Abstracts

Rationale: Inflammatory pathways are activated by seizures leading to production of the proinflammatory cytokines. For example, interleukin(IL)-1? has been shown to be synthesized in glia in response to proconvulsants. However, IL-1? levels are unlikely to be good biomarkers of seizures because reliably high cerebrospinal fluid concentrations from epilepsy patients have not been found. In somatic tissue, IL-1? induces an increase in GM-CSF granulocyte-macrophage colony-stimulating factor (GM-CSF) synthesis. There are published articles identifying brief peripheral elevation in other inflammatory cytokines such as IL6 in post ictal patients. A dependable plasma biomarker for seizures is yet to be discovered. We included GM-CSF in a global search for potential ictal biomarkers using peri-ictal plasma sampling of inflammatory cytokines. To mirror previous studies subjects were selected for likely temporal lobe seizures. Methods: Plasma samples were obtained from subjects in the epilepsy monitoring unit for routine care (prior to seizures through up to 37 hours post ictal.) Sodium citrate stabilized plasma was collected, immediately put on ice, and then spun in a cold centrifuge. Immediately after centrifugation, the samples were stored at -80 degrees C. The samples were then assayed using a multiplexed ELISA platform (meso scale, Gaithersburg, MD) and compared to historical normal controls. There were seven samples from 5 unique temporal lobe epilepsy subjects. Results: In this pioneer experiment, the ratio of plasma GM-CSF levels to the average of the control patients showed consistent marked elevation in concentration of 4 of 5 temporal lobe epilepsy patients. The average ratio of subject to control for these four was 3.6. In these subjects, the average GM-CSF concentration in pg/ml was 1.46 /- 0.16 [SD], 95% CI [1.30-1.63] compared to control 0.76 /- 0.19 [SD], 95% CI [0.61-0.91] p<0.0002. One temporal lobe epilepsy patient had a substantially low ratio of -2.7. A syncope subject had a ratio similar to that of the historical controls. The IL-6 levels and IL-1? were similar to controls in most cases. Conclusions: In four out of five temporal lobe epilepsy patients, plasma concentrations of GM-CSF were markedly elevated compared to controls. The last temporal lobe epilepsy subject had a very low ratio. IL-1? levels were not as consistent as predicted from previous literature. IL-6 levels were not as consistent, likely because of the small sample size and sampling times outside the optimal time window. Only GM-CSF showed characteristics that indicate it could be a good biomarker for frequency of temporal lobe seizures. A prominent potential confounder is that the origin of the GM-CSF is unknown. It could be peripheral, or it could have leaked from the CSF though compromise of the blood brain barrier during a seizure.