Neural network activity in the CNS is abolished or strongly reduced by gap junction (GJ) blockers. Indeed neural GJs are thought to contribute to network activity. However, when a GJ blocker is applied it exerts its action on neurons but also on astrocytes. Astrocytes are extensively coupled through GJs and have been recently shown to modulate several neuronal properties including synaptic transmission. We therefore tested whether GJ communication in astrocytes affects neural network activity. To this end, we used different primary culture models of hippocampal or cortical neurons and astrocytes, where neuronal GJs are not thought to be present.

Primary cultures were prepared from rat or mouse embryos, plated on astrocytes and grown for 8-15 days. Animals were humanely killed by cervical dislocation. Single or double whole-cell patch recordings from neurons were performed at -60 mV. Calcium signals were monitored in cells loaded with Fluo-4 AM (3 µM). Data were averaged between different experiments and presented as means ± S.E.M. Statistical analysis was carried out by Student’s paired t test.

Using patch clamp and calcium imaging techniques, we investigated the effects of the GJ blocker carbenoxolone (CBX) on spontaneous and bicuculline-induced (10 µM) network activity. In co-cultures of neurons and astrocytes, application of 100 or 20 µM CBX induced a strong and reversible decrease in the frequency of synaptic activity and calcium oscillations associated with spontaneous (from 3.3 ± 0.3 to 0.2 ± 0.1 min^-1) or bicuculline-induced (from 2.1 ± 0.2 to 0.5 ± 0.2 min^-1) burst activity, without altering the frequency or amplitude of miniature events recorded in the presence of TTX. Dye or electrical coupling between neurons was not detected in pair of neurons and immunocytochemical staining for major neuronal connexins was faint or absent, suggesting the effect of CBX is not mediated through neuronal GJs. To further test the involvement of astrocyic GJs we used other mice culture models where astrocytes were not coupled. Since neurons do not express receptors for endothelin 1 (ET-1), we used this peptide to block GJs in astrocytes. ET-1 (200 nM) did not mimic or occlude the effect of CBX, which reduced burst frequency from 11.3 ± 1.5 to 0.6 ± 0.2 min^-1. Furthermore CBX exerted its effect (from 10.3 ± 2.4 to 1.1 ± 0.2 min^-1) in enriched neuronal cultures where only a few, isolated astrocytes were present. We then tested if CBX affected basic neuronal properties and we found that 100 µM CBX increased the spike threshold and strongly decreased the firing rate in response to depolarising stimuli, suggesting it could affect some neuronal currents.

Thus, since astrocytic GJs are not involved in the generation of spontaneous network activity and neuronal GJs are not detected in culture, it is likely that CBX exerts its action through other mechanisms.