Behavioural Brahl Research, 49 (1992) 205-212 9 1992 Elsevier Science Publishers B.V. All rights reserved. 0166-4328/92/$05.00

205

BBR01306

Concurrent measurement of serotonin metabolism and single neuron activity changes in the lateral hypothalamus of freely behaving rat Kazuhiko Aoyagi, Yutaka Oomura I and Nobuaki Shimizu Department of Physiology, Faculty of Medicine, K)~tsh,t University, Fukuoka (Japal0 (Received 3 January 1991) (Revised version received 20 April 1992) (Accepted 22 April 1992)

Key ,tvrds: Serotonin; Feeding behavior; Lateral hypothalamus; In vivo voltammetry; Single neuron activity; Chronic recording; Rat

To further investigate the activity of serotonin neurons in relation to feeding behavior, the metabolic activity of the serotonergic system and single neuron activity changes in the lateral hypothalamic area (LHA) were investigated concurrently in freely behaving rats. The extracellular concentration of 5-hydroxyindoleacetie acid (5-ttlAA), a metabolic product of serotonin in the LHA, began to increase concomitantly with the early stage of nocturnal eating. The increased 5-HIAA returned to the basal level within 3 or 4 h. In conjunction with the increase in serotonin metabolism, activity of 12 out of 30 LHA neurons (40~o) increased, whereas it decreased in 7 (23~), and in 11 (37~o) it showed no change. An intracerebroventricular injection of lisufide suppressed the increased activity in 7 of the 12 neurons, but had no effect on the others. These results suggest that the concurrent increase in serotonin metabolism and neuron activity changes in the LHA may occur in the early portion of the nocturnal eating period, and may be important in controlling feeding behavior.

INTRODUCTION

A substantial number of behavioral studies suggest that serotonergic neurons in the central nervous system have a suppressive effect in the regulation of food intake. Procedures believed to activate serotonin receptors reduce food consumption, and decreased serotonin receptor activity brings about the opposite effect3"9'23. Recent data indicate that fenfluramine-induced anorexia is mediated through the serotonin-i receptor subtype, which is consistent with other evidence showing that the serotonin-1 receptor mechanism may be involved in the control of feeding behavior7'13'17. Further, 8-OHDPAT, a potent and selective serotonin-iA receptor agonist, induced feeding in satiated rats by decreasing serotonin synthesis and release via activation of serotonin autoreceptors 1'8. Considering the central site(s) and mechanism of se-

Present address: Department of Central Nervous Function Control Systems, Research Institute for Wakan-Yaku, Toyama Medical and Pharmaceutical University, Toyama 930-01, Institute of Bio-Active Science, Nippon Zoki Pharmaceutical Co. LTD, Yashiro 673-14, Japan. Correspondence: N. Shimizu, Department of Physiology, Faculty of Medicine, Kyushu University 60, Fukuoka 812, Japan.

rotonin's action, the hypothalamus is among the structures most probably involved in the suppressive effects. Direct injection of fenfluramine into regions of the hypothalamus, such as the ventromedial hypothalamic nucleus (VMH) and the paraventricular nucleus (PVN), have been investigated, and the anorectie effect was more potent in the PVN than in the VMH ~6'3~ In contrast to these hypothalamic nuclei, the suppressive effect of serotonin in the lateral hypothalamic area (LHA) is unclear. Contributions of the LHA to feeding behavior have been well established 19'2~ and neurons in the LHA, which have been characterized as chemosensitive, are postulated to be important in the regulation of food intake. It has been reported that about 75~o of LHA chemosensitive neurons receive inhibitory serotonin inputs from the dorsal raphe through serotonin1 receptors ll. In addition, during chronic recording of neuronal activity in the LHA, a large proportion of neurons were inhibited by acute immobilization, and this inhibition was significantly antagonized by preinjection of the serotonin receptor antagonist, methysergide27. Thus, behavioral and electrophysiological studies suggest that the lateral hypgthalamic serotonergie system is also important in the regulation of feeding behavior. To further investigate correlations between the cir-

206 cadian fluctuation of serotonin metabolism and LHA neuron activity changes in freely behaving conditions, we measured the lateral hypothalamic serotonin metabolite, 5-hydroxyindoleacetic acid (5-HIAA), by invivo voltammetry4"1~ A concurrent chronic recording of single neuronal activity was carried out in the same area of the hypothalamus to examine the correlation between Jocal serotonin metabolism and neuronal activity changes. MATERIALS AND METHODS

Subjects and mahttenance In this experiment, 40 male Wistar rats, weighing 250-300 g at the beginning of the experiment, were used. Before the experiment, they were individually housed at 23 + 1 ~ in a 12 h/12 h light-dark schedule (light on 08.00 h-20.00 h). They had ad libitum access to conventional rat chow (Japan Clea, CE2) and tap water. For 1 week preceding surgery, they were handled for several minutes daily to habituate them to handling. In some experiments, the rats were fed a tryptophandeficient diet (a gift from Dr. K. Torii, Central Res. Lab. Ajinom0to Co.) for 2 days before testing. The composition of the tryptophan deficient diet is shown in Table I. Microprobe assembly for #l vivo voltammetry measuremerit and chronic ttnit recording The working electrode for in vivo voltammetry measurements was prepared as follows: 2-3 carbon fibers (Torayca, Type M-40, outer diameter (o.d.)= 7 Fm) were passed through a glass capillary tube (o.d. = I00 Fm, length = 12 mm) and extended 300 Fm from the tip of the capillary. The tip of the capillary was sealed by applying a drop of low viscosity resin and the fibers were sealed into the capillary. Electrical connection between the carbon fibers and an electric wire was made using electrically conducting silver paint. After drying the resin, the electrode was electrochemically treated with a triangular wave (0-3 V vs. Pt, 10 Hz for 30 s) in 0.1 M H 2 S O 4 solution 26. As illustrated in Fig. 1A, one carbon fiber working electrode and 5 Teflon-coated Pt-Ir wires (Medwire, New York, o.d. = 25 Fm) for chronic recording of hypothalamic neuronal activity were supported together in a glass capillary (o.d. = 300 Fro). The Pt-Ir wires were also extended 300 llm from the capillary and exposed at the cut end.

Surgery The microprobe for concurrent measurement of 5HIAA levels and unit activity was implanted into the

lateral hypothalamic area (LHA; A = 4 . 6 , L = 1.5, H = - 2.7) according to the atlas of Ktinig and Klippel TM under pentobarbital anesthesia (50 mg/kg, i.p.). The Ag/AgCI reference electrode and auxiliary electrode were placed on the dura surface contralateral to the microprobe. After positioning the microprobe, the wires were connected to the terminal of a socket, which was then fixed to the surface of the skull with dental acrylic resin. A stainless-steel tube (29-gauge) was also implanted into the third ventricle at an angle of 30 ~ anterior to the vertical plane to avoid interference with insertion of the microprobe. The coordinates of the cannula were A, 6.6; L, 0.0; H, -2.0. To prevent leakage ofcerebrospinal fluid, a 100 Fm diameter stainless-steel wire was inserted into the cannula.

11l vivo voltammetry After 3 days of recovery from the surgery, the experiments were conducted within the test chamber placed in a sound proof box. Differential pulse voltammetry (BAS, DPV-5) was used for in vivo voltammetry measurements. The following parameters were used: potential range -0.2 V to 0.5 V; scan rate 25 mV/s; modulation amplitude 25 mV; pulse frequency 10 Hz. Chronic unit recording The activity of single neurons in the L H A was recorded with concurrent measurement of 5-HIAA changes. A dual-channel, field-effect transistor (2SK18A, Toshiba) was mounted directly on a plug that fitted into the socket on the animal's head. To eliminate artifacts, the input to the FET was derived from two of the five electrodes, one for recording single neural activity and the other to serve as an indifferent electrode. The incidence of feeding behavior was also monitored. An infrared-light-emitting diode (LED) and matched phototransistor were mounted on the wall of the feeding box, and breaking of the light beam by the animal's head during eating provided a pulse which was led to a pen-recorder. huracerebroventricular (i.c. v.) injection A 29-gauge stainless steel canula, which was previously implanted into the third-ventricle~ was connected through a long polyethylene tube (PE10) to a 50F1 Hamilton microsyringe, and the assembly was filled with freshly prepared test solution. The tubes were connected 30 min before starting an experiment. The infusion tubes were long enough to permit manipulation of the syringe outside of the test chamber, and minimize disturbance of the animal. To minimize intraventricular

207 pressure changes, 10 td of lisuride hydrogen maleate (Schering; 0.22 nmol]l), or artificial cerebrospinal fluid was injected through the cannula at a rate of 1 id/min. Histology and statistics After completing all experiments, the brain tissue was electrically microcoagulated under pentobarbital anesthesia for identification of the recording sites. A negative direct current (10 llA, 10 s) was passed through the Pt-Ir recording electrode. The rats were then perfused intracardially with.neutralized 10~ formalin. Each brain was cut into 100 Itm serial sections on a freezing microtome and counterstained with neutral red, for identification of the recording sites. All data were expressed as mean + S.E.M. A statistical analysis was done by one-way analysis of variance (one-way ANOVA) followed by Student's t-test, and a P-value of less than 0.05 was considered to indicate significance. Fisher's exact probability test was also used, and is indicated where appropriate.

TABLE I Composition of the tr)ptophan deficientdiet Contents

Asp Asn Thr Ser GIu Gin Pro Gly Ala (Cys)2 Val Met lie Leu Phe His Base Arg Base Tyr Lys Starch Cellulose Minerals Vitamins Choline CI Corn oil Impurities Total

%

0.62 0.62 0.60 0.96 0.79 0.95 0.53 0.41 0.75 0.15 0.82 0.36 0.71 1.20 0.70 0.30 0.87 0.50 1.01 71.34 4.00 4.00 1.01 0.02 5.00 1.78 100.00

RESULTS A typical differential pulse voltammogram in a solution of ascorbic acid (AA, 10/tM), 3,4-dihydroxyphenylacetic acid (DOPAC, 1 tiM) and 5-hydroxyindoleacetic acid (5-HIAA, 1 ttM)is shown in Fig. 1B(a). The carbon fiber working electrode produced three distinct peaks which corresponded to AA (P1), DOPAC (P2) and 5-HIAA (P3). In contrast, the voltammogram obtained in the LHA displayed only one peak, which appeared at an oxidation potential of about +190 mV (Fig. IB(b)). This peak has been further confirmed, by pharmacological manipulation 26, to reflect the extracellular level of 5-HIAA. Fig. 2 shows the circadian change in the extracellular levels of 5-HIAA in the LHA. The peak area of the voltammogram which is proportional to the concentration of 5-HIAA around the working electrode was calculated by a chromatogram data analyzer (Chromatocorder-1 I, SIC). The changes in the level of 5-HIAA were expressed as percent of the peak area in the voltammogram (P3) measured immediately before the beginning of the dark period. Measurements of the in vivo voltammog~am were made every 10 min, and the mean values + S.E.M. for 5 rats were plotted at 60 min intervals for clarity. As shown in Fig. 2A, a significant increase in the extracellular level of 5-HIAA occurred almost in synchronism with the ordinary feeding periods of the rat. The 5-HIAA concentration began to increase about 1 or 2 h prior to the dark period when the animals started ordinary feeding. About 2 h after

A

B

(a) P3

chemical recording

(b) glass pipette !i (~3oo==)~

t

!'I epoxyresin

impulse recording

\

pt wire

(r

,---'~;~1 r

/ IVV carbonfiber electrode

I. ' ' ' 0 ' 2 l ' I -02 0 . 0.4 Potential(V vs. Ag/AgCI)

Fig. 1. A: recording microprobe for in vivo voltammetry and chronic

single neuronal activity.Carbon fiber working electrode for in vivo voltammetryand fiveTeflon-coatedPt-lr wires are installed in one glass pipette (o.d. 300 ~um)with resin. Each Pt-lr wire is connected to one terminal of a ten-pin IC socket. A dual-channel,field-effect transistor (FET) was mounted directly on a plug that fit into the socket on the animal's head. B: typicaldifferentialpulse voltammograms of DOPAC (P2) and 5-tlIAA (P3), measuredin vitro (a) and in the LttA (b).

208 the initiation of feeding, the 5-HIAA level reached a maximum at 2.7 times that measured at 18.00 h. Similar changes in the 5-HIAA levels during the early dark period were observed even when the rats were fed a tryptophan deficient diet for 2 days before the experiment. As shown in Fig. 2B, the mean extracellular levels of 5-HIAA, obtained from the 5 rats, began to increase at about 17.00 h, and a maximum was observed at 20.40 h. The peak differences in the 5-HIAA level were statistically significant ( P < 0.01, Student's t-test). These results suggest that the circadian fluctuations of 5-HIAA might not be a result of increased serum tryptophan due to the consumption of a diet containing tryptophan. Correlation between serotonin metabolism and neuronal activity changes in 30 neurons of freely moving

]

,oo A

9

Standard Diet

5o

o-

.L

1'2 1~

1'6 &

2'o 2'2

b

~

~

~]

;

/

~ lb~2

rats was further studied by in vivo voltammetry and chronic unit recording. A typical pattern of the circadian fluctuation of extracellular 5-HIAA, neuronal activity changes and feeding episodes is shown in Fig. 3. The extracellular concentration of 5-HIAA began to increase at around 19.00 h and reached a maximum level at 240% of the basal level at 20.20 h. The neuronal activity began to increase in approximately the same fashion. The mean firing rate, 3.3 + 1.4 (impulses/s) at 19.00 h, increased to 11.0 + 2.8 at 20.40 h and continued this increase until about 23.00 h. Feeding behavior started at around 19.20 h. When artificial cerebrospinal fluid was injected into the cerebral third ventricle no change in the firing rate occurred. In Fig. 4, the extracellular concentration of 5-HIAA began to increase at around 19.20 h and reached maximum level at 240% of the basal level at about 21.00 h. The mean firing rate, 4 . 2 + 1.9 measured at 19.20 h began to increase at 19.50 h and reached a high level of 9.8 + 1.3 at 20.30 h. Feeding behavior began about 20 min before the increase of 5-HIAA began. Increased neuronal activity was observed 30 min after 5-HIAA elevation. An intrathird ventricular injection of lisuride at around 20.50 h suppr'essed neuronal activity with a latency of 7 min, and the mean firing rate measured at 20 min after the injection was 4.2 + 2.4. The suppression of the firing rate continued for about 30 min, and then returned almost tO the pre-injection level of 14.5 + 2.9. Fig. 5 shows a typical pattern of another type of neuron; one which decreased its firing rate after the

Time of Day

200

B

200" 160"

Tryptophan Deficient Diet

9