Abstracts

Rationale: Neuronal cyclooxygenase-2 (COX-2) catalyzes rapid production of prostaglandins consequent to increased excitatory activity in the brain. Results from animal models suggest that this may intrinsically subdue acute seizure activity and epileptogenesis. T-cell intracellular antigen-1 (TIA-1) is an RNA-binding protein that suppresses COX-2 protein expression in embryonic fibroblasts. It is constitutively expressed by glutamatergic neurons suggesting that it may also negatively influence COX-2 protein expression in the brain, thus contributing to a lower seizure threshold. It is hypothesized herein that mice lacking TIA-1 will have higher levels of COX-2 and a correspondingly higher seizure threshold.Methods: To test this possibility, TIA-1 homozygous (-/-) and heterozygous (+/-) null mice were subjected to pentylenetetrazole (PTZ)-induced acute seizures or kindling. Mice were treated with a single dose of PTZ (42 mg/kg, ip) in the acute seizure model and the incidence of convulsions was determined for each genotype group. In the kindling paradigm, PTZ (36mg/kg, ip) was administered daily for 21 days and the reduction in seizure threshold was record for each mouse. The PTZ dose for this paradigm did not elicit convulsions in naive mice from any genotype. Mice were considered kindled after 3 consecutive convulsive seizures. This was confirmed in each case by PTZ (36mg/kg)-induced convulsions 7 days after cessation of the paradigm. TIA-1 and COX-2 protein expression in brain samples were examined using immunofluorescence microscopy. In all cases, results from TIA-1 null mice were compared to wild-type (+/+) littermate controls tested in parallel.Results: TIA-1 protein was constitutively present in adult +/+ brains and absent in -/- tissue. However, COX-2 expression was not altered by deletion of TIA-1 as posited. Moreover, the incidence of acute convulsions was not different in the -/- (6/18) or +/- (11/30) genotype groups compared to their +/+ controls (6/20). This procedure was associated with a low incidence of mortality, which did not differ between genotype groups (5.6%, 6.7%, and 5.0% for -/-, +/-, and +/+, respectively). Similarly, acquisition of kindling was unchanged by TIA-1 deletion (p=0.9166, Mantel-Cox test). However, a higher mortality rate was observed in this paradigm for -/- and +/- genotype groups relative to the +/+ controls (29.4%, 15.0%, and 5.9%, respectively) and -/- was statistically different from +/+ (Chi square test, p=0.036).Conclusions: These results indicate that, in contrast to non-neuronal cells, TIA-1 does not appear to modulate COX-2 expression in neurons of the brain, nor does it contribute to maintenance of acute seizure threshold and epileptogenesis. Nevertheless, its loss did appear to predispose a fraction of mice to mortality in the kindling model of epileptogenesis, raising the possibility that TIA-1 may function as a molecular suppressor of sudden unexpected death in epilepsy (SUDEP). Source of funding: NIH/NINDS #R15NS082982