Brain on Fire: How Brain Infection and Neuroinflammation Drive Worldwide Epilepsy Burden

Abstract

Roughly 80% of the global burden of epilepsy resides in low- and middle-income countries (LMICs; WHO, 2022). Despite numerous new therapies for the treatment of epilepsy, the number of patients who remain resistant to available medications is unchanged. Additionally, no therapy has yet been clinically proven to prevent or attenuate the development of epilepsy in at-risk individuals. Unfortunately, access to next generation therapies in LMICs is low, the stigma associated with epilepsy remains high, and access to adequate resources is unchanged. Thus, the global epilepsy burden disproportionately falls on LMICs such that strategies to conscientiously integrate global epilepsy risk factors into preclinical research may meaningfully advance 21st century epilepsy therapies. Brain infections are one of the main risk factors for epilepsy in resource-poor settings. Further, both infection- and autoimmune-associated encephalitis contribute to worldwide epilepsy risk and remain relatively understudied. For example, clinical SARS CoV-2 infection can induce rare instances of encephalopathy and acute seizures. Among viruses known to cause acute brain infection, enteroviruses increase risk for encephalitis-induced epilepsy, but are not associated with risk for other neurodevelopmental disorders (eg, autism spectrum or attentional deficit hyperactivity disorders). Naturally occurring models of viral infection-induced epilepsy therefore provide an exquisite opportunity to uncover novel contributors to epileptogenesis. Moreover, the convergent neuroinflammatory pathways that are associated with viral infection-induced encephalitis and autoimmune encephalitis reflect an untapped therapeutic opportunity to meaningfully reduce the global burden of epilepsy. This review summarizes the latest advances in translational research integrating encephalitis-induced seizure and epilepsy models, in tandem with progress in clinical diagnosis of inflammation and virally mediated epilepsy. This improved awareness of the shared biological underpinnings of epileptogenesis following brain infection or autoimmune encephalitis is anticipated to beneficially impact the global burden of epilepsy.

Introduction

Epilepsy does not discriminate across gender, race, or income status, however there is a disproportionately higher incidence (>75% of worldwide cases) in developing versus developed countries. ¹ One potential contributing factor is the greater incidence of infections of bacterial and/or viral origins in low- and middle-income countries. Epilepsy is an under-recognized long-term complication of central nervous system (CNS) infection. ¹ Brain infections can unleash an immune system storm that provokes acute symptomatic seizures, and ultimately leaves lasting neurological damage to induce acquired epilepsy. Further, autoimmune encephalitis is an immune-mediated disease of the CNS in which the body’s own autoantibodies or immune cells attack the brain resulting in a wide range of neurological and psychiatric symptoms, including acute symptomatic seizures and epilepsy. ^1,2 Thus, infection-induced and autoimmune encephalitis are similarly associated with aberrant immune system responses within the brain, offering critical insight into the convergent epileptogenic mechanisms that may beneficially uncover novel disease-modifying therapies for the 21st century. Immune system contributions to epileptogenesis are an under-considered driver of epilepsy risk such that increased basic science investigation and clinical translation of these findings may be a relevant strategy to reduce global epilepsy burden. This review summarizes the latest advances in preclinical studies of viral infection-induced epilepsy, autoimmune epilepsy, and advances in clinical diagnosis and treatment of seizures associated with aberrant immune system activation. We conclude by discussing the remaining knowledge gaps to further mitigate disease burden and impact epilepsy risk on a global scale.

Inflammation and Immune System Activation Leading to Seizures and Epileptogenesis

The peripheral immune system plays a critical role in promoting viral infection-induced seizures, including innate immune system activation in epileptogenesis. The innate immune response is consistently associated with epileptogenesis in rodent models, implicating inflammatory events in the development and maintenance of epilepsy. ^3,4 Although the precise mechanisms leading to epilepsy remain unclear, neuroinflammation clearly contributes. ⁴ Inflammation can alter the excitatory and inhibitory balance among neurons, evoking neuronal excitation and seizure onset. However, the exact role and mechanisms of encephalitis in seizure generation remain unclear. Animal models that faithfully replicate human epilepsy and demonstrate reproducible seizure onset and epileptogenesis after brain insult offer an invaluable resource for hypothesis testing and profiling novel therapeutic targets for epilepsy.

Theiler’s Murine Encephalomyelitis Virus Model of Viral Infection-Induced Acute Seizures and Epilepsy in Mice

The Theiler’s murine encephalomyelitis virus (TMEV) model of infection-induced acute seizures and epilepsy has emerged in recent decades as a unique rodent model to simultaneously assess mechanistic contributors to seizure and epilepsy risk, as well as perform preclinical drug profiling to uncover agents with novel molecular mechanisms. ^3,5,6 Intracerebral infection of C57BL/6J mice with TMEV leads to the onset of many characteristics of encephalitis-induced human temporal lobe epilepsy (TLE), including neuroinflammation, neurodegeneration, and behavioral comorbidities, including anxiety-like behavior and cognitive deficit. Moreover, mice infected with TMEV that experience acute symptomatic seizures display lowered seizure threshold and a greater propensity for abnormal EEG activity and spontaneous recurrent seizures weeks to months later. Neuropathology is consistent with TLE (extensively reviewed by DePaula-Silva ³ and Barker-Haliski et al ⁶ ). Notably, this is the only mouse model of viral-induced TLE. Consequently, this model is uniquely poised to assess novel mechanisms through which viral-induced inflammation promotes seizures and epilepsy.

Theiler’s murine encephalomyelitis virus has a tropism for hippocampal pyramidal neurons. During the acute phase of the infection, seizures are accompanied by the infiltration of innate and adaptive immune cells from the periphery into the CNS (reviewed by DePaula-Silva ³ ). Since acute seizures manifest as early as 3 days postinfection, the innate rather than the adaptive immune response likely mediates seizure onset. This hypothesis is substantiated by data indicating that TMEV infection of mice deficient in B and T cells shows no impact on seizure generation. ^7,8 Similarly, antibody depletion of either natural killer cells or neutrophils has no significant effect on seizure incidence. However, mice treated with clodronate liposomes to deplete macrophages exhibited a significant reduction in seizures, ⁸ suggesting these cells influence TMEV-induced seizure generation.

Macrophages are myeloid cells that infiltrate the CNS hours after TMEV infection, ³ and mice experiencing seizures have increased infiltration of inflammatory macrophages into the brain versus non-seizing mice. ^3,8 Within the CNS, these cells can adopt various reactive states depending on the stimuli, resulting in the secretion of cytokines that can act through cytokine receptors to influence the function of glial cells and neurons. ⁹ Inflammatory macrophages secrete pro-inflammatory cytokines, inducing and intensifying brain inflammation. Conversely, these cells can also adopt an anti-inflammatory state, secreting anti-inflammatory cytokines, which can decrease CNS inflammation and promote healing and homeostasis.

Regarding peripheral immune cells, seizures induced following TMEV infection are associated with the infiltration of interleukin (IL)-6-producing macrophages into the brain (reviewed by DePaula-Silva ³ ). IL-6 plays a critical role in neuroinflammation, and IL-6 expression is highly induced in the cerebral spinal fluid (CSF), brain, and blood of epilepsy patients. ^10,11 Notably, IL-6-deficient mice exhibit a significant decrease in seizure generation following TMEV infection. ¹² Additionally, beneficial effects can occur by blocking the IL-6 receptor using tociluzimab in patients with refractory status epilepticus, but due to IL-6 pleiotropic nature, caution in blocking IL-6 pathways is needed. ^{13 -16} Further, emerging evidence suggests that TMEV infection-induced symptomatic seizures are highly influenced by extrinsic factors, ¹⁷ including evidence that intestinal microbiome modulation via diet formulation and sterilization may dramatically alter disease trajectory after TMEV infection in C57BL/6J male mice, ¹⁸ potentially through modulation of the immune system response. While more collaborative study is necessary, modulation of macrophages and their reactive state during neurotropic viral infection may reveal therapeutic opportunities to mitigate neuroinflammation and downstream impacts on seizure development.

Contribution of Perineuronal Nets to TMEV Infection-Induced Seizures

Over 100 years ago, Camillo Golgi described a peculiar coreset-like structure surrounding a population of neurons in the cortex and hippocampus, which he called perineuronal nets (PNNs). Initially dismissed by many researchers as a fixation artifact, PNNs are now recognized as extracellular matrix aggregates of chondroitin sulfate proteoglycans (CSPGs) tethered to the neuronal membrane via strands of hyaluronic acid. Functionally they can stabilize synapses, particularly in the somatosensory system where they prevent synaptic plasticity thereby stabilizing topographic maps. The study of tumor-associated epilepsy offered the unexpected discovery that PNNs on inhibitory neurons residing near the tumor are destroyed by proteolytic enzymes, notably MMP-2, 3, and 9 released by the tumor. ¹⁹ As a result, such neurons slowed their firing rate. Further studies showed this effect to be causally linked to PNN loss, which act as an insulator of the neuronal membrane, akin to myelin, thereby reducing the effective membrane capacitance. ^{20 -22} Upon digestion, capacitance increases and firing rate decreases.

With this finding, effort was made to examine other forms of acquired epilepsy where proteolysis may contribute. As noted above, viral brain infection happens to be among the leading causes of epilepsy in the developing world, significantly contributing to global epilepsy burden. Interestingly, as was the case with glioma-associated epilepsy, PNNs around inhibitory neurons in hippocampus and cortex of mice infected with TMEV were enzymatically destroyed. ²³ This was also found to be due to increased activity of MMPs, which could be inhibited by a broad spectrum MMP inhibitor. Interestingly, the dentate gyrus of hippocampus and amygdala both showed abnormal extracellular depositions of PNNs surrounding excitatory neurons. These were only present in TMEV-infected animals that developed seizures. Animals that failed to develop seizures did not show these abnormal PNNs. To question whether these de novo formed PNNs are causal for seizures, their formation was disrupted either pharmacologically or using a transgenic mouse in which the Acan gene, encoding for a necessary PNN constituent aggrecan, was deleted. Preventing PNNs from forming significantly reduced seizure frequency and severity, suggesting that abnormal PNN deposition in amygdala and hippocampus following TMEV infection is causally involved in seizure development in these animals. Preventing PNN development in these structures may thus also prevent seizures, and which may be achieved using minocycline, an agent known to inhibit MMPs. From a functional perspective, the presence of interstitial PNNs around excitatory neurons, owing to the negative charges of CSPGs, bind and accumulate K⁺ ions, causing a ∼7 mV depolarization of these neurons, sufficient to increase neuronal excitability. ²³

Neuro-Immune Interactions in Autoimmune-Associated Epilepsies

In the last 2 decades, various autoantibodies against neuronal cell structures of the CNS have been associated with autoimmune encephalitis. ²⁴ These include onconeuronal autoantibodies, which are often associated with tumors, but also autoantibodies against neuronal surface antigens, in particular against receptor-regulated ion channels, ion channel-associated proteins or neurotransmitter receptors, which are mainly expressed at excitatory and inhibitory synapses. ²⁵ Exposure to certain bacteria or viruses or tumors (eg, teratomas, lung carcinomas, seminomas) may trigger the immune system to produce these specific autoantibodies in the sense of cross-antigenicity mechanisms. ^{26 -29} However, in many patients the cause remains unknown.

Autoantibodies directed against extracellular targets play a direct role in neuronal dysfunction and show increased responsiveness to immunotherapy, leading to improved clinical outcomes. ^{30 -32} In contrast, the impact of autoantibodies directed against intracellular antigens remains uncertain in terms of therapeutic response and clinical significance. Although many of the intracellular autoantibodies may not have direct pathogenic effect, they serve as effective diagnostic biomarkers. Additionally, cytotoxic T lymphocytes with specificity for these antigens may contribute to autoimmune-induced neuronal damage. In a large proportion of patients clinically suspected of autoimmune encephalitis, no known autoantibodies can be detected. ^24,33 This may indicate a possible pathogenic role of previously unknown or an exclusively T cell-mediated disease in these patients.

Recently, an antibody against the intracellular scaffold protein, drebrin, was identified in patients with adult-onset seizures and suspected autoimmune encephalitis, which has significant effects on synaptic structure and neuronal activity. ³⁴ There were 2 relevant drebrin protein fragments identified that serve as epitopes for antibody recognition, the detection of specific banding patterns in immunoblots and distinct binding patterns to dendritic spines of primary neurons. Complementary functional in vitro data suggest a direct pathogenic effect of patient-derived anti-drebrin autoantibodies, as they bind to drebrin, impair postsynaptic drebrin abundance and distribution, and induce neuronal network hyperexcitability. The clinical course, neuroradiological and neuropathological findings in anti-drebrin autoantibody-positive patients indicate that this is a severe disease in the spectrum of autoimmune encephalitis. Anti-drebrin autoantibody-positive limbic encephalitis thus represents a novel autoimmune syndrome associated with epilepsy and characteristic clinical presentation.

Clinical Features of Febrile Infection-Related Epilepsy Syndrome and New Onset Status Epilepticus

Published consensus definitions of new onset status epilepticus (NORSE) define it as a clinical presentation, not a specific diagnosis, in a patient without active epilepsy or other relevant neurological disorder, with new onset of refractory status epilepticus (RSE) without a clear acute or active structural, toxic, or metabolic cause. ³⁵ Febrile infection-related epilepsy syndrome (FIRES) is a subcategory of NORSE that requires a prior febrile infection, with fever starting between 2 weeks and 24 hours prior to RSE onset. Both terms now apply to all ages (previously, FIRES was preferred for pediatrics, and NORSE for adults). If no cause is found after extensive evaluation, it is considered cryptogenic NORSE (or “of unknown etiology”). Note that these definitions include viral infections and autoimmune conditions, which can present as NORSE/FIRES. The most common identified cause is anti-NMDA encephalitis.

There is extensive evidence that inflammation significantly contributes to the pathogenesis of NORSE/FIRES, with upregulation of many pathways in the serum and CSF, particularly those involving cytokines related to innate immunity, including IL-6, CXCL-8/IL-8, IL-1, and CCL2. ³⁶ Most genetic studies have been negative, but some have shown mutations in genes related to inflammation/immunity or mitochondrial function. As in the TMEV model, activation of the peripheral innate immune system seems to underlie refractory seizures. ³⁶ Elevated cytokines correlate with higher seizure burden and worse long-term outcomes. Animal models confirm a proconvulsant effect of these molecules, ³⁷ and their role in disrupting the blood–brain barrier, which is also seen in NORSE patients. ³⁸ Pediatric patient studies suggest that elevated CSF neopterin and quinolinic acid might be early biomarkers of neuroinflammation and neurotoxicity in patients with FIRES and other infection-triggered encephalopathies. ³⁹

International consensus recommendations stress seizure control and early immunotherapy. ⁴⁰ By day 3, intravenous (IV) steroids and/or IV immunoglobulins are recommended. If seizures remain poorly controlled by day 7, ketogenic diet should be considered, and second line immunotherapy is suggested. If an antibody-mediated cause is likely, rituximab is recommended; otherwise, based on animal models, small series of patients, and the above cytokine findings, the recommended second line immunotherapy for cryptogenic NORSE/FIRES is anakinra (an IL-1 receptor antagonist) or tocilizumab (an IL-6 antagonist). Outcomes remain disappointing, with about ¼ not surviving, and the majority ending up with neurologic dysfunction of varying degrees, often with refractory epilepsy. ⁴¹

As current evidence is limited, collaborative research is crucial. International collaborative efforts are underway, including biobanks and investigations into clinical risk factors, autoimmunity, inflammation, genetics, infections, epidemiology, microbiomics, metabolomics, determinants of long-term outcome, and more. Basic scientists, clinicians, and patients/families are all working together. Information on a funded international biobank, as well as clinical resources and support for clinicians, researchers, and families, can be found via the NORSE Institute website (https://www.norseinstitute.org/).

Clinical Findings From Critically Ill Patients Infected With SARS-CoV2

SARS-CoV-2/COVID-19 infections may have multifaceted presentations, from asymptomatic to mild or more severe respiratory or systemic symptoms, including central or peripheral neurological manifestations. ^42,43 In the early pandemic months of 2020, case reports of SARS-CoV-2/COVID-19 encephalitis and prior animal studies demonstrating entry of SARS coronaviruses in the brain of mice transgenic for human angiotensin converting enzyme 2 (ACE2) ⁴⁴ raised concerns about possible neuroinvasiveness of SARS-CoV-2/COVID-19. A first case series of acutely ill patients hospitalized between March and April 2020 with COVID-19 infections, who had medically indicated EEG studies (mostly brief 8-channel headband EEGs) revealed high incidence of largely frontal origin epileptiform discharges (∼41%), including in patients who had no prior history of seizures (∼39%), among an encephalopathic, slow and disorganized background (>80%). ⁴² New onset encephalopathy (68%) and seizure-like events (54.5%) were among the most common indications for EEG. ⁴² Significant confounders however were present, including respiratory failure (96%) often requiring intubation (64%), concomitant metabolic disturbances (renal 46%, hepatic 77%), administration of anti-seizure medications or sedatives (86%). ⁴² In a subsequent case–control study using a different hospital inpatient population that underwent routine EEGs for evaluation of seizures or altered mental state, high rate of epileptiform or periodic discharges or seizures were seen in both COVID-19-positive (21%) and COVID-19-negative (9%) critically ill patients and both groups had high mortality (34.5%-37.5%), reflecting the severity of underlying medical conditions. ⁴⁵ Numerous other studies reported similar EEG background abnormalities but with variable rates of epileptiform discharges (0%-62.5%, mean 20.3%) or acute symptomatic seizures (0%-40%, mean 2%; reviewed in study by Kubota et al). ⁴⁶ A larger cohort monitored with continuous EEG, mostly consisting of intubated patients (81.7%), showed epileptiform abnormalities in 48.7%, electrographic seizures in 9.6%, and nonconvulsive status epilepticus in 5.6%; seizures were a predictor of mortality. ⁴⁷

The available studies support overall the higher risk for epileptiform abnormalities and seizures in critically ill patients with COVID-19 infections; however, it is less clear whether this risk directly reflects the impact of COVID-19 infection rather than the combined impact of the multifactorial medical disturbances in these cases. Furthermore, among the patients admitted with COVID-19 infections, 12% have neurological symptoms that require neuroimaging investigation. ⁴⁸ A new CNS pathology (ischemic or hemorrhagic cerebrovascular complication or infectious/inflammatory) is seen in 15% of patients with severe COVID-19 and acute encephalopathy, which could further increase seizures susceptibility. ⁴⁹ Despite initial concerns for SARS-CoV2 neuroinvasiveness, it has rarely been detected in CSF; when it is infrequently detected in neuropathology studies, it is limited and not suggestive of viral encephalitis. ⁵⁰ Postmortem studies demonstrate pronounced neuroinflammation and neuronal loss in the brainstem, cerebellum, meningeal cytotoxic T cell infiltration, microvascular injury or inflammation, acute hemorrhagic or ischemic lesions, or hypoxic-ischemic injury, but no direct CNS injury from SARS-CoV-2 infection. ^51,52 Compartmentalized intra-CNS immune responses were proposed to underlie the neurological manifestations of certain COVID-19 patients. ⁵³ This was supported by the distinct cytokine responses in the CNS versus plasma, as well as the diverse single cell RNA sequencing T cell activation patterns and antibody responses. ⁵³ Furthermore, brain but not pulmonary infection with SARS-CoV-2 in animals produced intrathecal SARS-CoV-2 antibodies. ⁵³

A review of 860 934 electronic health records from TriNetX Analytics compared two cohorts of 152 754 patients with COVID-19 or influenza, excluding those with prior epilepsy or recurrent seizures, about seizure or epilepsy development over 6 months after infection. ⁵⁴ Among COVID-19 patients, 0.81% developed seizures and 0.3% epilepsy, and had higher hazard risk (HR) for new-onset seizures (1.55) and epilepsy (1.87) versus influenza patients. ⁵⁴ The HR for epilepsy was greater among those who were not hospitalized and those <16 years old. ⁵⁴ Bioinformatics analyses of gene expression datasets from COVID-19 and epilepsy patients suggested common gene expression patterns in pathways involved in immune response or lipid metabolism. ⁵⁵ Further controlled and prospective studies are warranted to rigorously test these associations and clarify whether, and under which conditions, these associations might be mechanistically relevant for epileptogenesis or to advance therapeutic discovery.

Conclusions

Infection-induced encephalitis and neuroinflammation are significant contributors to the global burden of epilepsy. Despite numerous new therapies for the treatment of epilepsy, the number of patients who remain resistant to available medications is unchanged. To date, the only intervention that may beneficially modify neuroinflammation and disease burden is adherence to the high-fat, low-carbohydrate ketogenic diet. Ketogenic diet has been proposed to modulate the innate immune system and shift brain metabolic demands, indicating that strategies to modulate neuroinflammation may be relevant to epilepsy management and prevention. Accordingly, novel models of epileptogenesis that more accurately reproduce how epilepsy develops following infection-induced encephalitis may improve understanding of therapeutic strategies for epilepsy prevention, and advance seizure control strategies across all patient populations worldwide, not simply in high-income countries. Further, the simultaneous investigation into immune system-mediated drivers of seizure risk and epileptogenesis may prove useful to identify novel molecular targets and therapeutic strategies that are broadly applicable across a range of epilepsy conditions. Greater collaborative and interdisciplinary research into both mechanistic origin and clinical course of immune system-mediated seizure onset and epilepsy are desperately needed. Immune system activation-induced epilepsy models that exhibit a high level of fidelity with clinical epilepsy carry high translational value to meaningfully impact the global burden of epilepsy.