Adenosine is a neuromodulator that has been proposed to be a major endogenous anticonvulsant acting via A1 receptors. We tested if implementation of kindling through stimulation of the amygdala affected A1 receptor-mediated neuromodulation in hippocampal slices taken from rats 4 weeks after the last stage 5 seizure. The A1 receptor agonist, N6-cyclopentyladenosine (CPA) (6–100 nm), inhibited field excitatory postsynaptic potential (fEPSP) slope with an EC50 of 19.1–19.5 nm in control and sham-operated rats, but was less potent in kindled rats (EC50 = 42.7 nm). This might result from a decreased number of A1 receptors in hippocampal nerve terminal membranes, because A1 receptor immunoreactivity decreased by 28 ± 3% and the binding density of the A1 receptor agonist [3H]R-PIA decreased from 1702 ± 64 to 962 ± 78 fmol/mg protein in kindled compared with control rats. The tonic inhibition of hippocampal synaptic transmission by endogenous adenosine was also lower in kindled rats, because A1 receptor blockade with 50 nm 1,3-dipropyl-8-cyclopentyladenosine (DPCPX) enhanced fEPSP slope by 23 ± 3% and θ-burst-induced long-term potentiation by 94 ± 4% in control rats but was virtually devoid of effects in kindled rats. The evoked release of adenosine from hippocampal slices or nerve terminals was 56–71% lower in kindled rats probably due to the combined decrease in the capacity of adenosine transporters and decreased release of adenosine 5′-triphosphate (ATP), which was partially compensated by a higher extracellular catabolism of ATP into adenosine in kindled rats. These results indicate that, although adenosine might inhibit the onset of epileptogenesis, once kindling is installed, the efficiency of the adenosine inhibitory system is impaired.