High mobility group box 1 in the inflammatory pathogenesis of epilepsy: profiling circulating levels after experimental and clinical seizures

Abstract

BackgroundHigh mobility group box 1 (HMGB1) is a chromatin binding protein that is passively released by necrotic cells and actively secreted in response to inflammatory stimuli in a hyperacetylated form. HMGB1 is upregulated in the brain after injury and can regulate localised inflammatory reactions resulting in seizures. Antiepileptic drugs prevent seizures but have no disease modifying effect or influence on natural history. Immunomodulatory antiepileptic drugs have huge potential, but can have serious adverse effects. Patient stratification is required to maximise the benefit to risk ratio. We investigated serum and brain concentrations of HMGB1 in rodent models of seizures and epilepsy and in people with refractory epilepsy. MethodsWe used three experimental models in this study. Adult male C57BL/6J mice received repeated intraperitoneal injections of kainic acid (KA) until the onset of convulsive status epilepticus; status epilepticus was terminated after 2 h with diazepam, with brain and blood samples obtained at 3, 6, 24, and 72 h and 7 and 14 days thereafter. Tonic seizures were induced in adult male CF1 mice by maximal electroshock (MES) delivered via corneal electrodes, with brain and blood samples obtained at 1, 4, 8, 16, and 24 h after seizure induction. Finally, serial samples were obtained at baseline, 1, 4, 8, and 12 h after an observed seizure from people with drug refractory epilepsy undergoing videoelectroencephalograph telemetry. Total HMGB1 expression was measured by western blot and immunohistochemistry in brain and by ELISA in serum. Expression of individual HMGB1 isoforms was determined by liquid-chromatography tandem mass-spectrometry (LC-MS/MS). FindingsHMGB1 expression was significantly raised in hippocampus and cortex (p<0·001) at 24 h after KA-induced status epilepticus. Increased cytoplasmic HMGB1 staining was observed at 24 h and associated with an increase in the hyperacetylation of HMGB1 in the hippocampus. In the MES model, there was a significant increase in hippocampal (p<0·05) and cortical (p<0·05) HMGB1 expression at 24 h after a single tonic seizure. Serum concentrations of HMGB1 peaked 3 h after KA-induced status epilepticus (baseline mean 5·0 ng/mL [SD 0·6] vs 3 h 16·5 [1·5], p<0·0001) and 4 h after MES seizures (baseline 9·2 [2·5] vs 4 h 19·7 [11·3], p<0·05). 14 days after status epilepticus, serum HMGB1 remained significantly raised (18·0 [11·5], p<0·05) and the isoform was hyperacetylated. An analogous rise in the total HMGB1 serum concentration occurred after seizures in patients with epilepsy (baseline 7·1 [1·4] vs 4 h 15·2 [7·0], p<0·05). Mean serum HMGB1 was significantly higher in patients with drug refractory epilepsy than in healthy controls (8·6 ng/mL [3·5] vs 0·7 [0·3], p<0·002) and in patients with epilepsy who had been seizure free for more than 6 months (8·6 [3·5] vs 1·25 [0·71], p<0·0001). InterpretationThese findings suggest that blood and brain HMGB1 concentrations are increased as a result of seizures in both animal models and patients with epilepsy. Increases in HMGB1 might occur as a consequence of seizures. Patients with refractory epilepsy have higher baseline HMGB1, which might predispose them to recurrent seizure activity. The association between HMGB1 and seizures requires further exploration. FundingUK Medical Research Council, ICON, GlaxoSmithKline, AstraZeneca, the Medical Evaluation Unit.