Therapeutic role of targeting mTOR signaling and neuroinflammation in epilepsy

Abstract

Existing therapies for epilepsy are primarily symptomatic and target mechanisms of neuronal transmission in order to restore the excitatory-inhibitory imbalance in the brain after seizures.However, approximately one third of individuals with epilepsy have medically refractory epilepsy and do not respond to available treatments. There is a critical need for the development of therapeutics that extend beyond manipulation of excitatory neurotransmission and target pathological changes underlying the cause of the disease. Epilepsy is a multifaceted condition, and it could be that effective treatment involves the targeting of several mechanisms. There is evidence for both dysregulated PI3K/Akt/mTOR (mTOR) signaling and heightened neuroinflammatory processes following seizures in the brain. Signaling via mTOR has been implicated in several epileptogenic processes, including synaptic plasticity mechanisms and changes in ion channel expression following seizures. Inflammatory signaling, such as increased synthesis of cytokines and other immune molecules, has also shown to play a significant role in the development of chronic epilepsy. mTOR pathway activation and immune signaling are known to interact in normal physiological states, as well as influence one another following seizures. Simultaneous inhibition of both processes could be a promising therapeutic avenue to prevent the development of chronic epilepsy by targeting two key pathological mechanisms implicated in epileptogenesis.

1. Introduction

Epilepsy is one of the most common neurological disorders, affecting approximately 65 million people in the world (Devinsky et al., 2018). It is characterized by recurrent, spontaneous seizures and is often associated with comorbidities, including cognitive, psychiatric, and other neurobiological consequences, which can significantly affect quality of life. Several causes of epilepsy have been identified, including structural causes (brain injury, stroke, tumors), gene mutations, infections, metabolic disorders, autoimmune conditions, as well as unknown etiologies (Stafstrom and Carmant, 2015). Despite the variety of pharmaceutical therapies available, approximately one third of those with epilepsy do not respond to anti-epileptic drugs (AEDs), and have medically refractory epilepsy (Dalic and Cook, 2016).

Due to seizures resulting in uncontrolled neural excitation in the brain, many AEDs primarily target ion channels or other components of neuronal transmission in order to restore the excitatory/inhibitory (E/I) imbalance (Cook and Bensalem-Owen, 2011). However, many of these pharmaceuticals are symptomatic, and function to reduce seizures but do not influence the underlying cause of the disorder. In addition, while AEDs suppress seizure frequency and ictal discharges, they often do not abolish interictal epileptiform discharges which is of concern (D’Antuono et al., 2010). Abnormal electroencephalogram activity is associated with neurological damage, such as impairments in the blood brain barrier (BBB) and also in cognitive abilities, including declarative memory (Milikovsky et al., 2019; Reed et al., 2019). Recent efforts have been aimed at identifying disease-modifying targets, that go beyond modulation of the E/I imbalance and focus on other mechanistic underpinnings of the disorder. The pathophysiological mechanisms that underlie the development of chronic epilepsy, or epileptogenesis, are for the most part still unknown. Epilepsy is a multifaceted disorder, and the progression of the disease most likely involves a combination of neurological changes in the brain (Curia et al., 2014).

Neuroinflammatory signaling has been demonstrated to play a significant role in the development of epilepsy, both in animal models of acquired and genetic epilepsies, and in clinical populations (Vezzani et al., 2019). The heightened neuroinflammatory responses following seizures can impact hyperexcitability, thus contributing to epileptogenesis (Iori et al., 2016). Therapeutics that target inflammatory mechanisms and attenuate seizure-induced neuroinflammation have shown significant promise in both pre-clinical rodent models and clinical studies. In addition, cell signaling cascades that are disrupted following seizures have been another alternative target for anti-epileptogenic therapies (Hodges and Lugo, 2018; Wong, 2013b). Specifically, phosphoinositide 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) (PI3K/Akt/mTOR) signaling activity has been found to be upregulated in brain tissue of individuals with epilepsy, as well as in several genetic and acquired models of epilepsy in vivo (Ostendorf and Wong, 2015). Activation and signaling via the mTOR pathway regulates several seizure-induced processes, including synaptic plasticity mechanisms, cellular proliferation, and changes in ion channel protein expression that could potentially underlie epileptogenesis (Citraro et al., 2016). Inhibition of mTOR signaling has shown to be effective in reducing seizure frequency in patients with Tuberous sclerosis complex (TSC) caused by mutations in TSC1 or TSC2, suggesting that mTOR inhibition could be a promising target for other mTOR-related epilepsies (Cardamone et al., 2014).

Epilepsy is a complex neurological condition, and effective treatments that target the underlying cause of the disorder potentially involves a combinatorial approach targeting multiple mechanisms (Curia et al., 2014). Pre-clinical studies that use therapeutics aimed at alternative targets beyond manipulation of the E/I imbalance often only focus on changes pertaining to the specific target and mechanism of interest of the therapeutic in use. For example, treatments that inhibit seizure-induced neuroinflammation identify key mechanisms related to the immune system that have been impacted by the treatment but may ignore other molecular changes such as cell signaling pathways that are also affected (Marchi et al., 2011). Signaling via the mTOR pathway and inflammatory processes interact in normal physiological states, and additionally there is evidence that both processes influence each other following seizures (Thomson et al., 2009; Weichhart et al., 2015). This review will highlight the role both mTOR signaling and neuroinflammation play in epilepsy, as well as review evidence of how both of these processes interact in epileptic states in a variety of models (Fig. 1).

Figure 1. Relationship between epilepsy, neuroinflammation, and hyperactive Mtor signaling.

Epileptic events in the brain can be caused by several etiologies, including gene mutations, trauma, CNS infections, SE, ischemia and hypoxia, febrile seizures, and stroke. Seizures are associated with both neuroinflammation and hyperactive mTOR signaling, which can lead to a cascade of downstream pathophysiological effects. These changes can result in hyperexcitability, altered synaptic transmission, and increased seizure susceptibility, which may underlie the development of chronic epilepsy. Many pharmacological (ex. rapamycin, mycophenolate) and non-pharmacological (ex. curcumin, ketogenic diet, resveratrol) compounds have effects on both neuroinflammatory signaling and mTOR pathway activity. These therapies could be beneficial in inhibiting two key processes that potentially underlie epileptogenesis. In addition, there are several anti-inflammatory treatments that have shown efficacy in reducing seizures and associated pathological damage, such as cytokine inhibitors, HMGB1-TLR4 signaling cascade inhibitors, and COX inhibitors. CNS, central nervous system; SE, status epilepticus; mTOR, mammalian target of rapamycin; HMGB1, high- mobility group box protein 1; TLR4, toll-like receptor 4; COX, cyclooxygenase; BBB, blood brain barrier; E/I, excitatory/inhibitory.

2. Role of mTOR signaling in epilepsy

2.1. The mTOR pathway in normal physiological conditions

The mTOR signaling pathway is critical for proper functioning of several physiological processes throughout the lifespan. Activation of the pathway has shown to modulate cell growth, proliferation, protein synthesis, neuronal morphology, cortical development, and immune responses (Laplante and Sabatini, 2012). Beyond these classical roles of mTOR, activation of the pathway in the central nervous system (CNS) can influence neuronal signaling and excitability, such as axonal and dendritic morphology, neurotransmitter expression, synaptic plasticity, and cognition and behavior (Bekinschtein et al., 2007; Jaworski et al., 2005; Tang et al., 2002).

Central to the pathway is the highly conserved serine/threonine protein kinase mTOR, which is part of two larger complexes, mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2) (Jhanwar-Uniyal et al., 2019). The mTORC1 is comprised of a group of binding proteins including the regulatory-associated protein of mTOR (raptor) and stimulates cell growth and proliferation via modulating protein synthesis. It is also highly sensitive to rapamycin, an inhibitor of the mTOR pathway, and by which TOR received its name following identification of the kinase being the target of rapamycin’s antiproliferative effects (Wong, 2013a). In contrast, mTORC2 complexes with a separate group of accessory proteins including rapamycinin-sensitive companion of mTOR (rictor), modulates processes related to cell structure and metabolism, and is relatively insensitive to rapamycin without prolonged exposure of the inhibitor (Sarbassov et al., 2006). Dysregulation of signaling mechanisms via both mTORC1 and mTORC2 have been shown to be involved in the pathogenesis of several disorders, including cancer, diabetes, cardiovascular disease, as well several neurodevelopmental and neurodegenerative disorders (Costa-Mattioli and Monteggia, 2013; Lee et al., 2017; Perluigi et al., 2015; Pópulo et al., 2012; Rosina et al., 2019; Saxton and Sabatini, 2017; Zhu et al., 2019). This review will primarily focus on the impacts of mTORC1 disruption, as this complex has been more heavily implicated in mechanisms underlying epilepsy pathogenesis (Citraro et al., 2016; Ostendorf and Wong, 2015).

In normal physiological conditions, PI3K will be activated by receptor tyrosine kinases following activation from a variety of extracellular and intracellular stimuli, including amino acids, growth factors, neurotransmitters, intracellular stress signals, and immune molecules (Costa-Mattioli and Monteggia, 2013; Saxton and Sabatini, 2017). PI3K functions to convert phosphatidylinositol 4,5-bisphosphate (PIP₂) to phosphatidylinositol 3,4,5-trisphosphate (PIP₃). Through the kinase phosphoinositide-dependent kinase-1 (PDK1), phosphorylation and activation of Akt leads to inactivation of the tuberous sclerosis complex (TSC1/TSC2), which functions to antagonize pathway activity. Phosphorylation of the TSC1/TSC2 dimer leads to inhibition of its’ GTPase-activating protein activity, allowing the active form of Rheb to activate mTORC1. mTOR functions in a complex with other regulatory and binding proteins including raptor, mLST8, and inhibitory protein proline-rich Akt substrate of 40 kDa (PRAS40) to create mTORC1. mTORC1 directly phosphorylates the p70S6 kinase (p70S6K), which itself phosphorylates many downstream targets (S6, eEF2K, eIF4B) to promote ribosomal protein recruitment and initiation of translation. In addition, mTOR phosphorylates and inhibits the activity of 4E - binding protein 1 (4E-BP1), causing it to dissociate from eukaryotic initiation factor 4E (eIF4E) and complex with other initiation factors to further stimulate the downstream translation of proteins (Fig. 2).

Figure 2. The PI3K/Akt/mTOR signaling pathway.

The mammalian target of rapamycin (mTOR) is a serine-threonine protein kinase, which forms two complexes, rapamycin-sensitive mTORC1 and relatively rapamycin-insensitive mTORC2 (not shown). The mTOR pathway can be activated by several extracellular (neurotransmitters, amino acids, growth factors, toll-like receptor ligands) and intracellular stimuli (energy deprivation), resulting in activation of a series of downstream effectors. Signaling via the mTOR pathway is initiated with activation of PI3K by receptor tyrosine kinases, which leads to conversion of PIP₂ to PIP₃. Akt is phosphorylated by PDK1, which leads to phosphorylation and inhibition of the TSC1/TSC2 complex. This allows the active form of GTP-binding protein Rheb to activate mTORC1. mTORC1 is comprised of mTOR, in addition to regulatory protein raptor, mLST8, and inhibitory protein PRAS40. mTORC1 activation leads to phosphorylation of several targets involved in ribosomal biogenesis (p70S6K, S6) and protein translation (eEF2K, eIF4B, 4E-BP, eIF4E). Beyond protein translation, mTOR pathway activity regulates cell proliferation, cell growth, synaptic plasticity, ion channel expression, and apoptosis. The mTOR pathway can be negatively regulated by stimuli, such as PTEN, which antagonizes pathway activity by dephosphorylating PIP₃ to PIP₂. In addition, rapamycin binds FKBP12 to inhibit mTORC1 activity. PI3K, phosphoinositide 3-kinase; PIP₂, phosphatidylinositol 4,5-bisphosphate; PIP₃, phosphatidylinositol 3,4,5-trisphosphate; PDK1, phosphatidylinositol-dependent kinase 1; TSC, tuberous sclerosis complex; mLST8, mammalian lethal with SEC13 protein 8; PRAS40, proline-rich Akt substrate of 40 kDa; eEF2K, eukaryotic elongation factor-2 kinase; eIF4B, eukaryotic translation initiation factor 4B; eIF4E, eukaryotic translation initiation factor 4E; 4E-BP1, (eIF4E)-binding protein 1; PTEN; phosphatase and tensin homolog on chromosome 10; FKBP12, 12-kDa FK506 and rapamycin-binding protein; IRS-1, insulin receptor substrate 1; NF1, neurofibromatosis type I.

2.2. mTOR hyperactivation in epilepsy and impacts of pathway inhibition

Many physiological functions of the mTOR pathway, including synaptic plasticity mechanisms, cellular proliferation, and ion channel expression related to neuronal excitability, could underlie the development of epilepsy in pathological conditions (Wong, 2013a). There is evidence for hyperactive mTOR signaling in both genetic and acquired epilepsies, as demonstrated in animal models of epilepsy and in human tissue samples resected from individuals with epilepsy. For instance, several neurodevelopmental disorders with epileptic phenotypes stem from mutations in components of the mTOR pathway, the most notable of these being TSC1/TSC2 and phosphatase and tensin homolog (PTEN) mutations. A heterozygous mutation in either TSC1 or TSC2 results in the autosomal dominant disorder TSC, characterized by cortical malformations known as tubers and subependymal giant cell astrocytomas (SEGAs) in the brain (Jülich and Sahin, 2014). The respective protein products of TSC1 (hamartin) and TSC2 (tuberin) form a complex that antagonizes mTOR activity, such that loss of these proteins results in hyperactive mTOR signaling, along with often refractory epilepsy, intellectual impairment, and an autistic-like phenotype (Chu-Shore et al., 2010; Vignoli et al., 2015). PTEN is a tumor suppressor gene that also negatively controls mTOR activity, along with regulating cellular proliferation and survival (Bonneau and Longy, 2000). Similar to mutations in TSC1/TSC2, loss of PTEN results in hyperactive mTOR signaling accompanied by seizures and significant intellectual and behavioral impairments (Cupolillo et al., 2016). Other conditions associated with disrupted mTOR activity that have epileptic phenotypes include polyhydramnios, megalencephaly, and symptomatic epilepsy (PMSE), neurofibromatosis type I, Fragile X syndrome, focal cortical dysplasia, and hemimegaloencephaly associated with mutations in various elements of the mTOR pathway (Berry-Kravis, 2002; Crino, 2015; D’Gama et al., 2015; Hsieh et al., 2011; Nguyen et al., 2019; Orlova et al., 2010; Santoro et al., 2018).

Several genetic animal models of epilepsy exhibit mTOR hyperactivation and have been useful in exploring the potential of mTOR inhibitors as a rational anti-epileptogenic therapy. Various treatment paradigms with rapamycin has shown to suppress seizures, as well as attenuate many of the molecular and pathological changes that underlie epileptogenesis. For example, rapamycin administration has shown to reduce epileptiform activity, decrease the duration and severity of seizures, and prolong survival in both models of Tsc1/2 and Pten deletion (Way et al., 2012; Zeng et al., 2008). Additionally, rapamycin reduces mTOR hyperactivity and is protective against seizure-induced molecular changes such as ameliorating neuronal disorganization, dendritic abnormalities, aberrant mossy fiber sprouting, cellular hypertrophy, macrocephaly, and astrogliosis (Brewster et al., 2013; Ljungberg et al., 2009; Nguyen et al., 2015; Sunnen et al., 2011; Zhou et al., 2009). Unlike with the mechanism of traditional anti-seizure medications, rapamycin has minimal impact on neuronal excitability (Hartman et al., 2012). Instead, rapamycin most likely exerts its effects on the electrical activity of neurons indirectly, by modulating the expression of voltage-gated ion channels and neurotransmitter expression (Huang et al., 2012b; Niere and Raab-Graham, 2017; Raab-Graham et al., 2006; Tyan et al., 2010). Despite the beneficial effects of rapamycin on both seizure-induced molecular and behavioral changes, seizures often return with cessation of treatment and continuous intermittent rapamycin treatment may be necessary to sustain the anti-seizure effects (Sunnen et al., 2011).

Hyperactivity of the mTOR pathway has also been demonstrated in animal models of acquired epilepsy, following various insults that induce status epilepticus (SE) in the brain. Specifically, pre-clinical models of temporal lobe epilepsy (TLE) utilizing pilocarpine and kainate chemoconvulsants, as well as models of infantile spasms and hypoxic seizures results in mTOR hyperactivation (Buckmaster et al., 2009; Chachua et al., 2012; Huang et al., 2010; Raffo et al., 2011; Talos et al., 2012; van Vliet et al., 2012; Zeng et al., 2009). Unlike with genetic models of mTOR hyperactivation, treatment with rapamycin in acquired models of epilepsy has demonstrated significant variability. In some cases, mTOR inhibition has attenuated chronic seizures, and in other models rapamycin treatment has had minimal effect on epilepsy development (Buckmaster and Lew, 2011; Hartman et al., 2012; Zeng et al., 2009). Some studies have shown beneficial effects of mTOR inhibition on attenuating seizure-induced cognitive and behavioral impairments. For example, treatment with rapamycin following pilocarpine seizures has shown to protect against hippocampal-dependent spatial learning and memory deficits in a Morris water maze task and with recognition memory in a novel object recognition task (Brewster et al., 2013). Furthermore, rapamycin treatment immediately prior and following acute hypoxia-induced seizures attenuated seizure-induced impairments in sociability in adulthood (Talos et al., 2012). The efficacy of rapamycin has shown be to time, age, and model dependent, suggesting that mTOR activation in acquired models is most likely associated with specific pathological changes that vary across models (Chachua et al., 2012; Ostendorf and Wong, 2015).

Inhibition of mTOR with rapamycin and its analogs has shown therapeutic potential in reducing seizure frequency in clinical populations. The rapamycin analog, everolimus, is approved by the FDA for the treatment of SEGAs, kidney tumors, and partial epilepsy in individuals with TSC (Kim and Lee, 2019). There are several ongoing phase I, II, and III clinical trials investigating the effects everolimus has across multiple age groups and with varying treatment durations (Krueger et al., 2010; Krueger et al., 2016). One open-label, phase I/II clinical trial found that everolimus treatment for 12 weeks resulted in significant seizure reduction in 17 of the 20 TSC patients, along with improvements in patient-reported behavior and quality of life (Krueger et al., 2013). In the initial core phase of a prospective, multi-center phase 3 trial, adjunctive everolimus therapy for 18 weeks in individuals with treatment-resistant focal-onset seizures in TSC resulted in 28.2% of individuals experiencing at least a 50% reduction in seizure frequency when on the low-exposure regimen (3–7ng/ml), and 40% of individuals on the high-exposure regimen (9–15ng/ml) (French et al., 2016). Following the core phase, the efficacy of everolimus in reducing seizure frequency in patients persisted throughout the extension phase of the trial, in which individuals received everolimus (3–15ng/ml) for up to 48 weeks (EXIST-3 trial) (Curatolo et al., 2018). While there is accumulating evidence of the benefits of mTOR inhibition for seizure control in individuals with TSC, additional controlled studies are necessary to determine the efficacy of inhibition as a definitive treatment for epilepsy.

3. Role of neuroinflammation in epilepsy

3.1. Inflammation in the central nervous system

An increasing body of clinical and experimental evidence suggests that the immune system plays an important role in epilepsy. Seizures induce a cascade of biological events, characterized by the release of several inflammatory molecules, microglial and astrocyte activation, and disrupted BBB permeability leading to the infiltration of peripheral immune cells (Aronica et al., 2017; Rana and Musto, 2018). It is possible the upregulation in neuroinflammatory processes following seizures contributes to the hyper-excitable neuronal network and dysregulated neural connectivity that underlies epileptogenesis (Musto et al., 2016; Vezzani et al., 2019).

In normal physiological conditions, cells of the innate immune system will activate receptors following injury in order to stimulate a variety of pathways that function to repair damage to the CNS and provide protection from any additional injury. Cytokines are small, secreted and membrane-bound proteins that play an important role in mediating inflammatory signals in the CNS (Fig. 3). In addition to acting as chemical messengers, they regulate several other cellular activities including the survival and differentiation of cells, and mechanisms involved in neuronal development such as the pruning and refining of synapses (Deverman and Patterson, 2009; Paolicelli et al., 2011; Schafer et al., 2012). Cytokines include interleukins (IL), interferons (IFN), and tumor necrosis factors (TNF), and are typically considered to be primarily pro-inflammatory (IL-1β, IL-6, TNFα, IFNγ) or anti-inflammatory (IL-10, IL-4), although some can also have dual functions dependent on the physiological state (Cavaillon, 2001; Tanaka et al., 2014). They can be produced by a number of cells in the CNS, including glial cells such as microglia and astrocytes, as well as neurons and endothelial cells that make up the BBB (Galic et al., 2012). Cells of the innate immune system detect conserved pathogen-associated molecular patterns (PAMPs) in pathogens and damage-associated molecular patterns (DAMPs) released from damaged cells by the expression of germ-line encoded receptors known as pattern recognition receptors (PRRs) (Mogensen, 2009). In normal physiological states, these processes help to recruit cells to fight infection and aid in tissue repair and recovery, however, following an insult these signaling mechanisms can be disrupted and lead to a chronic neuroinflammatory state associated with cellular toxicity.

Figure 3. Cytokine signaling cascades.

Following seizures there is an upregulation in the synthesis of pro-inflammatory cytokines (IL-1β, TNFα, IL-6, i.e.) in the brain which can lead to activation of several downstream signaling cascades. Toll-like receptor 4 (TLR4) is also activated following seizures by its endogenous ligand high-mobility group box 1 (HMGB1) protein or can be exogenously stimulated via administration of lipopolysaccharide (LPS). Activation of cytokine receptors and their respective signaling cascades can initiate downstream transcription of cytokines and other inflammatory molecules, further enhancing the inflammatory environment after seizures. Together these signaling cascades can interact to enhance neuronal excitability and contribute to mechanisms underlying epileptogenesis. IL-1β, interleukin-1 beta; IL-1R1, IL-1 receptor 1; IL-1RA, IL-1R antagonist; IRAK, IL-1 receptor-associated kinase; MAPK, mitogen-activated protein kinases; TNFα, tumor necrosis factor alpha; TNFR1, TNF receptor 1; TRAF2, TNF receptor-associated factor 2; TRADD, TNFR1-associated death domain; RIP1, receptor-interacting protein 1; NF-kB, nuclear factor kappa-light-chain-enhancer of activated B cells; IL-6, interleukin 6; IL-6R, IL-6 receptor; JAK, janus kinase; STAT, signal transducer and activator of transcription; TIRAP, Toll/IL-1 receptor (TIR) domain containing adaptor protein; TRIF, TIR-domain-containing adaptor-inducing interferon-β; TRAM, translocating chain-associating membrane protein; IRF, interferon regulatory factor.

3.2. Neuroinflammatory targets and potential treatments in epilepsy

Cytokines are typically expressed at low concentrations in the brain, however, following an insult such as with seizures, a variety of cytokines (IL-1, IL-6, TNFα, IFNγ) and chemokines (MCP-1/CCL2, CCL3, CCL5) are rapidly produced and released in various brain regions (Arisi et al., 2015; Kosonowska et al., 2015; Uludag et al., 2015). The upregulation in cytokines can induce several downstream signaling events post-seizures, ultimately leading to neuronal hyperexcitability and a reduced threshold for future seizure events. For example, proinflammatory cytokine IL-1β has been shown to stimulate the release of glutamate from astrocytes, reduce glutamate re-uptake leading to elevated glutamate availability, and reduce inhibitory GABA-mediated neurotransmission (Alyu and Dikmen, 2017; Roseti et al., 2015). In addition, cytokine IL-1β has found to be upregulated in the cerebrospinal fluid of children with idiopathic epilepsy compared to individuals without epilepsy (Shi et al., 2017). Pro-inflammatory cytokine TNFα has similar effects, by increasing microglial glutamate release via upregulation of glutaminase, upregulating AMPA receptor expression, and inducing GABA receptor endocytosis, resulting in changes in excitability networks (Galic et al., 2012; Stellwagen et al., 2005; Takeuchi et al., 2006). Another pro-inflammatory cytokine often elevated post-seizures is IL-6, in which upregulations have shown to reduce long-term potentiation and hippocampal neurogenesis, as well as enhance gliosis (Levin and Godukhin, 2017; Vezzani et al., 2002).

One mechanism in which cytokines are produced following seizures is through activation of Toll-like receptors (TLRs). Toll-like receptors are a well-defined class of PRRs that induce inflammatory signaling events by activating a diverse set of molecules, including transcription factors (NF-kβ, AP-1) to stimulate production of immune molecules, and other receptors that activate additional downstream kinases (Kawai and Akira, 2007; Kawasaki and Kawai, 2014). Specifically, signaling via TLR4 and its endogenous ligand, high mobility group box 1 protein (HMGB1), play a significant role in initiating neuroinflammatory responses following seizures (Aronica et al., 2017). Both TLR4 and HMGB1 expression is upregulated in neurons, astrocytes, microglia, and endothelial cells compromising the BBB in both human and experimental epilepsies (Henneberger and Steinhauser, 2016; Hiragi et al., 2018; Maroso et al., 2010; Matin et al., 2015). Stimulation of TLR4 via exogenous activators, such as with agonist lipopolysaccharide (LPS) derived from the cell wall of gram negative bacteria, can also influence seizure progression by enhancing the excitability of neurons and decreasing seizure threshold in experimental models (Eun et al., 2015; Galic et al., 2008; Rodgers et al., 2009). Toll-like receptor 4 activation is interconnected with IL-1 receptor 1 (IL-1R1) signaling, as synthesis and release of their respective ligands, HMGB1 and IL-1β, both involve NLR family pyrin domain containing 3 (NLRP3) inflammasome and caspase 1 activation (Lopez-Castejon and Brough, 2011; Lu et al., 2012). Elevated IL-1R1-TLR4 signaling in both human epilepsies and experimental models of epilepsy has shown to contribute to mechanisms underlying epileptogenesis, along with increased neuronal hyperexcitability (Maroso et al., 2010; Ravizza et al., 2006; Ravizza et al., 2008).

Given the involvement of the immune system in epilepsy, targeting neuroinflammatory processes has become a potential therapeutic strategy as demonstrated in both animal models and recent clinical trials. Compounds that function to attenuate seizure-induced neuroinflammation have shown therapeutic potential in diverse seizure models. For example, direct inhibition of components of IL-1β biosynthesis, along with the HMGB1-TLR4 signaling cascade, have demonstrated beneficial effects. Pharmacological blockade or genetic inactivation in knockout (KO) models of IL-1R1 or TLR4 can significantly reduce seizure susceptibility and prevent seizure-induced damage (Bertani et al., 2017; Maroso et al., 2010; Semple et al., 2017). This effect can be partially attributed to the reduction in neuronal excitability due to the blockade of IL-1β-induced glutamate NMDA receptor-mediated neuronal Ca⁺⁺ influx via signaling of downstream Src kinases (Balosso et al., 2014; Viviani et al., 2003).

Several anti-inflammatory treatments are currently in clinical trials to treat drug-resistant epilepsies. For example, oral compound VX-765 (Belnacasan) inhibits IL-1β biosynthesis and subsequent HMGB1 release by inhibiting the inflammasome via blockade of interleukin converting enzyme (ICE)/caspase-1 (Maroso et al., 2011). In a phase II clinical trial, VX-765 was found to be safe, well tolerated, and efficacious in individuals with drug-resistant partial epilepsy (Kaur et al., 2016). The human recombinant form of IL-1R1 antagonist, Anakinra, has also been shown to be clinically effective in controlling seizures in pediatric and adolescent patients with refractory epilepsy, as well as in children with febrile infection-related epilepsy syndromes (FIRES) (DeSena et al., 2018; Dilena et al., 2019; Jyonouchi and Geng, 2016; Kenney-Jung et al., 2016). The TNFα monoclonal antibody, Adalimumab, has found to be well tolerated and to reduce seizure frequency in an open label pilot study in a small cohort of adults with Rasmussen’s encephalitis, a severe inflammatory neurological disorder that leads to pharmacoresistant epilepsy (Lagarde et al., 2016). In addition, inhibition of IL-6 biosynthesis with an IL-6 receptor inhibitor (Tocilizumab) terminated seizures in 6 of 7 adults with new onset refractory SE (NORSE). However, 2 individuals in the study experienced severe adverse events that were related to infections during treatment (Jun et al., 2018). While heightened proinflammatory cytokine expression is consistently observed in preclinical models, findings have been variable across different human epilepsies (Bauer et al., 2009; Li et al., 2011; Peltola et al., 2000). Certain conditions, such as in FIRES and NORSE, in which there is significant upregulation in inflammatory cytokines, have shown to be responsive to monoclonal antibody treatments, but further research is needed to ascertain whether these treatments can be utilized in other human epilepsies (Choi et al., 2011; Jun et al., 2018; Sakuma et al., 2015).

Beyond inhibition of the HMGB1-TLR4 signaling cascade and direct antagonism of proinflammatory cytokine synthesis, several other targets that reduce neuroinflammation have been investigated. For example, targeting of cyclooxygenase enzymes (COX1, COX2) to interfere with prostaglandin synthesis has been examined extensively in epilepsy (Rojas et al., 2019). Several animal studies have shown that blockade of COX2 significantly reduces seizure frequency and severity, as well as is neuroprotective against seizure-induced mossy fiber sprouting and neurogenesis (Jung et al., 2006; Ma et al., 2012). However, some studies have shown that inhibition of COX2 worsened chronic seizures and produced adverse effects, suggesting the effects of inhibition are time and model dependent, and could also depend on the type of prostaglandins that are generated following seizures (Holtman et al., 2010; Holtman et al., 2009; Kim et al., 2008; Polascheck et al., 2010; Yoshikawa et al., 2006). Clinical trials with Ibuprofen and Aspirin, both COX inhibitors, have shown variable effects. For example, in pediatric populations with Sturge-Weber syndrome and adults with focal-onset epilepsy, COX inhibition reduced seizure frequency, while in a randomized placebo-controlled trial in children with febrile seizures COX inhibition had no preventative effect on seizure reoccurrence (Bay et al., 2011; Godfred et al., 2013; Lance et al., 2013; van Stuijvenberg et al., 1998).

4. Interaction between mTOR and neuroinflammatory signaling in epilepsy

4.1. Link between the mTOR pathway and inflammation in normal physiological conditions

Immune signaling and PI3K/Akt/mTOR pathway activity are known to interact both in normal physiological conditions and in diseased states. Activation of mTOR has shown to be critical for the development of immune cells in the CNS, as well as many cellular processes of immune cells are regulated by mTOR signaling. There is evidence for both immune-stimulatory and immunosuppressive functions mediated by mTOR activation and the exact mechanisms underlying the complex interplay have yet to be fully elucidated. Inhibition of pathway activity has been shown to impair the maturation and function of dendritic cells (DCs) and inhibit T cell proliferation, demonstrating the importance of mTOR signaling in early immune system development (Thomson et al., 2009). Activation of mTOR also influences cytokine production and release, especially in DCs. For example, mTOR activation suppresses caspase-1 activation and therefore IL-1β production, as well as upstream PI3K negatively regulates TLR-mediated IL-12 biosynthesis from DCs (Fukao et al., 2002; Schmitz et al., 2008). The mTOR pathway can also be activated by several TLR ligands and cytokines, including in both mouse and human monocytes, DCs, and neutrophils (Haidinger et al., 2010; Lehman et al., 2003; Weichhart et al., 2015).

While rapamycin is known for its immunosuppressive effects, particularly in organ transplantations, inhibition via rapamycin or its derivative everolimus have also shown to have immunostimulatory effects in individuals (Touzot et al., 2012; Weichhart et al., 2015). For example, transplant recipients administered mTOR inhibitors have presented with proinflammatory side effects, including elevated NFkB-mediated inflammatory responses in blood leukocytes and increased IL-12-induced signaling (Gallon et al., 2015). Despite the role both neuroinflammatory signaling and mTOR pathway activity have in seizure development in epilepsy, how these two processes influence one another during the development of epilepsy has only recently begun to be investigated. The rest of the review summarizes the evidence to date demonstrating how the two influential processes interact in epilepsy, and potential therapeutics that target both mTOR signaling and seizure-induced neuroinflammation.

4.2. Therapeutic potential of pharmacological mTOR inhibitors to reduce seizure-induced neuroinflammation

Several studies have demonstrated that the anti-epileptogenic effects of rapamycin could potentially be due to its immunosuppressant and anti-inflammatory properties. In one of the first studies to demonstrate the beneficial effects of rapamycin, treatment was examined in a mouse model of TSC (Tsc1^GFAPCKO), a common genetic cause of epilepsy (Zeng et al., 2008). Zeng et al. (2008) found that rapamycin treatment starting on postnatal 14 prevented the development of epilepsy and premature death in Tsc1^GFAPCKO mice compared to vehicle-treated mice. In addition, treatment starting at a later developmental timepoint (6 weeks of age), was able to suppress seizures and prolong survival in KO mice. The prevention and suppression of seizures was associated with a reduction in astrogliosis in the neocortex and hippocampus of Tsc1^GFAPCKO mice, along with rapamycin-treatment prevented the enlarged brain size in KO mice (Zeng et al., 2008). Van Vliet et al. (2012) examined the effects of rapamycin treatment given daily for 7 days, beginning 4 h following SE onset induced by angular bundle stimulation, followed by administration every other day for 6 weeks after SE. In addition to reducing seizure frequency, treatment with rapamycin exhibited neuroprotective properties by reducing SE-induced hilar cell loss and mossy fiber sprouting but had no effect on microglial or astrocyte activation. Blood-brain barrier leakage following SE was minimally detected in rapamycin-treated rats, suggesting rapamycin may exert its anti-epileptogenic effects via restoration of BBB integrity (van Vliet et al., 2012). Van Vliet et al. (2016) further investigated the role of BBB leakage in the chronic phase after SE by examining glial and vascular markers both in vitro and in vivo. Six weeks following kainic acid (KA)-induced SE, rapamycin-treated rats had significantly reduced BBB leakage specifically in the piriform cortex and amygdala, which was associated with a milder form of epilepsy. In this study, rapamycin-treatment was also associated with a reduction in microglia and astrocyte activation, brain blood vessel density, and mossy fiber sprouting in the chronic phase of epilepsy in rats (van Vliet et al., 2016).

Astrocytes are implicated in promoting epileptogenesis by diverse mechanisms, including contributing to BBB disruption, the release of inflammatory molecules, and dysregulated astrocyte-neuronal synaptic signaling in epilepsy (Michinaga and Koyama, 2019; Seifert et al., 2010). During the chronic phase of epileptogenesis in an intrahippocampal mouse model of KA seizures and in individuals with drug-resistant TLE, mTOR activation has been found to be upregulated in reactive astrocytes (Sha et al., 2012). In addition, deletion of mTOR in reactive astrocytes prevented increases in seizure frequency and astrogliosis in mice with chronic epilepsy following KA-induced SE, however, did not have an effect on associated mossy fiber sprouting (Wang et al., 2017). In neuronal-specific Depdc5 KO mice that exhibit seizures and mTOR hyperactivation, rapamycin treatment normalized the increases in mTOR activity, along with the astrogliosis present in the brains of Depdc5 KO mice (Yuskaitis et al., 2019). Guo et al. (2017) utilized two-photon imaging to directly monitor in vivo the structural effects of KA seizures on cortical astrocytes. With both pre-treatment and post-treatment regimens, rapamycin (6 mg/kg) was able to prevent seizure-induced astrocyte vacuolization, astrogliosis, and decreases in astrocyte size (Guo et al., 2017). However, in both treatment regimens, rapamycin did not have an effect on astrocyte number following KA seizures.

The timing of rapamycin treatment has found to be critical in elucidating the seizure-suppressing effects of administration, and many studies have found the suppressant effects to be lost following cessation of treatment. A study by Drion et al. (2016) demonstrates the importance of timing when administering rapamycin in an electrogenic post-SE rat model of TLE. Rapamycin treatment (6 mg/kg) given to rats for 3 weeks following seizures significantly suppressed seizures, however, there was a gradual reappearance of seizures during the following 5 weeks once treatment was terminated and rapamycin blood levels decreased. When rapamycin (3 mg/kg) was given for 5 days in the chronic phase of epilepsy, treatment significantly decreased seizures, but as rapamycin washed out, seizure frequency began to increase again, although at still lower levels than baseline seizure frequency (Drion et al., 2016). In addition, pre-treatment with rapamycin (3 mg/kg) for three days prior to seizure induction did not reduce seizure frequency, concluding that rapamycin treatment was not able to prevent the development of epilepsy. In the same electrogenic rat model of TLE, rapamycin (6 mg/kg) administration starting 4 hours post-SE and continued only for 7 days did not reduce SE-induced increases in IL-1β, IL-6, TGFβ, Hmox-1, or HMGB1 one week following SE (Drion et al., 2018).

Other pharmacological therapies besides rapamycin that influence both inflammatory processes and mTOR signaling have begun to be investigated. One therapeutic commonly used as an immunosuppressant in organ transplantation surgeries, mycophenolate mofetil (MMF), has recently been examined in a rat model of lithium pilocarpine-induced TLE. The active form of MMF, mycophenolic acid (MPA), functions in nucleic acid biosynthesis pathways and is a rate-limiting enzyme in the synthesis of guanosine nucleotides which can result in inhibition of T-cell lymphocyte proliferation (Allison and Eugui, 2000). Following a latent phase (28 days) post-seizures in a mouse model, mice that were treated with MMF had reduced severity of spontaneous seizures, however treatment had no effect on seizure incidence compared to controls (Mazumder et al., 2019). Mycophenolate mofetil also had beneficial effects on seizure-induced behavioral impairments, as treatment reduced aggression and depressive-like behavior, and increased spatial and recognition memory in seizure rats. These behavioral changes could potentially be mediated by both inflammatory and cell signaling processes, as MMF was able to attenuate seizure-induced increases in glial fibrillary acidic protein (GFAP), mTOR, pS6, and S6 protein levels, as well as attenuate gene expression of many pro-inflammatory cytokines and mTOR pathway components (Mazumder et al., 2019).

4.3. Non-pharmacological agents that impact mTOR and neuroinflammatory signaling in epilepsy

Rapamycin treatment has shown to only suppress epileptogenesis with continual treatment, and prolonged treatment can also lead to many adverse side effects. For instance, mTOR inhibition with rapamycin or its analogs can lead to immunosuppression, mucositis, skin reactions, and dose-dependent increases in cholesterol and triglycerides (McDaniel and Wong, 2011; Tsang et al., 2007). As a result, other non-pharmacological compounds that have anti-epileptogenic properties by inhibiting mTOR activity and neuroinflammatory processes have been examined. One potential anti-epileptogenic therapeutic is curcumin, the primary compound present in turmeric from the Curcuma longa plant (Dhir, 2018; Rahmani et al., 2018).

The anti-inflammatory effects of curcumin can partially be attributed to both inhibition of the mTOR pathway and mitogen-activated kinase pathways (Beevers et al., 2013; Drion et al., 2018). In a study using a hippocampal-entorhinal cortex culture model of epileptogenesis, curcumin treatment reduced spontaneous seizure-like events, which was associated with a trending decrease in IL-1β and IL-6 and reduced mTOR activity, exemplified by low pS6 expression (Drion et al., 2019). In human astrocyte cultures, Drion et al. (2018) found that pre-treatment and simultaneous treatment with curcumin along with IL-1β stimulation, significantly reduced pro-inflammatory IL-6 expression, as well as simultaneous treatment significantly reduced COX-2 levels in vitro. However, curcumin administration for 7 days post-SE in a rat model of TLE did not reduce SE-induced elevations in IL-1β, IL-6, TGFβ, Hmox-1, or HMGB1 (Drion et al., 2018). While evidence suggests curcumin treatment could be a promising alternative to rapamycin that has similar anti-epileptogenic actions, additional studies are critical to determine the mechanisms underlying these effects.

Another non-pharmacological option that influences inflammation and mTOR activity could be resveratrol, a natural polyphenol compound found in a variety of plant species that has broad biological activity (Castro et al., 2017; Lu and Wang, 2015). It has been shown to have anti-epileptogenic effects in several seizure models, with administration of the compound prior to seizure induction preventing SE-induced neuroinflammation, neurodegeneration, and aberrant neurogenesis (Castro et al., 2017; Gupta et al., 2002; Lu and Wang, 2015; Shetty, 2011; Virgili and Contestabile, 2000; Wu et al., 2009). Its beneficial effects may partially involve suppression of NFkB and pro-inflammatory cytokine synthesis, which can be influenced by upstream mTOR signaling (Lu et al., 2010; Ren et al., 2013). Resveratrol administered 30 min prior to pilocarpine seizures in a rat model suppressed IL-1β, COX-2, and inducible nitric oxide synthase 3 hours after SE (Wang et al., 2013). Pre-seizure resveratrol treatment also increased expression of AMP-activated protein kinase (AMPK), which was associated with a reduction in phosphor-mTOR and phospho-S6, suggesting the anti-inflammatory properties of resveratrol following seizures involve regulation of AMPK/mTOR signaling (Wang et al., 2013).

Dietary therapies have also shown promise in controlling seizures in those with drug-refractory epilepsy (Bough and Rho, 2007; D’Andrea Meira et al., 2019; Lima et al., 2014; Marsh et al., 2006; Patel et al., 2010). The ketogenic diet (KD) consists of high-fat, low-carbohydrate, and adequate protein levels, which leads to the production of ketone bodies that function as the primary source of metabolic energy. The KD has both anti-inflammatory activity and may partially inhibit mTOR activity, contributing to its anticonvulsant effects. The diet’s anti-inflammatory and anti-convulsant effects may involve activation of peroxisome proliferator activated receptors (PPAR), which are involved in the regulation of anti-inflammatory and antioxidant genes (Boison, 2017; Lucchi et al., 2017). Activation of PPARγ is dependent on endogenous fatty acids, such as decanoic acid, that are increased in the KD diet. In a Kv1.1 KO mouse model that exhibits spontaneous epileptic activity, a PPARγ antagonist prevented the seizure-reducing effects of the KD diet in Kv1.1 KO mice (Simeone et al., 2017). In addition, the KD increased the latency to seizures, but was ineffective in PPARγ KO mice, further supporting the role of PPARs in the anticonvulsant mechanisms of the KD (Simeone et al., 2017).

Beyond the anti-inflammatory effects of the KD, it may also partially exert its effects through inhibition of mTOR signaling. Treatment with the KD following a single KA seizure has shown to reduce the frequency of spontaneous seizures, which was associated with less mossy fiber sprouting (Muller-Schwarze et al., 1999; Su et al., 2000). These anti-epileptogenic effects could be associated with a reduction in mTOR activity, as mTOR is known to be elevated following KA seizures and regulates pathological changes associated with epileptogenesis (Blazejczyk et al., 2017; Buckmaster et al., 2009; Zeng et al., 2009). The KD reduces insulin levels, which is an activator of PI3K/Akt/mTOR signaling and may indirectly reduce mTOR activity (Westman et al., 2008; Yoon, 2017). In addition, pro-inflammatory cytokine expression is reduced with KD therapy, and cytokines can activate the PI3K/Akt/mTOR pathway via TLR stimulation, thus serving as another potential link between the two mechanisms (Dupuis et al., 2015; Kim et al., 2012; Troutman et al., 2012). Mechanisms underlying the anti-seizure and anti-epileptogenic effects of the KD in epilepsy still requires additional investigation, especially regarding whether its effects on inflammation and mTOR could be associated.

4.4. In vitro evidence for link between mTOR signaling and neuroinflammation in epilepsy

The effect of rapamycin on components of immune function in epilepsy, specifically BBB permeability, has also been examined in vitro. To mimic inflammatory conditions observed during seizures, pro-inflammatory cytokines can be added to the medium of in vitro models which will result in disruption and reduced integrity of the BBB (Marchi et al., 2012; Pan et al., 2011; van Vliet et al., 2016). In the presence of pro-inflammatory cytokine TNFα, van Vliet et al. (2016) found that rapamycin treatment improved overall BBB function and demonstrated a beneficial role of rapamycin in restoration of the BBB following injury (van Vliet et al., 2016). However, this beneficial effect on the BBB could potentially be due to mechanisms other than a reduction in pro-inflammatory cytokine expression, as other studies have found contrasting results following rapamycin treatment. Drion et al. (2018) used astrocyte cultures to examine the effects of rapamycin on pro-inflammatory cytokine expression following an immune stimulus. Following both in vitro pre-treatment and simultaneous treatment of rapamycin along IL-1β stimulation, IL-1β-induced increases in pro-inflammatory IL-6 levels were not affected by treatment. Further, COX-2 mRNA expression was found to be significantly enhanced following rapamycin treatment compared to IL-1β-stimulated expression levels (Drion et al., 2018). In a hippocampal-entorhinal cortex slice culture model, Drion et al. (2019) examined the effects of rapamycin in slices obtained from 6-day old rat pups and found that in vitro administration of rapamycin from day 2 and onward, did not reduce spontaneous seizure-like events in cornus ammonis (CA) 1 neurons (Drion et al., 2019).

5. Perspectives and conclusions

Epilepsy is a complex neurological disorder that can be the result of several acquired and genetic etiologies. Current treatments are primarily symptomatic, and there is an unmet need for therapies that target the pathological mechanisms that underlie epileptogenesis. Both signaling via the mTOR pathway and neuroinflammatory processes have been shown to play a significant role in the development of chronic epilepsy (Aronica et al., 2017; Citraro et al., 2016; Wong, 2013a). In order to gain a better understanding of the intersection between mTOR signaling and the immune system in epilepsy, it is critical that studies examine the impact on both components when utilizing solely an mTOR inhibitor or anti-inflammatory agent. There is the possibility that several previously found benefits of one inhibitor, could also have been efficacious in controlling aspects of the other of the two non-targeted mechanisms. In regard to both inhibition of mTOR signaling and neuroinflammation, future studies are needed to determine the optimal timing of inhibition. For instance, many pre-clinical studies have found discontinuation of rapamycin treatment to be associated with reoccurrence of seizures, and for the timing of treatment to have differential effects on seizure-induced damage in the brain (Drion et al., 2016; Huang et al., 2010). In addition, neuroinflammation can be triggered at varying timepoints post-seizures dependent on the type of injury, so determining the optimal timing of intervention will most likely vary across diverse clinical populations with epilepsy. Understanding the temporal dynamics of epileptogenesis will clarify when the most beneficial time window of intervention will be to inhibit mTOR and neuroinflammatory processes simultaneously.

As a result of many current AEDs being primarily symptomatic, many individuals with epilepsy suffer from several comorbid cognitive, behavioral, and psychiatric conditions (Keezer et al., 2016; Srinivas and Shah, 2017). Targeting the pathological mechanisms that underlie epilepsy could potentially alleviate some of these comorbid impairments. Several studies have begun to elucidate the molecular impact of inhibition of mTOR or inflammatory mechanisms in epilepsy, however, there are a lack of studies examining the behavioral impact of these therapeutics in pre-clinical models. Due to these therapeutics targeting mechanisms underlying epilepsy, one would suspect a beneficial effect on associated seizure-induced cognitive and behavioral impairments, but without inclusion of behavioral components in studies this is difficult to determine. For example, studies have examined the impact of mTOR inhibition with rapamycin on learning and memory, activity levels, and aggression, yet there are few studies examining the effects of inhibition on autistic-like behaviors (Huang et al., 2012a; Raffo et al., 2011; Schneider et al., 2017; Talos et al., 2012; Yuskaitis et al., 2019). Many epilepsies are comorbid with other neurodevelopmental disorders, especially Autism spectrum disorder, and beneficial effects of these treatments in pre-clinical models may give rise to novel targets to attenuate impairments observed in autism as well (Bernard and Benke, 2015; Besag, 2017). In conclusion, the intersection between mTOR signaling and neuroinflammatory mechanisms in epilepsy serves as a promising avenue to pursue in the search for alternative therapeutic treatments that are anti-epileptogenic and target the pathological causes of the disorder.

Highlights.

• Hyperactive mTOR signaling is implicated in several seizure-induced brain changes

• Neuroinflammation can impact hyperexcitability and promote epileptogenesis

• mTOR and inflammatory signaling interact in both physiological and diseased states

• Dual inhibition of both processes could be a potential anti-epileptogenic therapy