Effects of intracisternal vs. intrahypothalamic 5,7-DHT on feeding elicited by hypothalamic infusion of NE

310

Brain Research, 597 (1992) 310-320 © 1992 Elsevier Science Publishers B.V. All rights reserved 0006.8993/92/$05.00

BRES 18311

Effects of intracisternal vs. intrahypothalamic 5,7-DHT on feeding elicited by hypothalamic infusion of NE D o n a l d V. C o s c i n a a.b.c a n d E l i z a b e t h C.H. de R o o y

c

Section of Biopsychology, Clarke Institute of Psychiatry and Departments of b Psychiatry and c Psychology, University of Toronto, Toronto, Ont. (Canada) (Accepted 30 June 1992)

Key words: Noradrenaline; Serotonin; Hypothalamus; Feeding; 5,7-Dihydroxytryptamine;Paraventricular nucleus; 8-OH-DPAT

A variety of evidence has led to suggestions that brain serotonin (5-HT) and norepinephrine (NE) interact within the medial hypothalamus to control food intake. To test the possibility that chronic decrements in 5-HT might enhance NE-induced feeding, adult male rats were prepared with permanently indwelling cannulae aimed at the paraventricular nucleus (PVN), then received either intracisternal (IC) or PVN injections of the 5-HT neurotoxin, 5,7-dihydro~tryptamine (5,7-DHT) vs. its vehicle, !% ascorbic acid. Over a 4-week period, IC-5,7-DHT rats showed no signs of enhanced daily feeding or drinking. However, in 40-min intake tests, feeding but not drinking was enhanced by injecting 20 nmol NE into the PVN commencing 2 weeks after neurotoxin treatment. Terminal monoamine assays confirmed that IC-50-DHT produced large (80-90%) depletions of brain regional 5-HT. A functional index of 5-HT terminal damage was also implied by the impaired short-term feeding responses IC-5,7-DHT rats showed to the systemic administration of the 5-HTIA agonist, 8-hydroxy-2.(di.n.propylamino) tetralin (8-OH-DPAT) when tested between 3 and 4 weeks after IC treatment. Over a comparable 4-week period, PVN-5,7-DHT rats also showed no tendencies to overeat or overdrink on a daily basis. However, in contrast to IC-5,7-DHT rats, they also showed no differences in their feeding or drinking responses to NE injections into the PVN. This was so despite reliable depletions of 5-HT in the hypothalamus (-28%) and hippocampus (-71%). These results support earlier work showing that neither widespread nor localized hypothalamic damage to brain 5-HT neurons produce chronic overeating. However, the data suggest that phasic enhancements of PVN NE activity may trigger enhanced feeding when there is widespread damage to brain 5-HT neurons, although the PVN does not appear to be the brain site mediating this effect.

INTRODUCTION Research efforts to define the brain neurochemicai systems which regulate food intake have identified a vast number of candidates that appear to participate in this behavior 54,6z7~. From among this multitude, brain norepinephrine (NE) and serotonin (5-hydroxytryptamine or 5-HT) systems are two for which there is clear evidence of important interactive relationships. In general, it has been proposed that endogenous NE and 5-HT within the medial hypothalamus operate in an antagonistic fashion wherein NE stimulates3s,Sa,59 and 5-HT inhibitss.9's6.57feeding. One line of evidence which supports this proposition derives from observations that the consumption of food elicited by deprivation or in otherwise satiated rats by injecting NE into the paraventricular nucleus (PVN) is attenuated by central or

peripheral injections of direct or indirect 5-HT agonistsS,9,s7,80. In an attempt to more fully characterize the functional boundaries within which brain 5-HT systems may interact with hypothalamic NE ones in controlling food intake, a variety of studies have recently been conducted in this laboratory. Specifically, we have been attempting to determine if treatments that impede 5-HT neurotransmission can enhance NE-induced feeding. As one test of that possibility, we recently reported 24 that systemic injections of the 5-HTtA agonist, 8-hydroxy-2-(di-n-propylamino) tetralin (8-OHDPAT), did not modify the amount of food consumed by ad libitum-fed rats stimulated to eat acutely by PVN injections of NE. Since previous work has shown that systemic 8-OH-DPAT can induce feeding in its own right 6'26'27'4s'81, and a primary effect of acute 8-OH-

Correspondence: D.V. Coscina, Section of Biopsychology, Clarke Institute of Psychiatry, 250 College Street, Toronto, Ont., M5T 1R8, Canada. Fax: (1) (416) 979-7871/979-2243.

311 DPAT is to reduce the release and tun.over of endogenous brain 5-HT 38'4°'41'46'7°'7t, our finding implied that diminishing endogenous 5-HT release does not enhance the capacity of NE applied to the PVN to elicit feeding. However, since systemic 8-OH-DPAT produces a variety of physiological and endocrine effects in the same dose range that stimulates food intake (see ref. 24 for discussion), it is possible that this agent and/or its route of administration did not provide a selective enough means by which to probe the effects of impeding brain 5-HT function on NE-induced feeding. Therefore, the present experiments were conducted to determine if chronic depletion of brain 5-HT with 5,7-dihydroxytryptamine (5,7-DHT) might enhance food intake stimulated by PVN injections of NE. This was accomplished in two ways: (1) through intracisternal (IC) injection of 5,7-DHT, in order to effect global depletion of b~-ain 5-HT, and (2) through localized injection of 5,7-DHT into the PVN. MATERIALs AND METHODS

Sub/ects Forty-one adult, male Sprague-Dawley rats (Charles River; Montreal, Que.) from an earlier study 24 were used. Animals had previously been implanted with unilateral 26-gauge guide cannulae targeted to terminate 1.0 mm dorsal to the PVN and tested for NE-induced feeding after pretreatment with systemic 8-OH-DPAT (see ref. 24 for details). At the time of this study, all rats were in excellent health, averaging about 450 g in body weight, and were still reliable feeding responders to NE infusion into the PVN (see below). As in the previous study, animals were housed in separate stainlesssteel cages within a temperature- (22+ 2°C) and humidity- (40-50% relative) controlled colony room illuminated 12 h daily beginning at 08.00 h. Throughout these investigations, rats had ad libitum access to fresh Purina Lab Chow pellets (4% fat by weight) on their cage floors as well as tap water.

Procedure General NE-induced feeding procedure. For both screening and test trials, rats were removed from tl~eir home cage during mid-light cycle and dummy cannulae were removed. While hand-held, a 33-gauge injector 1.0 mm longer than the outer guide cannula was inserted and 0.5 ~tl of control or test solution was infused over 10 s. Following each infusion, the injector was left in place for 30 s before removal, the dummy cannula reinserted, and rats returned to their home cages in which 3 pellets (about 15 g) of freshly weighed Lab Chow had been placed. The amount eaten over the next 40 rain was determined after corrections were made for spillage. Experiment 1: effect of IC 5, 7-DHT on NE-induced feeding. Seventeen rats were screened for eating in response to 20 nmol NE vs. saline over two separate 40-min tests conducted 2 days apart, counterbalancing across subjects for order of infusion. Since each rat responded to NE by eating at least two standard deviations more food than the mean amount eaten by the group after saline, all were included in this study. To render 5,7-DHT uptake more specific to 5-HT neurons ~, all rats were administered 25 mg/kg desmethylimipramine (DMI; Horpramine~ Lakeside Labs; in 2 ml/kg 0.9% saline) intraperitoneally 30-60 min prior to IC injection. Then, rats were lightly anesthetized with an inhalant (Metofane, Pitman-Moore, Miss., Ont.), and received a 20-/zl injection of 5,7-DHT (200 ttg free base; n = 10) or its vehicle, 1% ascorbic acid in deionized water (n = 7) IC using previously published methods ~4. To retard the development of seizures

which can occur after IC 5,7-DHT at this dosage 14, all rats received 15 mg/kg sodium pentobarbital (Somnotol) subcutaneously (SC) immediately after IC injections. Rats were then returned to their home cages to recover. To determine the time course effects of IC 5,7-DHT on animals' eating after PVN NE, all rats were tested for 40-rain food and water intakes following injection of 20 nmol NE vs. saline on days 3, 5, 8, 11, 14, and 21 after IC injections. On each test day, both treatments were given sequentially, half of each group receiving saline first, the other half receiving NE first. This order was alternated on successive test days. To provide a functional test of 5,7-DHT's ability to impede brain 5-HT function (see ref. 6 for similar strategy), all rats were tested for 40-min intakes after 250 ~g 8-OH-DPAT vs. saline SC on days 23 and 25 using a counterbalanced design for drug order across subjects within groups. A final 40-min intake test of 20 nmol NE vs. saline was conducted on day 28. In addition to these 40-min tests, daily intakes and body weights were recorded for the first 2 weeks of the study, then for 22-h periods following each 40-min feeding test. All short-term tests and daily maintenance procedures were conducted between 13.00 and 16.00 iL Experiment 2: effect of PVN 5, 7-DHT on NE.induced feeding. The remaining 24 rats were screened for 40-min feeding responses as described above and all found to be reliable eaters. However, in this group, 40 nmol NE was used rather than 20 nmol since previous work 24 with these animals revealed they were less consistent responders to the lower dose. The use of the higher dosage elicited statistically reliable enhancements of feeding comparable to that seen in the first group tested with 20 nmol. As in Expt. 1, all rats were pretreated SC with DMI 30-60 min prior to 5,7-DHT vs. ascorbate injection. Then, a 3-pA solution of 12 p~g 5,7-DHT (free base; n = 13) or its vehicle, 1% ascorbate in deionized water (n = 11), was infused into the PVN of awake animals over 60 s, with injectors being left in place for an additional 60 s. This dose and volume were chosen because previous work 79 had demonstrated their ability to reliably deplete hypothalamic 5-HT when injected at a slightly more lateral site. Since this previous work as well as our own observations indicated that no seizures developed following such infusions, animals did not receive systemic sodium pentobarbital after this treatment. To parallel the time course of feeding tests used in Expt. 1, all rats were assessed for their 40.min feeding and drinking responses to PVN NE as well as saline on days 3, 5, 8, 11, 14, 21 and 28 after PVN 5,7-DHT vs. ascorbate infusion, again counterbalancing for within-day treatments across days and within groups. In addition to 40-rain test intakes, daily intakes and body weights were measured for the first 3 weeks of this study in a fashion parallel to that described for Expt. 1.

Biochemical assays and histology One week after completing each experiment, rats were sacrificed by decapitation and their brains removed for dissection. For animals in Expt. I, the cortex, hippocampus and striatum were rapidly separated from each whole brain, frozen over dry ice, and stored at -70°C until high-performance liquid chromatographic (HPLC) assays of endogenous NE, dopamine (DA), 5-HT and 5-bydroxyindoleacetic acid (5-HIAA) were performed on each area using slight modifications of a previously published method 2s. The remainder of the brain was placed in 10% formalin for fixation so that the loci of hypothalamic injection sites could be determined after routine histological processing and staining of cell bodies with Cresyl violet. For animals in Expt. 2, the cortex, hippocampus and hypothalamus were removed and retained for HPLC assays of the same monoamines and metabolite described above.

Data and statistical analyses Because of temporarily occluded cannulae or injectors on any given test day, or the attrition of animals over time due to permanent loss of their implants or illness, the ns for which data could be analysed varied across time. However, ns for any particular test were never less than 6 per group in Expt. 1 and 8 per group in Expt. 2. Behavioral and biochemical data obtained were analyzed by t-tests,

312

Daily BWs and 22-hr Intakes

two-way or three.way analyses of variance (ANOVAs) as required, with repeated measures corrections used on ANOVA factors if appropriate (e.g. for time course analyses). Student's t-tests were also performed in cases where clarification of significant F values were desireable. All P levels reported represent two-tailed distributions. For ease of reading, F and t values have not been included in the text.

8 0 ~-

T

[ ~3-Controi 1

40

20 II

RESULTS

Cannula localization As described in our previous s t u d y 24, the placements of cannu!ae tips for rats in Expt. 1 were localized in the appropriate anterior/posterior as well as medial/lateral planes in terms of their proximity to the target site, the PVN. However, there was some variability in the dorsal/ventral plane whereby cannula tips terminated either in or just above the PVN. This variability was considered acceptable since previous work from this laboratory 24,65"67as well as others 34 has demonstrated tha~ exact positioning within the PVN is not critical in order to obtain reliable NE-induced feeding, and because we independently confirmed the efficacy of eliciting feeding at these injection sites both prior to and during this study. No histological data were available for rats in Expt. 2 since their hypothalami were retained for biochemical analyses. Itowever, since these rats were NE responders in a previous study 24 and here (see below), we felt they were valid subjects for the purposes of this work.

Experiment 1: effect of IC 5, 7-DHT on NE-induced feeding Fig. I depicts mean + S.E.M. body weights plus 22-h food and water intakes obtained from daily measures across the entire experimental period. At the time rats received IC injections, their body weights averaged about 430 g (422 + 6.2 g for controls, n = 8; 434 + 14.1 g for 5,7-DHT-treated rats, n = 9) and were not reliably different from one another. After these injections, the food and water intakes of both groups dropped to near zero levels the first day after injection, resulting in body weight declines of about 25 g. For the next 4 days, intakes of both groups showed steady increases so that by the end of the 5th post-injection day they had largely stabilized at normal levels where they remained essentially unchanged for the rest of study. As these data imply, separate two-way ANOVAs (Group × Time) on each intake variable revealed highly reliable ( P < 0.001) effects of time on each measure. However, no statistically significant differences were found between groups as main effects, although transient feeding and drinking suppressions produced significant interactions with time ( P < 0.001 and 0.03, respectively).

~ ~

.

I

,

I

,

I

,

I

~

I

,

I

30 25 20

5

500475

-

450 ~ 400 375 350 325 0

.

j . ~ i

' ' ' ~--~- ~ ~ ' 5 10 15 20 25 Days after IC Injection

30

Fig. 1. Mean + S.E.M. body weights (lower frame), 22-h food intakes (middle frame) and 22-hr water intakes (upper frame) across time after intracisternal (IC) 5,7-DHT (200 t~g) vs. control (ascorbate vehicle) injections. See inset for definition of group symbols.

The early, transient depressions of intake on the part of 5,7-DHT-treated rats, although not large or consistent enough separately to produce reliable main effects on either intake measure, were apparently sufficient when combined to produce small (i.e. 5-8%) yet reli. able suppressions of body weight. This conclusion derives from a two-way ANOVA (Group x Time) that revealed reliable main effects of time ( P < 0.001) and group ( P < 0 . 0 5 ) as well as an interaction between them ( P < 0.001). Fig. 2 summarizes the results across all 40-min NE intake tests. For ease of analysis, an intake score was devised for each subject per time point by subtracting the amount consumed after NE from that consumed after saline. A two-way ANOVA (Group x Time) on water intake scores revealed no significant effects. A separate two-way ANOVA on food intake scores also showed no significant main effects of either factor alone, but suggested a trend towards a significant interaction between them ( P < 0.06). Independent t-tests revealed significant differences between the feeding responses of 5,7-DHT rats and controls only on day 14 after IC injection (P < 0.05) with a trend (0.10 > P > 0.05) toward significance on day 21. This same pattern

313

40 Min Intake Scores

40 Min Intake Scores

(NE Response minus Saline Response)

2

(Mean NE Response minus Mean Saline Response)

*1

4

~ntrol I

_

[ • Control m 5,7-DHT

1

i

-2

,

l

,

l

,

l

t

l

,

l

l

,

0

l

t

5 10 15 20 25 Days after IC Injection

~2

I

Pro-Day 14

30

Fig. 2..Mean+S.E.M. food intake scores (lower frame) and water intake scores (upper frame) following 20 nmol NE tests at different times after intracisternal (!(2) 5,7-DHT vs. control (ascorbate vehicle) injections described in Fig. 1. See inset for definition of group symbols.

of results was confirmed by analyzing the absolute intake scores to saline vs. NE over time (not shown) in a three-way ANOVA (Group × PVN injection type × Time). That analysis revealed a main effect of PVN injection type (NE > saline; P < 0.001), no effect of group or time factors alone, but a reliable (P < 0.02) triple interaction among all three factors. To help clarify the meaning of these results, we considered that both groups showed improved daily feeding responses after IC injections which coincided with reattaining their pre-injection body weights around day 14 (see Fig. 1). Since the previous two-way ANOVA for food intake scores (see Fig. 2) assigned equal statistical weight to this repeated measure across all seven time points studied, this meant that the four feeding tests preceding day 14 contributed somewhat more to the time factor than did the three subsequent feeding tests. Therefore, in an attempt to reduce within-subject variability across days, we re-analyzed these data (Fig. 2) after averaging each rat's food intake scores for days 3, 5, 8 and 11 (i.e. pro-day 14) vs. their scores for days 14, 21 and 28 (i.e. post-day 14). These data are summarized in Fig. 3. A two-way ANOVA (Group × Time) on these data revealed no significant main effects of treatment or time. However, a significant (P < 0.02) interaction between factors was found.

Post-Day 14

Days after IC Injection Fig. 3. Mean+S.E.M. food intake scores from Fig. 2 re-plotted as avera~,e test responses before vs. after day 14 following intraeisternal treatwrients. See inset for definition of group symbols. Asterick indicates significant difference from control response.

Table I summarizes the results of the 8-OH-DPAT intake tests conducted on days 23 and 25. A two-way ANOVA (Group x Drug tested) on 40-min feeding responses revealed a significant (P < 0.005) main effect of group (IC 5,7-DHT < IC Control) and a significant (P <0.001) main effect of drug tested (SC 8-OHDPAT > SC Saline), but no interaction between these factors. A separate two-way ANOVA on 40-rain drinking responses (data not shown) revealed no group main effect, but a significant (P < 0.01) main effect of drug which paralleled the feeding responses (i.e. SC 8-OHDPAT > SC Saline). Results of an additional two-way ANOVA on the 22-h feeding which occurred subsequent to these 40-rain feeding tests (see Table I and Fig. 1) revealed opposite main effects of group (IC 5,7-DHT > IC Control; P < 0.001) and drug (SC 8OH-DPAT < SC Saline; P < 0.02). A separate two-way

TABLE I Mean + S.E.M 40 min vs. 22-h food intakes of rats screened for 8.0H-DPAT days 23-25 after receiving ascorbate vehicle (Control) or 5, 7-dihydroxytryptaminc (5, 7-DHT) injections intracisternally Group sizes: Control = 6; 5,7-DHT = 8.

Group Control 5,7-DHT

40-rain Saline

8-OH-DPAT

22-h Saline

8.0H-DPAT

1.87 + 0.47 0.97+0.39

4.67 + 0.97 2.18+0.45

26.2 + 0.25 31.3+1.50

24.9 + 0.44 28.1+0.87

314 ANOVA on 22-h water intake during that time (data not shown) revealed no significant effects. As implied from the opposite direction of these findings, when the data from both the 40-rain tests were combined with the 22-h intakes that followed them, separate two-way ANOVAs on total food intake or total water intake showed no significant main effects or interactions. Table II summarizes the results of the biochemical assays performed on the brain regions from these animals. Independent t-tests confirmed there were highly reliable (P < 0.005 at least) depletions of both 5-HT and 5-HIAA in all three areas approximating 90%. However, no reliable depletions of NE or DA were found.

Experiment 2: effect of PVN 5, 7-DHT on NE-induced feeding Fig. 4 shows the dat'ly measures of intake and body weight for PVN-treated rats. Like the IC-treated rats (see Fig. 1), both groups showed almost total suppression of food and water intake the first post-injection day, resulting in a similar decline of about 25 g in total body weight. However, intake patterns of these animals recovered much faster so that by the third post-injection day both intakes as well as body weight gains had stabilized at pre-injection levels. For this reason intakes and weights were followed for only three weeks. As these data imply, separate two-way ANOVAs performed on both forms of intake and for body weight confirmed reliable ( P < 0.01) time effects for each variable, but no reliable group differences or interactions between group and time. Fig. 5 shows the results across time for the 40-rain intake tests. As in Expt. 1, intake scores were calcu-

Daily BWs and 22-hr Intakes 80

1

60

_~- 5,7_-DHTI

40

20 i

i

i

35 re~

~

25 20 15 10 5 0 525 500 475 1 450 425

350

I

0

,

i

I

I

I

I

I

I

I

5

10

15

20

25

3O

Days after PVN Injection Fig. 4. Mean + S.E.M. body weights (lower frame), 22-h food intakes (middle frame) and 22-h water intakes (upper frame) across time after injecting 5,7-DHT (12/zg) vs. control (ascorbate vehicle) solutions into the paraventricular nucleus (PVN) of the hypothalamus. See inset for definition of group symbols.

40 Min Intake Scores (HE ResponseminusSalineResponse)

=e= Control [ --a=5,7-Dl-rri

1

TABLE II Mean + S.E.M monoamine or metabolite levels in brain regions from rats receiving ascorbate vehicle (Control) or 5,7-dihydroxytryptamine (5, 7-DHT) injections intracisternally Group sizes: Control = 7; 5,7-DHT -- 10. 5-HT ° Cortex Control 5,7-DHT Hippocampus Control 5,7-DHT Striatum Control 5,7-DHT

537±27 34± 4 * * 400~ 41 <25"

1

-2

I

i

i

l

l

i

I

5

5-HIAA

NE

iDA

312± 29 34+ 4 * *

180+17 182+12

6084.178 39+ 7 "

365±99 N.D. 243 4. 52 N.D.

718+ 48 78:t: I0 *

361+36 11,600+619 3014.23 11,174+477

2094- 29 189+ 13

2 1 0

3564.15 48+ 8 *

l

Values expressed in n g / g frozen tissue weight; N.D. = not detectable. * P < 0.005 compared to controls. ** P < 0.001 compared to controls.

'

0

i

5 10 15 20 25 30 Days after PVN Injection

Fig. 5. Mean 4. S.E.M. food intake scores (lower frame) and water intake scores (upper frame) following 40 nmol NE tests at different times after injecting 5,7-DHT vs. control (ascorbate vehicle) solutions described in Fig. 4. See inset for definition of group symbols.

315 TABLE !11

Mean + S.E.M monoamine or metabolite levels in brain regions from rats receiving ascorbate vehicle (Control) or 5,7-dihydroxytryptamine (5, 7-DHT) injections into the PIOV Group sizes: Control = 7; 5,7-DHT = 8.

5-HT a Cortex Control 5,7-DHT

5-HIAA

NE

DA

342+14 285+30

281+ 21 231+ 18

260+ 13 400+25 288+ 21 444+31

195+22 56+11 * * *

407+ 65 171+ 30 * *

370+ 27 N.D. 342+ 31 N.D.

Hippocampus Control 5,7-DHT

Hypothalamus Control 934+97 5,7-DHT 671+95 *

1,325+191 838+ 79 *

2,222+118 450+37 2,012+ 74 410+26

a Values expressed in n g / g frozen tissue weight; N.D. -- not detectable. * P < 0.05 compared to controls. * * P < 0.01 compared to controls. * * * P < 0.001 compared to controls.

lated for each animal per day by subtracting their responses to saline from that to NE. Separate two-way ANOVAs on each variable revealed no reliable differences across time or between groups and no interaction between factors for either measure. Table III summarizes the results of the HPLC assays performed on the three brain regions obtained from rats in this experiment. No group differences were found for any of the compounds measured in cortex. Furthermore, no group differences were found for NE or DA in the hypothalamus or hippocampus. However, PVN 5,7-DHT rats had reliable depletions of both 5-HT and 5.HIAA in the hypothalamus ( - 2 8 % and - 3 7 % , respectively; P < 0.05) an0 hippocampus ( - 71% and - 58%, respectively; P < 0.01) compared to controls. DISCUSSION

The main objective of these experiments was to determine if chronically depleting brain 5-HT might enhance the short episodes of 'feeding elicited by injecting NE into the PVN. The results imply that such feeding enhancement can, iedeed, occur, but only if 5-HT is broadly depleted throughout the neural axis. As such, these data raise important questions about the nature of 5-HT--NE interactions within the medial hypothalamus in their controls over feeding. The results of Expt. 1 showed that widespread depletion of brain 5-HT induced by IC 5,7-DHT did not produce animals who chronically overate or gained weight. Such findings agree with a large literature (see refs. 14, 43, 75 for data summaries and discussion)

which has failed to replicate an early s t u d y 68 suggesting that such treatment might lead to long-term hyperphagia and weight gain. Since global depletion of brain 5-HT by destroying the midbrain raphe nuclei 1a'~7does not produce hyperphagia or weight gain, and the overeating induced by central administration of the 5-HT synthesis inhibitor, p-chlorophenylalanine, produces hyperphagia by means other than through 5-HT depletion ls'43,~'sS, it seems that a tonic reduction in 5-HT release alone is an insufficient condition to enhance daily food intake a n d / o r weight accretion. Yet our IC-treated animals were found to be over-responsive to the feeding effects of PVN NE injections. Interestingly, this enhanced feeding only emerged 2 weeks after IC treatment, and even then was not so robust as to be statistically demonstrable on each successive day of testing from that point onward (see Fig. 2). However, accepting the fact that individual tests of NE-induced feeding can be somewhat variable on a day-to-day basis, the fact that averaged NE feeding scores were enhanced after vs. before day 14 (see Fig. 3) supports this general conclusion. Since the results of NE testing in Expt. 1 varied across time, it is important to note that we confirmed the efficacy of our IC 5,7-DHT treatment to impede serotonergic function in more ways than one. First, we showed that it selectively lowered endogenous 5-HT and 5-HIAA levels as determined from terminal monoamine assays of selected brain regions (see Table It). Second, and equally as important, we demonstrated behaviorally that the 5-HTIA agonist, 8-OHDPAT, was ineffective in eliciting a normal short-term feeding response in these animals but did not disrupt their 24-h intakes (see Table I and Results). Since 8-OH-DPAT is believed to exert its feeding effects by acutely inhibiting the release of endogenous 5-HT (see Introduction), the blunted feeding of 5,7-DHT-treated rats to this agonist provides additional evidence that our neurotoxin treatment impeded 5-HT neurotransmission 6. The fact that we screened our animals for such 8-OH-DPAT feeding after day 21 of the PVN NE time-course testing indicates that the brain 5-HT axis in these animals was still functionally impaired even after reliable increments in PVN-NE feeding had been demonstrated (see Fig. 2). In hindsight, though, we wonder if this test, which preceded the last PVN-NE probe on day 28, may have in some way contributed to the non-significant response evident on that last test day (see Fig. 2). Such a possibility arises from previous observations suggesting that 8-OH-DPAT might produce sub-sensitivity in post-synaptic 5-HT responsivehess 5,5a. However, it must be pointed out that we used a much lower dose of 8-OH-DPAT than was used in

316 that earlier work as well as the fact that subsequent research 39 using in-vivo microdialysis has shown that endogeous 5-HT release is still suppressed after 8OH-DPAT despite prior administration of that drug. In seeking an explanation for the delayed appearance of this enhanced feeding response, several possibilities present themselves. The simplest one is that some non-specific suppressant effects on feeding occurred after IC 5,7-DHT which reflected in PVN NE responses. However, examination of daily intake and body weight data (see Fig. 1) reveals that even control rats displayeu a generalized suppression of intake and weight gain which was statistically reliable. In all likelihood, this effect was due to the powerful effects of DMI pretreatment at the dosage employed 14.66. Such an interpretation is supported by the finding that there were no statistically reliable differences between groups on measures of daily intake and weight gain. That fact notwithstanding, 5,7-DHT-treated rats did remain marginally lower in their body weights throughout these investigations (see ref. 14 for similar observation). As a result, some might view our PVN NE time-course effects as reflecting a non-specific feeding enhancement in an attempt to regain weight following sufficient recovery from the combined DMI and 5,7-DHT treatments. We see this as unlikely since there were no signs of depressed responding to NE in the IC-5,7-DHT group prior to day 14 when their body weights were at their lowest levels. A more likely explanation is that the consequences of IC 5,7-DHT took time to develop. Such a conclusion has been reached by a number of investigators in terms of both biochemical and behavioral parameters that have been measured after administering this neurotoxin. It has been known for some time that endogenous levels of 5-HT in a number of discrete brain regions reach their nadir approximately 2 weeks after such treatment 4. This likely reflects the amount of time required for complete degeneration of affected 5-HT neurons to occur along the entire brain axis, plus the beginning of compensatory collateral sprouting, which in some brain regions may last up to one year before full reinnervation is attained 3. More recently, some investigators have documented that various sub-types of 5-HT receptors begin showing stable changes around this same time period ~,~.Ts. However, since there does not appear to be any comprehensive literature available that has examined all variants of 5-HT receptor change in the same time frames as reported here, it is not possible to suggest which subset(s) of receptors and in which specific brain regions might be responsible for mediating these time-dependent feeding effects. From a behavioral standpoint, evidence also exists indicating

that a 2-week time frame may be required before changes in activity-based measures stabilize 33, although earlier work TM suggested that only 2-4 days were needed before behavioral supersensitivity could be demonstrated if motor responses were driven by 5-HT precursor loadings. Since the brain 5-HT axis is exquisitely plastic and capable of diverse functional reorganization 3, it is possible that the exaggerated feeding which occurred in response to PVN NE 2-3 weeks after IC 5,7-DHT represents a rather narrow window in time during which maximal 5-HT depletory effects can be detected, but after which collateral sprouting a n d / o r post-synaptic receptor up-regulation are sufficient to compensate for pre-synaptic insufficiencies. Alternatively, the enhanced feeding response commencing day 14 may represent a relatively permanent change, but one which was not seen during the last test on day 28, either because of random error in testing or due to some unknown, enduring effect of the 8-OH-DPAT test conducted prior to it (see comments above). Further behavioral tests will be required to distinguish between these possibilities, as will separate work to clarify the neurobiological mechanisms responsible for these time-dependent effects. The results of Expt. 2 showed that PVN 5,7-DHT treatment was also insufficient in producing long-term overeating a n d / o r weight gain. This finding agrees with early work by Myers 63, who found no evidence that male rats receiving intrahypothalamic 5,7-DHT would overeat a chow diet or gain weight. In that same report, Myers also summarized other data that neither rats nor monkeys receiving intrahypothalamic injections of the structurally related 5-HT neurotoxin, 5,6dihydroxytryptamine, increased their feeding and weight gain, but instead showed initial dose-dependent anorexias and weight losses. All of these findings seemingly clash with other work 79 that employed the same dose of 5,7-DHT used here. In that report, injections into the perifornical region of female rats produced mild long-term overeating and obesity, but only if the animals were fed a high-fat diet. However, this effect took a month to develop~ prior to which there were no signs of overeating or weight gain when animals were fed a standard chow diet. Therefore, the feeding of a high-fat diet, which itself engenders weight gain 22, appears to have interacted with the animals' sex (i.e. the use of female rats) to produce thes~ ~unique findings. Not only did PVN 5,7-DHT treatment fail to alter the long-term feeding of our animals, but it also was ineffective in modifying the episodic feeding elicited by NE injections in the PVN. Based on the results of Expt. I, the inability of PVN 5,7-DHT treatment to enhance PVN NE feeding at all was particularly in-

317 triguing. Given the vast literature which has implicated the medial hypothalamus, in general, and the PVN, in particular, as the regions in which 5-HT and NE interact to control feeding 8,9,12,13,15,42,54-$7,60,73-75,80,we were surprised that the same pattern of feeding enhancement seen in Expt. 1 was not duplicated in Expt. 2. What might account for this disparity? Several possibilities present themselves. One is a purely behavioral one; that is, that PVN-treated rats, who ate somewhat more than IC-treated rats in response to PVN NE (compare Figs. 2 and 4), were incapable of consuming any more due to maximal gastric distention (i.e. a ceiling effect on food intake). This explanation seems unlikely, since infusions of more potent orectic agents like NPY or muscimo154'62 can elicit much higher levels of intake over comparable time periods than those reported here. Furthermore, previous work a4 has demonstrated that prior stomach filling does not prevent enhanced feeding by PVN NE unless very large preloads are involved that likely spill over into the upper intestines. Given the modest magnitudes of food intake elicited in our short-term tests, it is not likely that the basal satiety signals evoked were so strong as to prevent the possibility of additional feeding enhancements in PVN 5,7-DHT-treated rats. Another possibility is that 5-HT depletion within the hypothalamic site(s) responsible for NE-induced feeding was simply not as great as that achieved in IC-injected rats. While we have no direct data to address this issue, this seems highly unlikely. Although it is true that global levels of hypothalamic 5.HT were depleted by only 28% (see Table Ill), it actually seems probable that depletions at the site where NE was infused were even greater than those sustained in IC-treated rats, since these were the very same sites at which 5,7-DHT was infused. Therefore, inadequate depletion of hypothalamic 5-HT seems an unlikely explanation for these differences. However, direct confirmation of this likelihood must await further work in which more localized techniques (e.g. HPLC assays of micropunched dissections or dialysates of the PVN region) are brought to bear on this issue. Still another possibility is that the magnitude of 5-HT depletion in brain sites other than the hypothalamus were responsible for these differences. As shown in Table III, our PVN 5,7-DHT rats sustained large ( - 7 1 % ) depletions of hippocampal 5-HT in addition to their hypothalamic depletions. This pattern parallels almost exactly that reported previously by Waldbillig et al. 79, whose injection parameters were used here, albeit at a more medial injection site. However, it seems unlikely that large hippocampal depletions of 5-HT - - either alone or in combination with hypothalamic depletions - - can be viewed as

having prevented the appearance of enhanced PVN NE-induced feeding since (a) the hippocampus of ICtreated rats in Expt. 1 was also severely depleted (see Table If), and (b) the results of previous work using hypothalamic lesions 2°, knife-cuts2~ or 5,7-DHT injections 79 all produced comparable depletions of hypothalamic and hippocampal 5-HT, yet yielded animals that showed enhanced daily feeding and weight gains, not the suppressed ones reported here. One brain region that was highly depleted of 5-HT in our IC-treated rats was the striatum (see Table II). Although no comparable data were obtained for PVNtreated rats, the pathway of ascending 5-HT neurons to this structure, which are situated in the lateral aspects of the medial forebrain bundle 2, would not be expected to have been affected significantly by our medial injection placements. The reason this difference may be relevant in explaining our findings is that recent work by Fletcher 3°'3~ has suggested that feeding elicited by midbrain raphe injections of 8-OH-DPAT is substantially mediated by disinhibited striatal a n d / o r accumbens DA systems. If this hypothesis is correct, then our observation that IC-treated rats ate more in response to PVN NE could be accounted for by such enhanced intake resulting from the additivity of two separable processes: (1) the normal feeding induced by PVN NE, and (2) an exaggeration of that response by a disinhibited forebrain DA axis. Such an interpretation is in keeping with recent observations (unpublished) in this laboratory that simultaneous injections of 8-OHDPAT into the dorsal raphe nucleus along with NE into the PVN can produce dose-dependent additive effects on short-term food consumption. Another observation which can be seen as supporting this view is that systemic29s2 or intraventricular ~9 but not PVN i~,s0 injections of the 5-HT antagonist, metergoline, can enhance feeding. If, indeed, 5-HT normally inhibits striatal a n d / o r accumbens DA release, then only the two feeding-effective routes of administrating this agent sited above would be expected to enhance DA release. One final line of evidence which supports this interpretation derives from work 36 showing that feeding induced by PVN NE is distinct from that elicited by electrical stimulation of the medial forebrain bundle, a procedure known to release DA in the forebrain 37'42 much as can non-painful tail-pinch !°'25'6x, which is also effective in eliciting feeding~'6k The only seeming difficulty with this hypothesis is additional data showing that systemic administration of 8-OH-DPAT does not elicit even additive feeding effects with either PVN N E 24 o r systemic clonidine 23, the latter being an a 2adrenergic agonist hypothesized to elicit feeding by preferentially stimulating post-synaptic receptors within

318

the PVN 35's9. However, the fact that these two negative findings derive from studies in which peripheral as well as wide-spread brain effects would have resulted from such drug treatments may nt)t be overly problematic given the multiplicity of other effects elicited by such wide-ranging manipulations, some of which could impede the expression of overeating 23. While additional work is clearly needed to test the validity of such an 'additivity' hypothesis, one fact remains clear: that is, chronic, relatively localized depletion of hypothalamic 5-HT in no way appears to alter the capacity of the neural mechanisms within the medial hypothalamic axis to elicit feeding in response to exogenous NE administration. This conclusion takes on even broader implications if one considers the now commonly held b e l i e f 42'St'52"54'55't'°'73`77 that release of endogenous NE at that site is a causal element in the initiation and/or maintenance of normal meals. This is of some concern, since it has recently been suggested 4:-49'55'~9 that one pathological basis of the hunger and/or overeating characteristic of certain human eating disorders is insufficient inhibitory controls exerted by inadequate release of endogenous 5-HT within the hypGthalamus. Although we agree with the general premise that enhanced states of hunger or eating play a major role in these disorders (see ref. 16 for an even broader perspective on this), the current data prov!.i,~ no obvious support for ~his notion in so far as deficient hypothalamic 5-HT is concerned. Furthermore, '~s discussed above, there is no consistent evidence that any technique which chronically depletes brain 5-HT alone can reliably enhance long.term feeding or weight gain in animals. Given this lack of direct evidence, it may be informative to point out that the majority of animal studies which have been interpre. ted 55 to support the above proposition in eating disordered patients derive, in fact, from work showing that enhancing ~-HT function can inhibit natural or NE-induced feeding (although see refs. 12 and 32 for very recent data at odds even with this position). As a basic tenet of research, it is not logically sound to assume that the inverse, relationship automatically holds, which in this case is that deficits in 5-HT function disinhibit either natural or NE-induced feeding. Therefore, the data reported here, along with the issues we have raised surrounding them, suggest that some revision of thinking may be in order to more accurately reflect the causal and/or sustaining factors that brain serotoner. gic metabolism may contribute to these troublesome clinical disorders. Acknowledgements. This research was supported by funds from the Natural Sciences and Engineering Research Council (NSERC) of

Canada and the Clarke Institute of Psychiatry. The authors thank Dr. John Chambers for running the HPLC assays of brain tissue. The behavioural data obtained constituted part of a M.A. thesis submitted by the second author to the Department of Psychology, University of Toronto. A preliminary report of this work was presented at the annual meeting of the Society for Neuroscience in 1987 (Soc. NeuroscL Abstr., 13 (1987) 14. Abstract no. 9.1).

REFERENCES Antelman, S.M. and Szechtman, H., Tail pinch induces eating in sated rats which appears to depend on nigrostriatai dopamine, Science, 189 (1975) 731-733. Azmitia, E.C., The serotonin-producing neurons of the midbrain median and dorsal raphe nuclei. In L.L. Iversen, S.D. Iversen amd S.H. Snyder, (Eds.), Handbook of Psychopharmacology, Vol. 9, Plenum, New York, 1978, pp. 233-314. Azmitia, E.C. and Whitaker-Azmitia, P.M., Awakening the sleep. ing giant: anatomy and plasticity of the brain serotonergic system. J. Clin. Psychiatry, 52 Suppl. 12 (1991) 4-16. Baumgarten, H.G., Lachenmayer, L. and Bj6rklund, A., Chemical lesioning of indoleamine pathways. In R.D. Myers (Ed.), Methods in Psychobiology, Vol. III, Academic Press, New York, 1977, pp. 47-98. Beer, M., Kennett, G.A and Curzon, G., A single dose of 8-OHDPAT reduces raphe binding of [3H]8-OH-DPAT and increases the effect of raphe stimulation on 5-HT metabolism, Fur. J. Pharmacol., 178 (1990) 179-187. Bendoni, C. and Samanin, R., 8-Hydroxy-2-(di-n-propylamino)tetralin (8-OH-Dpat) elicits eating in free-feeding rats by acting on central serotonin neurons, Fur. J. Pharmacol, 121, (1986) 147-150. Bjorklund, A.H., Baumgarten, H.G. and Rensch, A., 5,7-dihydroxytryptamine: improvement of its selectivity for serotonin neurons in the CNS by pretreatment with desipramine, J. Neurochem., 24 (1975) 833-835. 8 Blundell, J,E., Serotonin and appetite, Neuropharmacology, 23 (1984) 1537-1551. 9 BlundeU J.E., Serotonin manipulations and the structure of feeding behaviour, Appetite, 7 (Suppl.) (1986) 39-56. 10 Boutelle, M.G., Zenerstrom, T., Pei, Q., Svensson, L. and Fillenz, M., In vivo neuroehemical effects of tail pinch, J. Ne,~,osci, Methods, 34 (1990) 151-157. I1 Brown, L.M., Smith, D.L., Williams, G.M. and Smith, DJ., Alterations in serotonin binding sites after 5,7-dihydroxytrypta. mine treatment in the rat spinal cord, Neurosci. Lett., 102 (1989) 103-107. 12 Carruha, M.O., Nisoli, E., Garosi, V., Pizzi, M., Memo, M. and Spano, P., Characterization of monoaminergic receptors in hypothalamic nuclei mediating feeding behavior in rats, J. Pharmacol. Exp. Ther., 259 (1991) 124-130. 13 Coscina, D.V., Brain amines in hypothalamie obesity. In R.A. Vigersky (Ed.), Anorexic Nert,osa, Raven, New York, 1977, pp. 97-107. 14 Coseina D.V., Effects of central 5,7-dihydroxytryptamine on the medial hypothalamic syndrome in rats, Ann. N.Y. Acad. Sci., 305 (1978) 627-644. 15 Coseina, D.V., Qualitative similarities of serotonin-depleting midbrain lesions to vagotomy in altering weight gain in different models of obesity, in L.A. Cioffi, W.P.T. James and T.B. Van ltallie (Eds.), The Body Weight Regulatory System: Normal and Disturbed Mechanisms, Raven, New York, 1981, pp. 19-23, 16 Coscina, D.V. and Garfinkel, P.E., Animal models of eating disorders: clinical perspectives. In P. Willner (Ed.), Behm,ioural Models in Psychopharmacology, Cambridge Univ. Press, London, 1991, pp. 237-250. 17 Coscina, D.V. and Stancer, H.C., Selective blockade of hypothalamic hyperphagia and obesity in rats by serotonin-depleting midbrain lesions, Science, 195 (1977) 416-419. 18 Coscina, D.V., Daniel, J. and Warsh, J.W., Potential nonserotonergic basis of hyperphagia elicited by intraventricular p-

319 chlorophenylalanine, Pharmacol. Biochem. Behav., 9 (1978) 791797. 19 Coseina, D.V., Nobrega, J.N. and Feifel, D., Stimulation of feeding by CNS metergoline varies with injection site, Soc. Neurosci. Abstr., 16 (1990) no. 131.3. 20 Coscina, D.V., Warsh, LW. and Anderson, G.H., Brain regional serotonin depletion in hypothalamie obesity: lesion vs. overeating as contributory factors, Soc. Neurosci. Abstr., 6 (1980) 783. 21 Coscina, DN., Chambers, J.W., Chang, J., Chiu, A. and Warsh, J.J., Brain regional serotonin in hypothalamic obesity: different depletion profiles using lesions vs. parasagittal knife-cuts, Soc. Neurc.~ci. Abstr., 7 (1981) 853. 22 Coscina, D.V., Lacombe, S., Chambers, J.W., Dixon, L.M. and Nobrega, J.N., Intake of greasy diets in hypothalamic obesity: a reassessment, Appetite, 13 (1989) 15-24. 23 de Rooy, E.C.H., Effects of depressing brain serotonin fun:tion on adrenergically-induced feeding in rats, M.A. Thesis, Department of Psychology, University of Toronto, 1987, 58 pp. 24 de Rooy, E.C.H. and Coscina, D.V., Effect of systemic 8-OHDPAT on the feeding induced by hypothalamic infusion of NE, Pharmacol. Biochem. Behav., 36 (1990) 937-943. 25 D'Angio, M., Serrano, A., Rivy, J.P. and Scatton, B., Tail-pinch stress increases extracellular DOPAC levels (as measured by in vivo voltammetry) in the rat nucleus accumbens but not frontal cortex: antagonism by diazepam and zolpidem, Brain Res., 409 (1987) 169-174. 26 Dourish, C.T., Hutson, P.H. and Curzon, G., Low doses of putative serotonin agonist 8-hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT) elicit feeding in the rat, Psychopharmacology, 86 (1985) 197-204. 27 Dourish, C.T., Hutson, P.H. and Curzon, G., Characteristics of feeding induced by the serotonin agonist 8-hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT), Brain Res. BMi., 15 (1985)377384, 28 Felice, LJ., Felice, J.D. and Kissinger, P.T., Determination of catecholamines in rat brain parts by reverse-phase ion-pair liquid chromatography, Gen. Neurochem., 31 (1978) 1461--1465. 29 Fletcher, P.J., Increased food intake in satiated rats induced by the 5-HT antagonists methysergide, metergoline and ritanserin, Psyehopharmacology, 96 (1988) 237-242. 30 Fletcher, P.J., Dopamine receptor blockade in nucleus accumhens or cauda~.e nucleus differentially affects feeding induced by 8-OH-DPAT injected into dorsal or median raphe, Brain Res., 552 (1991) 181-189. 31 Fletcher, P.J. and Davis, M., A pharmacological analysis of the eating response induced by 8.OH.DPAT injected into the dorsal raphe nucleus reveals the involvement of a dopaminergic mechanism, Psychopharmacology, 100 (1990) 188-194. 32 Fletcher, PJ., Ming, Z.-H., Zack, M,H. and Coscina, D.V., A comparison of the effects of the 5-HT z agonists TFMPP and RU 24969 on feeding following peripheral or medial hypothalamic injection, Brain Res., 580 (1992) 265-272. 33 Gately, P.F., Segal, D.S. and Geyer, M.A., The behavioral effects of depletions of brain serotonin induced by 5,7-dihydroxytryptamine vary with time after administration, Behav. Neural Biol., 45 (1986) 31-42. 34 Gibson, E.L. and Booth, D.A., Feeding induced by injection of norepinephrine near the paraventricular nucleus is suppressed specifically by the early stages of strong postingestional satiety in the rat, Physiol. Psyehol., 14 (1986) 98-103. 35 Goldman, C.K., Marino L. and Leibowitz, S.F., Postsynaptic a2-noradrenergic receptors mediate feeding induced by paraventricular nucleus injection of norepinephrine and clonidine, Eur. J. Pharmacol., 115 (1985) 11-19. 36 Halperin, R., Gatchalian, C.L., Adachi, TJ., Carter, J. and Leibowitz, S.F., Relationship of adrenergic and electrical brain stimulation induced feeding responses, Pharraacol. Biochem. Behav., 18 0983) 415-422. 37 Hernandez, L. and Hoebel, B.G., Feeding and hypothalamic stimulation increase dopamine turnover in the accumbens, Physiol. Behav., 44 (1988) 599-606.

38 Hillegaart, V., Hjorth, S. and Ahlenius, S., Effects of 5-HT and 8-OH-DPAT on forebrain monoamine synthesis after local application into the median and dorsal raphe nuclei of the rat, J. Neural Transm., 81 (1990) 131-145. 39 Hjorth, S., Single-dose 8-OH-DPAT pretreatment does not induce tachyphylaxis to the 5-HT release-reducing effect of 5-HTIA autoreceptor agonists, Eur. J. Pharmacol., 199 (1991) 237-242. 40 Hjorth S. and Magnusson, T., The 5-HT1A receptor agonist, 8-OH-DPAT, preferentially activates cell body 5-HT autoreceptors in rat brain in vivo, Naunyn-Schmiedesberg's Arch. Pharmacol., 338 (1988) 463-471. 41 Hjorth, S. and Sharp, T., Effect of the 5-HT~A receptor agonist 8-OH-DPAT on the release of 5-HT in dorsal and median raphe-innervated rat brain regions as measured by in vivo microdialysis, Life Sci., 48 (1991) 1779-1786. 42 Hoebel, B.G. and Leibowitz, S.F., Brain monoamines in the modulation of self-stimulation, feeding, and body weight. In H. Weiner, M.A. Hofer and AJ. Stunkard (Eds.), Brain, Behavior, and Bodily Disease, Raven, New York, 1981, pp. 103-142 43 Hoebel B.G., Zemlan F.P., Trulson M.E., MacKenzie R.G., DuCret R.P. and Norelli C., Differential effects of p-ehlorophenylalanine and 5,7-dihydroxytryptamine on feeding in rats, Ann. N.Y. Acad. Sci., 307 (1978) 590-594. 44 Holmes LJ., Storlien L.H. and Smythe G.A., Medial basal hypothalamie monoamine activity associated with intraeerebroventricular p-chlorophenylalanine-induced hyperphagia, Brain Res., 528 (1990) 269-272. 45 Hutson, P.H., Dourish, C.T. and Curzon, G., Neurochemical and behavioural evidence for mediation of the hyperphagic action of 8-OH-DPAT by 5-HT cell body autoreceptors, Fur. J. Pharmacol., 12 (1986) 347-352. 46 Invernizzi, R., Carli, M., Di Clemente, A. and Samanin, R., Administration of 8-hydroxy-2-(di-n-propylamino)tetralin in raphe nuclei dorsalis and medianus reduces serotonin synthesis in the rat brain: differences in potency and regional sensitivity, J. Neurochem., 56 (1991) 243-9.47. 47 Jimerson, D.C., Lesem, M.D., Kaye, W.H., Hegg, A.P. and Brewerton, T.D., Eating disorders and depression: is there a serotonin connection? Biol. Psychiatry, 28 (1990) 443-454. 48 Kaye, W.H. and Weltzin, T.E., Serotonin activity in anorexia and bulimia nervosa: relationship to the modulation of feeding and mood, J. Clin Psychiatry, 52 (SI2) (1991) 41-48. 49 Kaye, W.H. and Weltzin, T.E., Neurochemistry of bulimia nervosa, J. Clin. Psychiatry, 52 (SI0) (1991) 21-28. 50 Kennett, G.A., Marcou, M., Dourish, C.T. and Curzon, G., Single administration of 5-HTtA agonists decreases 5-HTIA presynaptic. but not postsynpatic receptor-mediated responses: relationship to antidepressant-like action, Eur. J. Pharmacol., 138 (1987) 53-60. 51 Kow, L.-M. and Pfaff, D.W., Responses of hypothalamic paraventricular neurons in vitro to norepinephrine and other feeding-relevant agents, Physiol. Behav., 46 (1989) 265-271. 52 Kruissink, N., van der Gugten, J. and Slangen, J.L., Short-term feeding- related changes in mediodorsal hypothalamic eatecholamine release, PharmacoL Biochem. Behav., 24 (1986) 575579. 53 Leibowitz, S.F., Paraventricular nucleus: a primary site mediating adrenergic stimulation of feeding and drinking, PharmacoL Biochem. Behav., 8 (1978) 163-175. 54 Leibowitz, S.F., Brain monoamines and peptides: role in the control of eating behavior, Fed. Proc., 45 (1986) 1396-1403. 55 Leibowitz, S.F., The role of serotonin in eating disorders, Drugs, 39 (1990), 33-48. 56 Leibowitz, S.F. and Shor-Posner, G., Brain serotonin and eating behavior, Appetite, 7 Suppl. (1986) 1--14. 57 Leibowitz, S.F., Weiss, G.F. and Shor-Posner, G., Hypothalamic serotonin: pharmacological, biochemical, and behavioral analysis of its feeding~guppressive action, Clin. Neuropharmacol., 11 (1988) 551-571. 58 Mackenzie R.G., Hoebel B.G., D~.cret R.P. and Trulson M.E., Hyperphagia following intravenUicular p-chiorophenylalanine-, leucine- or tryptophan-methyl esters: lack of correlation with

320 whole ~rain serotonin levels, Pharmacol. Biochem Behac., 10 (1979) 951-955. 59 McCabe, J.T., DeBellis, M. and Leibowitz, S.F., CIonidine-induced feeding: analysis of central sites of action and fiber projections mediating this response, BrabJ Res., 309 (1984) 85-104. 60 Minano F.J., Peinado, J.M and Myers, R.D., Profile of NE, DA and 5-HT activity shifts in medial hypothalamus perfused by 2-DG and insulin in the sated or fasted rat, Brabl Res. Bull., 22 (1989) 695-704. 61 Mittleman, G., Castaneda, E., Robinson, T.E. and Valenstein, E.S., The propensity for nonregulato~ ingestive behavior is related to differences in dopamiae systems: behavioral and biochemical evidence, Behac. Neurosci., 100 (1986)213-200. 62 Morley, J.E., Neuropeptide regulation of appetite and weight, Endocr. Rec.., 8 (1987) 256-287. 63 Myers, R.D., Hypothalamic actions of 5-hydro~tryptamine neurotoxins: feeding, drinking, and body temperature, Ann. N.Y. Acad. Sci., 305 11978) 556-575. 64 Nelson, D.L., Herbet, A., Bourgoin, S., Glowinski, J. and Hamon, M., Characteristics of central 5-HT receptors and their adaptive changes following intracerebral 5,7-dihydroxytryptamine administration in the rat, Mol. Pharmacol., 14 11978) 983-995. 65 Nobrega J.N. and Coscina, D.V., Chronic amitriptyline potentiates feeding induced by intrahypothalamic norepinephrine injections, Lift, Sci., 33 11983) 1249-1253. 66 Nobrega, J.N. and Coscina, D.V., Effects of chronic amitriptyline and desipramine on food intake and body weight in rats, Pharmacol. Biochem. Behat'., 27 (1987) 105- ! 12. 67 Nobrega, J.N. and Coscina, D.V., Regional changes in brain ~4C-2-deox'yglucoseuptake after feeding-inducing intrahypothalamic norepinephrine injections, Brah~ Res. Bull., 24 (1990) 249-255. 68 Sailer, C.F. and Stricker, E.M., Hyperphagia and increased growth in rats after intraventricular injection of 5,7-dihydrox'ytryptamine, Science, 192 11976) 385-387. 69 Samanin R, and Garattini S., Serotonin and the pharmacology of eating disorders, Ann. N.~ Acad. Sci., 575 11989) 194-208. 70 Schecter,L,E., Bolanos, F.J., Gozlan, H., Lanfumey, L,, Haj-Dahmane, S., Laporte, A.-M., Fattaccini, C.-M. and Hamon, M., Alterations of central serotoninetgic and dopaminergic neuro. transmission in rats chronically treated with ipsapirone: biochemical and electrophysiolo/~ical studies, J. Pltarmacol. Exp. Ther., 255 (1990) 1335-1347.

71 Sharp, T. and Hjorth, S., Application of brain microdialysis to study the pharmacology of the 5-HTtA autoreceptor, J. Neurosci. Methods, 34 11990) 83-90. 72 Stallone, D. and Nicolaidis, S., Increased food intake and carbohydrate preference in the rat following treatment with the serotonin antagonist metergoline, Neurosci. Lett., 102 (1989) 319-324. 73 Stanley B.G., Schwartz, D.H., Hernandez, L., Hoebel, B.G. and Leibowitz, S.F., Patterns of extracellular uo:'epinephrine in the paraventricular hypothaiamus: relationship to circadian rhythm and deprivation-induced eating behavior, Life Sci., 45 11989) 275-282. 74 Stanley B.G., Schwartz, D.H., Hernandez, L., Lcibowitz, S.F. and Hoebel, B.G., Patterns of extracellu!ar 5-hydroxyindoleacetic acid (5-HIAA) in the paraventricular hypothalamus (PVN): relation to circadian rhythm and deprivation, Pharmacol. Biochem. Behac., 33 (1989) 257-260. 75 Sugrue, M.F., Neuropharmacology of drugs affecting food intake, Pharmacol. Ther., 32 (1987) 145-182. 76 Trulson, M.E. and Jacobs, B.L., Behavioral evidence for denervation supersensitivityafter destruction of central serotonergic nerve terminals, Ann. N.Y. Acad. Sci., 305 (1978) 497-509. 77 Van der Gugten. J. and Slangen, J.L, Norepinephrine uptake by hypothalamic tissue from the rat related to feeding, PharmacoL Biochem. Behac., 3 11975) 855-860. 78 Verge, D., Daval, G., Marcinkiewicz, M., Patey, A., El Mestikaway, S., Gozlan, H. and Hamon, M., Quantitative autoradiography of multiple 5-HT I receptor subtypes in the brain of control or 5,7-dihydroxytryptamine-treated rats, J. Neurosci., 6 11986) 34743482. 79 Wald.billig,R.J., Bartness, T.J. and Stanley, B.G., Increased food intake, body weight, and adiposity in ra!s after regional neurochemical depletion of serotonin, J. Comp. Physiol. Psychol., 95 (1981) 391-405. 80 Weiss, G.F., Papadakos, P., Knudson, K. and Leibowitz, S.F., Medial hypothalamic serotonin: effects on deprivation and norepinephrine-induced eating, Pharmacol. Biochem. Behac., 25 (1986) 1223- 123O, 81 Wilkinson, L.O, and Dourish, C,T., Serotonin and animal behavior. In S.J. Peroutka (Ed,), Serotonin Receptors and Subtypes: Bask" and Clinical Aspects, Wiley-Liss, New York, 1991, pp, 147-210.