- Email: [email protected]
Effect of continuous infusions of dexfenfluramine on food intake, body weight and brain amines in rats
Life Sciences, Vol. 39, pp. 2581-2586 Printed in the U.S.A.
Pergamon Journals
EFFECT OF CONTINUOUS INFUSIONS OF DEXFENFLURAMINE ON FOOD INTAKE, BODY WEIGh~ AND BRAIN AMINES IN RATS Nell E. Rowland Department of Psychology and Center for Neurobiological Sciences, University of Florlda, Gainesville, Florida 32611
(Received in final form September 18, 1986) Sunmary The present experiments describe the effects of continuous SC infusion, via osmotic minipumD, of dexfenfluramine on food intake and body welght of male and female rats. It was found that the food intake of male rats was reduced by infusions of both 3 and 6 mg/kg/day although tolerance developed within 2-4 days at the lower dose. Further, these rats showed tolerance to an acute anorectic test dose of dexfenfluramine. Body weight loss was sustained by both groups. In older (6-8 mo old) female rats, scme of which had previously nursed three litters, the anorectic effects of dexfenfluramine (3 and 6 mg/kg/day) were sustained throughout the 6 day infuslon, and weight loss was substantial. The effects did not differ between bred and virgin rats of eomparable age. The lower dose of dexfenfluramine produced no depletion of brain serotonin (5HT), although 5HIAA was reduced. Both ccmpounds were depleted by the higher dose. The 3 mg/kg/day dose, in select rat populations, may be a close model for the mode of dexfenfluramine administration to humans. Fenfluranlne, an anorectlc agent, conslsts of d-(+) and I-(-) enantiomers of which the former is more potent in reducing food intake. The present studies use the d-enanticmer of fenfluramine which will be referred to as dexfenflur6mdne. The actions of both the racemic fenfluramine mixture and dexfenfluramine have been studied extenslvely and reviewed recently (1-3). Most studies have used rats as subjects. However, the pharmacokinetlcs of fenfluramine are substantially different between rats and man. Thus, the plasma half life of fenfluraamne is only 2-4 hr in rats, but is of the order of 18-24 hr in man (1,4) . This difference produces sc[ne potentlal problems in comparisons between studies of chronic administration. In man, the treatment is usually chronic and the long half llfe of fenfluramine and its active metabolite norfenfluramlne ensure that, even if the drug is taken only once daily, the plasma levels of the agents are stable across the entire day/night cycle (I). On the other hand, in chronic administration studies in rats (e.g., 5), onceor twice-daily injections will result in w±dely-fluctuating drug levels because of the short half life. It appears in rats that this sort of chronic drug regimen is assoclated with the rapid development of tolerance to its anorectic action, whereas in humans tolerance may not be a major proble~ (1,5) . In the present study, we have exanlned whether chronic infusions ~ y osmotic minipump) of dexfenfluramlne to rats, whlch pres~nably would result in steady drug concentratlons in the body, have effects which differ frc~ daily injection regimens. 0024-3205/86 $3.00 + .00 Copyright (c) 1986 Pergamon Journals Ltd.
2582
Dexfenfluramlne
Infusions and Feeding
Vol. 39, No. 26, 1986
Our previous work also indlcates that the effects of fenfluranlne may be more marked in obese or older as opposed to lean or younger rats (5). I have thus used rats whlch fall into the former category in the present studies. In the first study male Sprague-Dawley rats were used when they had grown close to asymtotic hody weight. In the second study, both vlrghn and bred 6-8 mo old female rats were used. This study was based on our previous observation that daily fenfluramine in]ectlons caused much more anorexia and welght loss in a group of retired breeder females than in any rats that we had previously tested (5). However, age-matched virgin controls were not used in that study, so we were unable to dissoclate factors of sex, age, or prior parturitions. Method Animals and Housing Study A: Adult male Sprague-Dawley rats, purchased from ZlVlC Miller Labs (Alllson Park, PA) or bred in our lahoratory from that stock, were approximately 4-6 mo old at the time of the experLment. They were ralsed in group cages wlth ad lJ01tum access to Purina Rodent Chow pellets (#5001) and tap water. For a few weeks prior to, and throughout the experiment they were housed singly in stamnless steel cages wlth ad llblt~n access to a powdered version of the same #5001 diet, presented in a glass jar inside the cage. The vivarium lights were on from 0700-1900 hr, and the temperature was 23 _+ 2°C. Study B: Thlrty-sLx female Sprague-Dawley rats, weighing near 200 g at the start of the study, were purchased from Z1vlc M111er Labs. Eighteen of the rats were housed in palrs in solid-bottom stainless steel cages. Pine shavings were placed in the cages and were changed twlce a week. Purina Rodent Chow pellets and tap water were available ad libitt~. The remaining 18 rats were housed singly in the same klnd of solid bottom cage. Stud male rats were introduced for about 1 week. The females were checked for delivery, and lltters culled to 8-9 pups within 2 days of birth. The pups were weaned at 22 days of age, the female allowed a 1 week rest, and then a male was agaln introduced and the breeding cycle repeated. All rats retained in the study gave birth to and weaned three litters. Thus, approxlmately 4-5 mo after the start of the study, when the rats were 6-8 mo old, two age-matched groups of rats were obtained: unbred vlrgln and bred (multlparous) . At this time all rats were housed singly in regular hanging cages. All other details of care and condition are as in Study A. Procedure Measurements of food intake were made every 1-2 days throughout the study, and body weight was recorded every 3 days. After a period of baseline measures, the rats were randomly assigned to three treatn~nt groups. All were anesthetlzed with ether and an osmotlc minipump (Alzet model #2001, 7 day, 1 ul/hr) was inserted subcutaneously over the scapulae. The incision was closed wlth a wound clip and the rat returned to its cage. The mlnlpumps were preloaded with either nothing (these blanks were prevlously-used pumps that had been washed and sterilized) , or dexfenfluramine hydrochloride. In Study A, the concentration of dexfenfluranlne was 80 mg/ml or 160 mg/ml in saline vehicle, which resulted in deliver 3, of 2 or 4 mg drug 24 hr, respectively. In Study B, the concentration of dexfenflurammne was 50 mg/ml or I00 mg/ml. In both studies, these concentrations resulted in the deliver], of approxlmately 3 or 6 mg/kg/day dexfenflurammne. For sL~oliclty, these will be referred to as the low and hlgh doses. The food intake and body weight were recorded for 6 days.
Vol. 39, No. 26, 1986
Dexfenfluramine Infusions nad Feeding
2583
Tolerance Test (Study A) : At the end of the 6th full day of infusion to the male rats, the minipumps were r~moved (again using ether anesthesia), and the rats were returned to their home cage without food for 24 hr. At the end of this time they were injected acutely with dexfenfluramine (2 mg/kg, IP) and 30 min later were glven 3 weighed food pellets. Their food intake was recorded over the next hour. Neurochemlstry (Study B): At the end of the 6th full day of infusion to the female rats, at about II00 hr and with the pumps still in place, the rats were killed by rapid decapitation. The whole brains were removed, weighed and m~nediately homogenized in 4 vol~nes of ice-coldO.2N perchloric acid wlth EDTA. The homogenate was centrifuged (15,000 rl~n, I0 min, 4°C), and an aliquot of the supernatant taken for analysis of seroton±n (5h'f) and 5hydroxylndoleacetlc acid (5HIAA) by reverse phase high performance liquid chrcmatography (HPLC) with electrochemtical detection. Details of our analysis conditions have been published elsewhere (6). Peak areas were normalized to an internal standard (isoproterenol). Statistics: D a t a w e r e treated by analysis of variance and post-hoc t-tests with significance level set at 0.05. Results Study A: The food intake data are shown in Figure IA (top) . Dexfenfluramine produced a dose-related inhlbition of intake [F(2,16) = 14.28, P<0.01]. Substantial tolerance developed to the anorexia across days of treatment [maln effect of days F(5,80) = 35.17, P<0.001; dose x days Interaction F (10,80) = 3.11, P<0.01]. Nonetheless, the intake in the high dose group remained 10-20% and slgnificantly lower than that of controls even after 6 days. The low dose group recovered control intakes within 2-4 days. All of the treatments caused body welght loss over the first 3 days, with some recovery in the second 3 day segment (F ratios were signiflcant for days but not dose). This recovery was mere complete in the low dose group, whlch weighed only 6 g below their starting weight (i.e., a 1% loss) , c(mpared with the high dose group whlch was 24 g (4%) below their initial weight. Thus the days x dose interactlon was slgniflcant [F(4,32) = 3.11, P<0.05]. The results of the postfast feedlng test for tolerance are shown in Table I. The control mlnipunp group showed a near total anorexia to the acute dose of dexfenfluramine, while the nun±pimp drug-treated groups showed only a small but sign±±±cant reduction in food intake cenpared to a namve vehicleinjected group. TABI.E I Effect of Acute Dexfenfluramlne (2 n~/kg, IP) on l-hr Food Intake of Food Deprlved Male Rats That Previously Received a 6-Day M1n±pump Infusion of Dexfenfluramlne Chronlc (minlpump, 6 day) Infusion
Acute Test Injection
Food Intake (g/hr)
None Blank Dexfenfluramine (3 mg/kg/day) Dexfenfluramine (6 mg/kg/day)
Vehicle Dexfenfluramine Dexfenfluramlne Dexfenfluramine
7.7 0.1 4.8 5.2
± ± ± ±
0.4 0. I** 0.7* 0.3*
Mean ± SE for groups of 4. F(3,12) = 41.96, P<0.001 main effect. *P<0.05 differs from vehicle injected group. **P<0.01 dlffers from all others.
2584
Dexfenfluramlne Infusions and Feedlng
Vol. 39, No. 2(~, 1986
The mean (f SE) body wezght of the unbred and bred groups were 398 ± 7 and 399 ± ii g at the start of the experiment. These groups also did not differ zn theLr response to dexfenfluramine treatment, and so they have been osmbined for all data analyses. The food zntake data are shown in Figure IB (top). Both low and hlgh doses of dexfenfluramine produced a marked and sustalned decrease in intake relatzve to the control group, with the high dose only slightly more effectzve than the low dose. Main effects of drug [F(2,28) = 40.10], days [F(5,140) : 11.07], and the drug x days interaction [F(10,140) : 4.34] all were highly szgnzfzcant (Ps<0.001) , DEXFENFLURAMINE
I
MINI P U M P
B FEMALE RATS
A MALE RATS
-~ + 4 0 -
uJ <
+-+ ~ ,- ,+, o - -
/
30-
÷--~. ÷ ,0
/
o "/
--0 • ~.
30-
0
÷
>-
-< a
~
,"
20-
,I
c
10-
/
/
¢:
•
• ~.
+ --+
Blank
~--A e
DF, Low DF, High
--
1
+
20-
2
3
4
5
6
,
~
/
10-
Days
1
2
3
4
5
8 Days
+10-
+10-
÷
O-
OLu -I 0 >
o -200 m Z -30 <
-20-
~
-40-
4o
-30-
• ~
_
~
FIG. 1 Effects of contlnuous SC znfusion, by minmpumps, of dexfenfluramine hydrochloride (DF) at either 3 mg/kg/day (low dose) or 6 r0g/kg/day (high dose). Left panels A for male Sprague-Dawley rats show mean dally chow intake (top) and body weight change from that on the day of ptmlo implantat±on (bottom). Right panels B show the sane measures for female Sprague-Dawley rats. Filled symbols indicate significant (P<0.05) differences from respective blank-pump control condition (+ --- +); open symbols are not sign±ficantly dlfferent. * indlcates body we±ght change in high dose DF groups is d±fferent (P~0.05) from that of low dose DF group. For clarity, standard errors have not been included, but ranged from 1-4 g for food intake and 3-10 g for weight change. The body weight data (Figure IB, bottom) show that the high dose group lost the most weight and sustained that loss during the second 3 day infusion
Vol. 39, No. 26, 1986
Dexfenfluramine Infusions and Feeding
2585
segment. The low dose group lost less weight initially, and regained some of that weight in the second 3 days. The main effects of drug [F(2,28) = 24.93] and days [F(I,28) = 8.85] were s±gnif±cant (Ps<0.01), but their interaction was not. The l o w a n d h i g h d o s e groups had lost about 2% and 7% of initial weight, respectively, by day 6. The control rats gained 2% weight, thus the losses of the dexfenfluramine groups relative to controls are correspondingly greater. The neurochemlcal data are shown in Table 2. Rats infused wlth the low dose of dexfenfluramine showed a nonsignificant 13% reduction in brain 5HT relative to the no drug controls. Rats infused wlth the high dose of dexfenfluramine had a large (67%) reduction ~n brain 5HT. With both doses, 5HIAA was reduced to a cc~mparably large extent, and norepinephrine was reduoed in a dose-related manner. Dopamine was unaffected by either dose. TABLE II Effect of Six Day Minipump Infuslon of DexfenfluramineonWhole Brain Concentrations, Expressed as % of Blank Ptmp Controls, of Indoleamines and Catecholamlnes in Female Rats. Dose of Dexfenfluramine
Compound
3 mg/kg/day Serotonin 5-hydroxyindoleacetic acid Norepinephrine Dopamine
87 43 81 93
± ± ± ±
5 6** 3** 3
6 mg/kg/day 33 39 64 104
± ± ± ±
5** 5** 3** 8
M + SE for Ns of 6-8. **P<0.01 lower than blank-pimp controls. Discussion This experiment shows that osmotic minip~mps Whlch deliver approximately 3 or 6 mg/kg/day dexfenfluramine to rats produce inltial anorexia and weight loss. These effects are largely reversed after 6 days at the low dose, and are only partly reversed at the high dose, in male rats (Study A) . The effects of dexfenfluramine in the female rats (Study B) appear to be greater and longer lasting than for the males in Study A (compare Panels A and B of Flgure l). The plasma half life of dexfenfluramine in rats is about 3 hr, or onesixth that in humans. At a crude approximation, then, to mimic a human dose in rats a oontinuous daily infusion of about 6 x that dose would be appropriate. Thus, a cllnical daily dose of 30 mg would typically be bet%L=en 0.3 and 0.5 m~/kg and would be mimicked by a rat infusion of 2-3 mg/kg/day. The low dose (approximately 3 mg/kg/day) was chosen on the basis of this reasoning, and clearly has anorectic effects in rats. This was beth greater and longer-lasting in females than males. The females were 6-8 mo old (i.e., "middle-aged") , and no differences were noted as a result of previous parturitions. Thus, although partial tolerance developed to the anorexia and weight loss, thls was less complete in these older females than in other groups of rats given dally injections of d, l-fenfluramine (5) . In other studies (i0), we have found that a 4-week infusion of dexfenfluramine (3 r~7/kg/day) into female rats made obese by access to a palatable diet (sweet milk and cookies), does not maintain absolute weight loss
2386
Dexfenfluramlne [nfuslons and Feedlng
Vol. 39, No. 26, ]986
beyond the second week, although weight is ma~ntalned below control-infused group levels. This sort of phenc~enon often has been ±nterpreted as a "lowered set point" (11,12) , although we have pointed out that some aspects of the phenomenon are inconsistent wlth this theoretical attrlbutlon (13). We have no independent evidence that the mlnlpumps did in fact deliver a constant amount of drug throughout the 6 day perled, nor have we determined tissue or plasma levels of dexfenfluramlne or norfenfluramlne. However, the d~nonstratlon of substantial tolerance to the acute anorectlc effects of dexfenfluramine in the ptmlo-treated groups (Study A) strongly suggests that the pumps were effective. Some loss of tolerance is to be expected in light of our previous flndlng that after 8 dally injectlons of fenfluramine, tolerance to an acute anorectic probe dose is reduced as h t t l e as 2 days after the end of the chronic regimen (7). Further, both the extent and time course of the apparent anorectic tolerance (Flmure IA) mirrors qulte remarkably that seen with chronlc dally injections of either fenfluramine (5) or dexfenfluramlne (i). We do not presume that the effects of chron±c slngle in3ectlons are identical to mlnip~ps: for example, in the in3ectlon paradigm rats' food intake may vary reciprocally with the drug levels, while in the infuslon paradigm the rats may redlstribute their feedlng (we have not made any mlcrobehavioral measurements (cf. 8) to support or refute this or other possibilities) . While the mmniptmlo route may offer some advantages in terms of relevance to drug levels in humans, the gross behavioral results are not markedly different from single lnjectlon paradigms. The low dose of dexfenfluramine did not significantly deplete brain 5h'f levels, but may reduce 5HT turnover as indicated by the reduced 5HIAA. Both 5h~ and 5HIAA were depleted by the high dose. It is of some interest that, despite this difference on brain 5HT level, the effects of low and high doses on food intake (Figure IA & B) ~ r e not remarkably different. Consistent with these data, it has been reported that acute low doses (1.25-2 ~g/kg dexfenflur~nlne IP) produce no change or a small increase in brain 5HT levels, while hlgher doses have a depleting effect (3,9). This raises the question of whether low doses of dexfenfluramlne, whlch reduce 5HIAA but not 5}{9, produce anorexia by increasing brain 5HT actlVlty or by some different mechanism (I). The present data may indicate decreased 5HT activity, despite ongoing anorexia. Grant support and dexfenfluramlne from L'Instltut des Recherches Servler. Re ferences I. 2. 3. 4. 5. 6. 7. 8. 9. i0. ii. 12. 13.
N.E. ROWLAND AND J. CARLTON, Progress in Neuroblology, 24, (In press,1986). Y. ROLLAND, In: Disorders of Eating _Behavlour: A Psychoneuroendocrlne Approach, edited by E. Ferrari, (In press, 1986). S. GARA_qTINI, T. ~ I N I , C. BENDOTTI, R. INVERNIZZI AND R. SAMANIN, Appetite, 7 Suppl., 15-38, (1986). S. CACCIA, M. BALLABIO, G. GUISO, M. ROCHEqTI AND S. GARATrINI, Arch. Int. Pharmacodyn. Ther., 258, 15-28, (1982). N.E. ROWIAND AND J. CARLTON, Appetite, 7 Suppl., 71-83, (1986). J. CARLTON AND N. ROWLAND, Pharmac. B±ochem. Behav., 2_00, 739-745, (1984). N. RC~LAND, S.M. ANTELMAN AND D. KOCAN, Eur. J. Pharmaool., 8_!i, 57-66, (1982). J.E. BLUNDELL, Appetite, 7 Suppl., 39-56, (1986). N.E. ROWI2iND AND J. CARLTON, In: Disorders of Eating Behavior: A Psychoneuroendocrlne Approach, edited by E. Ferrari, (In press, 1986) . J. CARLTON AND N.E. ROWLAND, Soc. Neuroscl. Abstr., 12, (1986). A.J. STUNKARD, Life Sci., 30, 2043-2055, (1982). M. FANTINO, F. FAION AND Y. ROLLAND, Appetite, 7 Suppl., 115-126, (1986). J. CARLTON AND N.E. ROWLAND, Pharmac. Biochem. Behav., 23, 551-554, (1985).
Pergamon Journals
EFFECT OF CONTINUOUS INFUSIONS OF DEXFENFLURAMINE ON FOOD INTAKE, BODY WEIGh~ AND BRAIN AMINES IN RATS Nell E. Rowland Department of Psychology and Center for Neurobiological Sciences, University of Florlda, Gainesville, Florida 32611
(Received in final form September 18, 1986) Sunmary The present experiments describe the effects of continuous SC infusion, via osmotic minipumD, of dexfenfluramine on food intake and body welght of male and female rats. It was found that the food intake of male rats was reduced by infusions of both 3 and 6 mg/kg/day although tolerance developed within 2-4 days at the lower dose. Further, these rats showed tolerance to an acute anorectic test dose of dexfenfluramine. Body weight loss was sustained by both groups. In older (6-8 mo old) female rats, scme of which had previously nursed three litters, the anorectic effects of dexfenfluramine (3 and 6 mg/kg/day) were sustained throughout the 6 day infuslon, and weight loss was substantial. The effects did not differ between bred and virgin rats of eomparable age. The lower dose of dexfenfluramine produced no depletion of brain serotonin (5HT), although 5HIAA was reduced. Both ccmpounds were depleted by the higher dose. The 3 mg/kg/day dose, in select rat populations, may be a close model for the mode of dexfenfluramine administration to humans. Fenfluranlne, an anorectlc agent, conslsts of d-(+) and I-(-) enantiomers of which the former is more potent in reducing food intake. The present studies use the d-enanticmer of fenfluramine which will be referred to as dexfenflur6mdne. The actions of both the racemic fenfluramine mixture and dexfenfluramine have been studied extenslvely and reviewed recently (1-3). Most studies have used rats as subjects. However, the pharmacokinetlcs of fenfluramine are substantially different between rats and man. Thus, the plasma half life of fenfluraamne is only 2-4 hr in rats, but is of the order of 18-24 hr in man (1,4) . This difference produces sc[ne potentlal problems in comparisons between studies of chronic administration. In man, the treatment is usually chronic and the long half llfe of fenfluramine and its active metabolite norfenfluramlne ensure that, even if the drug is taken only once daily, the plasma levels of the agents are stable across the entire day/night cycle (I). On the other hand, in chronic administration studies in rats (e.g., 5), onceor twice-daily injections will result in w±dely-fluctuating drug levels because of the short half life. It appears in rats that this sort of chronic drug regimen is assoclated with the rapid development of tolerance to its anorectic action, whereas in humans tolerance may not be a major proble~ (1,5) . In the present study, we have exanlned whether chronic infusions ~ y osmotic minipump) of dexfenfluramlne to rats, whlch pres~nably would result in steady drug concentratlons in the body, have effects which differ frc~ daily injection regimens. 0024-3205/86 $3.00 + .00 Copyright (c) 1986 Pergamon Journals Ltd.
2582
Dexfenfluramlne
Infusions and Feeding
Vol. 39, No. 26, 1986
Our previous work also indlcates that the effects of fenfluranlne may be more marked in obese or older as opposed to lean or younger rats (5). I have thus used rats whlch fall into the former category in the present studies. In the first study male Sprague-Dawley rats were used when they had grown close to asymtotic hody weight. In the second study, both vlrghn and bred 6-8 mo old female rats were used. This study was based on our previous observation that daily fenfluramine in]ectlons caused much more anorexia and welght loss in a group of retired breeder females than in any rats that we had previously tested (5). However, age-matched virgin controls were not used in that study, so we were unable to dissoclate factors of sex, age, or prior parturitions. Method Animals and Housing Study A: Adult male Sprague-Dawley rats, purchased from ZlVlC Miller Labs (Alllson Park, PA) or bred in our lahoratory from that stock, were approximately 4-6 mo old at the time of the experLment. They were ralsed in group cages wlth ad lJ01tum access to Purina Rodent Chow pellets (#5001) and tap water. For a few weeks prior to, and throughout the experiment they were housed singly in stamnless steel cages wlth ad llblt~n access to a powdered version of the same #5001 diet, presented in a glass jar inside the cage. The vivarium lights were on from 0700-1900 hr, and the temperature was 23 _+ 2°C. Study B: Thlrty-sLx female Sprague-Dawley rats, weighing near 200 g at the start of the study, were purchased from Z1vlc M111er Labs. Eighteen of the rats were housed in palrs in solid-bottom stainless steel cages. Pine shavings were placed in the cages and were changed twlce a week. Purina Rodent Chow pellets and tap water were available ad libitt~. The remaining 18 rats were housed singly in the same klnd of solid bottom cage. Stud male rats were introduced for about 1 week. The females were checked for delivery, and lltters culled to 8-9 pups within 2 days of birth. The pups were weaned at 22 days of age, the female allowed a 1 week rest, and then a male was agaln introduced and the breeding cycle repeated. All rats retained in the study gave birth to and weaned three litters. Thus, approxlmately 4-5 mo after the start of the study, when the rats were 6-8 mo old, two age-matched groups of rats were obtained: unbred vlrgln and bred (multlparous) . At this time all rats were housed singly in regular hanging cages. All other details of care and condition are as in Study A. Procedure Measurements of food intake were made every 1-2 days throughout the study, and body weight was recorded every 3 days. After a period of baseline measures, the rats were randomly assigned to three treatn~nt groups. All were anesthetlzed with ether and an osmotlc minipump (Alzet model #2001, 7 day, 1 ul/hr) was inserted subcutaneously over the scapulae. The incision was closed wlth a wound clip and the rat returned to its cage. The mlnlpumps were preloaded with either nothing (these blanks were prevlously-used pumps that had been washed and sterilized) , or dexfenfluramine hydrochloride. In Study A, the concentration of dexfenfluranlne was 80 mg/ml or 160 mg/ml in saline vehicle, which resulted in deliver 3, of 2 or 4 mg drug 24 hr, respectively. In Study B, the concentration of dexfenflurammne was 50 mg/ml or I00 mg/ml. In both studies, these concentrations resulted in the deliver], of approxlmately 3 or 6 mg/kg/day dexfenflurammne. For sL~oliclty, these will be referred to as the low and hlgh doses. The food intake and body weight were recorded for 6 days.
Vol. 39, No. 26, 1986
Dexfenfluramine Infusions nad Feeding
2583
Tolerance Test (Study A) : At the end of the 6th full day of infusion to the male rats, the minipumps were r~moved (again using ether anesthesia), and the rats were returned to their home cage without food for 24 hr. At the end of this time they were injected acutely with dexfenfluramine (2 mg/kg, IP) and 30 min later were glven 3 weighed food pellets. Their food intake was recorded over the next hour. Neurochemlstry (Study B): At the end of the 6th full day of infusion to the female rats, at about II00 hr and with the pumps still in place, the rats were killed by rapid decapitation. The whole brains were removed, weighed and m~nediately homogenized in 4 vol~nes of ice-coldO.2N perchloric acid wlth EDTA. The homogenate was centrifuged (15,000 rl~n, I0 min, 4°C), and an aliquot of the supernatant taken for analysis of seroton±n (5h'f) and 5hydroxylndoleacetlc acid (5HIAA) by reverse phase high performance liquid chrcmatography (HPLC) with electrochemtical detection. Details of our analysis conditions have been published elsewhere (6). Peak areas were normalized to an internal standard (isoproterenol). Statistics: D a t a w e r e treated by analysis of variance and post-hoc t-tests with significance level set at 0.05. Results Study A: The food intake data are shown in Figure IA (top) . Dexfenfluramine produced a dose-related inhlbition of intake [F(2,16) = 14.28, P<0.01]. Substantial tolerance developed to the anorexia across days of treatment [maln effect of days F(5,80) = 35.17, P<0.001; dose x days Interaction F (10,80) = 3.11, P<0.01]. Nonetheless, the intake in the high dose group remained 10-20% and slgnificantly lower than that of controls even after 6 days. The low dose group recovered control intakes within 2-4 days. All of the treatments caused body welght loss over the first 3 days, with some recovery in the second 3 day segment (F ratios were signiflcant for days but not dose). This recovery was mere complete in the low dose group, whlch weighed only 6 g below their starting weight (i.e., a 1% loss) , c(mpared with the high dose group whlch was 24 g (4%) below their initial weight. Thus the days x dose interactlon was slgniflcant [F(4,32) = 3.11, P<0.05]. The results of the postfast feedlng test for tolerance are shown in Table I. The control mlnipunp group showed a near total anorexia to the acute dose of dexfenfluramine, while the nun±pimp drug-treated groups showed only a small but sign±±±cant reduction in food intake cenpared to a namve vehicleinjected group. TABI.E I Effect of Acute Dexfenfluramlne (2 n~/kg, IP) on l-hr Food Intake of Food Deprlved Male Rats That Previously Received a 6-Day M1n±pump Infusion of Dexfenfluramlne Chronlc (minlpump, 6 day) Infusion
Acute Test Injection
Food Intake (g/hr)
None Blank Dexfenfluramine (3 mg/kg/day) Dexfenfluramine (6 mg/kg/day)
Vehicle Dexfenfluramine Dexfenfluramlne Dexfenfluramine
7.7 0.1 4.8 5.2
± ± ± ±
0.4 0. I** 0.7* 0.3*
Mean ± SE for groups of 4. F(3,12) = 41.96, P<0.001 main effect. *P<0.05 differs from vehicle injected group. **P<0.01 dlffers from all others.
2584
Dexfenfluramlne Infusions and Feedlng
Vol. 39, No. 2(~, 1986
The mean (f SE) body wezght of the unbred and bred groups were 398 ± 7 and 399 ± ii g at the start of the experiment. These groups also did not differ zn theLr response to dexfenfluramine treatment, and so they have been osmbined for all data analyses. The food zntake data are shown in Figure IB (top). Both low and hlgh doses of dexfenfluramine produced a marked and sustalned decrease in intake relatzve to the control group, with the high dose only slightly more effectzve than the low dose. Main effects of drug [F(2,28) = 40.10], days [F(5,140) : 11.07], and the drug x days interaction [F(10,140) : 4.34] all were highly szgnzfzcant (Ps<0.001) , DEXFENFLURAMINE
I
MINI P U M P
B FEMALE RATS
A MALE RATS
-~ + 4 0 -
uJ <
+-+ ~ ,- ,+, o - -
/
30-
÷--~. ÷ ,0
/
o "/
--0 • ~.
30-
0
÷
>-
-< a
~
,"
20-
,I
c
10-
/
/
¢:
•
• ~.
+ --+
Blank
~--A e
DF, Low DF, High
--
1
+
20-
2
3
4
5
6
,
~
/
10-
Days
1
2
3
4
5
8 Days
+10-
+10-
÷
O-
OLu -I 0 >
o -200 m Z -30 <
-20-
~
-40-
4o
-30-
• ~
_
~
FIG. 1 Effects of contlnuous SC znfusion, by minmpumps, of dexfenfluramine hydrochloride (DF) at either 3 mg/kg/day (low dose) or 6 r0g/kg/day (high dose). Left panels A for male Sprague-Dawley rats show mean dally chow intake (top) and body weight change from that on the day of ptmlo implantat±on (bottom). Right panels B show the sane measures for female Sprague-Dawley rats. Filled symbols indicate significant (P<0.05) differences from respective blank-pump control condition (+ --- +); open symbols are not sign±ficantly dlfferent. * indlcates body we±ght change in high dose DF groups is d±fferent (P~0.05) from that of low dose DF group. For clarity, standard errors have not been included, but ranged from 1-4 g for food intake and 3-10 g for weight change. The body weight data (Figure IB, bottom) show that the high dose group lost the most weight and sustained that loss during the second 3 day infusion
Vol. 39, No. 26, 1986
Dexfenfluramine Infusions and Feeding
2585
segment. The low dose group lost less weight initially, and regained some of that weight in the second 3 days. The main effects of drug [F(2,28) = 24.93] and days [F(I,28) = 8.85] were s±gnif±cant (Ps<0.01), but their interaction was not. The l o w a n d h i g h d o s e groups had lost about 2% and 7% of initial weight, respectively, by day 6. The control rats gained 2% weight, thus the losses of the dexfenfluramine groups relative to controls are correspondingly greater. The neurochemlcal data are shown in Table 2. Rats infused wlth the low dose of dexfenfluramine showed a nonsignificant 13% reduction in brain 5HT relative to the no drug controls. Rats infused wlth the high dose of dexfenfluramine had a large (67%) reduction ~n brain 5HT. With both doses, 5HIAA was reduced to a cc~mparably large extent, and norepinephrine was reduoed in a dose-related manner. Dopamine was unaffected by either dose. TABLE II Effect of Six Day Minipump Infuslon of DexfenfluramineonWhole Brain Concentrations, Expressed as % of Blank Ptmp Controls, of Indoleamines and Catecholamlnes in Female Rats. Dose of Dexfenfluramine
Compound
3 mg/kg/day Serotonin 5-hydroxyindoleacetic acid Norepinephrine Dopamine
87 43 81 93
± ± ± ±
5 6** 3** 3
6 mg/kg/day 33 39 64 104
± ± ± ±
5** 5** 3** 8
M + SE for Ns of 6-8. **P<0.01 lower than blank-pimp controls. Discussion This experiment shows that osmotic minip~mps Whlch deliver approximately 3 or 6 mg/kg/day dexfenfluramine to rats produce inltial anorexia and weight loss. These effects are largely reversed after 6 days at the low dose, and are only partly reversed at the high dose, in male rats (Study A) . The effects of dexfenfluramine in the female rats (Study B) appear to be greater and longer lasting than for the males in Study A (compare Panels A and B of Flgure l). The plasma half life of dexfenfluramine in rats is about 3 hr, or onesixth that in humans. At a crude approximation, then, to mimic a human dose in rats a oontinuous daily infusion of about 6 x that dose would be appropriate. Thus, a cllnical daily dose of 30 mg would typically be bet%L=en 0.3 and 0.5 m~/kg and would be mimicked by a rat infusion of 2-3 mg/kg/day. The low dose (approximately 3 mg/kg/day) was chosen on the basis of this reasoning, and clearly has anorectic effects in rats. This was beth greater and longer-lasting in females than males. The females were 6-8 mo old (i.e., "middle-aged") , and no differences were noted as a result of previous parturitions. Thus, although partial tolerance developed to the anorexia and weight loss, thls was less complete in these older females than in other groups of rats given dally injections of d, l-fenfluramine (5) . In other studies (i0), we have found that a 4-week infusion of dexfenfluramine (3 r~7/kg/day) into female rats made obese by access to a palatable diet (sweet milk and cookies), does not maintain absolute weight loss
2386
Dexfenfluramlne [nfuslons and Feedlng
Vol. 39, No. 26, ]986
beyond the second week, although weight is ma~ntalned below control-infused group levels. This sort of phenc~enon often has been ±nterpreted as a "lowered set point" (11,12) , although we have pointed out that some aspects of the phenomenon are inconsistent wlth this theoretical attrlbutlon (13). We have no independent evidence that the mlnlpumps did in fact deliver a constant amount of drug throughout the 6 day perled, nor have we determined tissue or plasma levels of dexfenfluramlne or norfenfluramlne. However, the d~nonstratlon of substantial tolerance to the acute anorectlc effects of dexfenfluramine in the ptmlo-treated groups (Study A) strongly suggests that the pumps were effective. Some loss of tolerance is to be expected in light of our previous flndlng that after 8 dally injectlons of fenfluramine, tolerance to an acute anorectic probe dose is reduced as h t t l e as 2 days after the end of the chronic regimen (7). Further, both the extent and time course of the apparent anorectic tolerance (Flmure IA) mirrors qulte remarkably that seen with chronlc dally injections of either fenfluramine (5) or dexfenfluramlne (i). We do not presume that the effects of chron±c slngle in3ectlons are identical to mlnip~ps: for example, in the in3ectlon paradigm rats' food intake may vary reciprocally with the drug levels, while in the infuslon paradigm the rats may redlstribute their feedlng (we have not made any mlcrobehavioral measurements (cf. 8) to support or refute this or other possibilities) . While the mmniptmlo route may offer some advantages in terms of relevance to drug levels in humans, the gross behavioral results are not markedly different from single lnjectlon paradigms. The low dose of dexfenfluramine did not significantly deplete brain 5h'f levels, but may reduce 5HT turnover as indicated by the reduced 5HIAA. Both 5h~ and 5HIAA were depleted by the high dose. It is of some interest that, despite this difference on brain 5HT level, the effects of low and high doses on food intake (Figure IA & B) ~ r e not remarkably different. Consistent with these data, it has been reported that acute low doses (1.25-2 ~g/kg dexfenflur~nlne IP) produce no change or a small increase in brain 5HT levels, while hlgher doses have a depleting effect (3,9). This raises the question of whether low doses of dexfenfluramlne, whlch reduce 5HIAA but not 5}{9, produce anorexia by increasing brain 5HT actlVlty or by some different mechanism (I). The present data may indicate decreased 5HT activity, despite ongoing anorexia. Grant support and dexfenfluramlne from L'Instltut des Recherches Servler. Re ferences I. 2. 3. 4. 5. 6. 7. 8. 9. i0. ii. 12. 13.
N.E. ROWLAND AND J. CARLTON, Progress in Neuroblology, 24, (In press,1986). Y. ROLLAND, In: Disorders of Eating _Behavlour: A Psychoneuroendocrlne Approach, edited by E. Ferrari, (In press, 1986). S. GARA_qTINI, T. ~ I N I , C. BENDOTTI, R. INVERNIZZI AND R. SAMANIN, Appetite, 7 Suppl., 15-38, (1986). S. CACCIA, M. BALLABIO, G. GUISO, M. ROCHEqTI AND S. GARATrINI, Arch. Int. Pharmacodyn. Ther., 258, 15-28, (1982). N.E. ROWIAND AND J. CARLTON, Appetite, 7 Suppl., 71-83, (1986). J. CARLTON AND N. ROWLAND, Pharmac. B±ochem. Behav., 2_00, 739-745, (1984). N. RC~LAND, S.M. ANTELMAN AND D. KOCAN, Eur. J. Pharmaool., 8_!i, 57-66, (1982). J.E. BLUNDELL, Appetite, 7 Suppl., 39-56, (1986). N.E. ROWI2iND AND J. CARLTON, In: Disorders of Eating Behavior: A Psychoneuroendocrlne Approach, edited by E. Ferrari, (In press, 1986) . J. CARLTON AND N.E. ROWLAND, Soc. Neuroscl. Abstr., 12, (1986). A.J. STUNKARD, Life Sci., 30, 2043-2055, (1982). M. FANTINO, F. FAION AND Y. ROLLAND, Appetite, 7 Suppl., 115-126, (1986). J. CARLTON AND N.E. ROWLAND, Pharmac. Biochem. Behav., 23, 551-554, (1985).