Exp Brain Res (1992) 91:105-114

Experimental BrainResearch 9 Springer-Verlag1992

Electrical properties of neurons in the mediolateral part of the lateral septum: intracellular recordings from guinea-pig brain slices B. Carette, P. Poulain, and O. Doutrelant INSERM U 156, Place de Verdun, F-59045 Lille Cedex, France Received December 4, 1991 / Accepted April 8, 1992

Summary. Membrane properties of 174 neurons were studied in the mediolateral part of the lateral septum (LSml) using an in vitro slice preparation of guinea-pig brain. Intracellular recordings were correlated with morphological data obtained from 34 neurons intracellularly stained with horseradish peroxidase (HRP). Neurons were divided into three classes according to their electrical responses. Class A and B neurons displayed the common property of an overshooting of spikes in response to the direct application of weak depolarizing current pulses. Class A neurons (59.2% of the total) generated tetrodotoxin-insensitive, high-threshold Ca 2+ spikes in a control medium. Class B neurons (20.7% of the total) generated high-threshold Ca 2+ spikes only if tetraethylammonium was used to block delayed rectifying K + current. Features common to class A and B neurons included the inactivation of Na + conductance, the participation of high-threshold Ca 2+ conductance in the generation of spikes - when repetitive discharges were elicited by strong depolarizing current pulses - and Cs +sensitive, Ba 2+-insensitive anomalous rectification. Class C neurons (20.1% of the total) displayed discharges comprising small-amplitude Na + spikes followed by slow and large Ca 2+ spikes, suggesting a locus of impalement which was not the soma. HRP-filled class B neurons (n = 5) were characterized by small to medium perikarya with spindly dendrites. The majority of HRP-filled class A (15/21) and all class C (n= 8) neurons showed large perikarya with thick primary dendrites and spiny dendritic branches. Thus, class A and C neurons typify the guinea-pig LSml in their morphological characteristics and in their ability to generate high-threshold Ca 2+ spikes in a control medium. Key words: Lateral septum - Slice - Horseradish peroxidase - Intracellular recording - Guinea-pig

Correspondence to: B. Carette

Introduction The rat lateral septal nucleus (LS) is classically divided into distinct subnuclei (Swanson and Cowan 1979), on the basis of the dorsoventral organization of the neurons, their cytoarchitectonic properties, afferent and efferent connections, and chemical characteristics (review Alonso and Frotscher 1989). Recently, in the guinea-pig, tracttracing studies have also demonstrated the specificity of the inputs (Staiger and Niirnberger 1989) and the outputs (Staiger and Niirnberger 1991a; Staiger and Niirnberger 1991b) of the different LS dorsoventral subnuclei. The dorsal part of the LS has received considerable attention due to its reciprocal connections with the hippocampus, including a relay through the medial septal nucleus for the septo-hippocampal projection (De France 1976; Meibach and Siegel 1977; Raisman 1966; Swanson and Cowan 1977). The electrical properties of the neurons from this area have been described for the rat (Stevens et al. 1984) and the guinea-pig (Alvarez de Toledo and Ldpez-Barneo 1988; Ldpez-Barneo et al. 1985). The area of the caudal LS located between the dorsal part and the bed nucleus of the stria terminalis - termed the mediolateral part of the LS (LSml: Jakab and Leranth 1990b) - forms a distinct anatomical entity. There is growing evidence, for the rat, that the LSml is a target for hippocampal fibres and for numerous different neuroactive substance-containing fibres which converge towards the same neurons (Jakab and Leranth 1990a; Jakab and Leranth 1990b; Jakab et al. 1991; Leranth and Frotscher 1989). In the rat and the guinea-pig, the LSml receives a robust enkephalinergic innervation originating from the hypothalamus (Poulain et al. 1984; Sakanaka and Magari 1989). The LSml mainly projects to the hypothalamic neuroendocrine areas rather than to the medial septal nucleus in the rat (Staiger and Wouterlood 1990) and the guinea-pig (Staiger and Nfirnberger 1991a). These observations point to the potential physiological importance of the LSml as a target for neuroactive substances, including enkephalins, as well as a

106

hippocampo-septo-hypothalamic pathway. Except for a few neurons in the guinea-pig studied by Alvarez de Toledo and L6pez-Barneo (1988), no data are available on the electrophysiological properties of the LSml neurons. The present paper is devoted to the description of intracellular recordings performed on LSml neurons from guinea-pig brain slices. In some cases, electrophysiologically characterized units were injected with horseradish peroxidase (HRP) to provide Golgi-like images of the recorded neurons. l i n k in a

Materials and methods

Frontal slices through the septum were obtained from female guinea-pigs weighing 200-250 g. Slices were cut in cooled oxygenated artificial cerebrospinal fluid from a block of tissue attached to the stand of a Vibroslice. Routinely, three successive slices, 400 gm thick, were studied. They were selected, during cutting, in such a way that, in the most posterior slice, the anterior commissure had crossed the midline. Slices were transferred immediately after sectioning to a recording chamber based on the Oslo interface model chamber (Schwartzkroin 1975). Slices were maintained at 34 ~ C and continuously perfused at a rate of 2 ml/min. Artificial cerebrospinal fluid contained (in raM): NaCI, 124; KC1, 5; CaCI/, 2.4; NaHCO3, 26; KHzPO4, 1.24; MgSO4, 1.3; glucose, 10. It was continuously gassed (95 % O2, 5 % COz) to keep a pH of 7.4. Drugs were dissolved in the medium at known concentrations. Tetrodotoxin (TTX) was used at 10-6 M, tetraethylammonium (TEA) at 10 mM, caesium (Cs +) at 10mM, 4-aminopyridine (4-AP) at 2 m M and bicuculline at 0.1 mM. Cobalt (Co 2+) and barium (Ba z +) were substituted at equimolar concentrations for Ca z +. For this purpose, CaClz was replaced with 2.4 mM COC12 or BaC12 when preparing the medium, while MgSO4 was replaced with MgC12 and KH2PO4 was omitted to avoid precipitation. Intracellular recordings were performed with microelectrodes pulled with a Flaming-Brown apparatus. Microelectrodes were made from capillary glass (1 mm o.d. x 0.58 mm i.d.) and filled with KCI 3M (resistance: 60-80 Ms or 7% H R P dissolved in 0.5 M TRIS-HC1 buffer (resistance: 55-85 M ~ after beveling). Microelectrodes were also obtained from thin-wall capillaries (1 mm o.d. x 0.78 mm i.d.) and did not require beveling when filled with HRP (resistances: 40-80 M~). Microelectrodes were lowered under binocular control in the LSml. Current-clamp recordings were made with a bridge amplifier (Neurolog). Data were displayed using the internal plotter of a digital storage oscilloscope (Gould 1600). HRP was injected by means of 10 ms, 2 nA depolarizing pulses at 50 Hz for 10 min. Following iontophoretic injection, the slice was fixed overnight in a mixture of 2 % glutaraldehyde, 1.25 % paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, and then transferred to a 20% sucrose phosphate buffer solution before histological treatment. The slice

was cut serially with a cryotome into frontal sections, 40 gm thick. They were reacted in H20 2 with diaminobenzidine as chromogen and nickel ammonium sulphate for intensification (Metz et al. 1989). After mounting onto gelatin-wated slides, sections were counterstained with 0. 1% thionin, dehydrated in graded alcohols, cleared with xylene and cover-slipped with Eukitt. Drawings of the HRP-injected cells were made at a magnification of x 40 using a camera lucida.

Results

The following data are derived from intracellular recordings in 174 LSml neurons. Criteria for the selection of these neurons were a stable impalement and a resting membrane potential (taken after withdrawing the microelectrode) of at least - 50 mV. Two cell types (79.9% of the total, n = 139), called classes A and B, displayed overshooting spikes in response to a threshold depolarizing current pulse applied via the microelectrode (Table 1). A third cell type (20.1% of the total, n = 35), called class C, was considered separately because it was characterized, under similar conditions of stimulation, by the occurrence of small amplitude spikes (Table 1).

Class A and class B neurons

There was no striking difference between the passive membrane properties of class A and class B neurons. The mean resting membrane potential for these neurons was - 61 • 7 mV (mean• SD, n = 30), their mean input resistance (obtained from steady-state responses to pulses of hyperpolarizing currents) was 78.9 + 22 Mf~ (mean + SD, n = 30). Five neurons exhibited regular spontaneous activity at the resting membrane potential. In addition, these neurons showed spontaneous depolarizing potentials which were suppressed by TTX (not shown). Distinction between class A and class B neurons. Class A

and class B neurons were defined by the presence or absence of responses to depolarizing current pulses applied after TTX administration. Under these conditions, class A neurons (n = 103) displayed large spikes (Fig. 1C). These TTX-insensitive spikes showed higher thresholds, slower rates of rise and fall, lower amplitudes and were of larger duration than those of spikes recorded in control conditions (Table 1). The Ca 2+ nature of these TTX-

Table 1. 'Characteristics (mean_+ SD) of the spikes evoked by depolarizing current pulses for representative neurons of each class Class A neurons (n = 30) First spike Height threshold-peak (mV) Width at threshold (ms) Rate of rise (mV/ms) Rate of fall (mV/ms)

Ca 2 + spike

Class B neurons (n = 16)

Class C neurons (n = 26)

First spike

First spike

Ca 2 + spike

17.5_+4.5

31.1-t-7

5.1-+2.5

18.8_+6.2 9.4_+2.3 --6.9+_1.7

62.2+ 6.5

20.7_+2.2

66.6_+ 4.1

2.8_+ 0.3 146.7_+33.8 --41.1-+11.3

6.6-t- 1.2 8 -+1.9 --8.7+_2.4

2 -+ 0.4 171.5-+13.9 --88.1+_19.2

* *

First spike corresponds to the spikes evoked by a direct stimulation just above the threshold; Ca 2 + spike corresponds to the high-threshold (TTX-resistant) spikes observed in class A and class C neurons

107

30 mV

~JB

_

D

~

J

E

l

j L

C

L

J

F

TTX

~

G

0.5nA

. T TX+

L

Fig. 1

A-G. Basic membrane properties of class A and class B LSml neurons. A-C. Discharges of a class A neuron (resting membrane potential, RMP, - 58 mV) to illustrate the differencesin amplitude and time course of spikes in response to threshold (A) and suprathreshold (B) depolarizing current pulses. In C, TTX-resistant

spikes were elicited by a strong depolarizing pulse. D-G. Discharges of a class B neuron (RMP - 66 mV). Spikes were totally abolished by T T X (F). Application of T E A revealed large high-threshold Ca 2+ spikes (G)

insensitive spikes was demonstrated by application of Co 2+, which abolished responses (not shown). In class B neurons (n= 36), administration of TTX eliminated all responses even under strong depolarizing current pulses (Fig. 1F). However, addition of TEA to the medium containing TTX allowed depolarizing pulses to trigger large-amplitude, long-duration spikes (Fig. 1G). These spikes were eliminated by Co 2 + (not shown), indicating their Ca 2+ nature.

(cf. Table 1). Their elimination in the presence of Co 2+ indicated that they were Ca 2 + spikes (not shown). Following application of Co 2+, the amplitude and duration of the AHPs were reduced, suggesting that the origin of the AHPs was a Ca 2 +-dependent K + conductance. Reduction of the AHP duration produced an increased number of Na + spikes in response to moderate depolarizing current pulses (Fig. 2H). With higher intensity current pulses, Na + spikes progressively decreased in amplitude, until spike discharge was totally suppressed (Fig. 2I). This phenomenon, which tended to limit the number of Na + spikes during a strong depolarizing current pulse, was probably due to inactivation of Na + conductance. The significant participation of Ca 2 + conductance in the generation of the spikes, triggered by moderate depolarizing current pulses, is apparent when comparing the traces F and I of Fig. 2. As a result of Co 2 + application, only a much reduced discharge of Na + spikes subsisted (Fig. 2I). The pattern of discharge observed in a control medium in response to a similar depolarizing pulse strongly suggested that activation of Ca 2 + conductance occurred as fast as Na + conductance inactivation, this mechanism allowing a repetitive discharge in spite of Na + inactivation (Fig. 2F). In the same cell, measuring the rate of rise of the spikes in the control medium (Fig. 2E) and during application of Co 2+ (Fig. 2H), in response to a similar depolarizing pulse, showed that a Ca 2+ influx occurred during the depolarizing phase of the spike. In the presence of Co 2 +, spikes within the train presented a rate of rise which progressively decreased, a feature which may be attributed to Na + inactivation, whereas in the control medium, progressive slowing of the rate of rise was much more accentuated than that observed under Co 2 + (Fig. 2D). The weaker rate of rise of the spikes in normal conditions than in the presence of Co 2 + is explained by the activation of Ca z + conductance during the rising phase of the spike. Another example suggesting that Ca 2+ conductance strongly participates in the genesis of spikes under nor-

Repetitive discharges of class A neurons. Direct stimulation, just above threshold, consistently evoked one or two fast and full amplitude spikes (Fig. 1A, 2A). As the current strength was increased (Fig. 1B, 2B, C, E, F), the neurons produced an increasing number of spikes. Whereas the first spike remained relatively short in duration (cf. Table 1), successive spikes became progressively wider following repetitive discharge. Widening depended on the intensity of injected current and on the time of occurrence of the spike within the train (Fig. 2A, B, C, E, F). Progressive attenuation in spike amplitude was also observed during repetitive discharge (Fig. 2E, F). Application of strong depolarizing pulses elicited a characteristic response consisting of an initial fast and full amplitude spike followed by larger and smaller spikes which immediately, or progressively, reached a constant amplitude (Fig. 2F). As indicated by the graph Fig. 2D, the rate of rise of the spikes also changed during repetitive discharge. Measurements showed a progressive decrease in the rate of rise in terms of both the intensity of injected current and the order of the spike within the train. Rate of fall decreased in a similar manner. During application of depolarizing current pulses, the first fast and full amplitude spike was followed by an after-hyperpolarizing potential (AHP) of short duration, whereas, during the following train of spikes, AHPs progressively increased before reaching a constant value (compare Fig. 2B and 2F). When using current pulses of longer duration and higher intensity, large spikes were evoked (Fig. 2G), resembling those obtained in presence of TTX

108

~

250' m V I m s 200

100 50

J

A

J

B

=

E

LJ

F co

--o-

1.

--~--L

-J

C

G

+

L

D

Ot

1 2 3 4 5 6 7 8 spike order

J30 mV I

~ ~ . Xm -s

IIo.a.. nA /

50

/ H

---D--

~_J I

j

Fig. 2A-K. Discharge patterns of three class A neurons. A-I correspond to the same neuron (RMP - 6 0 mV) and show repetitive discharges in response to depolarizing pulses of increasing amplitudes: note the modifications in amplitude and duration of spikes and AHPs. In the presence of Co 2 +, note the increase in the number of spikes (H) and the Na + inactivation (I). The graph in D indicates the evolution of the rate of rise (mV/ms) of successive spikes in function of current pulse intensities, in control medium and in

presence of Co 2+. Values were measured on every spike of the 5 discharges marked by different symbols. The trace in G shows that

mal conditions is shown in Fig. 2J. In this example, successive spikes triggered by a depolarizing current pulse became progressively wider, whereas a shoulder appeared on their repolarizing phases. Concomitantly, the AHPs increased in duration and amplitude before reaching a constant value. Co 2 + application suppressed the shoulders and reduced duration and amplitude o f AHPs. This finding was an indication of the involvement of a Ca 2+ conductance during the repolarizing phase of the spikes. It further suggested that a Ca 2 +-dependent K + conductance was involved in the generation of the AHPs. In some neurons, the Ca z + participation in the repolarizing phase of the spike in response to a moderate current pulse was observed as a prolonged afterdepolarizing potential (Fig. 2K). In a small subset o f neurons (n = 6), as illustrated in Fig. 3B, only one spike was triggered at the beginning of a depolarizing current pulse after Co 2+ application, whereas a repetitive discharge was obtained under control conditions (Fig. 3A). A brief hyperpolarizing pulse applied during the depolarizing pulse allowed the resetting of a spike at the anodal break of the hyperpolarization (Fig. 3C). An explanation of this observation is that, in these cells, suppression by Co 2+ o f Ca 2 +-dependent AHPs prevented repetitive discharge from occurring. This was probably due to a strong and unusual N a +

K a pulse of longer duration and higher amplitude evoked larger and slower spikes (arrowheads), followed by 4 pure Ca z+ spikes. J In another neuron (RMP - 65 mV) under continuous depolarization, shoulders appeared in the repolarizing phase of the spikes. Two traces are superimposed: responses in control medium are indicated by arrowheads. Shoulders were eliminated by Co z+ (asterisk). K In another neuron (RMP - 5 5 mV), a prominent after-depolarizing

potential in the repolarizing phase of the spike was observed with a threshold depolarizing current pulse

inactivation which was removed, in this example, by application of a brief hyperpolarizing pulse. Under control conditions (Fig. 3A), AHPs resulting from Ca 2+ entry during the spike were likewise able to de-inactivate N a + conductance, and consequently to trigger a new spike. F o r another neuron (Fig. 3D), a strong N a + inactivation rapidly occurred after the onset of the depolarizing pulse, and the discharge o f spikes endowed with N a + conductance was replaced by a pure Ca 2 + spike. As in the previous example, it was observed that the A H P which followed the Ca 2 + spike was effective in de-inactivating N a + conductance so that N a + participated in the following spikes, as indicated by their fast rate o f rise.

Repetitive discharges of class B neurons. Repetitive discharge of class B neurons was investigated under similar conditions to those used for class A neurons. F o r 16 neurons, in response to a small depolarizing current pulse, a regular discharge of full N a + spikes was observed (Fig. 1D). An increase in current pulse evoked discharges similar to those observed in the class A neurons (Fig. 1E). It appeared that discharge was impeded by Na + inactivation and Ca 2+ participated in the generation of the spikes as judged by their slow rate of rise (Fig. 1E) and their evolution during application of Co 2 + (not shown). However, no pure Ca 2+ spike appeared

109

_~ J

Co

L__J

A

L_

B

-~

r

A

30 mV I 0.5hA 50 ms j

I .

-1,0-0,8-0,6-0,4-0,2 . . . . . . .

nA

~ V

U--L_

C

0,0 .

.

.

_~-I0

0

30 mV [0.5nA 100ms [3 O A

]'-30 20 I I i ! -40

-50 ,

C

60

Fig. 5A-D. Presence of anomalous rectification in two class A .A

b

D Fig. 3A-D. Properties of two class A neurons. A At RMP - 63 mV, tonic discharge in control medium. B In the presence of Co 2+, only one spike was evoked by a current pulse of similar intensity. C A brief hyperpolarizing pulse evoked a spike. D In another cell (RMP - 5 5 mV), when a long current pulse was applied, a rapid and pronounced spike attenuation was observed, and then a Ca z + spike (as suggested by its rate of rise, 11 mV/ms; arrowhead) was elicited. This spike was followed by a developed AHP which triggered a repetitive discharge of spikes displaying more rapid rate of rise (ranging from 26 mV/ms to 36 mV/ms)

neurons (A RMP - 5 4 m V , B RMP - 6 1 mV) and in a class C neuron (D RMP - 67 mV). A, B, D Responses to hyperpolarizing current pulses showed a depression in the membrane voltage deflection and a post-anodal depolarization, which persisted in the presence of Ba 2+ (A) but were eliminated by application ofCs + (B, D). C I-V curves for the neurons illustrated in B and D. Peak (circles) and steady-state (squares; trianglesunder Cs +) membrane voltage deflections were measured at RMP as shown in B and D. Filled labels correspond to the neuron illustrated in B, open labels to the neuron illustrated in D

depolarizing pulse. Nevertheless, with higher depolarizing pulses, these n e u r o n s exhibited N a + spikes which progressively decreased in amplitude (Fig. 4B) in a similar w a y to class A neurons.

Inward (anomalous) rectification. A l m o s t all class A and

Ai,

~-

~

.

_

130mV

B

I [0.5nA

~

100ms

-

Fig. 4A-C. Properties of a class B neuron (RMP - 7 2 mV). A Activity evoked by a depolarizing current pulse. Note the higher value of the first interspike interval. B A larger current pulse resulted in a decrease in spike amplitude until failure of spike generation. C A transient outward rectification was apparent when a hyperpolarizing pulse was applied at a depolarizing membrane potential (RMP indicated by arrowhead) either in c o n t r o l conditions or in presence o f T T X (Fig. 1E, F). In 20 other neurons, o f which a typical example is s h o w n in Fig. 4A, C a 2 + participation in spikes was n o t a p p a r e n t : s h a r p - s h a p e d spikes h a d short durations during repetitive discharge, even in response to a strong

class B n e u r o n s displayed a p r o n o u n c e d depression o f their electrotonic responses to hyperpolarizing current pulses which was a c c o m p a n i e d by a r e b o u n d depolarization u p o n ending o f the pulses (Fig. 5A, B). This indicated the presence o f a time-dependent a n o m a l o u s rectification, which is illustrated by the current-voltage curves presented in Fig. 5C. It can be observed that, in addition to the time-dependent rectification, the points c o r r e s p o n d i n g to the measures o f input resistances deviated f r o m linearity, indicating the presence o f an instantaneous a n o m a l o u s rectification. The depression in the electrotonic responses was n o t suppressed by the addition o f T E A , T T X , Co 2+ (not shown) or Ba 2+ (Fig. 5A), but disappeared during the application o f Cs + (Fig. 5B). C o n c o m i t a n t l y , the p o s t - a n o d a l depolarization was partially eliminated (Fig. 5B). D u r i n g the application o f Cs +, the instantaneous c o m p o n e n t o f the a n o m a l o u s rectification was also eliminated.

Transient outward rectification. T o reveal a transient outw a r d rectification, brief hyperpolarizing current pulses were applied to n e u r o n s that were held at m e m b r a n e potentials m o r e positive than the resting potential. In eight LSml n e u r o n s f r o m b o t h classes, following term i n a t i o n o f the hyperpolarizing current pulse, there was

110 an initial rapid phase followed by a slower phase in the return to the base-line (Fig. 4C). Addition o f 4 - A P did not block the transient outward rectification (not shown).

to causing the appearance o f two distinct kinds o f spikes, depolarizing pulses elicited slow depolarizing potentials of variable amplitude and duration. They were observed as arising on the repolarizing phase of the fast spikes (Fig. 6E). They reached the threshold for generation of large spikes or inactivated after a short plateau phase (Fig. 6E). Slow depolarizing potentials could also be observed in isolation in response to the depolarizing pulse (not shown).

Class C neurons

Description. The m e a n resting m e m b r a n e potential for class C neurons (n = 35) was - 60 + 6.9 m V (mean 4- SD, n = 20). Their input resistance was 117 4- 59 Mf~ (mean 4- SD, n = 31). At the m e m b r a n e resting potential, low intensity depolarizing current pulses typically elicited a single (Fig. 6A) or, m o r e rarely, several fast spikes (not shown) o f 2-10 ms duration (5.1 4- 2.5 ms, m e a n + SD, n = 2 6 ) and 8 - 2 5 m V amplitude ( 1 7 . 5 + 4 . 5 m V , m e a n 4- SD, n = 26). With increased intensity o f the depolarizing pulse, the threshold for the generation o f a repetitive discharge o f slower and larger spikes was reached (Fig. 6B, E, F). These spikes had long durations o f 10-35 ms (18.8 4- 6.2 ms, mean • SD, n = 26) and large amplitudes of 20-45 m V (31.1 4- 7 mV, mean + SD, n = 26) (Table I). In some cases, they displayed double peaks (Fig. 6B). They were followed by prominent A H P s (Fig. 6B, E, F). Pronounced A H P s followed the bursts of large spikes and developed at the end of the depolarizing pulses (not shown). F o r eleven neurons, in addition

Ionic specificity. The ionic specificity of the currents responsible for the fast and the slow spikes was investigated with T T X (5 neurons) and Co z§ (13 neurons). As illustrated in Fig. 6C, application of T T X suppressed fast spikes whereas slow spikes were not affected. This indicated the N a § dependence of the former. In the same cell, following T T X treatment, the repetitive discharge of slow spikes was eliminated by Co 2+, indicating their Ca / + nature (Fig. 6D). Slow depolarizing potentials were also eliminated by Co z§ (Fig. 61). The relative importance of K § conductances during Ca 2 + and N a 2 § spikes was assessed by the addition o f T E A (10 neurons). During application o f TEA, the amplitude o f the Ca z § spike was increased (Fig. 6G, H) and its multi-peaked appearance was usually eliminated (not shown). In addition, application of T E A elicited a slow depolarizing potential on the repolarizing phase of the first fast spike

TTX TTX+Co

i

,.r

I__

S

A

J

- -

t__

C

I

-f= -

L

__f

L_..

D

30 mV

[0.5nA .... J

100 ms

I

I

.,,......,,,.d

F

50 ms ....

H'

'

I'

'

Fig. 6A-I. Recordings from class C neurons. A-D correspond to the same neuron (RMP - 63 mV). A In control medium, a depolarizing current pulse of low amplitude only triggered a small fast spike. B A stronger depolarizing pulse evoked in addition a repetitive discharge of larger spikes, one displaying a double peak. Large spikes were followed by well-developed AHPs. C In the presence of TTX, the small spike was suppressed whereas the large spikes were unaffected. D TTX-resistant spikes were blocked when Co 2+ was substituted for Ca 2+. E In another neuron (RMP - 5 8 mV), response to a strong depolarizing pulse showed a small fast spike followed by a repetitive discharge of large spikes. Note the presence of a slow depolarizing potential evoked on the repolarizing phase of a small spike (arrowhead). F-I correspond to the same neuron (RMP -- 65 mV). F In control medium, this neuron exhibited a small

j

Bic

fast spike followed by a large spike which was apparently triggered by another small fast spike. G Application of TEA increased the size of the large spike and of the related AHP. The repolarizing phase of the first spike was slowed and a slow depolarizing potential developed (arrowhead). H One minute later, the slow depolarizing potential had increased in amplitude and was able to trigger a large spike. I After washing in a medium containing Co 2+ instead of Ca 2+, the discharge consisted of repetitive small spikes. J Action of bicucuUine (Bic) on another neuron. Spontaneous firing at the RMP (upper trace) showed postsynaptic potentials which triggered small Na § spikes (asterisks). Postsynaptic potentials were blocked after 2 min application of Bic (middle trace). Lower trace shows recovery

111

.~

(Fig. 6G). Progressively, T E A administration caused the slow depolarizing potential to increase in amplitude sufficiently to trigger a slow spike (Fig. 6H). This suggests that voltage- and/or Ca 2 +-dependent K + conductances participated in the repolarization o f Ca 2+ and N a + spikes. Following application o f TEA, AHPs following individual Ca 2+ spikes (Fig. 6G, H), and bursts of Ca 2+ spikes produced by depolarizing pulses (not shown) were considerably increased, suggesting the involvement of a Ca 2 +-dependent, K + conductance. Other characteristics. Ten neurons exhibited TTXsensitive spontaneous potentials at resting membrane potential. As illustrated in Fig. 6J, these events were positive-going with respect to the base line, their decay was multiphasic and they had maximal amplitudes o f 9-10 mV. They were sometimes associated with small (Na +) and large (Ca 2 +) spikes. The relatively high number of these events made it possible to study the effects of bicuculline on three neurons. Application of bicuculline completely suppressed all spontaneous potentials. This effect was reversible (Fig. 6J). These observations probably indicate that the TTX-sensitive spontaneous potentials are GABA-mediated inhibitory postsynaptic potentials, which were observed as depolarizing events due to the use of KC1 microelectrodes. Anomalous rectification was present during hyperpolarization for almost all neurons. Characteristics of time-dependent and instantaneous components o f the rectification were similar to those observed from class A and class B neurons (Fig. 5C, D).

50pm

VL~

!LSi / / 1 " \ Ls,

{1'

I Ms i

Morphological observations

Fig. 7A-C. Camera lucida drawings of 3 HRP-filled neurons in the LSml. Position of neurons is indicated in a frontal section through the guinea-pig septum. Topography is depicted according to Swanson and Cowan (1979): ac, anterior commissure; BST, bed nucleus of the stria terminalis; CC, corpus callosum; LSd, dorsal part of LS ; LSi, intermediate part of LS; LSv, ventral part of the LS; MS, medial septal nucleus; VL, lateral ventricle. Localization of LSml is indicated by the dotted area. A Class A neuron showing a robust dendritic trunk (arrowhead). Dendrites do not branch much and are covered with spines. Axon, as in following drawings, is indicated by small arrows. B Class B neuron with slender, non-spiny dendrites. C Class C neuron showing two large-based primary dendrites (arrowheads) and spines distributed on the perikaryon and on the dendrites

Some experiments were carried out with HRP-filled microelectrodes. These microelectrodes were selected for their excellent electrical properties which made it possible to classify the impaled neurons into the three classes ascribed before H R P injection. Thirty-four well-injected neurons were found in the LSml and belonged to class A (n=21), class B (n= 5) or class C (n= 8). Neuronal profiles of class A neurons had heterogeneous morphologies but generally, two main types o f neurons, differing consistently in structural features, emerged from our observations. Fifteen neurons exhibited common characteristics, although they possessed variable morphologies. Perikarya were large (30 x 17 gm average diameter), were ovoid, globular or multipolar and were typified by the presence of thick dendritic arbors exhibiting large bases, measuring 2-4 gm in diameter. For 3/15 cells, several thick primary dendrites radiated in all directions, whereas in the other cells (12/15), a single enlarged primary dendrite extended one pole o f the perikaryon. This thick primary dendrite branched near to its origin or, in some cases, stretched away at a greater distance, conferring on the perikaryon a characteristic stalked appearance. In all cases, large dimensions of the perikarya and thickened primary dendrites constituted morphological properties which gave these neurons a very typical appearance. Moreover, all cells but one showed

densely spiny branches. Generally, thickened primary dendrites were devoid of spines, whereas second and third order branches were profusely covered with various types of appendages. The morphology of one of these 15 large neurons is illustrated in Fig. 7A. Six class A neurons exhibited different morphological characteristics. They had small to medium perikarya (22.5 x 12 gm average diameter), with fusiform contours and two dendritic trunks originating from each end of the perikaryon. Dendrites were spindly, non-spiny (3 cells) or fairly spiny (3 cells). Perikarya of class B neurons were small to medium (20 x 10 gm average diameter), round or ovoid. Processes were slender and dendritic spines were rare, except for one cell which was distinguished from others because distal portions of the dendrites exhibited a profusion of appendages. A reconstitution of one cell of the 5 from this class is shown in Fig. 7B. Several features in c o m m o n with the largest class A neurons were apparent when examining the morphology of the 8 class C neurons. First, perikarya were large (28 • 15 gm average diameter) and ovoid. Second, neurons showed thick, large-based primary dendrites giving, in two instances, a characteristic stalked aspect to the perikarya. Third, dendrites possessed an abundance of

112

spines. An illustration of a representative neuron from this group is shown in Fig. 7C.

Discussion

In the present study of guinea-pig LSml neurons, we found three easily distinguishable classes of neurons. Unlike class B neurons, the prominent group of neurons representative of class A (59.2% of the total) exhibited h i g h - t h r e s h o l d C a 2 + spikes under control conditions. In class B neurons (20.7% of the total), Ca 2+ spikes were only observed after delayed rectifying K + currents were blocked with TEA. Moreover, our findings indicate that the morphologies of class A and class B neurons differ significantly. Despite these marked differences, class A and class B neurons possess some properties in common. First, in all neurons, a rapid decline in the amplitude of Na § spikes in response to strong depolarizing current pulses was observed. This feature, referred to in our study as an inactivation of Na § currents, was clearly observed in experiments carried out in the presence of Co 2§ Recent patch-clamp studies on guinea-pig septal neurons have clarified this peculiarity by demonstrating, for Na + channels, the presence of a fast inactivating component that recovers extremely slowly (Castellano and Ldpez-Barneo 1991). Second, in class A neurons and in a subpopulation of class B neurons, high-threshold Ca 2+ conductance participated in the generation of spikes when repetitive discharges were elicited by strong depolarizing current pulses. This was indicated in our study by the occurrence of smaller and slower spikes. Third, almost all neurons displayed an anomalous rectification with a prominent depression in the hyperpolarizing response, related to a time-dependent anomalous rectification. In our study, the conductance underlying the rectification was blocked by extracellular Cs + but not by Ba 2+, suggesting involvement of K+/Na § currents (Halliwel and Adams 1982). Fourth, a few neurons in the two classes presented a transient outward rectification. In our experiments, 4-AP did not alter this rectification, an observation reported for other preparations (Rudy 1988). For class A and class B neurons, during spike generation in response to strong depolarizing current pulses, involvement of C a 2 + currents combined with inactivation of Na + currents led to the appearance of smaller and slower spikes which we call Na+/Ca 2+ spikes. Generation of Na+/Ca 2+ spikes involves at least two components. First, Na + inactivation resulted in a decrease in the rate of rise of spikes and an increase in their duration. On physiological grounds, it can be speculated, as suggested by L6pez-Barneo et al. (1990), that these effects favour an increasingly maintained Ca 2§ entry for successive spikes. Second, as demonstrated by our findings, Na+/Ca 2+ spikes were followed by Ca2+-dependent AHPs. AHPs might serve to de-inactivate Na + conductance, allowing Na + channels to participate in the next spike (Barrett and Barrett 1976; Galarraga et al. 1989). In our experiments, variable degrees of interaction between participation of Ca 2+ currents and Na + inactiva-

tion were indicated by the different patterns of discharges observed in response to depolarizing current pulses. Differences in patterns of discharges were most striking in class A neurons, for which pure Ca 2+ spikes appeared. In their detailed study of the dorsal part of the guineapig LS, Alvarez de Toledo and L6pez-Barneo (1988) and L6pez-Barneo et al. (1985) recorded only one type of neurons. These neurons had electrical properties similar to our class A neurons. The few neurons recorded by these authors in the guinea-pig LSml exhibited the same characteristics. In the present report, we demonstrated that the prominent group of recorded LSml neurons belongs to class A. Thus, it is suggested that the majority of the guinea-pig LS neurons display similar electrical properties regardless of their localization throughout the anatomical subdivisions of the region. Nevertheless, our recordings in the LSml are typified by the presence of class B neurons, which were not observed by Alvarez de Toledo and L6pez-Barneo (1988) and L6pez-Barneo et al. (1985). Moreover, distinct electrical properties ha~ee not been reported by the above-mentioned authors, namely, occurrence of a transient outward rectification and of inferior olive-type discharges (Llin~s and Yarom 1981). It must be noted that, in the dorsolateral part of the rat LS, intrinsic properties of neurons seem to be different: the absence of a high threshold Ca 2+ current in the control medium, no Na + inactivation, no anomalous rectification, but the presence of a low-threshold Ca 2+ current (Stevens et al. 1984). A third class of neurons, class C neurons, corresponds to 20.1% of the total population of recorded LSml neurons. This finding extends to the LSml the observation of L6pez-Barneo et al. (1985) who obtained, in the dorsal part of the guinea-pig LS, recordings which greatly resemble those of our class C neurons (see their Fig. 2C, D). This type of recording has not been reported for the rat LS. Class C neurons displayed two types of spikes: (1) low-threshold, short duration, small amplitude Na +dependent spikes and (2) high-threshold, long duration and large amplitude Ca2+-dependent spikes. In our study, several lines of evidence would argue against th_e supposition that recordings displaying small amplitude Na + spikes are due to cell injury. High and stable resting membrane potentials and well-developed depressions in the electrotonic response (Mostfeldt-Laursen and Rekling 1989) of class C neurons, associated with no trace of HRP leakage around the injected cells, were indications of good impalements. Occurrence of small Na + spikes in apparently healthy neurons then raises the problem of the locus of recording. One hypothesis is that class C neurons recordings were obtained at the dendritic level. Two electrophysiological observations support this hypothesis. First, small Na + spikes were followed b y Ca 2+ spikes which showed multiple componentsdendrites offer multiple sites for the generation of u n i t a r y Ca 2+ events, which add up to produce large C a 2 + spikes (Llinfis and Nicholson 1971). Second, for a third of neurons, Ca2+-dependent slow depolarizing potentials were recorded, reminiscent of plateaus observed during intradendritic recordings (Llinfis and Sugimori 1980; Masukawa and Prince !984). For the hypothesis of in-

113 tradendritic recordings, the small N a + spikes recorded in these neurons should correspond to the electrotonic reflexion of somatic spikes arising at a distance f r o m the recording site. An argument against the possibility that our recordings o f class C neurons were performed at the dendritic level is their relatively high proportion. In the LSml, a laminar organization does not exist (Alonso and Frotscher 1989), making it difficult to obtain numerous and stable intra dendritic impalements. Nevertheless, all the HRP-labeled class C neurons exhibited p r i m a r y dendrites of large diameter which could give a stalked aspect to the soma. These large parts of the dendrites m a y represent preferential sites for dendritic penetrations. Another hypothesis is that impaling of class C neurons t o o k place at the junction between the soma and the primary dendrite. This was the suggestion of Castellano et al. (1991) to explain the recordings obtained in the dorsal p a r t o f the LS by Ldpez-Barneo et al. (1985). Indeed, there are focuses of high concentration of Ca 2 + channels at the bases of dendrites in h i p p o c a m p a l neurons (Westenbroek et al. 1990). Again, the thickening of the dendritic base in class C neurons m a y facilitate such impalements. Interestingly, our findings indicate that the m o r p h o l o g y of class C neurons is similar to that for the majority (15/21) o f class A neurons. This might suggest that class A and class C neurons correspond to the same type of neuron, and that these neurons were recorded either at the somatic level (class A) or at the dendritic or dendritic/somatic level (class C). As a third hypothesis, if we consider that class C neuron recordings were obtained f r o m the soma, the relative amplitude of the N a + and Ca 2+ spikes m a y reflect a special localization of voltage-dependent N a § conductances and high-threshold Ca 2+ conductances at the somatic level. This implies that class A and class C neurons, which exhibit similar morphologies, nevertheless have different electrical properties. In summary, data f r o m this study, focused on the guinea-pig LSml, provide evidence for the existence of three electrophysiologically distinct populations of neurons. Neurons of classes A and C represent the large majority and exhibit similar morphologies. In their electrophysiological profiles, they resemble neurons of the dorsal part of the guinea-pig LS. Class B neurons have electrophysiological and morphological features distinct f r o m those of class A and class C neurons, and have not been observed in the dorsal part o f the guinea-pig LS. It is now i m p o r t a n t to characterize the p h a r m a c o l o g y o f LSml neurons in light of the potential importance of this LS subnucleus.

Acknowledgements. We wish to thank Claude Bel for her excellent technical assistance. This work was supported by the Fondation Pour La Recherche M6dicale Frangaise and the University of Lille II.

References Alonso JR, Frotscher M (1989) Organization of the septal region in the rat brain: a Golgi/EM study of lateral septal neurons. J Comp Neurol 286:472-487

Alvarez de Toledo G, Ldpez-Barneo J (1988) Ionic basis of the differential neuronal activity of guinea-pig septal nucleus studied in vitro. J Physiol (Lond) 396:399-415 Barrett EF, Barrett JN (1976) Separation of two voltage-sensitive potassium currents and demonstration of a tetrodotoxin-resistant calcium current in frog motoneurons. J Physiol (Lond) 255: 737-774 Castellano A, Ldpez-Barneo J (1991) Sodium and calcium currents in dispersed mammalian septal neurons. J Gen Physiol 97 : 303-320 De France JF (1976) A functional analysis of the septal nuclei. In: De France JF (ed) The septal nuclei. Advances in behavioral biology, vol 20. Plenum Press, New York London, pp 185-227 Galarraga E, Bargas J, Sierra A, Aceves J (1989) The role of calcium in the repetitive firing of neostriatal neurons. Exp Brain Res 75:157-168 Halliwell JV, Adams PR (1982) Voltage-clamp analysis of muscarinic excitation in hippocampal neurons. Brain Res 250:71-92 Jakab RL, Leranth C (1990a) Catecholaminergic, GABAergic and hippocamposeptat innervation of GABAergic neurons in the rat lateral septal area. J Comp Neurol 302:305-321 Jakab RL, Leranth C (1990b) Somatospiny neurons in the rat lateral septal area are synaptic targets of hippocamposeptal fibers: a combined EM/Golgi and degeneration study. Synapse 6:10-22 Jakab RL, Naftolin F, Leranth C (1991) Convergent vasopressinergic and hippocampal input onto somatospiny neurons of the rat lateral septal area. Neuroscience 40: 413-421 Leranth C, Frotscher M (1989) Organization of the septal region in the rat brain: cholinergic-GABAergic interconnections and the termination of hippocampo-septal fibers. J Comp Neurol 289 : 304-314 Llimls R, Nicholson C (1971) Electrophysiological properties of dendrites and somata in alligator Purkinje cells. J Neurophysiol 34 : 534-551 Llinfis R, Sugimori M (1980) Electrophysiological properties of in vitro Purkinje cells dendrites in mammalian cerebellar slices. J Physiol (Lond) 305:197-213 Llin/ts R, Yarom Y (1981) Electrophysiology of mammalian inferior olivary neurones in vitro. Different types of voltagedependent ionic conductances. J Physiol (Lond) 315:549-567 Ldpez-Barneo J, Alvarez de Toledo G, Yarom Y (1985) Electrophysiological properties of guinea-pig septal neurons in vitro. Brain Res 347 : 358-362 Ldpez-Barneo J, Castellano A, Toledo-Aral J (1990) Thyrotropinreleasing-hormone (TRH) and its physiological metabolite TRH-OH inhibit Na § channel activity in mammalian septal neurons. Proc Natl Acad Sci USA 87:8150-8154 Masukawa LM, Prince DA (1984) Synaptic control of excitability in isolated dendrites of hippocampal neurons. J Neurosci 4:217-227 Meibach RC, Siegel A (1977) Efferent connections of the hippocampal formation in the rat. Brain Res 124:197-224 Metz CB, Schneider SP, Fyffe REW (1989) Selective suppression of endogenous peroxidase activity: application for enhancing appearance of HRP-labeled neurons in vitro. J Neurosci Meth 26:181-188 Mosfeldt-Laursen A, Rekling JC (1989) Electrophysiological properties of hypoglossal motoneurons of guinea-pigs studied in vitro. Neuroscience 30:619-637 Poulain P, Martin-Bouyer L, Beauvillain JC, Tramu G (1984) Study of the efferent connections of the enkephalinergic magnocellular dorsal nucleus in the guinea-pig hypothalamus using lesions, retrograde tracing and immunohistochemistry: evidence for a projection to the lateral septum. Neuroscience 11:331-343 Raisman G (1966) The connexions of the septum. Brain 89:317-348 Rudy B (1988) Diversity and ubiquity of K channels. Neuroscience 25 : 729-749

114

Sakanaka M, Magari S (1989) Reassessment of enkephalin (ENK)containing afferents to the rat lateral septum with reference to the fine structures of septal ENK fibers. Brain Res 479 : 205-216 Schwartzkroin PA (1975) Characteristics of CA1 neurons recorded intracellularly in the hippocampal in vitro slice preparation. Brain Res 85: 423-436 Staiger JF, Niirnberger F (1989) Pattern of afferents to the lateral septum in the guinea-pig. Cell Tissue Res 257:471-490 Staiger JF, Niirnberger F (1991a) The efferent connections of the lateral septal nucleus in the guinea-pig: projections to the diencephalon and brainstem. Cell Tissue Res 264:391-413 Staiger JF, Niirnberger F (1991b) The afferent connections of the lateral septal nucleus in the guinea-pig: intrinsic connectivity of the septum and projections to other telencephalic areas. Cell Tissue Res 264:415-426

Staiger JF, Wouterlood FG (1990) Efferent projections from the lateral septal nucleus to the anterior hypothalamus in the rat: a study combining Phaseolus vuloaris-leucoagglutinin tracing with vasopressin immunocytochemistry. Cell Tissue Res 261 : 17-23 Stevens DR, Gallagher JP, Shinnick-Gallagher P (1984) Intracellular recordings from rat dorsolateral septal neurons, in vitro. Brain Res 305:353-356 Swanson LW, Cowan WM (1977) An autoradiographic study of the organization of the efferent connections of the hippocampal formation in the rat. J Comp Neurol 172:49-84 Swanson LW, Cowan WM (1979) The connections of the septal region in the rat. J Comp Neurol 186:621-656 Westenbroek RE, Ahlijanian MK, Catterall WA (1990) Clustering of L-type Ca/+ channels at the base of major dendrites in hippocampal pyramidal neurones. Nature 347: 281-284