The Journal of Neurophysiology Vol. 79 No. 6 June 1998, pp. 3229-3237
Copyright ©1998 by the American Physiological Society

Synaptic Connectivity of Distinct Hilar Interneuron Subpopulations

Matteo Forti and Hillary B. Michelson

Department of Pharmacology, State University of New York Health Science Center at Brooklyn, Brooklyn, New York 11203

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Forti, Matteo and Hillary B. Michelson. Synaptic connectivity of distinct hilar interneuron subpopulations. J. Neurophysiol. 79: 3229-3237, 1998. Dual intracellular recordings of hilar interneurons and CA3 pyramidal cells were performed in transverse slices of guinea pig hippocampus in the presence of the convulsant compound 4-aminopyridine (4-AP) and ionotropic glutamate receptor antagonists. Under these conditions, interneurons burst fire synchronously, producing synchronized inhibitory postsynaptic potentials (sIPSPs) in pyramidal cells. Three different hilar interneuron subpopulations that contributed to the sIPSP were identified based on their projection properties and morphology. These three types were pyramidal-like stellate interneurons, spheroid interneurons, and oviform interneurons. Physiologically, pyramidal-like stellate interneurons could be differentiated from the other interneuron subpopulations because they generated short synchronized bursts of action potentials coincident with the hyperpolarizing and depolarizing -aminobutyric acid-A (GABA_A)-mediated inhibitory postsynaptic potentials (IPSPs) recorded in pyramidal cells. The bursts in pyramidal-like stellate cells were abolished by theGABA_A-receptor blocker, bicuculline. In contrast, spheroid interneurons of the dentate-hilus (D-H) border and oviform hilar interneurons exhibited prolonged bicuculline-resistant bursts that occurred coincident with the GABA_B pyramidal cell sIPSPs. Pyramidal-like stellate interneurons likely did not contribute to the generation of synchronized GABA_B responses in hippocampal pyramidal cells. Spheroid interneurons were unique among these subpopulations of interneurons in that the bicuculline-resistant bursts in spheroid interneurons were sustained by a synaptic depolarization that persisted in the presence of antagonists of ionotropic glutamate, GABA_A and GABA_B receptors [6-cyano-7-nitroquinoxaline-2,3-dione, 20 µM; 3-3(2-carboxipiperazine-4-yl)propyl-1-phosphonate, 20 µM; bicuculline, 10-15 µM; CGP 55845A, 20 µM]. This novel depolarizing potential reversed between 30 and 0 mV. No noticeable synaptic depolarization sustaining burst firing could be isolated in oviform interneurons, suggesting that firing in this interneuron subpopulation was synchronized by nonsynaptic mechanisms. The results of the present study indicate that the hilar inhibitory circuit is composed of at least three different subpopulations of interneurons, distinguishable by their morphological characteristics and synaptic inputs and outputs. These findings give further support to the hypothesis that there are distinct populations of interneurons producing GABA_A and GABA_B responses with defined functional roles within the hippocampal inhibitory circuit. Notably, we found that spheroid interneurons were unique among the hilar interneurons studied, in that the synchronized bursts observed in these cells are sustained by a novel ionotropic glutamate and GABA receptor-independent synaptic depolarization.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

In the hippocampus, GABAergic interneurons play an essential role in the control of normal network activity (Buzsáki et al. 1992; Cobb et al. 1995; Miles and Wong 1987; Traub et al. 1996). Interneurons can influence the efficacy of principal cell excitation and thus modulate population activity within the hippocampal circuit. The dentate gyrus, in particular, plays an important role in information processing in the hippocampus. As the first step in the trisynaptic intrahippocampal loop, the dentate receives significant input from entorhinal afferents that it then relays to the CA3 subfield (Blakstad et al. 1970; Hjorth-Simonsen 1971). The dentate also functions to gate the propagation of epileptiform activity into the hippocampus (Paré et al. 1992).

Recent studies have demonstrated that dentate gyrus interneurons have very extensive and regionally specific axonal arborizations, implying that single interneurons can exert widespread influence on the excitability of neurons throughout the hippocampus (Buckmaster and Schwartzkroin 1995a). Within the inhibitory circuit, interneurons generate chloride-dependent fast inhibitory postsynaptic potentials (IPSPs) in principal cells mediated by -aminobutyric acid-A (GABA_A) receptors, and potassium-dependent slow IPSPs mediated by GABA_B receptors (Alger and Nicoll 1982). However, it is not clear whether distinct subpopulations of interneurons differentially produce GABA_A- or GABA_B-mediated IPSPs. Some investigators have suggested that, in the CA1 region of the hippocampus, separate groups of inhibitory cells are responsible for activating postsynaptic GABA_A and GABA_B receptors (Segal 1990; Samulack and Lacaille 1993; Williams and Lacaille 1990; Williams et al. 1993). No such evidence exists for interneurons in the hilus.

Several studies have described a wide variety of interneurons in the dentate gyrus, according to their morphological features (Amaral 1978; Lorente de Nó 1934; Ramón y Cajal 1893; Ribak and Seress 1983). More recent studies have identified different classes of interneurons according to their immunocytochemical properties (Baskt et al. 1986; Gulyás et al. 1991; Sloviter and Nilaver 1987) or their axonal targets (Freund and Buzsáki 1996; Gulyás et al. 1993; Han et al. 1993; Mott et al. 1997). Correlations of morphology with intrinsic properties have been performed on hilar interneurons (Buckmaster and Schwartzkroin 1995a,b; Mott et al. 1997); however, no studies have investigated correlations between the morphological heterogeneity of hilar interneurons and their recurrent connectivity or inhibitory projections onto pyramidal cells.

In the present study, we use the previously characterized 4-aminopyridine (4-AP) model of interneuronal synchronization (Aram et al. 1989; Michelson and Wong 1991, 1994; Muller and Misgeld 1990; Perrault and Avoli 1991) to examine the relationship between morphological and functional aspects of inhibitory neurons. Previous studies with 4-AP have demonstrated that interneurons can recurrently excite other interneurons via a GABA_A-mediated depolarizing response and that subpopulations of interneurons can be differentiated according to their responsiveness to the GABA_A receptor antagonist bicuculline (Michelson and Wong 1994). The aim of the present study was to examine the recurrent connectivity and inhibitory projection properties of morphologically identified hilar interneuron populations.

View larger version (29K): [in this window] [in a new window]   FIG. 1. Synaptic inputs and morphology of pyramidal-like stellate interneurons. A: hyperpolarization and depolarization of a pyramidal-like stellate cell by DC current injection in control conditions, with 4-aminopyridine (4-AP, 75 µM), 3-3(2-carboxipiperazine-4-yl)propyl-1-phosphonate (CPP, 20 µM), and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 20 µM) in the bath. All events were spontaneously occurring. Resting membrane potential was 60 mV, input resistance was 44 M. PYR, pyramidal cell. , hyperpolarizing -aminobutyric acid-A (GABAA) input. - - -, depolarizing GABAA input. Note that the peak of the depolarization shifted toward the peak of the GABAA hyperpolarizing component in the pyramidal cells when the interneurons where hyperpolarized beyond the Cl reversal potential, suggesting that the pyramidal-like interneurons received also a GABAA hyperpolarizing input. Also note that electrical capacitance transients associated with the firing of simultaneously recorded cells are visible in most traces in this and subsequent figures. B: hyperpolarization of the same cell after bicuculline wash-in. Note that a GABAB-mediated hyperpolarization was apparent in 2 of 5 pyramidal-like stellate cells examined. - - -, GABAB input. C and D: camera lucida drawings of pyramidal-like stellate cells stained with biocytin. D is a composite of cells taken from 5 different slices. All the cells had functional properties similar to those shown in the figure. - - -, granule cell layer; a, axon. Calibration bars, C: 15 µm; D: 90 µm.

	METHODS

Abstract Introduction Methods Results Discussion References

Guinea pigs weighing 200-300 g were anesthetized with2-bromo-2-chloro-1,1,1-trifluoroethane and rapidly decapitated. Transverse slices of hippocampus, 400 µm in thickness, were prepared using a vibratome according to standard procedures (Michelson and Wong 1994). Slices were transferred onto the nylon mesh of a gas-fluid interface recording chamber, maintained at 35°C (pH 7.4), and exposed to a warm, humidified atmosphere saturated with a 95% O₂-5% CO₂ gas mixture. The lower surfaces of the slices were in contact with a perfusion solution containing (in mM) 124 NaCl, 5 KCl, 2 CaCl₂, 1.6 MgCl₂, 26 NaHCO₃, and 10D-glucose. Control solutions also contained 4-AP (75 µM);6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 20µM); and3-3(2-carboxipiperazine-4-yl)propyl-1-phosphonate (CPP, 20 µM). Bicuculline methiodide (10-15 µM) and CGP 55845A, (20 µM) were added to the solution in some experiments. CNQX and CPP were obtained from Tocris Cookson (St. Louis, MO), and CGP 55845A was kindly provided by CIBA-GEIGY (Basel, Switzerland). All other chemicals were obtained from Sigma Chemical (St. Louis, MO).

Dual intracellular recordings were performed using glass micropipettes (40-100 M resistance) filled with 2% biocytin in 1 M potassium acetate. Only one interneuron per slice was recorded in order to avoid multiple intracellular staining. Signals were amplified by a dual-channel Neurodata amplifier and stored on tape for off-line digital analysis. All measurements of signal amplitude and duration are expressed as means ± SD with significance set atP < 0.05.

At the end of the experiments, the slices were fixed overnight with 4% paraformaldehyde for biocytin processing. The slices were incubated in ABC complex (Vectastain, Vector Laboratories, Burlingame, CA) and were processed with diaminobenzidine for visualization. Finally the slices were dehydrated and cleared in a series of alcohol rinses before being mounted and examined using a Nikon microscope with appropriate photographic accessories.

	RESULTS

Abstract Introduction Methods Results Discussion References

Different subtypes of interneurons generating synchronized IPSPs

The results presented have been limited to include only cells in which data was obtained continuously before and throughout all drug administrations. Dual intracellular recordings (n = 15) were obtained from hilar interneurons and CA3 pyramidal cells in the presence of the convulsant compound 4-AP (75 µM). CNQX (20 µM) and CPP (20 µM) also were added to the bath to block ionotropic glutamatergic excitatory inputs (Honoré et al. 1988). Under these conditions, synchronized IPSPs (sIPSPs) were generated in pyramidal cells at a frequency of 0.125-0.25 Hz, which occurred simultaneously with spontaneous bursting activity in interneurons (Fig. 1A). The triphasic sIPSPs in pyramidal cells were composed of a GABA_A hyperpolarizing phase and a GABA_A depolarizing phase which were both blocked by bicuculline (see Fig. 2A), followed by a late GABA_B component blocked by CGP 55845A (see Fig. 6). Previous studies have demonstrated that sIPSPs are population events that occur simultaneously in all pyramidal cells and are generated by the synchronized discharge of interneurons (Aram et al. 1991; Michelson and Wong 1991, 1994).

View larger version (28K): [in this window] [in a new window]   FIG. 2. Contribution of pyramidal-like interneurons to the GABAB synchronized inhibitory postsynaptic potentials (sIPSPs) in pyramidal cells. A: spontaneous sIPSPs in 4 pyramidal cells were averaged before (n = 237) and after (n = 241) bicuculline wash-in to measure the amplitude of the GABAB responses before and after blockade of bursting activity in pyramidal-like interneurons. Events were detected as a negative peak in the sweep and aligned at the peak. B: mean GABAB sIPSP amplitude ± standard deviations were measured at the latencies from onset of the event, as shown by - - - in A, and plotted against time. Note that the amplitude of the GABAB event was significantly different only at 375 and 400 ms, at which time there was significant overlap of the depolarizing GABAA response in control conditions. Asterisks: P < 0.05 (Student's t-test).

View larger version (28K): [in this window] [in a new window]   FIG. 6. Reversal potential of the synaptic depolarizing input underlying bicuculline-resistant bursts in spheroid interneurons. A: spontaneous burst in a spheroid interneuron held at a membrane potential of 44 mV and corresponding GABAB sIPSP in a pyramidal cell (PYR). , GABAB input. B: hyperpolarization and depolarization of the same cell in the presence of CGP 55845A (CGP, 20 µM) to abolish the GABAB input. Resting membrane potential was 64 mV. C: amplitude of the depolarization underlying the burst was measured at the latency of the dashed line in B before spike onset and plotted against the membrane potential of the interneuron. Intersection of the dashed line with the x axis represents the estimated reversal potential of the synaptic input (15 mV).

Some of the recorded interneurons were stained with biocytin, and the morphology of each cell was correlated with its physiological properties. Three types of interneurons were identified following the classification of Amaral (1978): pyramidal-like stellate cells, spheroid cells of the dentate-hilus (D-H) border, and oviform cells.

Pyramidal-like stellate cells

Pyramidal-like stellate interneurons (n = 5) were morphologically identified as cells with large triangular somas and three to five primary dendrites (Fig. 1, C and D). Spines were not apparent on the dendrites of these interneurons.

Pyramidal-like stellate interneurons could be distinguished from the other interneuron populations by their bursting characteristics in the presence of 4-AP and ionotropic glutamate receptor blockers. All pyramidal-like stellate interneurons exhibited short burst discharges with a duration of 336.7 ± 16.8 ms (n = 60 in 3 cells) and a latency to peak of 152.1 ± 10.4 ms. These bursts occurred coincident with the GABA_A components of the spontaneous IPSPs recorded in pyramidal cells (Fig. 1A). The short synchronized bursts in all pyramidal-like stellate cells were sustained by a GABA_A depolarizing input that was abolished by bicuculline (10-15 µM; Fig. 1B). This depolarizing input had a reversal potential more positive than 45 mV (Fig. 1A). A hyperpolarizing GABA_A event was revealed when the cells were sufficiently depolarized (Fig. 1A, 50 mV). Two of five pyramidal-like stellate cells received a GABA_B input, which became evident after blockade of the bursting activity with bicuculline. Under these conditions, monophasic GABA_B IPSPs with a reversal potential around 90 mV occurred in the interneurons simultaneously with the sIPSPs in the pyramidal cells (Fig. 1B). These GABA_B events were blocked by CGP 55845A (not shown).

Although bicuculline blocked all bursting activity in pyramidal-like stellate cells, synchronized GABA_B events in pyramidal cells persisted in the presence of bicuculline, presumably sustained by synchronized activity in other interneurons. We therefore performed experiments to evaluate whether pyramidal-like stellate cells participate in the generation of the synchronized GABA_B event before bicuculline administration. These experiments compared the amplitude of the GABA_B sIPSPs before and after bicuculline-induced blockade of bursting activity in pyramidal-like stellate interneurons (Fig. 2). The amplitude of the GABA_B sIPSP was significantly larger after bicuculline wash-in when measured near the peak of the synchronized event in control conditions (at 375 ms: control = 6.1 ± 0.3 mV; bicuculline =9.1 ± 0.5 mV, P < 0.05). However, when measured at a delay of 525-725 ms from the onset of the sIPSP, the amplitude of the GABA_B event was not significantly different before and after bicuculline wash-in (at 700 ms: control = 2.2 ± 0.5 mV; bicuculline = 2.4 ± 0.8 mV, P > 0.05).

Because the early phase of the sIPSP in control conditions is a mixed event, with substantial overlap between the GABA_A depolarizing component and the GABA_B component, the amplitude of the late phase of the sIPSP, beyond the peak of the depolarizing GABA event, reflects a more reliable measurement of the pure GABA_B response (Perkins and Wong 1996). If pyramidal-like stellate interneurons contributed to the synchronized GABA_B response in pyramidal cells, the amplitude of the GABA_B event would be expected to be significantly smaller after bicuculline administration. Thus the lack of a significant change in the amplitude of the GABA_B response after bicuculline, measured at later latencies during the event, suggests that pyramidal-like stellate cells did not contribute significantly to the generation of GABA_B responses in pyramidal cell sIPSPs.

Oviform cells

Oviform cells (n = 5) were identified morphologically as cells with one or two main large spiny dendrites emerging from oviform somas on the opposite side of the axon. These cells also had three to four smaller primary dendrites that ramified extensively within the hilus (Fig. 3, C and D).

View larger version (33K): [in this window] [in a new window]   FIG. 3. Synaptic inputs and morphology of oviform interneurons. A: hyperpolarization and depolarization of an oviform interneuron in control conditions. Resting membrane potential was 65 mV, input resistance was 86 M. : hyperpolarizing GABAA input. Synchronized bursts in these cells were either monophasic and synchronous with the GABAB input in pyramidal cells or biphasic. - - -, biphasic burst in the interneuron corresponding to GABAA and GABAB components in the pyramidal cell (PYR) sIPSPs. B: hyperpolarization and depolarization of the same cell after bicuculline wash-in. , GABAB input.- - -, GABAB sIPSPs in the pyramidal cell and corresponding bicuculline-resistant burst in the interneuron. Note the nonlinear voltage dependence of the small depolarization underlying the burst. C and D; camera lucida drawings of oviform cells stained with biocytin. All the cells had functional properties similar to those shown in the figure. - - -, granule cell layer; a, axon. Calibration bars, C: 15 µm; D: 90 µm.

The synchronized bursts in the oviform interneurons shared some common features with bursts occurring in spheroid interneurons (see further). They were significantly longer than synchronized bursts recorded in pyramidal-like stellate cells (626.3 ± 27.4 ms; n = 60 in 5 cells; P < 0.05) and exhibited a longer rise time (293.7 ± 43.7 ms, P < 0.05). Unlike pyramidal-like stellate cells, the bursts in oviform interneurons were resistant to bicuculline (Fig. 3B, middle and right, n = 8) and occurred coincident with the pure GABA_B component of the sIPSP in pyramidal cells.

All oviform cells received both GABA_A inputs and aGABA_B input (Fig. 3, A and B, ). The bicuculline-resistant component of the synchronized burst in oviform interneurons usually originated abruptly from the baseline at resting membrane (Fig. 4B, 65 mV). When the cells were hyperpolarized, small truncated spikes occurred originating directly from the baseline (not shown). These characteristics were consistent with a nonsynaptic origin of the burst discharges, as previously described in hilar interneurons (see Michelson and Wong 1994).

View larger version (72K): [in this window] [in a new window]   FIG. 4. Dye coupling in oviform interneurons stained with biocytin. Three oviform cells were stained with biocytin after the recording and filling of a single interneuron in the hilus. A: camera lucida reconstruction of the 3 cells. Calibration bar: 150 µm. B: objective ×40 photo. Calibration bar: 30 µm. - - -, granule cell layer and CA3 pyramidal cell layer. Note the large extension of the axon collaterals (a). Homogeneity of biocytin staining intensity in the 3 cell bodies and their primary dendrites indicates that the multiple cell staining did not result from dye leakage from a single broken cell. Clear separation of the cell body of >1 of the cells also argues strongly against artifactual dye coupling among these cells.

Dye coupling of interneurons, an indication of the presence of electrotonic junctions, has been demonstrated previously (Michelson and Wong 1994) and was found again among oviform interneurons in one instance (n = 1 of 5). Figure 4 shows three oviform interneurons stained together during a single recording. The dye coupling among oviform cells suggests that electrical connections among oviform interneurons may play a role in their synchronization.

Spheroid cells of the D-H border

Spheroid cells (n = 3) were located on the border between the hilus and the granule cell layer. These cells had small round somas with six to nine long primary dendrites showing few ramifications (Fig. 5, C and D).

View larger version (28K): [in this window] [in a new window]   FIG. 5. Synaptic inputs and morphology of spheroid interneurons of the dentate-hilus (D-H) border. A: hyperpolarization and depolarization of a spheroid cell of the D-H border in control conditions. Resting membrane potential was 64 mV, input resistance was 34 M. PYR, pyramidal cell. , hyperpolarizing GABAA input. - - -, GABAB component in the pyramidal cell sIPSPs and corresponding burst in the interneuron. Note that the GABAB event was evident only when the cells were depolarized. B: hyperpolarization and depolarization of the same cell after bicuculline wash-in. , GABAB input. - - -, bicuculline-resistant burst and underlying depolarizing input corresponding to the GABAB sIPSPs in the pyramidal cell. Note the large spike afterhyperpolarizations during the burst discharges. C and D: camera lucida drawings of spheroid cells stained with biocytin. All the cells had functional properties similar to those shown in the figure. - - -, granule cell layer; a, axon. Calibration bars, C: 15 µm; D: 90 µm.

All spheroid cells of the D-H border received both a hyperpolarizing and depolarizing GABA_A input and a GABA_B input (Figs. 5, A and B, and 6A, ). Addition of bicuculline (10-15 µM) and CGP (20 µM) into the bath revealed a depolarization that occurred coincident with the monophasic GABA_B event in pyramidal cells (Figs. 5B and 6A). This depolarization increased linearly in amplitude when the cells were hyperpolarized, consistent with a chemical synaptic origin of the event (Fig. 6B).

The reversal potential of the excitatory synaptic input sustaining the prolonged burst in spheroid interneurons was estimated by measuring the amplitude of the depolarization at different membrane potentials in the presence of CGP 55845A (n = 2). CGP 55845A effectively blocked theGABA_B input in both pyramidal cells and interneurons and, in spheroid interneurons, further pharmacologically isolated the depolarizing potential. The monophasic depolarization exhibited a linear voltage dependence and showed an extrapolated reversal between 30 and 0 mV, suggesting the involvement of a mixed cationic current in the generation of the event (Fig. 6C). As this excitatory synaptic input occurred with CNQX, CPP, and bicuculline in the perfusion solution, it clearly was not mediated by ionotropic glutamate or GABA_A receptor activation.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Two important conclusions can be drawn from the results of the present study. First, it is clear that the inhibitory circuit is composed of at least three different subpopulations of interneurons, each sharing common morphological characteristics, functional connectivity and synaptic inputs. These findings give further support to the hypothesis that there are distinct populations of interneurons producing GABA_A and GABA_B responses. Second, one subpopulation of hilar interneurons, the spheroid cells, are unique among the interneurons studied because the synchronized bursts generated in these cells in 4-AP appear to be sustained by an ionotropic glutamate and GABA receptor-independent synaptic depolarization.

The experimental conditions, using 4-AP and ionotropic glutamate receptor antagonists, allowed for the characterization of spontaneous bursting activity of different interneuron subpopulations, and its correlation with the synchronized inhibitory responses produced in pyramidal cells in the absence of fast excitatory glutamatergic activity. Our experiments support the findings of others who have demonstrated that interneurons within the hilus exhibit physiologicaland morphological heterogeneity (Buckmaster andSchwartzkroin 1995a,b; Buhl et al. 1994; Han et al. 1993; Mott et al. 1997). These previous studies, however, could only speculate on the functional role of interneuron subpopulations based on their axonal arborization. The conditions of the present study allowed for the first time some insight into the functional role of different interneurons subtypes within the inhibitory circuit.

To morphologically classify the interneurons in the present study, we used the classification system of Lorente de Nó (1934) and Amaral (1978); this classification system correlated very well with the differences in synaptic connectivity observed in our experiments. This morphological classification also has been used recently by other investigators to classify hilar interneurons (Buckmaster and Schwartzkroin 1995a,b). The main morphological differences observed in the present study among these interneuron subpopulations were in the shape and dimension of the soma and the number of primary dendrites. Other differences were found in the degree of arborization, the length of the dendrites, and the spine density. Another recent classification, based mainly on the axon projections of interneurons, has been used by Han et al. (1993) and others (Mott et al. 1997; also see Freund and Buzsáki 1996). We did not perform an extensive study on the axon ramifications to compare the recorded interneurons with the subtypes proposed by these authors.

Because each morphologically identified interneuron subpopulation also could be differentiated by its synaptic connectivity, it is possible to speculate on functional roles of each interneuron subtype within the inhibitory circuit. With our experimental paradigm, we found that pyramidal-like stellate interneurons generated only GABA_A receptor-mediated responses in CA3 pyramidal cells, whereas GABA_B responses were generated only by spheroid and oviform interneurons.

Table 1 summarizes the characteristics of the three subtypes of interneurons. Pyramidal-like stellate cells were defined as GABA_A interneurons because synchronized firing of the population elicited only the GABA_A components in pyramidal cell sIPSPs. In addition, pyramidal-like stellate cells do not appear to participate in producing the GABA_B response because the amplitude of the GABA_B component of the sIPSP was not affected when the synchronized bursts in these cells were blocked by bicuculline.

Spheroid cells of the D-H border and oviform cells are likely candidates for the generation of GABA_B events because their prolonged burst corresponded to the GABA_B component in sIPSPs and because the bicuculline-resistant bursts were correlated temporally with pure GABA_B sIPSPs in the pyramidal cells. These results are suggestive of a distinct population of GABA_B-producing cells; however, because none of the pairs of interneurons and pyramidal cells recorded in this study were connected monosynaptically, these results are not unequivocal. Recording directly from monosynaptically connected pairs of spheroid or oviform interneurons and pyramidal cells should confirm this hypothesis. In addition, because it is not possible to pharmacologically isolate depolarizing and hyperpolarizing GABA_A-mediated responses in interneurons, the present data cannot exclude the possibility that the GABA_B oviform or spheroid interneurons also contribute to the generation of the hyperpolarizing GABA_A response.

An unexpected finding is that spheroid interneurons received a chemical excitatory synaptic input that was not mediated by ionotropic glutamate or GABA receptors. The chemical nature of the synaptic depolarization sustaining the bicuculline-resistant burst in spheroid interneurons is demonstrated by its linear voltage dependence. This input is sufficiently large at resting potential to generate suprathreshold firing and therefore is likely to be an important factor in the synchronization of these interneurons. The nature of the excitatory input is still under investigation, but neither metabotropic glutamate receptors nor -adrenergic or muscarinic cholinergic receptors appear to be involved, as(S)-(-methyl-4)-carboxyphenylglycine, propranolol, and atropine did not affect the GABA_B component of the sIPSPs (unpublished observations). We recently have described a similar glutamate and GABA-independent synaptic depolarization in dentate granule cells which appears to be generated by the synchronous firing of interneurons (Forti and Michelson 1997). Because the experimental conditions in the present study were identical to those used in the previous study on dentate granule cells (Forti and Michelson 1997), it is likely that the synchronized firing of spheroid and/or oviform cells contribute to the generation of this novel synaptic depolarization in the granule cells. In addition, because the granule cell depolarization and spheroid interneuron depolarization are elicited under identical conditions, it is probable that the glutamate and GABA-independent depolarizations in these two cell populations either share a common input or are generated by a similar mechanism.

In contrast to spheroid interneurons, the synchronized bursts in oviform interneurons occurred within a limited range of membrane potentials and were likely to be driven by nonsynaptic mechanisms. The observation of dye coupling selectively among this interneuron subtype suggests that these cells may be coupled electrotonically and is consistent with previous findings demonstrating dye coupling among interneurons with picrotoxin-resistant bursts that were not sustained by an underlying synaptic event (Michelson and Wong 1994).

In conclusion, three subpopulations of hilar interneurons have been identified in the present study, each with distinct morphological properties and synaptic connectivity. These findings indicate that hilar interneuron subpopulations are differentially recurrently connected via various excitatory pathways, including depolarizing GABA synaptic pathways and nonsynaptic pathways. Uniquely, spheroid interneurons are interconnected via a novel, and as yet, unidentified, excitatory synaptic mechanism that is not mediated by either glutamate or GABA. Together, the intrinsic and projectional differences among these interneuron subtypes suggest that each subpopulation of interneuron plays a specific physiological role within the hippocampal circuit.

	ACKNOWLEDGEMENTS

  We thank Drs. R.K.S Wong, R. Bianchi, and K. Perkins for helpful discussions and critical reading of the manuscript.

  This work was supported in part by grants from the National Institute of Neurological Disorders and Stroke Grant NS-33628 and the American Epilepsy Society with support from the Milken Family Medical Foundation to H. B. Michelson.

	FOOTNOTES

  Address for reprint requests: H. B. Michelson, Dept. of Pharmacology, Box 29, State University of New York Health Science Center at Brooklyn, 450 Clarkson Ave., Brooklyn, New York 11203.

  Received 24 October 1997; accepted in final form 13 February 1998.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• ALGER, B., NICOLL, R. A. Pharmacological evidence for two kinds of GABA receptor on rat hippocampal pyramidall cells studies in vitro. J. Physiol. (Lond.) 328: 125-141, 1982.[Abstract/Free Full Text]
• AMARAL, D. G. A golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol. 182: 851-914, 1978.[Medline]
• ARAM, J. A., MICHELSON, H. B., WONG, R.K.S. Synchronized GABAergic IPSPs recorded in the neocortex after blockade of synaptic transmission mediated by excitatory amino acids. J. Neurophysiol. 65: 1034-1041, 1991.[Abstract/Free Full Text]
• BAKST, I., AVENANO, C., MORRISON, J. H., AMARAL, D. G. An experimental analysis of the origins of somatostatin-like immunoreactivity in the dentate gyrus of the rat. J. Neurosci. 6: 1452-1462, 1986.[Abstract]
• BLAKSTAD, T. W. Commissural connections of the hippocampal region in the rat, with special reference to their mode of termination. J. Comp. Neurol. 105: 417-537, 1956.
• BUCKMASTER, P. S., SCHWARTZKROIN, P. A. Interneurons and inhibition in the dentate gyrus of the rat in vivo. J. Neurosci. 15: 774-789, 1995a.[Abstract]
• BUCKMASTER, P. S., SCHWARTZKROIN, P. A. Physiological and morphological heterogeneity of the dentate gyrushilar interneurons in the gerbil hippocampus in vivo. Eur. J. Neurosci. 7: 1393-1402, 1995b.
• BUHL, E. H., HAN, Z. S., LORINCZI, Z., STEZHKA, V. V., KARNUP, S. V., SOMOGYI, P. Physiological properties of anatomically identified axo-axonic cells in the rat hippocampus. J. Neurophysiol. 71: 1289-1307, 1994.[Abstract/Free Full Text]
• BUZSÁKI, G., HORVATH, Z., URIOSTE, R., HETKE, J., WISE, K. High-frequency network oscillation inthe hippocampus. Science 256: 1025-1027, 1992.[Abstract/Free Full Text]
• COBB, S. R., BUHL, E. H., HALASY, K., PAULSEN, O., SOMOGYI, P. Synchronization of neuronal activity in hippocampus by individual GABAergic interneurons. Nature 378: 75-98, 1995.[Medline]
• FORTI, M., MICHELSON, H. B. Novel glutamate- and GABA-independent synaptic depolarization in granule cells of guinea pig hippocampus. J. Physiol. (Lond.) 504: 641-648, 1997.
• FREUND, T. F., BUZSÁKI, G. Interneurons of the hippocampus. Hippocampus 6: 347-470, 1996.[Medline]
• GULYÁS, A. I., TÓTH, K., DANOS, P., FREUND, T. F. Subpopulations of GABAergic neurons containing parvalbumin, calbindin D28k, and cholecystokinin in the rat hippocampus. J. Comp. Neurol. 312: 317-378, 1991.
• HAN, Z. S., BUHL, E. H., LORINCZI, Z., SOMOGYI, P. A high degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci. 5: 395-410, 1993.[Medline]
• HONORÉ, T., DAVIES, S. N., DREJER, J., FLETCHER, E. J., JACOBSON, P., LODGE, D., NEILSEN, F. E. Quinoxalinediones: potent competitive non-NMDA glutamate receptor antagonists. Science 241: 701-703, 1988.[Abstract/Free Full Text]
• LACAILLE, J.-C., WILLIAMS, S. Membrane properties of interneurons in stratum oriens-alveus of the CA1 region of rat hippocampus in vitro. Neuroscience 36: 349-359, 1990.[Medline]
• LORENTE DE NÓ, R. Studies on the structure of the cerebral cortex. II. Continuation of the study of the ammonic system. J. Psychol. Neurol. 46: 113-177, 1934.
• MICHELSON, H. B., WONG, R.K.S. Excitatory synaptic responses mediated by GABA receptors in the hippocampus. Science 253: 1420-1423, 1991.[Abstract/Free Full Text]
• MICHELSON, H. B., WONG, R.K.S. Synchronization of inhibitory neurones in the guinea-pig hippocampus in vitro. J. Physiol. (Lond.) 477: 35-45, 1994.[Medline]
• MILES, R., WONG, R.K.S. Inhibitory control of local excitatory circuits in the guinea-pig hippocampus. J. Physiol. (Lond.) 388: 611-629, 1987.[Abstract/Free Full Text]
• MOTT, D. D., TURNER, D. A., OKAZAKI, M. M., LEWIS, D. V. Interneurons of the dentate-hilus border of the rat dentate gyrus: morphological and electrophysiological heterogeneity. J. Neurosci. 17: 3990-4005, 1997.[Abstract/Free Full Text]
• MULLER, W., MISGELD, U. Picrotoxin- and 4-aminopyridine-induced activity in hilar neurons in the guinea pig hippocampal slice. J. Neurophysiol. 65: 141-147, 1991.[Abstract/Free Full Text]
• PARÉ, D., DE CURTIS, M., LLINAS, R. Role of the hippocampal entorhinal loop in temporal lobe epilepsy: extra-and intracellular study in the isolated guinea pig brain in vitro. J. Neurosci. 12: 1867-1881, 1992.[Abstract]
• PERKINS, K. L., WONG, R.K.S. Ionic basis of the postsynaptic depolarizing GABA response in hippocampal pyramidal cells. J. Neurophysiol. 76: 3886-3894, 1996.[Abstract/Free Full Text]
• PERREAULT, P., AVOLI, M. Physiology and pharmacology of epileptiform activity induced by 4-aminopyridine in rat hippocampal slices. J. Neurophysiol. 4: 771-785, 1991.
• RAMÓN Y CAJAL, S. Structure of the Ammon's Horn. Translated by Elisabeth M. Kraft. Springfield, IL: Charles Thomas, 1955.
• RIBAK, C. E., SERESS, L. Five types of basket cells in the hippocampal dentate gyrus: a combined Golgi and electron microscopic study. J. Neurocytol. 12: 577-597, 1983.[Medline]
• SAMULACK, D. D., LACAILLE, J.-C. Hyperpolarizing synaptic potentials evoked in CA1 pyramidal cells by glutamate stimulation of interneurons from the oriens/alveus border of rat hippocampal slices. II. Sensitivity to GABA antagonists. Hippocampus 3: 345-358, 1993.[Medline]
• SEGAL, M. A subset of local interneurons generate slow inhibitory postsynaptic potentials in hippocampal neurons. Brain Res. 511: 163-164, 1990.[Medline]
• SLOVITER, R. S., NILAVER, G. Immunocytochemical localization of GABA-, cholecystokinin-, vasoactive intestinal peptide-, and somatostatin-like immunoreactivity in the area dentata and hippocampus of the rat. J. Comp. Neurol. 256: 42-60, 1987.[Medline]
• TRAUB, R. D., WHITTINGTON, M. A., COLLING, S. B., BUZSÁKI, G., JEFFERYS, J.G.R. Analysis of gamma rhythms in the rat hippocampus in vitro and in vivo. J. Physiol. (Lond.) 493: 471-484, 1996.[Medline]
• WILLIAMS, S., VACHON, P., LACAILLE, J.-C. Monosynaptic GABA-mediated inhibitory postsynaptic potentials in CA1 pyramidal cells of hyperexcitable hippocampal slices from kainic acid-treated rats. Neuroscience 52: 541-554, 1993.[Medline]

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society