Annamaria Vezzani PhD, Department of Neuroscience, Mario Negri Institute for Pharmacological Research, Via G. La Masa 19, 20156 Milano, Italy. (fax: +39-02-3546277; e-mail: email@example.com).
Abstract. Maroso M., Balosso S., Ravizza T., Liu J., Bianchi M.E., Vezzani A. (Mario Negri Institute for Pharmacological Research, Milano; and San Raffaele University and Research Institute Milano; Italy). Interleukin-1 type 1 receptor/Toll-like receptor signalling in epilepsy: the importance of IL-1beta and high-mobility group box 1 (Symposium). J Intern Med 2011; 270: 319–326.
Inflammatory processes in brain tissue have been described in human epilepsy of various aetiologies and in experimental models of seizures. This, together with the anticonvulsant properties of anti-inflammatory therapies both in clinical and in experimental settings, highlights the important role of brain inflammation in the aetiopathogenesis of seizures. Preclinical investigations in experimental models using pharmacological and genetic tools have identified a significant contribution of interleukin-1 (IL-1) type 1 receptor/Toll-like receptor (IL-1R/TLR) signalling to seizure activity. This signalling can be activated by ligands associated with infections (pathogen-associated molecular patterns) or by endogenous molecules, such as proinflammatory cytokines (e.g. IL-1beta) or danger signals [damage-associated molecular patterns, e.g. high-mobility group box 1 (HMGB1)]. IL-1beta and HMGB1 are synthesized and released by astrocytes and microglia in the rodent brain during seizures. Notably, a rapid release of HMGB1 from neurons appears to be triggered by proconvulsant drugs even before seizure occurrence and is involved in their precipitation of seizures. The activation of IL-1R/TLR signalling mediates rapid post-translational changes in N-methyl-d-aspartate-gated ion channels in neurons. A long-term decrease in seizure threshold has also been observed, possibly mediated by transcriptional activation of genes contributing to molecular and cellular plasticity. This emerging evidence identifies specific targets with potential anticonvulsant effects in drug-resistant forms of epilepsy.
Epilepsy is a brain disorder characterized by an enduring predisposition to generate epileptic seizures and by the neurological, cognitive and psychological consequences of this condition, International League against Epilepsy (ILAE definition, 2005). Seizures are transient alterations in behaviour owing to abnormal, synchronized and repetitive burst firing of neurons in the central nervous system (CNS). Epilepsy affects up to 1% of the general population; about 35% of cases are symptomatic, being associated with an identifiable CNS injury , whereas the remainder are because of either genetic or unknown causes.
About 30% of epileptic patients are considered to be drug resistant as they do not respond to the available anti-epileptic drugs (AEDs) . Moreover, these AEDs are essentially only able to alleviate symptoms, and thus, there is an urgent need for novel drugs with disease-modifying properties .
In the last decade, studies using experimental models of seizures have highlighted new targets for pharmacological intervention, with inflammatory molecules attracting much interest. Clinical and experimental studies have provided proof-of-concept evidence that brain inflammation is an important factor in epilepsy . In particular, high levels of proinflammatory cytokines [e.g. interleukin (IL)-1beta, tumour necrosis factor (TNF)-alpha], danger signals [high-mobility group box (HMGB)1, S100 beta] and downstream inflammatory mediators (e.g. prostaglandins, the complement system) have been measured in epileptogenic tissue resected at surgery from patients affected by epilepsy of various aetiologies . The major contributors to the synthesis of these inflammatory mediators are brain-resident cells such as activated microglia, astrocytes and neurons. Endothelial cells of the blood–brain barrier (BBB) and leucocyte extravasation into brain parenchyma also contribute to brain inflammation. Studies in experimental models have demonstrated that seizure activity per se induces long-lasting brain inflammation, even without concomitant brain pathology. On the other hand, the presence of cell loss or aberrant cells and displastic neurons, such as those observed in malformations of cortical development, appears to contribute to inflammation in brain tissue [4, 5].
Pharmacological findings indicate that brain inflammation in epilepsy is not a mere epiphenomenon but contributes to seizures. Thus, in the clinical setting, various drugs and treatments with wide-spectrum anti-inflammatory properties, such as steroids and adrenocorticotropic hormone, can control AED-resistant seizures in some epileptic conditions. In addition, pharmacological or genetic interference with specific proinflammatory pathways can drastically reduce seizures in various experimental models [4, 6]. Moreover, the induction of systemic and brain inflammation in rodents using molecules derived from bacteria or viruses results in both acute and chronic lowering of seizure threshold . Accordingly, various epileptogenic injuries (i.e. infection, stroke, trauma, febrile seizures) provoke long-lasting brain inflammation in experimental models and are associated with acute symptomatic seizures and a higher risk of developing epilepsy in humans [1, 8, 9].
These findings prompted further studies to investigate which of the proinflammatory pathways activated in epileptic tissue could be targeted pharmacologically to provide significant anticonvulsant effects.
Proinflammatory cytokines and danger signals: IL-1beta and HMGB1
Particular attention has been focused on the proinflammatory cytokine IL-1beta and the danger signal HMGB1. These two molecules are released by activated cells in epileptic tissue and contributed to increased neuronal excitability.
Immunocytochemical and biochemical investigations have provided evidence that cell-signalling pathways mediated by IL-1beta and HMGB1 are activated in human and experimental epilepsy.
Activated microglia and astrocytes are prominently involved in the production and release of IL-1beta in epileptic tissue. These cells produce elevated cytokine levels within 30 min from seizure onset, then the wave of inflammation propagates to endothelial cells of the BBB, and may also involve clusters of neurons in the areas in which seizures originate and spread . The induction of IL-1beta in perivascular astrocytes and endothelial cells of the microvasculature may contribute to BBB breakdown, which occurs together with inflammation in experimental and human epileptic tissue [10–14].
The IL-1 type 1 receptor (IL-1R1) mediates the biological actions of released IL-1beta and is overexpressed in neurons and astrocytes in epileptic tissue [10, 15–17]. The cells that produce IL-1beta also express its receptor, suggesting the involvement of both autocrine and paracrine signalling.
High-mobility group box 1 is ubiquitously expressed in neuronal nuclei in physiological conditions; however, after proconvulsant injuries, HMGB1 translocates from the nucleus to the cytoplasm, a phenomenon usually associated with its release from cells [18, 19]. Increased nuclear expression and cytoplasmic translocation of HMGB1 are also observed in activated microglia and astrocytes during seizure activity. A similar pattern of HMGB1 induction and cytoplasmatic translocation was observed in human epileptic tissue from patients with temporal lobe epilepsy and malformations of cortical development [19, 20].
The search for receptors activated by released HMGB1 led to the observation of prominent upregulation of Toll-like receptor 4 (TLR4) in neurons and in activated astrocytes during seizure activity, which led to further investigation into the role of HMGB1/TLR4-mediated signalling in seizures.
Proconvulsant activities of IL-1beta and HMGB1
IL-1beta. The first evidence of an active role for IL-1beta in seizures was provided more than 10 years ago by two studies showing that the intracerebral application of IL-1beta increases seizure activity provoked by different chemoconvulsant drugs in rats or mice [21, 22], and that the intracerebral application of IL-1 receptor antagonist (IL-1ra) induces powerful anticonvulsant effects . As IL-1ra is an endogenous competitive blocker of IL-1R1 and its specific action is to terminate or prevent the biological actions of endogenous IL-1beta, this finding provided compelling evidence that elevated brain levels of IL-1beta contribute to seizure activity. Accordingly, mice overexpressing IL-1ra in astrocytes or lacking IL-1R1 were intrinsically less susceptible to seizures . Subsequent studies showed that systemic administration of IL-1ra in rats reduced the incidence of status epilepticus , and this antagonist also delayed the development of kindling in young rats . Significant seizure reduction was also attained in rodents by blocking IL-1beta biosynthesis using specific IL-1beta-converting enzyme (ICE/caspase-1) inhibitors [25, 26], or in mice with a deletion of the ICE/caspase-1 gene . Notably, ICE/caspase-1 activity is increased in human epileptic tissue from drug-resistant patients . Based on these studies, a clinical trial has been initiated in the USA in patients with partial epilepsy and drug-resistant seizures to investigate the effect of an ICE/caspase-1 inhibitor that has shown powerful anticonvulsant properties in rodents [25, 26, 28] (http://clinicaltrials.gov/ct2/show/NCT01048255 (2010).
Because endogenous IL-1beta has proconvulsant actions, it is likely that this cytokine is released by cells following an inciting event. Activation of the inflammasome, a multiprotein complex that includes the cysteine protease ICE/caspase-1, is a crucial step required for IL-1beta release. A reduction in intracellular K+ and a rise in cytosolic Ca2+ and Na+ are required for inflammasome assembly and ICE/caspase-1 activation. Notably, these ionic modifications are induced in neurons and astrocytes by hyperexcitability underlying the generation of seizures. Stimulation of purinergic P2X7 receptors by extracellular ATP (released from activated or dying cells) can also activate the inflammasome ICE/caspase-1 complex .
Toll-like receptor 4. The involvement of TLR4 in seizures was first demonstrated by experiments showing that administration of lipopolysaccharide (LPS), a component of the bacterial cell wall and a powerful TLR4 ligand, resulted in acute and long-term decreases in seizure threshold (reviewed by Riazi et al. ). Subsequently, application of LPS to the rat somatosensory cortex was shown to rapidly evoke recurring epileptiform activity . The effects of LPS were prevented by selective TLR4 antagonists and by IL-1ra, indicating that IL-1beta release was involved.
High-mobility group box 1. Because seizures often occur in the absence of pathogens, we searched for a putative endogenous ligand of TLR4 that might contribute to seizure activity. We found that HMGB1 activates TLR4 and contributes to acute and chronic seizures in mouse models . This is supported by a number of observations. First, intracerebral injection of HMGB1 increases seizure activity that was induced either by stimulation of glutamatergic or by inhibition of gamma-aminobutyric acid (GABA)ergic receptors. Second, HMGB1 had no pro-convulsant effects when injected into Tlr4 mutant mice; moreover, Tlr4 mutant mice were instrinsically resistant to seizures. Finally, intracerebral administration of selective TLR4 antagonists or BoxA, a fragment of HMGB1 with receptor antagonist activity, had significant anti-seizure effects. Notably, these antagonists also inhibit epileptic activity resistant to classical AEDs .
It is worth noting that both genetic interference and pharmacological interference with the HMGB1/TLR4 axis not only reduced seizure frequency and duration, but also accelerated seizure onset, which usually occurs within minutes in kainate- and bicuculline-induced seizure models . This observation highlights the crucial role of HMGB1 in the precipitation of the first seizure after a proconvulsant challenge, thus providing an indirect demonstration of its fast release from pre-existing neuronal sources. The increased HMGB1 synthesis in glial cells observed during seizures represents a second wave of HMGB1 production and release which may contribute to sustain brain inflammation and play a role, in concert with IL-1beta, in seizure recurrence . Extracellular HMGB1 has a dual origin, demonstrated by the fact that it can be passively released by cells undergoing necrosis and actively secreted from cells in response to inflammatory stimuli or stressful events; macrophages engulfing apoptotic cells can also secrete HMGB1 . The precise mechanisms underlying HMGB1 release are not yet fully understood; however, increase in the level of intracellular Ca2+ and activation of protein kinase C may be required. In addition, HMGB1 methylation, acetylation and phosphorylation have been associated with increased cytoplasmatic localization and secretion of HMGB1 . We recently showed that neuronal cell cultures undergoing excitotoxicity during exposure to glutamate can release HMGB1 . Moreover, HMGB1 release can be evoked from human astrocyte cultures exposed to IL-1beta . Conversely, HMGB1 provokes the release of proinflammatory mediators from a variety of cell types including astrocytes .
Further studies are required to investigate the sequence of events promoting IL-1beta and HMGB1 release during aberrant excitability, cell injury or their combination. Knowledge of this process will be important to understand the molecular mechanisms underlying the induction and the persistence of brain inflammation in epilepsy. In this context, it is noteworthy that cooperation between HMGB1 and IL-1beta, which may require the binding of these two molecules, appears to be a crucial event to amplify the inflammatory response [34, 35].
Activation of IL-1R/TLR signalling alters neuronal excitability via both rapid nontrascriptional actions and slower transcriptional activation of genes under the control of nuclear factor-κB (NF-κB) and activator protein-1 (AP-1) .
Rapid effects. Rapid actions of IL-1beta have been reported in hippocampal slices and in cultured pyramidal neurons: this cytokine reduces synaptically mediated GABA inhibition in the CA3 area via unidentified kinases [37, 38] and increases neuronal excitability in the CA1 area by reducing N-methyl-d-aspartate (NMDA)-induced outward current via activation of P38 MAPK, mitogen-activated protein kinases kinases . IL-1beta also potentiates NMDA receptor function in cultured hippocampal neurons by enhancing within minutes the NMDA-mediated neuronal Ca2+ influx; this effect is the downstream molecular event induced by sphingomyelinase-dependent and Src kinase-catalysed NR2B subunit phosphorylation . A similar mechanism of rapid modification of ion channel conductance was described in the study by Davis et al. in hypothalamic neurons [41, 42]. The same cascade of molecular events appears to be induced in neurons by HMGB1 . Pharmacological interruption of this rapid nontranscriptional pathway, using sphingomyelinase or Src K antagonists, or ifenprodil which blocks NR2B-NMDA receptors, prevents the proconvulsant activities of both IL-1beta and HMGB1 [19, 43]. Thus, increased NMDA function is a crucial mechanism by which endogenous activation of IL-1R1 and TLR4 contributes to seizures, in accordance with the crucial role of NMDA receptors in seizure generation and recurrence .
Long-term effects. The activation of IL-1R1/TLR4 signalling may also trigger transcriptional changes that could perpetuate inflammation via NF-κB-dependent transcription of inflammatory genes and lead to a chronically decreased seizure threshold by inducing the expression of genes involved in neurogenesis, cell death, synaptic molecular reorganization and plasticity . These processes occur during epileptogenesis after an initial precipitating event and are considered to contribute to the transformation of normal brain tissue to seizure-generating tissue .
Amongst the long-term processes that are likely to be affected by brain inflammation, and in particular by IL-1beta and HMGB1, is the transcriptional activation of endothelial cells and perivascular astrocytes leading to damage of the BBB [13, 14, 45]. The role of changes in BBB permeability in epileptogenesis has become increasingly recognized; in particular, serum albumin extravasation into the brain parenchyma has been shown to activate a transforming growth factor β-mediated gene programme in astrocytes leading to downregulation of both Kir4.1 K+ channels and the glutamate transporter. These changes result in impaired ionic buffering and increased extracellular glutamate levels which contribute to chronic hyperexcitability [46, 47].
Lipopolysaccharide-induced long-term modifications of neuronal excitability have been shown to depend on the release of TNF-alpha and IL-1beta from activated microglia cells . Notably, long-term increased excitability induced by LPS in rodents was associated with specific and persistent changes in the expression pattern of NMDA receptor subunits in the cerebral cortex and hippocampus. Increased transcript levels of these receptor subunits predict transcriptional activation of the related genes .
The activation of IL-1R/TLR signalling following a pathogenic threat represents a homeostatic response of the brain and may cause pathological neuronal hyperexcitability only when the extent or duration of signalling activation exceeds the homeostatic threshold. This situation can occur after excessive production and release of IL-1beta and HMGB1 and/or because of receptor upregulation in brain cells. An efficient resolution of inflammation is required in order to avoid detrimental effects in brain tissue, and this can be achieved in multiple ways, including through the production of anti-inflammatory mediators, endogenous receptor antagonists, soluble receptors and decoy receptors. In the case of IL-1beta, it has been shown that IL-1ra is produced during seizures with some delay and never in greater amounts than IL-1beta [48, 49]; thus, the brain, in contrast to the periphery, may be unable to rapidly terminate the effects of elevated IL-1beta levels.
The identification of IL-1R/TLR signalling as an important contributor to acute and long-term changes in brain excitability and to seizures (Table 1) highlights the possibility that innate inflammatory responses to microbial pathogens (pathogen-associated molecular patterns) or to endogenous danger signals (damage-associated molecular patterns) may converge on common molecular targets (Fig. 1). This may explain the causal link between CNS infections or brain damage and epilepsy [4, 8].
Table 1. Pharmacological targeting of IL-1R/TLR signalling in seizure models
See text for details. Chemoconvulsant drugs include kainic acid, bicuculline, penthylenetetrazol. Data reviewed in Ref. [7, 51, 52].
Given the role of IL-1R/TLR signalling in seizures, the preclinical findings support the hypothesis that pharmacological inhibition of this signalling may be useful to treat seizures in AED-resistant human epilepsy. Of note, a number of drugs targeting this signalling are undergoing clinical evaluation for treatment of sepsis and inflammatory diseases, and some, such as the IL-1 receptor antagonist Anakinra, are already in clinical use showing anti-inflammatory and therapeutic effects in chronic peripheral inflammatory conditions [45, 50]. Considering the constraints imposed by the BBB for systemic drug access to the brain, and the safety issues related to chronic or long-term drug administration in epilepsy, these drugs might provide therapeutic potential by reducing inflammatory processes in the epileptic brain and thereby counteracting neuronal hyperexcitability.
The authors are grateful to the Cariplo Foundation for supporting part of this study. Mattia Maroso received a fellowship from NeuroGlia (EU-FP7-project 202167) and Jaron Liu from the EU-FP7 training program ‘International Graduate Programme in Molecular Medicine’.
Conflict of interest statement
Dr Vezzani owns intellectual property on a patent on the use of ICE/caspase-1 inhibitors in epilepsy issued from Vertex Pharmaceuticals, Cambridge, USA, that does not involve any financial interest or profit. M.E. Bianch is founder and part-owner of HMG Biotech, a biotech that sells products and idea based on the functions of HMGB proteins.