Discussion

In the present study, we demonstrate that functional nAChRs are present in visually identified pyramidal neurons and interneurons of the CA1 field of the hippocampus of developing rats. Application of nicotinic agonists to CA1 neurons evoked action potentials as well as whole-cell currents that were either fast decaying (in interneurons) or slowly decaying (in pyramidal neurons and interneurons). Of major interest, however, was the finding that the predominant response of CA1 neurons to nicotinic agonists is the release of GABA, evidenced by the ability of nicotinic agonists to induce GABA-mediated PSCs in pyramidal neurons and interneurons.

Characterization of the nicotinic responses recorded from CA1 neurons in hippocampal slices.

Similarly to hippocampal neurons in culture, neurons in the CA1 field of hippocampal slices showed pharmacologically and kinetically distinct responses upon rapid application of nicotinic agonists. To compare the kinetics of the nicotinic responses recorded from hippocampal neurons in slices with that of the nicotinic currents recorded from hippocampal neurons in culture, it was necessary to take into account that the diffusion barriers in the slices slow down the agonist access to the neurons and that the agonist-delivery systems used to investigate the nicotinic responses of neurons in the slices were much slower than those used to obtain the responses of the neurons in culture. These factors contribute to slower kinetics of receptor activation and inactivation. Thus, the fastest nicotinic response of CA1 neurons in the slice (see fig. 7A) would correspond to the type IA currents recorded from the cultured neurons, and the slower decaying current recorded from the neurons in the slice (see figs. 7 and 11) would be equivalent to the type II currents recorded from hippocampal neurons in culture. The fast-decaying current, which was recorded predominantly from interneurons of the CA1 field of the hippocampal slices, was unveiled only when the α-7 nAChR-specific agonist choline was used, which suggests that such response is most likely subserved by an α-7 nAChR. Most of the neurons that showed a fast-decaying current in response to choline responded to nonspecific nicotinic agonists with a whole-cell current that had a fast- and a slow-decay component, or with a current that decayed very slowly. This finding could be explained by the ability of such nonspecific agonists to activate both rapidly and slowly desensitizing nAChRs, and by the fact that, depending on its magnitude, a slowly decaying current can easily mask a rapidly decaying current. Moreover, this finding indicates that a single neuron can express more than one subtype of functional nAChR.

The slowly decaying current recorded from interneurons most likely is subserved by an α4β2 nAChR because of the sensitivity of these currents to blockade by DHβE; and the slowly decaying currents recorded from the pyramidal neurons most likely are subserved by a different nAChR subtype, because cytisine, a poor α4β2 nAChR agonist (Alkondon and Albuquerque, 1995), was as effective as ACh as an agonist in evoking the slowly decaying currents in pyramidal neurons.

In addition to the nicotinic agonist-evoked whole-cell currents, a high frequency of PSCs was consistently evoked by application of nicotinic agonists to pyramidal neurons and could be recorded even from neurons that did not exhibit a “direct” nicotinic current. It should be mentioned that in interneurons PSCs were fewer in number during nAChR activation. Many nicotinic agonists such as (+)AnTX, choline and (−)-nicotine, at the concentrations used, were able to elicit PSCs without evoking substantial direct nicotinic currents in CA1 pyramidal neurons and in some interneurons, which suggests that the two events may be mediated by different subtypes of nAChRs.

The PSCs evoked by nicotinic agonists were blocked by picrotoxin, but not by CNQX, which indicates that these currents were mediated by GABA released from presynaptic GABAergic neurons upon nAChR activation. The finding that TTX prevented nicotinic agonist-induced GABA release from CA1 neurons supported our earlier suggestion that nAChRs controlling release of GABA may be located on the axons of GABAergic CA1 neurons (Albuquerque et al., 1997). Such “preterminal” nAChRs have also been found in neurons of rat interpeduncular nucleus (Léna et al., 1993) and avian lateral spiriform nucleus (McMahon et al., 1994). Although activation of muscarinic receptors on CA1 neurons has been shown to increase the release of GABA (Pitler and Alger, 1992), muscarinic receptors did not account for any of our results, because 1) the muscarinic receptor antagonist atropine (1 μM) was present in all solutions, and 2) compounds that are known to activate nicotinic but not muscarinic receptors (e.g., AnTX, (−)-nicotine, cytisine and (+)-epibatidine) were also capable of evoking responses that resembled those evoked by ACh in the presence of atropine.

Bursts of fast current transients, appearing like bursts of action potentials, were also induced by application of nicotinic agonists to interneurons, which suggests that nAChR-mediated fast current transients are characteristic nicotinic responses of nonpyramidal neurons.

The sensitivity of the nicotinic responses recorded from neurons in the CA1 field of hippocampal slices to blockade by d-TC supported the notion that these responses were mediated by an nAChR, but could not provide definitive clues regarding the nature of the nAChR subtype involved because of lack of selectivity ofd-TC to any particular nAChR subtype. Considering that in CA1 hippocampal neurons, the levels of mRNA coding for the α-7 subunit are very high (Séguéla et al., 1993) and that α-7 nAChR-mediated currents are the prevalent responses of cultured hippocampal neurons to nicotinic agonists (Alkondon and Albuquerque, 1993), we analyzed the sensitivity of the nicotinic responses recorded from CA1 neurons to MLA and α-BGT, which specifically inhibit the activation of α-7 nAChRs (Alkondon et al., 1992, 1994; Gray et al., 1996; Albuquerqueet al., 1997).

MLA up to 1 μM, a concentration 1000-fold higher than that required to completely block α-7 nAChR-mediated type IA currents in cultured hippocampal neurons, did not block the slowly decaying nicotinic currents recorded from pyramidal neurons. Likewise, α-BGT did not affect nicotinic agonist-induced fast current transients in interneurons. The insensitivity to MLA or α-BGT suggests a priori that these nicotinic responses may not be subserved by α-7 nAChRs. Taking into account the high levels of mRNA coding for α-7 subunit and the high levels of α-BGT binding sites (which are likely to represent the levels of α-7 subunits) in hippocampal neurons (Pauly et al., 1989; Séguéla et al., 1993), it is possible that the slowly decaying currents and the fast current transients recorded from the CA1 neurons were evoked by the activation of nAChRs made up of a combination of α-7 and other subunits. In fact, it has been reported that nAChRs containing α-7 and other subunits have kinetics and sensitivities to α-BGT and MLA that are different from those of the homomeric α-7 nAChRs (Yumet al., 1996).

As mentioned previously, however, the fast-decaying currents recorded from CA1 interneurons are likely to be mediated by an α-7 nAChR, because these responses could be evoked by the α-7 nAChR-specific agonist choline and resembled kinetically the α-7 nAChR-mediated type IA currents recorded from cultured hippocampal neurons. Considering the discrepancy between the proportion of cultured hippocampal neurons exhibiting α-7 nAChR-mediated nicotinic responses and the proportion of CA1 neurons showing responses that are presumably mediated by an α-7 nAChR, it is feasible that our culture conditions might favor the survival of the hippocampal neurons that express α-7 nAChRs. Alternatively, it is also possible that the lack of the major cholinergic input to the hippocampal neurons in culture favors the expression of α-7 nAChRs in these neurons. In fact, it has been shown that the cholinergic innervation to the muscle endplate plays a key role in the developmental changes of the muscle nAChRs in vivo (Betz and Osborne, 1977; Schuetze and Vicini, 1984; Kueset al., 1995).

TTX-sensitive nicotinic responses recorded from hippocampal neurons in culture and from neurons in the CA1 field of hippocampal slices.

In voltage-clamped cultured hippocampal neurons, ACh evoked fast current transients similar to those recorded from CA1 interneurons exposed to nicotinic agonists. When the same cultured hippocampal neurons were current clamped, ACh evoked action potentials. These action potentials could also be induced by the α-7-selective agonist choline, and were sensitive to blockade by the α-7-selective antagonist MLA, which indicates that in cultured hippocampal neurons, activation of α-7 nAChRs can result in depolarization sufficient to trigger action potentials.

Irrespective of the nAChR subtype present on hippocampal neurons in slices and in cultures, it is likely that the action potentials triggered by activation of these receptors on the soma or dendrites of the neurons (Alkondon et al., 1996a) are generated in a fashion similar to those triggered by the injection of current in the distinct areas of the neurons (Stuart and Sakmann, 1994). In some experiments, fast current transients or action potentials were not accompanied by large nicotinic currents or depolarization, which suggests that even a small depolarization, which may not have been detected in our experiments, is sufficient to initiate action potentials.

The action potentials initiated by nAChR activation could affect the neuronal function in two ways. First, the action potential might invade the dendrites of the same neuron by back-propagation, causing Ca⁺⁺ influx in the dendrites and affecting the synaptic efficacy. This has taken place in neocortical neurons and in CA1 pyramidal neurons (Markram et al., 1997; Sprustonet al., 1995). Second, the action potentials might invade the axon terminals and trigger the release of neurotransmitters onto neighboring neurons. The elicitation of PSCs by nAChR activation, and their blockade by TTX, is strong evidence that activation of nAChRs present on CA1 GABAergic neurons can initiate action potentials that, by invading the axon terminal, lead to the release of GABA.

The involvement of nAChRs in a local neuronal circuitry in the CA1 field of the hippocampus.

Our findings provide direct evidence that functional nAChRs are present on the surface of both CA1 pyramidal (P, fig. 14) and interneurons (I, fig.14) and that the receptors present in the interneurons modulate GABA release.

• Download figure
• Open in new tab
• Download powerpoint

Figure 14

Proposed scheme of nicotinic actions on CA1 hippocampal neurons. Three pathways (A–C) are proposed to explain the nicotinic responses recorded from neurons in the CA1 field of the rat hippocampus. In A, the neuron under study receives an autaptic input from its own axon, and a GABAergic input from another neuron. In B, the neuron under study receives a GABAergic input from another neuron. In C, the neuron under study gives out a glutamatergic output to a GABAergic neuron, which in turn provides an inhibitory connection to the primary neuron. P, pyramidal neurons; I, interneurons; a, axon. Open triangles represent GABAergic and filled triangle represents glutamatergic boutons. Arrows indicate the main axonal pathway for the neuron under study. The tip of a patch pipette is shown on each neuron under study.  

Different pathways could have been involved in the nAChR-induced, GABA-mediated PSCs depending on the neuronal type studied (pyramidalversus interneuron) in the hippocampal slice. For instance, when an interneuron was being studied (A, fig. 14), there could have been at least two sources for the nAChR-induced PSCs. First, the PSCs could have been the result of GABA release induced by the depolarization triggered by activation of nAChRs on another interneuron which synapses onto the neuron under study. Second, the interneuron under study could have made an autaptic synapse with its own cell body or dendrites. This being the case, because of the space-clamp problems that arise from the needle clamp in the whole-cell patch, nAChR-mediated depolarization of unclamped areas of the axon could initiate action potentials that invade the autaptic terminals and induce the PSCs. Such autaptic connections are very common in cultured hippocampal neurons (Bekkers and Stevens, 1991; Segal, 1991), and may very well explain the PSCs induced by nAChR activation in these neurons. However, the extent to which autaptic connections exist in the interneurons of the developing hippocampus is unknown. A recent study carried out on layer V pyramidal neurons of the developing rat neocortex indicated that about one third of their synaptic contacts could come from autapses (Lübke et al., 1996).

In pyramidal neurons, PSCs elicited by nAChR activation could arise from two mechanisms as illustrated in schemes B and C of figure 14. Similar to the mechanism mentioned above, the activation of nAChRs present on various interneurons could result in the release of GABA at the synapses made onto the pyramidal neurons. Then, GABA, in turn, by activating GABA_A receptors on the pyramidal neurons leads to the activation of PSCs. Alternatively, action potentials generated in the axon of pyramidal neurons by nAChR activation could release glutamate onto an interneuron viaan axon collateral. Glutamate, then, could lead to the increased excitability of the interneuron, which, in turn, would release GABA at its synapses onto the pyramidal neuron under study, resulting in the PSCs. We favor scheme B over C, because, in our experiments, nicotinic agonists were not seen to evoke fast current transients, reminiscent of action potentials, in pyramidal neurons in slices. Further, the inability of CNQX, a glutamate receptor antagonist, to block the PSCs makes scheme C unlikely.

Physiological significance of nAChR-mediated modulation of GABA release in the hippocampus.

After exposure to nicotinic agonists, the frequency of GABA-mediated PSCs was always much higher in recordings obtained from pyramidal neurons than in those obtained from interneurons, which thus indicates that nicotinic agonists may have a profound inhibitory influence on the activity of CA1 pyramidal neurons. According to our findings, if a neuronal network composed of interneuron ⇒ interneuron ⇒ pyramidal neuron exists, nAChR activation in the second interneuron should be very effective in increasing the inhibitory tonus to the pyramidal neuron, thereby overriding a possible functional disinhibition of the pyramidal neuron (Tóth et al., 1997). This disinhibition could occur in such a neuronal network only if nicotinic agonist-induced GABA release from the first interneuron in the circuitry were sufficient to silence the interneuron that synapses onto the pyramidal neuron. Thus, the frequency and timing of cholinergic stimuli arriving at the various CA1 interneurons may be essential in determining the magnitude of the output signals of the CA1 pyramidal neurons.

In the hippocampus, pyramidal neurons outnumber interneurons such that each interneuron, by virtue of its extensive axonal arborization, is capable of inhibiting many pyramidal neurons. Further, each CA1 pyramidal neuron can be innervated by as many as 25 basket cells, and each basket cell can form up to 12 synaptic contacts on the soma and proximal dendrites of the pyramidal neurons (Buhl et al., 1994), which thus ensures an effective inhibition of the activity of these neurons. Moreover, axoaxonic cells, by innervating predominantly the initial axon segments, and bistratified cells, by innervating predominantly the dendrites of pyramidal cells, constitute additional inhibitory pathways within the hippocampal circuitry (Buhl et al., 1994). GABAergic interneurons have been effective in synchronizing pyramidal cell activity (Cobb et al., 1995). This was attributed to the θ oscillatory activity induced by the perisomatic GABA_A-receptor-mediated synaptic events (Fox, 1989; Soltesz and Deschênes, 1993). θ-Frequency oscillations can also be brought about in the hippocampus by the cholinergic neurons (Bland, 1986), and such oscillations induced by cholinergic agonists have been shown to enhance the synaptic efficacy in the CA1 neurons (Huerta and Lisman, 1993). The θ rhythm is frequently recorded from the hippocampus of animals subjected to learning tasks (Winson, 1978; Otto et al., 1991), which suggests that nAChR-mediated modulation of GABA release onto CA1 pyramidal neurons may be a mechanism underlying the ability of nicotinic agonists to improve learning and memory.

The finding that activation of a neuronal nAChR on CA1 hippocampal neurons modulates the release of GABA may also provide the basis for the understanding of the mechanism by which an intact forebrain cholinergic innervation is essential to suppress kindling epileptogenesis in the hippocampus (Kokaia et al., 1996). Recent studies in rats have shown that selective immunolesioning of basal forebrain cholinergic neurons with the immunotoxin 192IgG-saporin results in a marked facilitation of the initial stages of seizure development in hippocampal kindling (Kokaia et al., 1996), which appears to be directly associated with a reduction of GABAergic inhibition in the CA1 field of the hippocampus (Lothman et al., 1993). Because activation of muscarinic receptors (Pitler and Alger, 1992) and nAChRs increases the release of GABA from CA1 interneurons, it is conceivable that facilitation of the development of seizures caused by the destruction of the major cholinergic input to the hippocampus is a result of the reduced activation of both muscarinic and nicotinic receptors in this brain area. In addition, our finding that the nAChR controlling GABA release from CA1 hippocampal neurons is insensitive to MLA sheds new light on the mechanism underlying the ability of nicotine to induce seizures in experimental animals, given that this effect of nicotine, which is believed to initiate in the hippocampus (Brown, 1967; Stumpf and Gogolak, 1967), cannot be prevented by MLA (Gasior et al., 1996). It is possible that nicotine-induced desensitization of the MLA-insensitive nAChRs present on CA1 hippocampal neurons accounts for its ability to induce seizures in the experimental animals.

A recent study has also implicated α-7 nAChRs in schizophrenia by providing linkage data supporting the involvement of the α-7-nAChR gene in the pathophysiological aspect of the illness (Freedman et al., 1997). The authors of that study attributed the attentional deficit seen in patients with schizophrenia to a decrease in the inhibitory control to the hippocampal pyramidal neurons. Our present results provide direct evidence for the existence of a mechanism by which nAChRs may control the inhibitory tonus in the hippocampus.

In conclusion, this study demonstrates that diverse subtypes of functional nAChRs (such as α-7-containing nAChRs, α4β2-containing nAChR and possibly nAChRs containing other subunits) are expressed in the pyramidal neurons and interneurons of the CA1 field of the rat hippocampus, and that activation of nAChRs present on the interneurons can control the excitability of pyramidal neurons and possibly of other interneurons by modulating the release of GABA.