Depolarization of nerve terminals with 30 mM KCl opens voltage-dependent Ca2+ channels and initiates neurotransmitter release (Barrie et al. 1991). Glutamate release after 5 min depolarization of cerebrocortical nerve terminals was 3.1 ± 0.1 nmol of glutamate/mg of protein, mean ± SEM, n = 16. This release was reduced with the GABAB receptor agonist baclofen (22.8 ± 2.8%, p < 0.001, n = 11) (Fig. 1a and b). The GABAB receptor antagonist CGP4638-1 suppressed the inhibition of release by baclofen (−0.6 ± 1.8%, p < 0.01, n = 6). In addition, we found that pertussis toxin (PTX) abolished the baclofen response (2.3 ± 2.6%, p < 0.00, n = 4) (Fig. 1b), suggesting that Gi/o proteins are involved in this pre-synaptic mechanism. Tertiapin-Q is a derivative of tertiapin, a peptide isolated from Apis mellifera bee venom that blocks a range of inward rectifier K+ channels including GIRK channels (Jin and Lu 1999). Blocking GIRK channels with tertiapin-Q did not reverse the baclofen effect (23.8 ± 5.0%, p > 0.05, n = 3) (Fig. 1a and b), indicating that GIRK channels do not mediate the inhibition of release by baclofen under 30 mM KCl stimulation conditions. We also found that the specific protein kinase C (PKC), and cAMP-dependent protein kinase (PKA), inhibitors bisindolylmaleimide and H-89, did not alter the inhibition of release by baclofen (25.1 ± 3.2%, p > 0.05, n = 4) and (19.5 ± 1.6%, p > 0.05, n = 3), respectively (Fig. 1b). Hence, these kinases do not appear to participate in the signalling pathway that leads to the inhibition of release. Although PKC activity does not mediate in the signalling initiated by GABAB receptors, the suppression of pre-synaptic G protein-coupled receptors, GPCRs, responses by PKC activation is a widespread phenomenon that occurs at many glutamatergic synapses (Swartz et al. 1993; Herrero et al. 1996). Consistent with these observations, we found that the prior activation of PKC with the phorbol ester 4β-phorbol dibutyrate, PDBu, completely suppressed (−0.1 ± 0.4%, p < 0.001, n = 5) the inhibition of evoked release by baclofen. Finally, the activation of adenylyl cyclase with forskolin did not alter the baclofen action (18.9 ± 0.4%, p > 0.05, n = 3) (Fig. 1b).
Figure 1. GIRK-channel independent inhibition of release by GABAB receptors. (a) The Ca2+-dependent release of glutamate was evoked by 30 mM KCl at 1.3 mM CaCl2 in the absence and presence of agonists and antagonists added at the indicated times prior to depolarization (arrows). Baclofen (20 μM, 40 s); CPG46381 (100 μM, 60 s); tertiapin-Q (100 nM, 60 s); ω-CgTx-GVIA (2 μM, 100 s) and ω-Aga-IVA (200 nM, 100 s). (b and c) histograms show the % inhibition of Ca2+-dependent release after a 5 min depolarization under the aforementioned conditions and in the presence of pertussis toxin, PTx, (1.5 μg/ml, 2 h); bisindolmaleimide (1 μM, 30 min); phorbol esters, PDBu, (1 μM, 60 s); H-89 (10 μM, 30 min); and forskolin (100 μM, 1 min). (d) Change in the [Ca2+]c evoked by 30 mM KCl at 1.3 mM CaCl2 in fura-2-loaded nerve terminals and after blockade of N and P/Q-type Ca channels with ω-CgTx-GVIA (2 μM) and ω-Aga-IVA (200 nM), respectively, in the absence and in the presence of baclofen (20 μM, 40 s). (e) Histograms indicate the baclofen-induced changes in the extent of Ca2+ influx sensitive to the Ca2+ channels toxins. Data represent the mean ± SEM (n = 3–16). NSp > 0.05; **p < 0.01; ***p < 0.001 (students t-test) when compared with control values, unless indicated otherwise.
Download figure to PowerPoint
The partial inhibition of release by baclofen could be the result of a restricted expression of GABAB receptors to a subpopulation of the cerebrocortical nerve terminals. In this preparation, glutamate release relies on N and P/Q type voltage-dependent Ca2+ channels (Turner and Dunlap 1995; Vázquez and Sánchez-Prieto 1997; Millán et al. 2002) each of which is expressed in a different subpopulation of nerve terminals in the adult rat (Millán et al. 2003). As GABAB receptors reduce synaptic transmission by inhibiting pre-synaptic Ca2+ channels (Dittman and Regher 1996; Isaacson and Hille 1997;Takahashi et al. 1998), we sought to establish which type of Ca2+ channel was inhibited by GABAB receptors, by performing occlusion experiments with the Ca2+ channel antagonists ω-CgTx-GVIA and ω-Aga-IVA which selectively blocks N and P/Q channels, respectively (Olivera et al. 1985; Mintz et al. 1995). The blockage of N-type Ca2+ channels with 2 μM ω-CTx-GVIA reduced glutamate release (25.4 ± 1.1%, p < 0.01, n = 5) and largely occluded a further inhibition of release by balofen (29.6 ± 2.0%, p > 0.05, n = 11) (Fig. 1a and c). In contrast, blocking P/Q type Ca2+ channels with 200 nM ω-Aga-IVA strongly reduced glutamate release (66.4 ± 5.2%, p < 0.001, n = 5) but still allowed further inhibition by baclofen (84.6 ± 2.6%, p < 0.01, n = 12) (Fig. 1a and c). Thus, GABAB receptor inhibition of release evoked by high KCl largely relies on N-type Ca2+ channels.
To further address that GABAB receptors inhibit N-type Ca2+ channels we determined changes in the cytoplasmic free Ca2+ concentration, [Ca2+]c, in nerve terminals loaded with the Ca2+ indicator fura-2. In fura-2 loaded synaptosomes, 30 mM KCl evoked a partially transient rise in [Ca2+]c followed by fall to a plateau within 1 min (Fig. 1d). The net rise in [Ca2+]c at the plateau (129.5 ± 3.2 nM, n = 5) was reduced with ω-CgTx-GVIA by (33.6 ± 2.7 nM, p < 0.01, n = 5) and with ω-Aga-IVA by (52.5 ± 4.5 nM, p < 0.001, n = 5) (Fig. 1d and e). Baclofen, largely occluded the response to ω-CgTx-GVIA, (10.0 ± 2.7 nM, p < 0.001, n = 7) but not that to ω-Aga-IVA (50.6 ± 5.4 nM, n = 6, p > 0.05) (Fig. 1d and e). Hence, a reduction of Ca2+ influx mediated by N-type Ca2+ channels seems to be associated to GABAB receptor inhibition of glutamate release at high KCl stimulations.