Three types of inhibitory postsynaptic potentials generated by interneurons in the anterior thalamic complex of cat

1. These experiments were carried out to study how thalamic interneurons generate inhibitory postsynaptic potentials (IPSPs) in relay cells. Intracellular recordings were performed in the anterior thalamic (AT) nuclei, a nuclear group in which interneurons constitute the only intrathalamic source of γ-aminobutyric acid (GABA). 2. In the AT complex, as in most dorsal thalamic nuclei, interneurons can influence relay cells through their presynaptic dendrites (PSDs) and their axons. This dual mode of action is paralleled by a different termination pattern of prethalamic fibers and cortical axons on interneurons. Prethalamic fibers, which in the AT nuclei arise in the mammillary bodies (MBs), end mostly on PSDs, whereas cortical terminals usually synapse on the parent dendrites of PSDs. We therefore took advantage of the differential mode of termination of cortical and MB afferents on interneurons to infer the respective roles of the axons and PSDs of interneurons in the genesis of the IPSPs recorded from relay cells. 3. In all responsive AT cells, cortical stimuli delivered at low frequency (≤0.5 Hz) evoked a biphasic IPSP, with an early and a late phase, having a total duration of 221.96 ± 8.18 ms (mean ± SE). The early part of the IPSP (termed A) had a reversal potential (E(R)) close to the equilibrium potential for Cl^- ions: -79.25 ± 2.14 mV. Furthermore, it reversed in polarity after impalement of the cells with KCl-filled pipettes. The late IPSP (termed B) always began before the end of the early IPSP, 45.93 ± 2.50 ms after the onset of the A-IPSP. The B-IPSP had an E(R) of -109 ± 2.4 mV and was not affected by Cl^- injection. 4. By contrast, MB stimuli delivered at low frequency (≤0.5 Hz) evoked a triphasic IPSP having a total duration of 220.5 ± 9.42 ms in most (61.2%) AT cells. The IPSP with the shortest latency (termed a) was evoked only by MB stimuli. Before the return of the membrane potential to the resting level, a second hyperpolarizing potential began (7.41 ± 0.46 ms after the onset of the a-IPSP). This second inhibitory phase was biphasic and had electrophysiological characteristics similar to the biphasic A- and B-IPSP evoked by cortical stimulation. Both the MB-evoked a- and A-IPSPs had an E(R) close to the equilibrium potential for Cl^- ions (-72.22 ± 0.68 and -72 ± 0.82 mV, respectively) and reversed in polarity after impalement of the cells with KCl-filled pipettes. The MB-evoked B-IPSP had an E(R) close to the equilibrium potential for K⁺ ions (-111 ± 2.8 mV) and was not affected by Cl^- injection. 5. Gradual increments in the intensity of MB and cortical stimulation produced augmentations in the amplitude of the A- and B-IPSPs. Although the amplitude increments of these two components were usually highly correlated (r ≥ 0.8), there was generally no parallel between the amplitude changes of the MB-evoked a-IPSP and the fluctuations of A and B components. 6. In contrast to the a-IPSP, the components A and B of the MB- and cortically evoked IPSPs exhibited a marked dependency on the frequency of stimulation. At low stimulation rates (≤0.5 Hz), all the components of the MB- and cortically evoked IPSPs were of stable amplitude. At higher stimulation rates (≥2 Hz), the MB-evoked A- and B-IPSPs were usually abolished, whereas the component a was preserved. By contrast, the cortically evoked A- and B-IPSPs were usually not abolished at these stimulation rates but only reduced in amplitude. 7. At intermediate stimulation rates (~1 Hz), the components A and B of the MB- and cortically evoked IPSPs exhibited marked fluctuations in amplitude from one stimulus to the next (up to 8-10 mV). Usually, MB- and cortically evoked A- and B-IPSPs assumed two to four discrete amplitude levels. In spite of these important fluctuations, their amplitude remained highly correlated (r > 0.8). 8. In light of previous electrophysiological, pharmacological, and morphological findings, a number of tentative conclusions can be reached as to the nature of the cortically and MB-evoked IPSPs. First, the differential response of the various inhibitory phases to membrane hyperpolarization and Cl^- injection suggests that the a- and A-IPSPs are mediated by GABA(A) receptors, whereas the B-IPSP is mediated by GABA(B) receptors. Second, the fact that the a-IPSP is evoked selectively by MB stimulation suggests that it is mediated by the PSDs of interneurons, whereas the A- and B-IPSPs, which can be evoked by both MB and cortical stimuli, are mediated by the axons of interneurons. This interpretation is also supported by the differential response of the a-IPSP and the longer-latency A and B components to changes in the intensity and frequency of stimulation.