vitro and In vivo [reviewed in Levite and Goldberg (2021) and summarized also in the next paragraphs].

Regretfully, we could not affinity-purify sufficient amounts of each type of the patient’s Glutamate rector antibodies from their own small blood samples, to perform all the functional studies also with them, and compare their effects to those induced by their total IgG antibodies, and by control antibodies of healthy subjects.

Various types of autoimmune antibodies to Glutamate receptors were identified in various diseases and shown to have pathological effects (Levite and Goldberg 2021; Levite 2014; Levite et al. 2020). Of all types of Glutamate receptor antibodies, the AMPA-GluR3B peptide antibodies seem to be the most exclusive to Epilepsy since they were found so far only in some patients with Epilepsy. Elevated levels of GluR3 antibodies, especially GluR3B peptide antibodies, were found so far in the serum of about 35% of > 370 persons with severe, intractable and enigmatic Epilepsy of various types (Rogers et al. 1994; Twyman et al. 1995; Andrews and McNamara 1996; He et al. 1998; Ganor et al. 2004; Levite 2014; Lai et al. 2022).

Recently, we also found elevated GluR3B peptide antibodies in 86% of young South Sudanese patients with the devastating and often fatal pediatric Epilepsy: Nodding syndrome (NS). Based on our own findings in this study (Levite et al. 2020) (summarized in the next section), and those of Johnson et al. (2017), we conclude that NS is most probably an ‘Autoimmune Epilepsy’. GluR3B peptide antibodies were also found in the CSF of some Epilepsy patients [see for example (Ganor et al. 2004)]. The GluR3B peptide antibodies are very pathogenic autoimmune antibodies and can on their own bind and then kill neural cells by three mechanisms: excitotoxicity, reactive-oxygen-species (ROS), and complement-fixation. The GluR3B peptide antibodies can also induce and/or facilitate brain damage, seizures and various types of behavioral impairments (Levite and Goldberg 2021; Levite 2014; Levite et al. 2020).

Autoimmune Glutamate receptor antibodies in Nodding syndrome.We recently published a study on autoimmunity in NS patients that consists of In vitro and In vivo investigations (Levite et al. 2020). NS is a catastrophic type of Epilepsy of unknown etiology, affecting children aged 3–18 years in three sub-Saharan countries including South Sudan, Tanzania and Uganda. The first clinical symptom is often an involuntary nodding of the head in a previously healthy child, potentially triggered by eating. The nodding episodes are thought to be one manifestation of a syndrome that includes various pathological features, among them the following: repetitive head nodding, seizures-usually GCTSs, neurological deterioration and cognitive impairments, stunted growth, weak muscle tone (hypotonia), frequent falling down, breathing problems, wasting and additional problems (Levite et al. 2020; Johnson et al. 2017; Dowell et al. 2013; Colebunders et al. 2023, 2017; Idro et al. 2018; Mwaka et al. 2018; Spencer et al. 2019, 2013, 2022). Eventually, NS leads to death.

In our recent study on 30 South Sudanese epileptic NS patients and healthy subjects (Levite et al. 2020) we found that most NS patients have elevated levels of autoimmune antibodies to three extracellular peptides of Glutamate receptors peptides: AMPA GluR3B peptide antibodies (elevated in 86% of patients), NMDA-NR1 peptide antibodies (77%) and NMDA-NR2 peptide antibodies (87%). These Glutamate receptor antibodies found in NS patients (Levite et al. 2020), are the same types of GluR antibodies we found in the serum of IE-3 and IE-15 in the present study.

We further found that the affinity-purified GluR3B antibodies of the NS patients by themselves (which we managed to purify from their blood, but not from the Epilepsy patients IE-3 and IE-15 which we investigated in the present study), induced reactive oxygen species (ROS) in human neural cells, and killed these cells in tissue culture (Levite et al. 2020).

Furthermore, video EEG experiments in normal mice, revealed that when the NS patient’s purified IgG antibodies were released continuously (24/7 for 1 week) in naive mouse brain, they induced all the following effects: (1) outburst of seizures, (2) cerebellar purkinje cell loss, (3) degeneration in the hippocampus and cerebral cortex, and (4) elevation of CD3+ T cells, and of activated Mac-2+ microglia and GFAP+ astrocytes in both the gray and white matter of the cerebral cortex, hippocampus, corpus calossum and cerebellum of naïve mice. These effects, induced In vivo by the NS patient’s IgG antibodies in naïve mice (Levite et al. 2020), are in line with the effects induced by the IgG antibodies of the Epilepsy patients IE-3 and IE-15 induced in naïve rats in the present study.

Together, these findings support ‘Autoimmune Epilepsy’ in very different Epilepsy patients.

Autoimmune GluR3 antibodies can cause profound behavior abnormalities

GluR3B antibodies can cause profound brain damage in addition to Epilepsy, including significant changes in recognition memory and impairments in social behavior and in social cognitive functions. This conclusion is based on findings in three different studies.

First, we previously found that GluR3B antibodies associate with some cognitive/ psychiatric/behavioral abnormalities in Epilepsy patients (Goldberg-Stern et al. 2014). In that study, we revealed that among the 21 patients with GluR3B antibodies, 19 patients (90%) had learning problems, 16 (76%) attention problems, and 15 (71%) psychiatric problems. In contrast, among the 20 patients without GluR3B, only 6 (30%) had learning problems (p < 0.0001), 5 (25%) attention problems (p = 0.0017), and 2 (10%) psychiatric problems (p < 0.0001) (Goldberg-Stern et al. 2014).

Second, in another published study we found that GluR3B-immunized mice produced high titers of GluR3B antibodies, that these mice were significantly more susceptible to seizures, compared to all the groups of control mice, and that the seizure scores associated significantly with the GluR3B antibody levels (Ganor et al. 2014), Furthermore, the GluR3B-immunized mice were also more anxious in Open-Field test, fell faster in RotaRod test, and fell more in Grip test, compared to all the groups of control mice. Thus, the GluR3B antibodies induced abnormal behavior (Ganor et al. 2014).

Third, Scheggia et al. (2021) performed intracerebroventricular infusion of GluR3 antibodies purified from the serum of Frontotemporal Dementia (FTD) patients into mice, and found that the GluR3 antibodies caused a reduction of synaptic levels of GluR3-containing AMPARs in the prefrontal cortex (PFC). In addition, the animals injected with the patient’s GluR3 IgG antibodies showed significant changes in recognition memory and impairments in social behavior and in social cognitive functions. As visualized by confocal imaging, the functional outcomes were paralleled by profound alterations of dendritic spine morphology in the PFC. All the observed behavioral, molecular and morphological alterations were transient and not detected 10–14 days from GluR3 antibodies injection (Scheggia et al. 2021).

Diagnosis of Glutamate receptor antibodies in Epilepsy patients should be preferably done by testing if patient’s antibodies bind isolated GluR extracellular antigenic peptides

An important diagnosis-related practical fact to note is that in our opinion the most sensitive and reliable In vitro method for detecting autoimmune Glutamate receptor antibodies in blood or CSF of patients seems to be ELISA that tests for binding patient’s antibodies to the GluR3B peptide, NR1 peptide and NR2 peptide. It seems that the other methods often fail to detect the autoimmune antibodies to AMPA and NMDA receptors, and the results of the tests are in fact false negative.

There are many more enigmatic intractable Epilepsy patients who may have pathogenic autoimmune antibodies

Our current study focused only on two intractable and enigmatic epileptic patients IE-3 and IE-15, and revealed a wealth of findings about them, that support the assumption that they suffer from ‘Autoimmune Epilepsy’. As such, the study cannot teach us how many more enigmatic Epilepsy patients have similar pathogenic autoimmune antibodies that can on their own induce seizures and neural damage.

However, from the clinical point of view, and as evidenced by Prof. H.G. (2nd author herein) the two Epilepsy patients studied and described in this paper represent a significant proportion (~ 20%) of all the young enigmatic Epilepsy patients treated in the Epilepsy Center, in Schneider Children's Medical Center. In this clinic, ~ 1500 to 2000 Epilepsy patients are diagnosed and treated each year. Therefore, and based on the findings of this paper and other published so far in the field of ‘Autoimmune Epilepsy’, it is important to keep in mind that when an Epilepsy patient has a combination of symptoms including intractable seizures and psychiatric abnormalities, behavioral impairments and some other neurological abnormalities and/or deficits, and a comprehensive work-up (MRI, metabolic) is normal, it is very important to test for autoimmune etiology, by performing all possible In vitro and In vivo tests that can reveal functional active pathological autoimmune antibodies.

The absolute essential tests are the In vitro ones that test the effects of patient’s serum, and preferably purified IgG, on the survival of neural cells.

These functional tests should be done even if the standard passive diagnosis of known autoimmune antibodies yield negative results.

Appropriate treatments for ‘Autoimmune Epilepsy’ can decrease seizures in intractable patients who have pathological autoimmune antibodies

Patients diagnosed with ‘Autoimmune Epilepsy’ or with encephalitis and seizures, can benefit from various therapeutic strategies, as described and discussed in several original papers and reviews (Dalmau et al. 2019; Britton et al. 2016; Levite and Hart 2002; Quek et al. 2012; Tan et al. 2020; Gao et al. 2016; Dubey et al. 2017, 2018; Bien and Holtkamp 2017; Feyissa et al. 2017, 2018; Cui et al. 2018; Pruss et al. 2012; Greco et al. 2016; Husari and Dubey 2019; Bruijn et al. 2019; Lancaster et al. 2010; Hoftberger et al. 2013; Petit-Pedrol et al. 2014; Carvajal-Gonzalez et al. 2014; Sonderen et al. 2016; Spatola and Dalmau 2017; Finke et al. 2012). Overall, immunotherapy administered early after the first diagnosis of ‘Autoimmune Epilepsy’ seems to be particularly effective, and few optional therapeutic strategies exist, either as a monotherapy or a combinatorial therapy.

These therapeutic options can be classified into three therapeutic groups.

Group 1: immunosuppressive and anti-inflammatory chemical and biological drugs.

(A) Intravenous Immunoglobulin (IvIg). IvIg is a ‘cocktail’ product made up of antibodies of thousands of people, to make a super-concentrated and very diverse collection of antibodies against many possible infectious organisms that the body might encounter. IvIg is given intravenously. IvIg can play the role of immunomodulatory and/or even life-saving therapy to modify the course of the underlying disease in patients with autoimmune diseases. (B) Methylprednisolone, (C) Cyclophosphamide, (D) Tacrolimus, (E) Natalizumab—monoclonal antibody (mAb) against the cell adhesion molecules α4-integrins. The α4-integrins are cell surface receptors that mediate cell-extracellular matrix (ECM) and cell–cell adhesions by interacting with fibronectin (FN) and vascular cell adhesion molecule 1 (VCAM-1), respectively. (F) Rituximab—mAb against CD20—a protein expressed primarily on B cells. Anti-CD20 mAbs can act through several mechanisms, including complement-dependent cytotoxicity, antibody-dependent cellular cytotoxicity (ADCC), antibody-dependent cellular phagocytosis, and direct apoptosis induction (for review see Pavlasova and Mraz 2020), (G) Adalimumab—mAb against TNFα. The primary role of TNFα. is in the regulation of immune cells. TNFα induces dozens of effects on target cells that express its receptors.

Group 2: Immune purging procedures: IgG absorption and plasmapheresis.

Group 3: Some AEDs, or a combination of AEDs with immunotherapy.

Additional therapeutic strategies for ‘Autoimmune Epilepsy’ may be developed in the future, using animal models of this disease, such as the one we developed and used in the present study in rats, or in our previous study in mice (Levite et al. 2020).

The study limitations 1.

This study and its many results are on two Epilepsy patients only. Of course, many more Epilepsy patients need to be investigated in a similar way.

2.

Although we found that the IgG antibodies of the two epileptic patients we studied bound to both neural cells of human origin In vitro, in tissue culture (Figs. 5, 6, 7, 8), and to rat neural cells in brains of live rats In vivo (Figs.