Abstracts

Rationale: COX-2 catalyzes the initial step in the metabolism of free arachidonic acid to lipid mediators, such as prostaglandins (PGs). In the normal central nervous system (CNS), COX-2 is localized to the cell soma and dendritic arbors of certain glutamatergic neurons and its level of expression and activity are coupled to excitatory neurotransmission. Under conditions of excess excitation, such as that associated with acute convulsive seizure activity, COX-2 activity appears to serve a suppressive role. On the other hand, COX-2 activity facilitates synaptic plasticity and, thus, has been posited to play a permissive role in epileptogenesis. However, this notion remains controversial. Previous studies have employed pharmacological inhibitors or traditional gene-deletion approaches to address the role of COX-2 in epileptogenesis. The goal of this study was to reassess this possibility using a targeted transgenic (TG) over-expression approach. Methods: High-level constitutive expression of a human COX-2 transgene was targeted to neurons using a well-characterized Thy-1 promoter construct. The TG line employed herein (L300, C57BL/6) exhibits broad neuron-specific expression of human COX-2 protein and has ~10-fold higher basal PG levels in the CNS than non-TG littermates (Vidensky et al., NeuroMol Med 3:15-27, 2003). Acute seizure activity and kindling acquisition induced by the chemoconvulsant, (PTZ), were compared in TG and non-TG littermate controls. Results: Consistent with the conclusion that PGs suppress excessive neuronal activity during excessive acute seizure activity, the incidence of PTZ-induced convulsive seizures was markedly less in TG mice over-expressing human COX-2 in neurons compared to non-TG littermate controls. Only 2/17 TG vs. 14/20 non-TG mice exhibited convulsive seizure behavior after 32mg/kg PTZ, i.p. With regard to epileptogenesis, both the median time to onset of kindling (i.e., first convulsive seizure) and latency to kindling acquisition (i.e., forth convulsive seizure) induced by daily administration of 28mg/kg PTZ i.p. were significantly delayed in TG mice. Thus, kindling onset was 5 vs. 3 days and acquisition was 9 vs. 7 days for TG vs. non-TG mice, respectively (Mann-Whitney test P = 0.039 and 0.028, respectively; N = 18-21). Although kindling acquisition was delayed at this dose of PTZ, all TG mice became kindled within 21 days. However, results from a preliminary study using a kindling dose of 24 mg/kg PTZ showed that 0/3 TG mice became kindled compared to 3/4 non-TG littermates. Conclusions: These results suggest that neuronal COX-2 serves to suppress excessive excitation induced by an acute convulsive stimulus and offer compelling evidence that this elevation of acute seizure threshold antagonizes kindling acquisition; hence, neuronal COX-2 activity may function as an important endogenous neuromodulator that suppresses the process of epileptogenesis. (Supported by NIH NS056304)