The expression of mRNA for connexin 43, a gap junction protein putatively found in astrocytes, is studied in two experimental models of epilepsy: the electrically kindled rat and the tetanus-toxin-injected rat. Rats were kindled by electrical stimulation of the amygdala to Racine class 5 seizures and divided into cohorts of three to undergo 3, 6, or 10 such events, respectively. Another two cohorts of rats received injections of tetanus toxin at strengths of 3 and 9 MLD50, respectively, into the amygdala. Features of epileptogenicity were identified electrographically in both cohorts during the first 4 wk following toxin injection with spontaneous ictal events recorded in the latter cohort. All rats were sacrificed 4 wk after electrode or cannula implantation, except for two toxin-injected cohorts that were sacrificed at wk 8 or 10. The epileptogeonic area in the region of the amygdala was harvested and pooled by cohort for Northern blot analysis. These were compared with control nonimplanted tissues. In the tetanus-toxin-injected animals, at time-points of 4, 8, and 10 wk, connexin 43 mRNA expression in epileptogenic tissues is found to be decreased or unchanged relative to control cases. Kindled rats demonstrated reductions of connexin mRNA with a trend toward normalizing levels with increasing numbers of stimulations when compared to control animals. Connexin 43 immunostained sections of the basolateral amygdala showed a similar trend in protein expression. Both experimental models of epilepsy show no connexin 43 mRNA upregulation despite varying degrees of epileptogenicity. This study therefore does not support the hypothesis that an increase in transcription is the basis for any proposed increase in gap junction communication involving connexin 43 in the context of epileptogenicity or as a reaction to increased neuronal excitability.