Combined drug treatment for obesity

MISSING IMAGE

Material Information

Title:
Combined drug treatment for obesity
Physical Description:
viii, 105 leaves : ill. ; 29 cm.
Language:
English
Creator:
Roth, Jonathan D., 1972-
Publication Date:

Subjects

Subjects / Keywords:
Psychology thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Psychology -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1999.
Bibliography:
Includes bibliographical references (leaves 98-104).
Statement of Responsibility:
by Jonathan David Roth.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 030478721
oclc - 42584436
System ID:
AA00014293:00001


This item is only available as the following downloads:


Full Text












COMBINED DRUG TREATMENT FOR OBESITY


By

JONATHAN DAVID ROTH



















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1999















ACKNOWLEDGMENTS


It has been an honor and pleasure to work with my

advisor and mentor, Neil Rowland, as a graduate student at

UF. His scientific insight and dedication to research have

truly served as a source of inspiration to me. I gratefully

acknowledge Alan Spector for always going above and beyond

the call of duty to assist me in my research endeavors.

Thanks for the many thoughtful discussions and assistance

with analyses along the way; you have been a terrific

scientific role model. The rest of my committee members

(Mark Lewis, Tim Hackenberg, and Satya Kalra) also deserve a

great deal of appreciation; your input and intellectual

challenges along the way have been outstanding. A special

thanks to Mark Lewis and Randall Sakai for their generous

pharmaceutical contributions; without them some of these

studies would not have been possible.

Several individuals contributed invaluable technical

instruction and/or assistance during the course of these

experiments: Gloria Smith, thank you for all of your help

and expertise along the way. Thanks to Jessica Couch, Nicole

Cook, Roxanne Khuri and Misty Marshall for helping with data

collection.











I would like to especially thank my dear friend and

colleague Stacy Markison. Graduate school can be a long,

difficult road; great friends along the way make it a more

enriching and endurable experience. I look forward to

collaborating with you at University of Pennsylvania.

Finally, I dedicate this work to my parents, Jerome and

Beatrice Roth, whose love and support have let me see this

through to the end.

This work was partially supported by a National

Institute of Mental Health pre-doctoral fellowship (MH11982-

01A1) awarded to J. Roth.

















TABLE OF CONTENTS




ACKNOWLEDGMENTS.................
ACKNOWLEDGMENTS . . ii


ABSTRACT ........


CHAPTERS


1 INTRODUCTION ........ . ... 1

2 EFFECTS OF FEN/PHEN ON INGESTIVE BEHAVIOR 8

Introduction . . .. 8
General methods .............. 10
Animals and Housing . .. 10
Drugs and Minipumps . .. 11
Statistics . . 11
Procedures . . 11
Results . . ... ... 15
Discussion . . .. 27

3 DFEN/PHEN AND PAIR FEEDING ... . .35

Introduction . . 35
General methods .. . 36
Animals and Housing . .. 36
Drugs and Minipumps . .. .37
Statistics . . .. .37
Results ... . 38
Discussion ... . .42


4 ANORECTIC EFFICACY OF DFEN/PHEN ADDITIVE OR
SYNERGISTIC? . .


Introduction .
General methods
Animals and
Drugs .
Statistics
Procedures
Results ..
Discussion .


Housing


. 44

. 44
. 52
. 52
. 52
. 52
. 53
. 56
. 62















5 A NOVEL ANORECTIC COMBINATION: FLUOXETINE+DESIPRAMINE 71


Introduction .
General methods .
Animals and Housing .
Drugs .
Statistics .
Procedures .
Results . .
Discussion .


6 GENERAL DISCUSSION .


REFERENCES .


BIOGRAPHICAL SKETCH .


. . 71
. . 74
. . 74
. . 75
. . 75
. . 76
. . 78
. . 85


.. 91


.. 98


105















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

COMBINED DRUG TREATMENT FOR OBESITY

By

Jonathan David Roth

August 1999


Chairperson: Neil E. Rowland
Major Department: Psychology

Recently a combination of the anorectics fenfluramine

(FEN) and phentermine (PHEN) was used to treat obesity. FEN

and PHEN are believed to act via serotonin (5-HT) and

dopamine/noradrenaline (DA/NA), respectively. Despite the

recent and widespread combined off-label use of these

appetite suppressants, it has not been well established to

what extent these drugs interact.

The present experiments show: (1) the effects of once-

daily injections of dexfenfluramine (DFEN) combined with

PHEN on dessert test intake and body weight in non-deprived

rats; (2) the effects of sustained release (minipump)

DFEN/PHEN on 24 h intake and body weight and (3) the results

of an isobolographic analysis of the effects of DFEN/PHEN on

sweetened milk intake in non-deprived versus 24 h food-

deprived rats. The interactive effects of the 5-HT reuptake












inhibitor fluoxetine (FLU) and the NA reuptake inhibitor

desipramine (DMI) was also investigated in an effort to

examine other potential anorectic combinations.

Acute and chronic administration of drug combinations

had profound effects on ingestive behavior in rats.

DFEN/PHEN consistently interacted to suppress intake to a

greater extent than their individual administration. The

anorectic effects of DFEN and PHEN were greater than what

would be predicted from their individual actions; the

isobolographic analysis revealed that DFEN and PHEN interact

in a synergistic manner.

Reuptake inhibitors can also combine to suppress

intake, however, they are not as potent as the DFEN/PHEN

combination. An isobolographic analysis of the interactive

anorectic effects of FLU and DMI on acute intake tests in

non-deprived rats revealed dose-additive anorectic effects.

Similarly, in chronic, once daily injections of the

combination of FLU and DMI also appeared to suppress 24 h

food intake and reduce body weight in a dose-additive

manner.

These findings suggest that combined drug treatment for

obesity is a theoretical approach that merits future

investigation. However, future research addressing the

specificity, safety and efficacy of anorectic combinations












is warranted before combined drug treatment for obesity

becomes a clinical reality.


viii















CHAPTER 1
INTRODUCTION


Obesity can be characterized as a chronic disease,

arising in part because of positive caloric balance.

Optimally, excess weight can be lost by increasing energy

expenditure from exercise combined with dieting to decrease

caloric intake. For a variety of reasons, many individuals

suffering from obesity do not respond well to this kind of

regimen alone (e.g., failure to adhere to diet, excessive

snacking, lack of motivation). An additional or alternative

form of treatment involves long-term administration of anti-

obesity medications that suppress appetite and/or promote

thermogenesis.

Many anti-obesity drugs exert their anorectic effects

via different brain neurotransmitter pathways. Historically,

clinically used anti-obesity drugs can be subdivided into

two pharmacological classes: catecholaminergics (e.g.,

phentermine, amphetamine, amfepramone, and mazindol) and

serotonergics (e.g., fenfluramine, dexfenfluramine,

fluoxetine, and sertraline). The first class exerts effects

by promoting the release of catecholamines (norepinephrine,

NE and dopamine, DA) from nerve terminals and/or inhibiting














re-uptake processes. Similarly, serotonergic agents also

serve as indirect agonists and stimulate the release of

serotonin (5-HT) and/or prevent its re-uptake.

While diet drugs are initially effective at reducing

body weight, in the long term they have fallen short of

earning the distinction of being "magic bullets." Several

reviewers point out that individual drug treatment for

obesity does not seem to be very effective. For example,

results from a long-term, multi-center study on the effects

of dexfenfluramine (DFEN) showed that while DFEN-treated

individuals lost significantly more weight than individuals

given a placebo at six months of treatment, these

differences had diminished by the one-year mark (Guy-Grand,

Apfelbaum, Crepaldi, Gries, Lefebvre & Turnier, 1989).

Similarly, clinical trials with fluoxetine (FLU) have been

disheartening. Patients taking high doses (>60 mg/day) lost

an unimpressive 5 kg and bodyweight had almost returned to

placebo-treated control levels by 52 weeks (Goldstein,

Rampey, Roback, Wilson, Hamilton, Sayler & Tollefson, 1995).

In a study on the adjunctive effects of appetite

suppressants and behavior therapy, Atkinson, Greenway, and

Bray (1977) found that patients receiving mazindol (1

mg/day) together with behavior therapy lost little more














weight than patients receiving a placebo treatment and

behavior therapy.

In an effort to optimize and maintain drug-induced

weight loss, clinicians and researchers have begun

investigating the utility of administering combinations of

appetite suppressants, especially those that act on

different transmitter pathways. On a theoretical level this

approach is attractive for several reasons. First, feeding

behavior is mediated my multiple (and probably redundant)

neurotransmitter and neuropeptide systems. An anorectic that

operates only on an individual neurotransmitter/neuropeptide

is unlikely to have robust or long-lasting effects because

the remaining systems may be sufficient to maintain intake

and or "compensate" (e.g., up- or down-regulate) for the

affected system. Targeting multiple systems simultaneously

may less susceptible to effective compensation and as a

consequence, promote more substantial weight loss. Second,

treatment with multiple antiobesity drugs opens a range of

possibilities for drug interactions. Treatment regimens

could be developed whereby, under idealized conditions, the

drug combinations would exert additive or synergistic

anorectic and weight reducing effects. A mixture of lower

doses of each drug could be prescribed, potentially














minimizing some of the undesirable side effects associated

with larger doses of an individual drug, while maintaining

the desired effect upon appetite suppression and weight

loss. Finally, multiple drug treatment has proven to be

effective for other chronic diseases. For example,

individuals with hypertension typically take several drugs

to control their blood pressure (Atkinson, Blank, Loper,

Schumacher & Lutes, 1995). Combined drug treatment for

obesity may also be a viable option.

While the effects of individual anorectic agents on

intake and body weight in humans and non-humans have been

established, very little is known about the efficacy and

safety of these drug combinations. The most popular

combination of appetite suppressants consisted of

phentermine (PHEN), a catecholaminergic agonist, combined

with fenfluramine (FEN), a 5-HT agonist. The combination of

these two drugs was first implemented by Weintraub, Hasday,

Mushlin and Lockwood (1984) in a clinical trial (24 weeks)

where the weight reducing effects of FEN (60 mg/day) alone

or PHEN (30 mg/day) alone, were compared to FEN (30 mg/day)

+ PHEN (15 mg/day). Combining a half dose of each drug

produced similar weight reducing effects as full doses and

patients taking the combination reported fewer adverse side














effects. In a long term follow up study (Weintraub, 1992),

FEN/PHEN gained even greater acclaim when it was

demonstrated that patients reduced their initial body weight

by 15-20% after 28 weeks of treatment. There were also

improvements in many of the co-morbidities associated with

obesity; patients reported a reduction in blood pressure,

serum cholesterol and triglycerides. Following the Weintraub

(1992) study, the FEN/PHEN combination became a staple in

obesity management; it is estimated that in 1996 the total

number of prescriptions for FEN/PHEN exceeded 18 million

(Connolly, Crary, McGoon, Hensrud, Edwards, Edwards &

Schaff, 1997). It appeared that clinicians and their

patients finally had access to their long sought after

"magic bullet" of weight loss.

The FEN/PHEN combination fell from grace, however, when

there was an unexpected association of valvular heart

disease associated with the use of the fenfluramines

(Connolly et al., 1997), effects potentially exacerbated by

the addition of PHEN. Soon thereafter, the FEN and DFEN were

withdrawn from the worldwide market by their manufacturers.

The withdrawal has since left many overweight individuals

and their clinicians without effective pharmacological

tools.














In spite of the potential health risks associated with

FEN/PHEN, the combination was highly effective at promoting

weight loss as demonstrated by Weintraub and colleagues

(1981, 1992). Although FEN/PHEN will no longer be used

clinically for the treatment of obesity, it has set a new

"gold standard" for anti-obesity drug treatments. Future

anorectics and/or combinations of anorectics should be

developed that reduce body weight and food intake to the

same extent (or greater) as FEN/PHEN but without the

problematic side effects.

While the FDA does regulate the approval of the

clinical use of an individual drug, they do not regulate the

use of drug combinations. As a result, no preclinical

studies were conducted on the safety and efficacy of the

FEN/PHEN combination. Thus, we began to formally examine the

interaction between these two drugs on ingestive behavior

and body weight in rats. Contained within this dissertation

are a collection of experiments that characterize the

anorectic and weight-reducing efficacy of the FEN/PHEN

combination relative to the individual drugs and include:

(1) the effects of chronic administration (once-daily

injections) of a combination of dexfenfluramine (DFEN) and

PHEN on sweetened milk intake and body weight; (2) the














effects of chronic minipump administration of DFEN/PHEN on

24 h intake and body weight; and (3) an isobolographic

analysis of the effects of DFEN/PHEN on sweetened milk

intake in ad-libitum-fed versus food-deprived rats. As a

preliminary examination of the effectiveness of other

potential drug combinations we also: (1) conducted an

isobolographic analysis of the acute anorectic effects of a

combination of fluoxetine (FLU) and desipramine (DMI) and

(2) examined the effects of chronic administration (once

daily injections) of FLU/DMI on 24 h intake and body weight.















CHAPTER 2
EFFECTS OF FEN/PHEN ON INGESTIVE BEHAVIOR IN RATS

Introduction


From an evolutionary perspective, a reasonable

theoretical assumption is that a behavior like feeding, that

is both complex and critical to survival, is going to be

modulated by multiple redundant genes, neural pathways,

hormones and transmitters. It is therefore unlikely that any

single gene, brain region or peptide will predominate in the

control of ingestive behavior. As a consequence, searching

for a single pharmacologic agent that will miraculously and

continually turn on or off a hunger signal is futile

(Rowland, Li & Morien, 1996). Historically, pharmacotherapy

for obesity has almost exclusively employed individual

agents. It should not be surprising that on average weight

loss is, at best, modest (< 10 kg) and is reversed upon

discontinuation of drug treatment. One explanation may be

that other systems not targeted by the anorectic agent may

be sufficient to maintain intake, especially during extended

treatment. Within this context, a more fruitful approach

would appear to involve simultaneously targeting multiple














neurotransmitter systems with combinations of appetite

suppressants.

When these studies were conducted the clinical use of

FEN/PHEN was on the rise and there was a general lack of

basic research on the effects of diet drug combinations.

Thus, we opted to study the combination of DFEN/PHEN. The d-

isomer was selected because with its recent FDA approval

(April 29, 1996) we believed that it would replace the d,l

mixture in the FEN/PHEN combination. The following studies

characterize the effectiveness of the combination relative

to the individual drugs on both ingestive behavior and body

weight in rats. Experiment 2.1 shows the effects of acute

administration of individual doses of DFEN, PHEN and

DFEN/PHEN on daily intake of sweetened milk in non-deprived

rats and on change in body weight. Experiment 2.2 shows the

effects of chronic administration (via minipump) of these

drugs when given alone and in combination on 24 hr food

intake and body weight in rats. Additionally, Experiment 2.2

incorporated two additional groups of rats that received

(via minipump) the 5-HT2, agonist 1-[3-(trifluoromethyl)-

phenyl] piperazine (TFMPP) either alone or in combination

with PHEN. The purpose of this combination was to determine

whether combining PHEN with a direct agonist for












10

postsynaptic 5-HT2, receptors would be as effective as when

PHEN was combined with DFEN, an indirect 5-HT agonist

exerting both pre- and post-synaptic effects. Additionally,

a rating scale was used to quantify activity levels. In

Experiment 2.3 the selectivity of the intake suppressing

effects of the DFEN/PHEN combination was evaluated by

examining its effects on one hour water intake in 24 hr

water-deprived rats.

General Methods


Animals and Housing


Experiment 2.1 used naive adult, male (350-400 g)

Sprague-Dawley (n=22) rats. Experiment 2.2 was conducted on

experimentally naive retired female breeder (300-350 g)

Sprague-Dawley (n=36) rats. Experiment 2.3 used naive adult,

male (400-500 g) Sprague-Dawley (n=24) rats. All rats were

purchased from Harlan (Indianapolis) and housed singly in

suspended steel mesh cages. Unless otherwise specified, rats

had ad lib access to Purina Rodent Chow (#5001 powder) and

tap water. The vivarium was maintained at 23 1 C, with

lights on 0700-1900 h. All procedures were conducted during

the daytime, between 0900-1100 h. All experiments were












11

conducted in accordance with the "Principles of Laboratory

Animal Care set forth by NIH.

Druos and Minioumps


Drugs were made up daily prior to testing. All drugs

were dissolved in saline except in Experiment 2.2 where they

were dissolved in distilled water. Dexfenfluramine was a

gift from Servier (Nevilly, France). Phentermine

hydrochloride was purchased from Sigma (St. Louis, MO) and

TFMPP was obtained from RBI (Natwick, MA). Osmotic minipumps

were Alzet #2001 (Alza Corp; Palo Alto, CA) and delivered

drug at the rate of lul/h for 7 days.

Statistics


Unless otherwise specified, data were analyzed with

one- and two-way analyses of variance (ANOVA) with post-hoc

Bonferroni adjusted t-tests computed when necessary.

Significance level for all tests was set at p < 0.05. For

clarity, standard error bars have not been plotted in

Figures 1-6.

Procedures


Experiment 2.1: Effects of acute DFEN. PHEN. and DFEN/PHEN on
milk intake and body weight in non-deprived rats














Experiment 2.1 evaluated the effects of these drugs

using a "dessert" test paradigm. Dessert tests are

advantageous because rats can be tested in a non-deprived

state (intake is spontaneous), and obese individuals report

carbohydrate or sweet cravings (Rowland & Carlton, 1988).

Rats were habituated to consume a sweetened milk solution

(100 g commercial sugar and 100 g powdered milk per liter;

yielding approximately 0.8 kcal/ml) for 1 hr/day until their

intakes stabilized. Stabilization was defined as three

consecutive days of consistent intake + 1-2 ml and was

achieved after 7 days. Next, rats were assigned to four

injection groups, matched as closely as possible for

baseline intake. Injection groups included: vehicle, DFEN

(2mg/kg), PHEN (5mg/kg) and DFEN (2mg/kg) + PHEN (5mg/kg).

All drugs were dissolved in saline solution and were

administered 30 min prior to the presentation of sweetened

milk, to allow sufficient time for the onset of action of

the drugs. The half-life for DFEN delivered i.p. in rats is

2 hr (Jenden, Hinsvark & Cho, 1978). Although we know of no

available data on the half-life of i.p. PHEN, the half-life

for intravenous PHEN is 1.5 hr (Jenden, et al. 1978) and

other investigators have also used a 30 min delay following

injection (Cox & Maickel, 1972). Body weight was recorded














daily before drug administration. In total there were 11

days of drug administration and a final test day on which no

drugs were administered. Chow was available at all times

except during the milk intake tests.

Experiment 2.2: Chronic administration of DFEN. PHEN. TFMPP.
DFEN/PHEN and TFMPP/PHEN

Experiment 2.2 examined the effects of delivering these

drugs in a more continuous manner (via osmotic minipump) to

more closely approximate the extended release tablets used

clinically. Baseline bodyweight and 24 hr food intake were

collected for three days prior to drug administration. Rats

were briefly sedated (Metofane) and a 7-day minipump was

implanted subcutaneously in the back; minipumps were left in

place for the duration of the experiment. A small incision

was made in the scapular region, the pump inserted and the

incision was closed with stainless steel wound clips. The

minipumps were loaded to deliver either DFEN (3 mg/kg/day),

PHEN (10 mg/kg/day), TFMPP (1.5 mg/kg/day), PHEN

(10mg/kg/day) + DFEN (3 mg/kg/day), or PHEN (10 mg/kg/day) +

TFMPP (1.5 mg/kg/day). The dose of TFMPP was selected based

on pilot work (results not shown) in which we equated a dose

of TFMPP that resulted in a similar suppression in intake as

DFEN (3mg/kg) in a 1 hr intake test in 24 hr food deprived












14

rats. Control rats received an identical surgical procedure

except that no pump was implanted. Body weight and 24 hr

food intake were recorded daily for a total of 25 days (7

days minipump + 18 days recovery).

On the third day of testing activity levels were

quantified by a rater who was unaware of the treatment

conditions. Each rat was rated for a 10 s interval, 3 times

during the day (1130, 1230, 1330). During each 10 sec block,

activity was rated as follows: 0 "asleep" or no movement,

1 occasional normal movement (e.g., eating, drinking or

grooming), 2 2 5 sec repetitive movements (e.g.,

locomotion/sniffing/rearing) and 3 9 sec repetitive

movements. The scores across the three time blocks were

summed together as an overall activity score. The group

means for the activity scores were analyzed using

nonparametric statistics.

Experiment 2.3: Effects of DFEN. PHEN and DFEN/PHEN on water
intake in 24 hr water deprived rats

This experiment assessed the behavioral specificity of

the combination relative to the individual drugs by

comparing their ability to suppress water intake in water

deprived rats. Rats were water-deprived for 24 h with food

present. Thirty minutes before testing they were injected












15

(i.p.) with either vehicle, DFEN (2mg/kg), PHEN (5 mg/kg) or

DFEN (2mg/kg) + PHEN (5 mg/kg). Food was removed and rats

were given a 1 hr water intake test.

Results


Experiment 2.1: Effects of Acute DFEN. PHEN. and DFEN/PHEN
on Sweetened Milk Intake and Body Weight in Non-
Deprived Rats

Milk intake

Figure 2.1 shows 1 hr daily milk intake in non-deprived

rats. A group by day ANOVA revealed significant main

effects for group [F(3,18)=27.63, P<.01] and day

[F(11,198)=12.69, P<.01] and a significant interaction

[F(33,198)=4.07, P<.01]. PHEN (5 mg/kg) did not reduce

intake; in fact, there was a significant increase in intake

across days [F(11,55)=2.05, P<.05]. While there was a trend

for the control rats to increase their intake across days,

the differences were not statistically significant. DFEN

(2mg/kg) was initially effective at suppressing intake, its

effectiveness diminished following repeated administration

[F(11,55)=6.24, P<.01]; DFEN-treated rats only differed from

the vehicle group on days 1, 2, 3, 6, and 7 (all Ps<.05).

Combined administration of DFEN/PHEN was the most effective

at suppressing milk intake [F(11,55)=12.38, P<.01], and


























-0- CON

-A DFEN

--W PHEN


-*- DFEN/PHEN


1 2 3 4 5 6 7 8 9 10 11 No Drug

Days









Figure 2.1. Effects of DFEN (2 mg/kg), PHEN (5 mg/kg),
DFEN (2 mg/kg) + PHEN (5 mg/kg) on mean 1 h sweetened
milk intake in non-deprived rats. Rats treated with
DFEN/PHEN consumed significantly less than other groups
on all test days except 8-11 when they did not differ
from DFEN-treated rats.


27

24

21
c
18

E 15

aU
r 12
a)
a 9



3












17

these rats consumed significantly less than other groups on

all test days (all Ps<.05) except on days 8 through 11 when

they did not differ from DFEN-treated rats. Note that

DFEN/PHEN completely inhibited intake on the first six test

days and only in the latter several milk intake tests did

some of the DFEN/PHEN-treated rats begin to drink. On the

final day of testing, when no drug was administered, a one-

way ANOVA revealed no significant differences between groups

[F(3,18)=1.54, P=.24].

Body weight

Figure 2.2 shows percent of initial body weight plotted

across the 11 test days. A group by day ANOVA revealed

significant main effects for group [F(3,18)=7.06, P<.01] and

day [F(1l, 198)=15.65, P<.01] and a significant interaction

[F(33,198)=5.32, P<.01]. From day 5 and thereafter only the

rats treated with DFEN/PHEN weighed significantly less than

the vehicle injected group (all Ps <0.05).

Experiment 2.2: Chronic Administration of DFEN. PHEN. TFMPP.
DFEN/PHEN and TFMPP/PHEN

Food intake: minioumD administration

For clarity, results are shown in two graphs. In Figure

2.3, the effects of combined DFEN/PHEN versus DFEN and PHEN

alone on food intake are shown. In Figure 2.4, a comparison

































--- CON
--M- PHEN
-a- DFEN
--- DFEN/PHEN


1 I I I I I I I7 10
1 2 3 4 5 6 7 8 9 10 11


Days





Figure 2.2. Effects of DFEN (2 mg/kg), PHEN (5 mg/kg),
DFEN (2 mg/kg) + PHEN (5 mg/kg) on change in bodyweight
in non-deprived rats over the 11 day experiment period.
Body weights expressed in terms of % initial body weight.
From day 5 and thereafter, only rats treated with
DFEN/PHEN weighed significantly less than controls.


L-+--T -tf-f-














of the effects of DFEN/PHEN (replotted) versus TFMPP/PHEN

are shown. TFMPP alone is not plotted because it had no

effect. A group (6) by day (7) ANOVA during minipump drug

administration revealed significant main effects for group

[F(5,29)=13.35, P<.01] and day [F(6,174)=20.68, P<.01] and a

significant interaction [F(30,174)=2.24, P<.01]. One-way

ANOVAs on each test day revealed that groups significantly

differed on days 2-5 (.01
revealed that on day 2 all drugs significantly reduced 24-hr

food intake compared with the control group (all Ps<.01).

By day 3, rats treated with either PHEN, TFMPP, or TFMPP +

PHEN no longer differed from controls (P>.05). On day 4,

rats treated with DFEN no longer differed from controls

(P>.05). On days 3-5, rats treated with DFEN/PHEN consumed

significantly less than all other groups (P<.05).

Food intake: post-treatment recovery

A group (6) by day (18) ANOVA during post-treatment

indicated significant main effects for group [F(5,30)=5.16,

P<.01] and day [F(17,510)=14.09, P<.01] and a significant

interaction [F(85,510)=1.40, P<.05]. One-way ANOVAs

conducted on each day showed significant differences between

groups on days 14-21 (.01
revealed that on these days only rats treated with DFEN/PHEN




















Minipump


Recovery



/


-e- CON

--- DFEN

--- PHEN


4-- DFEN/PHEN


I I I I I I I I I I I I I I i I I I I I I I
Prel 2 3 4 5 6 7 8 9 10111213141516171819202122232425

Days






Figure 2.3. Mean daily food intake in rats with minipumps
containing DFEN (3 mg/kg/day), PHEN (10 mg/kg/day) or DFEN
(3 mg/kg/day) + PHEN (10 mg/kg/day). Minipumps expired on
day 7. On days 3-5 rats treated with DFEN/PHEN consumed
significantly less than all other groups. Note the marked
anorexia in these rats on days 1-5. By day 3, rats treated
with PHEN no longer differed from controls. By day 4, rats
treated with DFEN no longer differed from controls.





















Minipump


-- CON


--- TFMPP/PHEN

-*-- DFEN/PHEN


I I I I I I I I i I I i I I j I i I I I
Prel 2 3 4 5 6 7 8 9 10111213141516171819202122232425

Days







Figure 2.4. Mean daily food intake in rats with minipumps
containing TFMPP (1.5 mg/kg/day) + PHEN (10 mg/kg/day) or
DFEN/PHEN (same as Figure 2.3). By day 3, rats treated with
TFMPP/PHEN no longer differed from controls.


Recovery














consumed significantly more than controls (P<.05; Figure

2.3), except on day 15 when DFEN-treated rats also differed

significantly from controls (P<.05).

Body weight: minioump administration

Figures 2.5 and 2.6 depict the percent of initial body

weight plotted across days. A group (6) by day (7) ANOVA

revealed significant main effects for group [F(5,30)=20.16,

P<.01] and day [F(6,180)=5.03, P<.01] and a significant

interaction [F(30,180)=3.89, P<.01]. One way ANOVAs

conducted for each test day showed that drug significantly

affected body weight on each of the 7 days (.01
Post-hoc Bonferroni comparisons showed that rats treated

with DFEN/PHEN had significantly lower body weight compared

to controls on all days (all Ps<.01), and were the only

group to differ from controls on the final day of minipump

treatment. Additionally, from days 3-7, DFEN/PHEN-treated

rats differed significantly from all groups (all Ps<.05). On

days 6 and 7 rats treated with DFEN or PHEN no longer

differed from controls. TFMPP alone did not cause a

significant reduction in body weight on any of the days of

minipump administration. Combined treatment of TFMPP/PHEN

resulted in a significant weight loss only on day 2 (P<.01;

Figure 2.6).




















Minipump






aC


Recovery


-0-- CON


-A- DFEN


--- PHEN

-4-- DFEN/PHEN


1 2 3 4 5 6 7 8 9 10111213141516171819202122232425

Days







Figure 2.5. Change in body weight across days (expressed
as % initial body weight) in rats with minipumps containing
DFEN (3 mg/kg/day), PHEN (10 mg/kg/day) or DFEN (3 mg/kg/day)
+ PHEN (10 mg/kg/day). Note the maintenance of a markedly
reduced body weight in rats that received DFEN/PHEN for the
7 day minipump period. On days 3-7, DFEN/PHEN-treated rats
differed significantly from all groups.




















Minipump Recovery


-e- CON

--- TFMPP/PHEN

-4- DFEN/PHEN


I I I I I I I I I I I I r I I I I I I I I I I
1 2 3 4 5 6 7 8 9 10111213141516171819202122232425

Days







Figure 2.6. Change in body weight across days (expressed
as % initial body weight) in rats that received TFMPP
(1.5 mg/kg/day) + PHEN (10 mg/kg/day) and DFEN/PHEN
(replotted from Figure 2.5) via 7 day minipump. TFMPP alone
(not shown) did not cause a significant weight loss on any
of the days tested; combined treatment of TFMPP/PHEN resulted
in a significant weight loss only on day 2.














Body weight: post-treatment recovery

A group (6) by day (18) ANOVA during post-treatment

indicated a significant main effect for day

[F(17,510)=53.99, P<.01] and a significant interaction

[F(85,510)=7.10, P<.01]. One-way ANOVAs conducted on each

test day showed significant differences in body weight

between the groups for days 8-12 (P<.05). Post hoc

Bonferroni tests revealed that only the DFEN/PHEN-treated

rats had significantly lower body weight relative to

controls (all Ps<.05). These rats, however, rapidly regained

lost bodyweight and no longer differed from controls after

day 12. While there was a trend for rats that received DFEN,

PHEN and DFEN/PHEN to gain excess weight during recovery,

they did not differ from controls on any of these days.

Activity levels

Figure 2.7 shows activity levels in all groups of rats.

A Kruskal-Wallis one-way ANOVA revealed a significant effect

of treatment (P<.05). Mann-Whitney tests were then used to

compare each group to the controls. Rats treated with DFEN

and TFMPP did not differ from controls with respect to

activity score. In contrast, PHEN, DFEN/PHEN and

TFMPP/PHEN-treated rats all displayed significantly enhanced

activity relative to control rats (P<.05).


















*











6
*5 *







3 -


2







CON DFEN TFMPP PHEN DFEN/PHEN TFMPP/PHEN







Figure 2.7. Activity scores obtained from rating scale
expressed as group mean total scores + SE. Rats that
received PHEN alone or in combination exhibited
enhanced activity; these effects were not counteracted
by DFEN or TFMPP. indicates significant differences
from control group.












27

Experiment 2.3: Effects of DFEN. PHEN and DFEN/PHEN on water
intake in 24 hr water deprived rats

Figure 2.8 shows that all of the anorectics

significantly attenuated 1 hr water intake [F(3,20)=30.53,

P<.01]. All drug-treated groups drank significantly less

water than the vehicle group (all Ps<.05). Additionally,

rats that received DFEN/PHEN also consumed significantly

less water than rats that received either DFEN or PHEN alone

(all Ps<.01).

Discussion


The present experiments evaluated the effects of the

DFEN/PHEN combination on ingestive behavior and body weight

using several paradigms. In the first of these, the effects

of repeated administration of DFEN, PHEN and DFEN/PHEN on

daily sweetened milk intake in non-deprived rats were

assessed. A dessert test paradigm was selected because it is

a sensitive assay for evaluating anorectics and is thought

to reflect "snacking" behavior or intake above normal meals

(Rowland & Carlton, 1988). The particular doses of DFEN

(2mg/kg) and PHEN (5mg/kg) were chosen because they produce

a 50-80% suppression in food intake in 24 hr food-deprived

rats (Rowland & Carlton, 1986; Cox & Maickel, 1972), that

diminishes following repeated treatments. In accordance with




















18

16-

S14 *

12 -

S10

8


6

4**

2-



VEH DFEN PHEN DFEN/PHEN








Figure 2.8. One hour water intake in 24 h water-deprived
rats treated with DFEN (2 mg/kg), PHEN (5 mg/kg) or DFEN
(2 mg/kg) + PHEN (5 mg/kg). Combined treatment with
DFEN/PHEN almost completely suppresses intake. Water
intakes are expressed as mean + SE. Indicates
significant differences from controls and significant
from all other groups.












29

previous findings (Carlton & Rowland, 1984; Kandel, Doyle &

Fischman, 1975), while DFEN was initially effective at

suppressing intake, its anorectic efficacy diminished

following repeated administration. Surprisingly, PHEN did

not suppress milk intake on any of the test days. However,

when this ineffective dose of PHEN was coadministered with

DFEN it resulted in a highly effective suppression of

intake. In fact, milk intake was abolished for the first 6

test sessions and only in the latter several test days did

DFEN/PHEN-treated rats begin to consume approximately 25% of

their baseline levels.

In Experiment 2.1 rats had ad lib access to food.

While the drug-treated rats were initially expected to incur

some weight loss, it was surprising that DFEN/PHEN treated

rats maintained a reduced body weight, given the relatively

short half-lives of these anorectics. In other words,

DFEN/PHEN treated rats might have compensated for the

transient anorectic effects of these drugs by increasing

their food intake when the effects wore off. Apparently, the

combination exerted persistent effects on daily food intake

after the dessert test. A meaningful interpretation of daily

food intake in a dessert test paradigm would be difficult,

because food intake is confounded with consumption of milk.














To examine the effects of DFEN/PHEN on daily food

intake rats were implanted with 7-day minipumps. Osmotic

minipumps offer an advantage because drug delivery more

closely approximates the way these drugs are used clinically

(i.e., as extended release tablets). In contrast, acute

injections in Experiment 2.1 probably resulted in large and

rapid fluctuations of plasma drug levels. In Experiment 2.2,

chronic administration of DFEN/PHEN initially resulted in a

severe anorectic response; for the first 4 days of minipump

infusion DFEN/PHEN-treated rats consumed only 2-5 g/day.

This was coupled with a rapid, significant and sustained

weight loss of about 14-16% or 40-50 g, suggesting that

minipump delivery of DFEN/PHEN is more effective than acute

administration (Figure 2.2), when body weight decreased by

only 5-6%. However, it is also possible that female rats may

show greater plasma and brain levels of DFEN following acute

administration (Datla & Curzon, 1997). It is noteworthy that

in Experiments 2.1 and 2.2, where either adult male or

retired female breeder rats were used, DFEN/PHEN clearly

produced remarkably more weight loss than either agent

alone.

In Experiment 2.2 there was a distinct trend for some

of the groups to overeat and surpass their initial body














weight once the minipumps expired. This effect was most

prominent in the DFEN/PHEN-treated group. One explanation is

that simultaneously targeting multiple transmitter systems

also has repercussions on weight regain following treatment.

In other words, when only one system has been affected by a

chronic treatment the other systems can still compensate and

weight regain will be "normal". However, when multiple

systems have been compromised, the other systems cannot

compensate and the rebound may be exaggerated. To evaluate

whether this was the case, an additional experiment on

DFEN/PHEN was conducted incorporating a pair-fed control

group for comparison (Chapter 3).

DFEN likely exerts it anorectic effects both by

promoting the release of 5-HT and preventing its reuptake

(Samanin & Garratini, 1993). DFEN then, exerts both pre- and

post-synaptic effects. Experiment 2.2 also determined

whether combining PHEN with a direct agonist for the 5-HTc

(formerly 5-HTi; Pompeiano, Palacios, & Mengod, 1994)

postsynaptic receptor (TFMPP), a subtype thought to play a

potential role in feeding (Kennett, Dourish, & Curzon,

1987), would produce a similar enhanced anorexia and

reduction in body weight. When TFMPP (1.5mg/kg) was

administered by minipump its anorectic efficacy rapidly














diminished. While our pilot work (results not shown) using

acute injections of TFMPP (1.5 mg/kg) suggested that this

dose was as potent as the dose of DFEN (3mg/kg) in

suppressing food intake in 24 hr food deprived rats, TFMPP

did not significantly reduce body weight and its anorectic

effects were attenuated by the second test day.

Additionally, the anorectic and weight-reducing effects of

PHEN were not significantly enhanced when combined with

TFMPP. It is possible that the half life of TFMPP is less

than the half life of DFEN and hence, at these doses,

minipump administration of a 5-HT,, agonist combined with an

indirect CA agonist was not sufficient to induce even an

additive effect.

Overall, results from Experiments 2.1 and 2.2 suggest

that DFEN and PHEN may interact to exert additive or maybe

supra-additive effects with respect to reduction of intake

and bodyweight. However, these experiments only

characterized the efficacy of an individual dose combination

of DFEN/PHEN. In an effort to determine whether this truly

represented either an additive or synergistic inhibitory

effect was examined in subsequent experiments (Chapter 4)

where the effects of a range of doses of DFEN and PHEN were












33

systematically compared to a range of doses of the DFEN/PHEN

combination using an isobologram technique.

Although PHEN is not thought to have as strong an abuse

potential as amphetamine, it maintains some stimulant

effects. For example, PHEN has been shown to significantly

enhance overall activity levels (Garratini, Borroni,

Mennini, & Samanin, 1978). In contrast, FEN has a mild

sedative effect (Rowland, 1986). One rationale behind the

FEN/PHEN combination was that because each drug exerts

opposing effects on overall activity, their coadministration

should result in a cancellation of these effects (Weintraub,

1992). In Experiment 2.2 a rating scale was employed to

determine whether this occurs. While neither the number of

observations of activity, nor the activity scale employed

can be considered exhaustive, the results suggest that at

these doses the enhanced activity induced by PHEN is not

reduced by DFEN. Rats were tested in the middle of their

light cycle when they would normally be asleep or at least

inactive. Groups that received PHEN alone or as part of a

combination however, remained highly active and may have

slept less than the control group. Interestingly, while

hyperactivity and sleeplessness did occur in some patients














described by Weintraub (1992), the majority remained

unaffected.

Perhaps the simplest, albeit not the best way to test

whether an anorectic is exerting selective effects on food

intake versus other ingestive behaviors is to test its

effects on water intake in water deprived rats. For example,

norfenfluramine (2.5 mg/kg) depressed food intake in food

deprived rats by 81% and water intake in water deprived rats

by only 54% (Rowland, Antelman, & Bartness, 1985).

Experiment 2.3 determined whether DFEN/PHEN would suppress

water intake in water-deprived rats to a greater extent than

either drug alone. Individually, the anorectics suppressed

water intake by about 30%; combined, there was an almost

complete suppression of intake. These results, combined with

the general increase in activity observed in Experiment 2.2,

suggest that in rats, the DFEN/PHEN combination does not

exert selective effects.















CHAPTER 3
DFEN/PHEN AND PAIR FEEDING

Introduction


Several observations from the previous experiments

suggested that the combination of DFEN/PHEN was exerting

enhanced metabolic effects. First, while PHEN (5 mg/kg) was

ineffective at suppressing acute milk intake in non-deprived

rats (Figure 2.1), these rats still significantly reduced

their body weight for at least the first 5 test days. This

result is in accordance with Arch (1981) who reported that

PHEN (25 mg/kg/day, PO) increased energy expenditure in

normal mice (by 35%) relative to untreated controls. In our

study, PHEN appeared to amplify the weight-reducing effects

of DFEN. Second, whereas rats treated with the individual

drugs lost approximately 6% of their initial body weight,

rats treated with the DFEN/PHEN combination lost

approximately 16%. This reduction appeared to be greater

than additive, and would be consistent with enhanced

metabolic effects. Third, during the recovery period there

was a trend for rats treated with the combination to overeat

and surpass their initial body weight once the minipumps

expired.














Based on these findings, it was of interest to

determine to what extent the reduction in body weight was

due solely to either the hypophagic effects of the

combination or to enhanced metabolic effects (e.g.,

thermogenesis). One way to distinguish between these

alternatives is to compare the body weight of drug-treated

animals to a pair-fed control (PFC) group that is only

allowed to consume as much food as the drug-treated animals.

If both groups lose a comparable amount of weight then it is

likely that the weight-reducing effects of the drug can be

attributed to its hypophagic effects. On the other hand, if

drug-treated rats lose significantly more weight relative to

PFCs it is likely that the weight reducing effects of the

drug can be attributed to a combination of hypophagia and

enhanced metabolism. The present experiment compared food

intake and body weight in DFEN/PHEN-treated rats (DFEN 3

mg/kg/day + PHEN 10 mg/kg/day; osmotic minipump), PFCs and

untreated controls.

General Methods


Animals and Housing


Retired Sprague Dawley female breeders (n=6 per group;

300-350 g) were purchased from Harlan (Indianapolis, IN,

USA) and housed singly in suspended steel mesh cages.














Control rats and DFEN/PHEN-treated rats had ad-libitum

access to powdered Purina Rodent Chow (#5001). Pair-fed

controls (PFC) had restricted food access; they were only

allowed to consume as much food as DFEN/PHEN treated rats.

Tap water was available ad-lib for all groups. Food intake

(corrected for spillage) and body weight were recorded

daily.

Druas and MinioumDs


Dexfenfluramine was a gift from Servier (Neuilly,

France). Phentermine hydrochloride was purchased from Sigma

(St. Louis, MO). As in the previous experiment, osmotic

minipumps were Alzet #2001 (Alza Corp; Palo Alto, CA) and

delivered drug (DFEN 3 mg/kg/day + PHEN 10 mg/kg/day) at the

rate of ldl/h for 7 days.

Statistics


Data were analyzed with one- and two-way ANOVA with

post-hoc Bonferroni comparisons computed when necessary.

Significance level for all tests was set at p < 0.05.














Results


Food Intake: Minipump Administration


Figure 3.1 shows daily food intake in DFEN/PHEN treated

rats, PFC rats and control rats during minipump

administration (7 days) and during a recovery period. A

group (3) by day (7) ANOVA revealed significant main effects

for group [F(2,15)=27.7, P<.01], day [F(7,105)=59.4, P<.01]

and a significant interaction [F(14,105)=11.8, P<.01].

DFEN/PHEN-treated rats and PFC rats consumed significantly

less food relative to control rats from days 1-4 (all

P<0.01). From day 5 and on there were no significant

differences between the groups.

Body Weiaht: Minipump Administration


Figure 3.2 shows percent initial bodyweight in

DFEN/PHEN treated rats, PFC rats and control rats. A group

(3) by day (7) ANOVA revealed significant main effects for

group [F(2,15)=59.0, P<.01], day [F(6,90)=77.2, P<.01] and a

significant interaction [F(12,90)=18.2, P<.01]. DFEN/PHEN-

treated rats and PFC rats lost significantly more weight

relative to control rats from days 2-7 (all P<0.01).





















Minipump Recovery


T i


- DFEN/PHEN
-A- Pair-fed
--- Control


prel 2 3 4 5 6 7 8 9 101112131415161718192021

Days









Figure 3.1. Daily food intake in controls, DFEN/PHEN-
treated rats and a pair-fed control group. As in the
previous experiment DFEN/PHEN exerted powerful
anorectic effects.

















Minipump


1 2 3 4 5 6 7 8 9 10111213141516171819202122

Days





Figure 3.2. % change in body weight in controls,
DFEN/PHEN-treated rats and a pair-fed control
group. While body weight DFEN/PHEN treated rats
and pair-fed controls were significantly reduced
relative to control rats; DFEN/PHEN and pair-fed
controls did not differ on any of the test days.


Recovery














Food Intake: Post-treatment Recovery


A group (3) by day (18) ANOVA during post-treatment

indicated a significant main effect for group [F(2,15)=16.0,

P<.01]. Subsequent analyses showed that:(1) on day 12-13

DFEN/PHEN differed both from PFC rats and controls, and PFC

rats differed relative to controls, and (2) on day 14-15

DFEN/PHEN treated rats were the only significantly different

group. From day 16 and on, there were no significant

differences between the groups.

Body Weight: Post-treatment Recovery


A group (3) by day (18) ANOVA revealed a significant

main effect for day [F(14,210)=21.9, P<.01] and a

significant day by group interaction [F(28,210)=6.5, P<.01].

Subsequent analyses revealed that: (1) on day 8-10 DFEN/PHEN

and PFC rats differed relative to controls, (2) on day 11

and 12, DFEN/PHEN treated rats differed from PFC both from

PFC and from controls. From day 13 and on there were no

significant differences between the groups.














Discussion


The present experiment assessed whether the DFEN/PHEN

combination was inducing weight loss on the basis of its

anorectic effects or whether it was inducing additional

enhanced metabolic effects. In accordance with the previous

experiments with minipump delivery (Experiment 2.2),

DFEN/PHEN exerted significant anorectic effects and

DFEN/PHEN-treated rats lost between 12-14% of their initial

bodyweight. While these effects were not as profound as in

the previous study (tolerance appeared to develop more

rapidly) they were still quite impressive.

For comparison, the present experiment also included a

control group that was only allowed access to as much food

as DFEN/PHEN-treated rats consumed. While DFEN/PHEN and

pair-fed rats both lost significantly more weight than the

ad-lib fed controls, DFEN/PHEN did not induce significantly

more weight loss in comparison to the PFC-group. Although

energy expenditure was not directly measured, nor was

carcass composition quantified at the conclusion of the

study, these results do suggest that weight loss incurred by

DFEN/PHEN-treated rats can likely be attributed to its

hypophagic effects.

In Experiment 2.2 there was a trend (though not

statistically significant) for some of the groups to overeat













and surpass their initial body weight, once the minipumps

expired. This effect was most pronounced in the DFEN/PHEN

treated group. On this basis we speculated that

simultaneously targeting multiple transmitter systems might

influence weight regain following cessation of treatment.

For example, when only one system has been affected by

chronic treatment, the other systems could still compensate

and weight regain will be "normal". However, when multiple

systems are compensated, the other systems may not be able

to compensate and the rebound may be exaggerated (Roth and

Rowland, 1998). This hypothesis was not supported by the

present experiment; while DFEN/PHEN-treated rats did

temporarily overeat during recovery, their body weight did

not surpass baseline levels (though weight regain was


somewhat delayed relative to PFC rats).















CHAPTER 4
ANORECTIC EFFICACY OF DFEN/PHEN: ADDITIVE OR SYNERGISTIC?

Introduction


Although the DFEN/PHEN combination did not appear to be

exerting enhanced metabolic effects, it was consistently and

impressively effective at suppressing intake and reducing

bodyweight in a variety of test paradigms. For example, in

some of our early pilot work we found that a mixture of a

half dose of DFEN with a half dose of PHEN was at least as

effective as a full dose of either drug alone (Figure 4.1).

Experiments in Chapter 2 showed that in non-deprived rats

habituated to consume a sweetened milk solution, combining a

modestly effective dose of DFEN (2 mg/kg) with an

ineffective dose of PHEN yielded a highly effective

suppression of intake (Figure 2.1). Similarly, when DFEN (3

mg/kg/day) or PHEN (10 mg/kg/day) were administered alone

they each reduced bodyweight by around 6%; when given in

combination they yielded a 16% reduction in bodyweight

(Figure 2.3). These results are suggestive of, but do not

prove, synergistic anorectic effects.

What constitutes drug synergy? When both drugs produce

the same effect being measured (e.g., two drugs that


















8 D Vehicle

|7 -:: DFEN (1 mg/kg)

S6 i PHEN (2.5 mg/kg)

5 -- DFEN (2 mg/kg)

4 0 PHEN (5 mg/kg)


SDFEN (1 mg/kg) +
2-- PHEN (2.5 mg/kg)

1 -


Group




Figure 4.1. Individual and combined effects
of DFEN and PHEN on 2 hr food intake in 24 h
food-deprived rats. Rats (n=4 per group) were
treated with half- and full-doses of DFEN and
PHEN. An additional group received a combination
of DFEN and PHEN at half-doses; this group showed
the greatest suppression of food intake.
of intake.














suppress intake), quantifying an interaction is difficult

because the effect of one agent creates a new baseline which

must be accounted for when evaluating the effect of the

second agent (see reviews by Dews, 1976; Schuster, 1976).

Two different predictive models are widely used for

evaluating the interactions of drugs which alone produce

similar effects. The first model is called the effect-

addition model. This model predicts that the combination of

dose a of drug A and dose b of drug B will produce an effect

equal to the sum of each effect alone. Such an interaction

is termed effect-additive and deviations from additive

effects are considered to be synergistic. Because of its

intuitive nature, this is the most frequently applied model

to the study of drug interactions (Wessinger, 1986). The

major limitation of the effect-addition model is that it

assumes that the dose-effect data are linear and predicts

that the effects of combining 1 mg drug A and 1 mg drug B

would be equivalent to administering 2 mg drug A. However,

consider the more likely scenario of dose-response data

where the dose axis is plotted logarithmically, and only the

area between 20-80% Emax is linear. In this case, doubling

the dose of drug A does not result in merely an














additive effect but rather in a greater-than-additive

effect.

An alternative model is called the dose-addition model;

this model takes both dose and effect into account. An

example of this is the isobolographic ("iso" means equal and

"bol" means effect) analysis. Isobolographic analyses were

introduced by Loewe and Muischnek (1926) and have since been

widely employed in the study of interactive effects of

anticancer agents, investigations of lethal effects of

combinations of analgesics and barbiturates, and

interactions of a- and P-adrenoceptor agonists (Berenbaum,

1989; Wessinger, 1986).

The methodology of the analysis can briefly be

described with the following hypothetical example: First,

dose response curves are generated for the individual drugs

(e.g., drug A and drug B) and the ED5s (in our case

effective doses for reducing intake by 50%) values are

computed using regression analysis. Next, a fixed dose of

drug A is combined with a range of doses of drug B.

Likewise, a fixed dose of drug B is combined with a range of

doses of drug A. Again, using regression analysis the ED50

values for the combinations are computed. The results from

these experiments are then plotted on an isobologram. On the













hypothetical isobologram (Figure 4.2) the x-axis contains

increasing doses of drug A and the y-axis contains

increasing doses of drug B. The ED5 values (and the 95%

confidence intervals) for the individual drugs are plotted

on the isobologram and then connected to form a

"theoretical" dose additive line. The rationale behind this

is to determine where the ED50 value of the drug

combinations fall relative to the theoretical dose additive

line (95% CI). If the ED5, values fall: (1) above these

boundaries then the combination is infra-additive, (2)

within these boundaries then the combination is dose-

additive or (3) below the boundaries then the combination is

supra-additive.

There are several conceptual and experimental

advantages in testing for drug synergy in this manner.

First, as mentioned above, the model takes both dose and

effect into account and does not make assumptions about the

shapes of the agent's dose response curves (Berenbaum,

1989). Second, the isobologram facilitates data reduction,

providing an easily interpretable graphical representation

of the results with an unambiguous terminology (Gessner,

1995). Third, the isobolographic analysis is an empirical

model and not a mechanistic one, an advantage highlighted by



















8.0 ED50 individual drugs
-- Theoretical dose additive line
7.0

6.0 -
\Infra-additive
5.0

4.0 \

3.0

2.0

1.0\\\
Supra-additive
0.0


0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Drug A (mg/kg)






Figure 4.2. Hypothetical example of an isobolographic
plot. ED50 values (+ 95% CI) of individual drugs are
plotted to yield a dose-additive region. ED50 values
of the combinations that fall above this region are
infra-additive,within this region are dose-additive
and below this region are supra-additive.














the following example (adapted from Berenbaum, 1989).

Mechanistic approaches are based on an analysis of the

mechanisms of action of the agents and on this basis

calculating the expected effect of the combination. Imagine

that you had quantitative information (e.g., reversibility

of binding, number of binding sites) on two agents, drug A

and drug B. Based on the available information you

hypothesized that both drugs bind different sites and

constructed an equation to test the combined effects of the

agents. Hypothetically, the experiments revealed that (based

on your equation) the combined effects of the drugs were

greater than that predicted by the model and you concluded

that drug A and drug B interact in a synergistic manner. The

fundamental difficulty with this approach is that as an

increasing understanding of the agents develops, the model

is progressively modified. In the above example, imagine if

a novel, common binding site is suddenly discovered for drug

A and B. When the equation is modified to reflect this, it

is determined that drugs A and B interact in only an

additive manner. In other words, "interaction" defined in

terms of mechanism may vanish as knowledge expands. An

interaction that is said to be synergistic one day may be

said to be additive based on the same set of observed

data. Interactions should ideally be quantified on the basis












51

of the observed effects of the agents and their combination

(e.g., an empirical model) without reference to mechanisms.

The isobolographic analysis was selected as an

analytical tool to determine to what extent DFEN and PHEN

were interacting. Specifically, when two anorectic drugs

that target different neurotransmitter systems are co-

administered, will they exert infra-additive (e.g.,

antagonistic), dose-additive or greater than additive (e.g.,

synergistic) anorectic effects? In the following experiments

we systematically examined the efficacy of DFEN alone, PHEN

alone and several DFEN/PHEN combinations for suppressing 90

min sweetened milk intake. These intake data were then

plotted on isobolograms to determine whether DFEN and PHEN

were interacting in an additive or synergistic manner.

Because the anorectic effects of DFEN/PHEN have not been

evaluated under deprived versus non-deprived conditions and

in light of previous reports that PHEN and other

catecholaminergic receptor agonists may interact with

deprivation state (Papasava et al., 1985; Papasava et al.,

1980 and Carroll et al., 1981), groups of rats were tested

under ad-lib fed conditions and following 24 h food

deprivation.














Materials and methods


Animals and Housing


Adult female (350-450 g) Sprague-Dawley rats (n=144)

were purchased as retired breeders (~6 mo old) from Harlan

(Indianapolis, IN). All rats were housed singly in hanging

steel mesh cages in a vivarium maintained at 23 1 C and

illuminated from 0600-1800. Intake tests were performed

during the daytime between 0900 and 1100 hours. Unless

otherwise specificed, rats had ad-lib access to Purina Chow

pellets (5001) and tap water. All experiments were conducted

in accordance with "Principles of Laboratory Animal Care"

set forth by NIH.

Drugs


As in the previous experiments drugs were dissolved in

saline and administered i.p., 30 min prior to intake tests.

When drug combinations were tested they were mixed in the

same solution and administered in the same injection volume

as the individual drugs.

Statistics


All statistics were conducted with SYSTAT for Windows

software package. Data were first analyzed with ANOVA and

Bonferroni adjusted t-tests were applied when appropriate.

Significance level for all tests was set at P < 0.05. Next,














the ED50 values and their 95% CI were calculated from the

dose response curves of the individual drugs and the

combinations using nonlinear regression. The ED50 values (+

95% CI's) for the individual drugs were plotted on an

isobolographic plot. A theoretical dose additive region was

constructed by connected the ED5, values for the individual

drugs by a dose-additive line and by connecting the the 95%

confidence intervals to form boundaries around the line. The

rationale behind this was to determine where the ED,0 value

of the drug combinations fell relative to the theoretical

dose additive line (95% CI). If the ED50 values for the

combination fell: (1) above these boundaries region then the

combination was considered infra-additive, (2) within these

boundaries then the combination was dose-additive or (3)

below the boundaries then the combination was supra-

additive.

Procedures


Experiment 4.1: Isobolographic analysis of the DFEN/PHEN
combination on milk intake in non-deprived rats


Rats (n = 70) were adapted to drinking a sweetened milk

solution (100 g commercial sugar and 100 g powdered milk per

liter) in the home cage for 90 min a day until their intakes

stabilized. Stabilization was defined as three consecutive












54

days of consistent intake 2-3 ml. Once stable intakes were

achieved (after 10 days) rats were divided into injection

groups (n = 5-6 per group) that were matched as closely as

possible for baseline intake. On test day (day 11), rats

were injected with either saline, DFEN (0.5, 1.0, 1.5, 2.5

and 3.5 mg/kg) or PHEN (1.0, 3.0, 5.0, 7.0, 10.0 and 13.0

mg/kg). Thirty minutes later, sweetened milk solution was

presented and total intake ( 0.5 ml) was recorded after 90

min. On days 12-16 rats had daily access to sweetened milk

but did not receive any drugs. This time period enabled

their intakes to restabilize. Rats were then divided into

groups that were balanced as closely as possible both for

previous injection group and for their new baseline intakes.

On day 17 rats were injected with one of the following drug

combinations 30 min before testing: a fixed dose of PHEN

(2.5 mg/kg) was combined with a range of doses of DFEN

(0.25, 0.5, 0.75, 1, and 1.5 mg/kg), and a fixed dose of

DFEN (0.5 mg/kg) was combined with a range of doses of PHEN

(1, 2, 3, 4, 6 mg/kg). In sum, on separate occasions, all

rats were injected with drug and tested twice once with an

individual drug (on day 11) and once with a drug combination

(on day 17).














Experiment 4.2: Isobolographic analysis of the DFEN/PHEN
combination on milk intake in food-deprived rats


As in the previous experiment, naive rats (n = 72) were

adapted to drinking a sweetened milk solution (100 g

commercial sugar and 100 g powdered milk per liter) in the

home cage for 90 min a day. The major difference between

Experiments 4.1 and 4.2 was that when rats were being

habituated to consuming sweetened milk in Experiment 4.2

they were tested following 24 h food (but not water)

deprivation on days 8, 11 and 14. The rationale was to

obtain a stable baseline of food-deprived sweetened milk

intake against which to measure the effects of the

individual drugs and the combinations. Food-deprived milk

intake on day 14 served as baseline intake for statistical

comparisons. The days of non-deprived milk intake

interspersed between the periods of 24 h food deprivation

allowed for the rats' bodyweight to return to baseline.

Thus, on day 17, 24 h food deprived rats (n=5-6 per group)

were injected with either DFEN (0.5, 1.0, 1.5, 3.0 and 4.0

mg/kg) or PHEN (1.0, 3.0, 5.0, 7.0, 10.0 and 13.0 mg/kg). On

day 20, 24 h food deprived rats were injected with either a

fixed dose of PHEN (2.5 mg/kg) combined with a range of

doses of DFEN (0.5, 0.75, 1 and 1.5 mg/kg), or a fixed dose

of DFEN (0.5 mg/kg) combined with a range of doses of PHEN

(1.0, 2.0, 3.0, 4.0, and 6.0 mg/kg).














Results


Experiment 4.1: Isobolographic analysis of the DFEN/PHEN
combination on milk intake in non-deprived rats


Mean baseline intake was 21.3 0.5 ml over a 90 min

session. Figure 4.3 shows the dose response curves for the

anorectic effects of DFEN, PHEN and the DFEN/PHEN

combinations on milk intake in non-deprived rats. Intake

data were analyzed as follows: (1) milk intakes were

transformed by dividing test intake by baseline intake, (2)

analysis of variance and post-hoc tests were performed on

these scores, and (3) non-linear regression was used to fit

curves to the dose-response data to determine one-half

maximum asymptote (ED5,), the asymptotic standard error and

the 95% confidence intervals (CI). These results (except

asymptotic standard errors) are summarized in Table 4.1.

To determine whether DFEN and PHEN interacted in an

additive or supra-additive manner, these results were

plotted on an isobolographic plot. Drug interactions were

considered to be significantly different from dose-additive

if the EDs, values for the drug combinations did not overlap

the 95% CI of the theoretical dose additive line. Figure 4.4

shows that DFEN and PHEN appear to interact in a supra-

additive manner; the EDs values for the two combinations













1.6
1.4
1.2
1.0
0.8
0. 6
0.4
0.2








1.6-
1.4 -
1.2 -
1.0 -
0.8-
0.6-
0.4-
0.2-


I I I I I
0 1 2 3 4
Dexfenfluramine (mg/kg)
combined with
Phentermine 2.5 mg/kg


I I I T I I I I
0 2 4 6 8 10 12 14
Phentermine (mg/kg)
combined with
Dexfenfluramine 0.5 mg/kg


Figure 4.3. Dose response curves for DFEN, PHEN, and
DFEN/PHEN combinations. Rats were non-deprived at time
of testing. ED50 values were estimated using non-linear
regression.


0 1 2 3 4 0 2 4 6 8 10 12 14
Dexfenfluramine (mg/kg) Phentermine (mg/kg)
- :^

- --

- ,^K





















8.0- ED, individual drugs
A ED,0 drug combinations
7.0 -- Dose additive line

S6.0 \
S0


S4.0
-Hl
i 3.0

a 2.0
i.o \
1.0\ \

0.0-


0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

Dexfenfluramine (mg/kg)










Figure 4.4. Isobolographic plot for the anorectic
effects of the DFEN/PHEN combination in non-deprived
rats drinking sweetened milk. Notice that the doses
of both of the combinations fall outside of the dose-
additive region.














both fall outside of the 95% confidence intervals of the

dose-additive line.



Table 4.1. Effects of DFEN and PHEN alone and in combination
on 90 min sweetened milk intake in non-deprived rats.


Drug and doses ANOVA EDs5 (95% CI)
(mg/kg) mg/kg

DFEN (0.5-3.5) F(5,30)=12.2, 2.5 (1.3-3.6)
P<0.01

PHEN (1.0-13) F(6,35)=30.4, 6.3 (5.3-7.3)
P<0.01

PHEN (2.5) + F(5,33)=25.4, PHEN (2.5) + DFEN
DFEN (0.25-1.5) P<0.01 0.4 (0.2-0.7)

DFEN (0.5) + F(5,33)=23.3, DFEN (0.5) + PHEN
PHEN (1.0-6.0) P<0.01 2.3 (1.6-3.0)




Experiment 4.2: Isobolographic analysis of the DFEN/PHEN on
milk intake in 24 h food-deprived rats

During adaptation rats consumed significantly [T(69)=

7.5, P<0.01] more milk following 24 h food deprivation

(mean=25.3 ml/90 min) compared with under non-deprived

conditions in the previous study (x=21.3 ml/90 min).

Intake data were analyzed in an identical manner for

Experiment 4.2. Figure 4.5 shows the dose response curves

for DFEN, PHEN and the DFEN/PHEN combinations in 24 h food

deprived rats. These results are summarized in Table 4.2.















1.6-

1.4 -

1.2 -

1.0 -

0.8 -

0.6-

0.4 -

0.2 -


0 1 2 3 4 0 2 4 6 8 10 12 14
0 1 2 3 4 0 2 4 6 8 10 12 14


Dexfenfluramine (mg/kg)


Phentermine (mg/kg)


0 1 2 3 4 0 2 4 6 8 10 12 14


Dexfenfluramine (mg/kg)
combined with
Phentermine 2.5 mg/kg


Phentermine (mg/kg)
combined with
Dexfenfluramine 0.5 mg/kg


Figure 4.5. Dose response curves for DFEN, PHEN, and
DFEN/PHEN combinations. Rats were 24 h food deprived
at time of testing. EDs, values were estimated using
non-linear regression.












61

Table 4.2. Effects of DFEN and PHEN alone and in combination
on 90 min sweetened milk intake in 24 h food-deprived rats.

Drug and doses ANOVA ED,5 (95% CI)
(mg/kg) mg/kg

DFEN (0.5-4.0) F(5,29)=11.7, 2.4 (2.1-2.7)
P<0.01
PHEN (1.0-7.0) F(6,31)=ll.1, 4.9 (4.1-5.7)
P<0.01
PHEN (2.5) + F(4,26)=8.9, PHEN (2.5) + DFEN
DFEN (0.75-1.5) P<0.01 0.8 (0.5-1.2)
DFEN (0.5) + F(5,34)=15.4, DFEN (0.5) + PHEN
PHEN (1.0-6.0) P<0.01 2.4 (1.5-3.3)



When rats were tested following 24 h food deprivation

it required approximately 2.4 mg/kg of DFEN alone to

suppress 90 min sweetened milk intake by 50%. When DFEN was

combined with PHEN (2.5 mg/kg), the EDso value was reduced

to DFEN 0.8 mg/kg. These results are similar to those

obtained when rats were tested under non-deprived

conditions. In contrast, it required less PHEN to suppress

90 min sweetened milk intake in food-deprived rats (4.9

mg/kg) relative to non-deprived rats (6.3 mg/kg; Experiment

4.1). When a range of doses of PHEN were combined with a

fixed dose of DFEN (0.5 mg/kg) the EDs, was reduced to PHEN

2.4 mg/kg. These ED50 values were also almost identical to

those obtained in non-deprived rats; the DFEN/PHEN














combination appears to be equipotent in non-deprived and

deprived animals.

As in Experiment 4.1, these results were plotted on an

isobologram (Figure 4.6). In this case, the ED5s value for

one of the drug combinations fell outside of the 95%

confidence intervals of the dose-additive line, while the

EDso value for the other drug combination fell just within

the 95% confidence intervals.

Discussion


Recently, the anorectic efficacy of DFEN, PHEN and the

DFEN/PHEN combination were compared under acute and chronic

conditions in multiple paradigms of stimulated intake (Roth

and Rowland, 1998). In acute intake tests, combining a full

dose of DFEN with a full dose of PHEN produced at least

double the anorectic effect of the individual drugs.

Additionally, it took a considerably longer period of time

for the rats to develop anorectic tolerance to the DFEN/PHEN

combination. The present experiments extended these single

dose studies by comparing the anorectic effects of a range

of doses of the individual drugs and combinations using an

isobolographic analysis.

When both drugs are capable of producing the effect

under study (e.g., both DFEN and PHEN produce anorexia),

single dose studies provide limited information regarding





















8.0

7.0

, 6.0-
-
01
E Sn


0.0 1.0 2.0 3.0


4.0 5.0 6.0 7.0 8.0


Dexfenfluramine (mg/kg)










Figure 4.6. Isobolographic plot for the anorectic
effects of the DFEN/PHEN combination in 24 h food
deprived rats drinking sweetened milk. One of the
combinations fell outside of the dose-additive
region and the other combination fell just within
the dose-additive region.


* ED0, individual drugs
A EDs0 drug combinations
-- Dose additive line














the magnitude of a drug interaction. In other words, it

becomes difficult to interpret the degree of the interaction

as additive or synergistic because the expected outcome of

the combination is always expected to exceed either of the

individual drugs. The isobolographic analysis circumvents

this difficulty by comparing equieffective doses of the

individual drugs and the combinations; dose and effect are

taken into account (Gessner, 1974; Wessinger, 1986;

Berenbaum, 1989).

Do combinations of anorectic drugs that target

different neurotransmitter systems exert additive or

synergistic effects? The results from the isobolographic

analysis suggest that, at least under acute conditions, DFEN

and PHEN do exert synergistic anorectic effects. When these

drugs were administered to non-deprived rats consuming a

highly palatable milk solution the EDo0 values for both

DFEN/PHEN combinations fell outside the 95% CI of the dose

additive line, within the synergy region of the plot.

Similarly, when rats were tested following 24 h food

deprivation the EDso for one combination fell below the 95%

CI of the dose additive line and the other combination fell

just within the boundaries of the 95% CI. In other words,

the DFEN/PHEN combination produced anorectic effects that












65

were greater than those predicted by their separate effects,

at least at these doses.

Wellman et al. (1995) used an isobolographic analysis

to examine the combined effects of FEN (a serotonergic) and

phenylpropanolamine (an a-1 adrenoceptor agonist) on 60 min

food intake in 16 h food-deprived rats. Their results

supported a dose-additive interaction between FEN and

phenylpropanolamine. Recently, Bhasvar, Watkins and Young

(1998) conducted an isobolographic analysis on the combined

effects of (i.p) cholecystokinin (CCK) and amylin (a peptide

hormone co-secreted with insulin) and demonstrated that

these satiety peptides may interact to synergistically

suppress intake. Collectively, this lends further support to

the notion that simultaneously targeting multiple

transmitter or peptide systems with various agents may be a

promising approach that will probably yield at least dose-

additive anorectic effects. The constituents of an ideal

drug combination could conceivably be prescribed in low

enough doses that would still yield powerful hypophagic

effects but have a lower incidence of adverse side effects

relative to larger doses of individual drugs.

It is noteworthy that in the present experiments and in

other published work (e.g., Wellman et al., 1995 and Foltin

et al., 1983), the ED,, values for the mixtures were













obtained from dose response curves that were generated by

combining a fixed dose of one drug with a range of doses of

the other drug. A more valid method for obtaining EDs,

values for the combinations is for each point on the

response curve to reflect a fixed ratio of drug A to drug B.

For a more in-depth explanation on the use of fixed-ratio

drug combinations in the study of drug interactions we refer

the reader to Tallarida and Raffa (1996). Future analyses on

the interactive effects of anorectic agents should consider

this statistical methodology.

In a pilot experiment on the effects of PHEN on 90 min

chow intake in 24 h food deprived rats the EDso for PHEN was

estimated to be 3.3 mg/kg and intake was suppressed by -90%

at 5 mg/kg (Roth and Rowland, unpublished observations).

However, recall from Experiment 2.1 that a dose of 5 mg/kg

did not significantly attenuate 1 h sweetened milk intake in

non-deprived rats. Based on this it was hypothesized that

PHEN might be a more potent anorectic under conditions of

food deprivation. However, comparing chow intake in food

deprived rats to milk intake in non-deprived rats may be

confounded because both deprivation state and the properties

of the food are different. The present experiments attempted

to hold the food stimulus constant and rats were tested

under deprived versus non-deprived conditions. Ad-lib fed

rats consumed ~21 ml/90 min while 24 h food deprived rats














consumed -25 ml/90 min. Because intake is spontaneous and

robust under non-deprived conditions, and deprivation does

not excessively elevate consumption, comparisons between

non-deprived and deprived sweetened milk intake may prove

useful in evaluating whether an agent interacts with

deprivation state. In this case, deprivation state did not

appear to have a major impact on the efficacy of the

anorectics. The dose response curves for DFEN and the

DFEN/PHEN combinations were almost identical across

deprivation state. There was a slight shift in the dose

response curve for PHEN, such that the ED50 95% CI value

was 4.9(4.1-5.7) in deprived rats versus 6.3 (5.3-7.3) mg/kg

in nondeprived rats. PHEN appears to be exert somewhat more

effective anorectic effects under food deprived conditions,

although the 95% confidence intervals do overlap. Consistent

with these findings, the reinforcement efficacy of PHEN has

also been shown to vary as a function of deprivation state

(Papasava et al. 1981; Papasava et al. 1985). At 80% of

free-feeding weight, rats will self-administer significant

amounts of phentermine while their free-feeding counterparts

do not.

What are the neurochemical mechanisms that mediate the

anorectic effects of the DFEN/PHEN combination? DFEN is an

indirect 5-HT agonist that promotes the release of 5-HT and

prevents its reuptake (Samanin and Garratini, 1993). DFEN is














metabolized in vivo into a highly active metabolite,

dexnorfenfluramine (DNFEN). DNFEN appears to have both

direct agonist (5-HT2c) and indirect effects on brain 5-HT

systems (Curzon et al., 1997). The present experiments used

short-term (90 min) intake tests and most of the anorexia

is probably attributable to DFEN's actions (although DNFEN's

effects were not ruled out). Nonetheless, it is important to

point out that under conditions of long-term administration,

it is likely that PHEN would interact both with DNFEN and

DFEN. Whether a combination of DNFEN/PHEN (e.g., more of a

direct 5-HT agonist relative to DFEN with a

catecholaminergic) would be sufficient to yield synergistic

anorectic effects remains to be examined.

PHEN's anorectic effects are thought to be mediated by

noradrenaline and dopamine. 6-hydroxydopamine lesions have

been shown to strongly attenuate the anorectic effects of

lower doses of PHEN (2.5-5.0 mg/kg), but they do not reduce

the anorectic effects of higher doses (10 mg/kg; Samanin et

al., 1975). One explanation may be that high doses of PHEN

also elevate levels of brain 5-HT and this is enough to

support PHEN's anorectic effects in lesioned rats. Indeed,

recent studies using in vivo microdialysis suggest that PHEN

enhances extracellular levels of dopamine and 5-HT in the

nucleus accumbens (Shoaib et al., 1997), but only dopamine














in the striatum (Balcioglu and Wurtman, 1998). In the

present experiments, a potential neurochemical correlate for

DFEN/PHEN's synergistic anorectic effects may involve

enhanced brain 5-HT levels.

Another possible explanation for the synergistic

effects of the DFEN/PHEN combination may be related to their

effects on meal pattern. Catecholaminergic and serotonergic

drugs have been shown to differentially effect the profile

of intake (Asin, Davis and Bednarz, 1992). A microstructural

analysis of ingestive behavior revealed that drugs that

enhance 5-HT (FEN and FLU) seem to reduce intake by reducing

the "size" of the meal whereas drugs that act via

catecholaminergic systems (amphetamine and

phenylpropanolamine) reduce intake by reducing the "number"

of meals. As a "behavioral corollary" to the notion that

targeting multiple transmitters might be particularly

efficacious is that it may also be easier to suppress

ingestive behavior by targeting both its initiation (e.g.,

with PHEN) and its termination (e.g., with DFEN).

While the d-isomer of FEN is believed to be responsible

for the anorectic properties of the compound (reuptake

inhibition and 5-HT release; Garratini et al. 1978; Samanin

et al., 1972), the levo-isomer is not without effects.

Levofenfenfluramine has been shown to have anti-dopaminergic

properties and at higher doses increase DA concentration in












70

striatal dialysates (Bettini, Ceci, Spinelli and Samanin,

1987). To what extent using the d-isomer versus the racemic

mixture (d,l-FEN) impacts upon the behavioral synergy

observed in the present experiments remains to be

determined.















CHAPTER 5
A NOVEL ANORECTIC COMBINATION: FLUOXETINE + DESIPRAMINE

Introduction


Because the fenfluramines have recently been implicated

in primary pulmonary hypertension (PPH), the combination of

FEN and PHEN is no longer clinically viable. It remains

important to find safe and effective alternatives to replace

FEN/PHEN. Serotonin reuptake inhibitors (SRIs) have been

prescribed as antidepressant drugs for many years (Wong,

Bymaster & Engleman, 1995); more recently they have been

proposed as potentially useful antiobesity agents (Samanin &

Garattini, 1993). One advantage of SRI's over 5-HT releasing

agents like the fenfluramines, is that SRI's have not been

implicated in PPH (Heal, Aspley, Prow, Jackson, Martin &

Cheetham, 1998). Thus, they remain potential candidates for

future diet drug combinations.

Clinical trials with SRI's, however, have been as

disheartening as most single drug treatments for obesity.

For example, patients taking high doses of FLU (60-80

mg/day) lost an unimpressive 5 kg and body weight returned

to control levels by 52 weeks (Goldstein et al., 1995). It

is noteworthy that doses of FLU (>60 mg/day) required to












72

induce significant weight loss are considerably higher than

those that seem to be effective for alleviating symptoms of

depression (20 mg/day; Wong et al., 1995). While FLU is

well-tolerated in the antidepressive dose range (20 mg/day),

as the dose is increased to 60-80 mg/day adverse side

effects like nausea (reported in 25% of patients; Wong et

al., 1995; Goldstein et al., 1995), insomnia, anxiety and

urinary abnormalities begin to emerge.

A recent study in non-humans (Jackson, Needham,

Hutchins, Mazurkiewicz & Heal, 1997), however, suggests that

FLU may hold utility as part of an anorectic combination.

The mixture of nisoxetine (NIS; 30 mg/kg), a selective

norepinephrine reuptake inhibitor (Wong & Bymaster, 1976)

and FLU (30 mg/kg), was highly effective at suppressing

spontaneous intake of powdered chow at 2 h and 8 h into the

dark cycle. Neither FLU (3, 10 and 30 mg/kg) nor NIS (3, 10,

and 30 mg/kg) alone significantly reduced intake at either

time point. Although these investigators administered only

individual doses, these findings are compelling because they

document what appears to be at least an additive and

possibly a synergistic interaction between reuptake

inhibitors for noradrenaline and 5-HT.












73

On the basis of these findings we decided to conduct an

isobolographic analysis on the acute anorectic efficacy of a

combination of FLU with desipramine (DMI), a clinically used

tricyclic anti-depressant with predominantly NE inhibiting

properties (Snagdee & Franz, 1979; Hytell, 1982).

Additionally, because the effects of FLU/DMI on ingestive

behavior had never been evaluated, a test for behavioral

specificity was included. This was done by administering

doses of FLU, DMI and FLU + DMI that reduced milk intake by

50% (EDs, values from the dose response curves) to rats that

were 24 h water deprived and given a water intake test.

The combination of DMI and FLU is not without

precedent; these agents have been co-administered to

individuals suffering from depression and have failed to

respond to treatment with individual drugs (Weilburg,

Rosenbaum, Biederman, Sachs, Pollack & Kelly, 1989; Nelson,

Mazure, Bowers & Jatlow, 1991). With combined antidepressant

treatment, therapeutic benefits were recognized more rapidly

and to a greater extent relative to individual drug

treatment. The effectiveness of the combination was

attributed to its ability to more rapidly down- regulate P-

adrenoceptors relative to the individual drugs (a putative

mechanism of action of antidepressants). Indeed, this














hypothesis was substantiated by several experimenters

examining down-regulation of cortical P-adrenoceptors in

rats (Goodnough & Baker, 1994; Baron, Ogden, Siegel,

Stegeman, Ursillo & Dudley, 1994) treated with DMI, FLU or

FLU/DMI. Repeated administration of either FLU (Rowland,

Antelman & Kocan, 1982; McGuirk, Muscat & Willner, 1992) or

DMI (Nobrega & Coscina, 1987; Durcan, McWIlliam, Campbell,

Neale & Dunn, 1988) alone has been shown to reduce long-term

intake and body weight in rats. However, neither of the

aforementioned reports that quantified the effects of the

mixture of FLU + DMI on P-adrenoceptors examined ingestive

behavior. Thus, in the present experiments the effects of

repeated, once-daily administration of FLU, DMI, and the

combination of FLU + DMI on 24 h food intake and body weight

were evaluated.


Materials and Methods


Animals and Housina


Retired Sprague Dawley female breeders (300-350 g) were

purchased from Harlan (Indianapolis, USA) and housed singly

in suspended steel mesh cages. Rats in Experiment 5.1 (n=60)

had ad-lib access to pelleted Purina Rodent Chow (#5001) and












75

tap water, except during milk intake tests when their water

bottles were removed and replaced with graduated tubes

containing sweetened milk. Rats in Experiment 5.2 (n=30)

were 24 h water deprived at the time of testing. Food was

available at all times. Rats (n=24) in Experiment 5.3 had

ad-lib access to powdered Purina Rodent Chow (#5001) and tap

water. The vivarium was maintained at 23 10C, with lights

on 0700-1900 hours. All procedures were conducted during the

daytime, between 0900-1100 hours.

Drugs


Drugs were mixed daily prior to testing. All drugs were

dissolved in distilled water. Desipramine hydrochloride was

purchased from Sigma (St Louis, MO., USA) and FLU

hydrochloride was a gift from Eli Lilly (Indianapolis, IN).

Dosages refer to the weight of the salt.

Statistical Analysis


All statistics were conducted with SYSTAT for Windows

software package. Data from Experiment 5.1 were first

analyzed with ANOVA and post-hoc Bonferroni comparisons

computed when appropriate. Next, the ED,, values and their

95% confidence intervals (CI) were calculated from the dose

response curves of the individual drugs and the combination














using nonlinear regression. Data from Experiment 5.2 were

analyzed by ANOVA with post-hoc tests where appropriate.

Data from Experiment 5.3 were analyzed with one- and two-way

ANOVA, with post-hoc Bonferroni comparisons computed when

necessary. Significance level for all tests was set at

P<0.05.

Procedures


Experiment 5.1: Isoboloaraohic analysis of FLU + DMI

Rats were adapted to consume a sweetened milk solution

in their home cages for 30 min a day until their intakes

stabilized, as in the previous experiments. Once stable

intake was achieved (three days of 2-3 ml) they were

divided into injection groups (n = 6 per group) that were

matched for baseline intake. On test day, rats were injected

with either saline, FLU (2.0, 4.0, 8.0 or 16.0 mg/kg) or DMI

(2.0, 4.0, 8.0 or 16.0 mg/kg). Thirty minutes later,

sweetened milk was presented and total intake ( 0.5 ml) was

recorded after 30 min. On days 12-16 intakes were allowed to

restabilize; rats were not injected with any drugs during

this period. On day 17 rats were injected with a range of

fixed dose ratios of the combination. Half the rats were

treated with doses of DMI:FLU in a 1:2 ratio (0.5:1.0,












77

1.0:2.0, 2.0:4.0 or 4.0:8.0 mg/kg) while the other half were

treated with doses of DMI:FLU in a 2:1 ratio (1.0:0.5,

2.0:1.0, 4.0:2.0 or 8.0:4.0 mg/kg). Thus, all rats received

only two injections once with an individual drug (on day

11) and once with a ratio of the combination (on day 17).

Experiment 5.2: Effects of FLU, DMI and FLU/DMI on water
intake in 24 h water deprived rats

Rats were water but not food deprived for 24 h. Rats

(n=6 per group) were injected with doses of the individual

drugs and combinations that produced a 50% suppression of

milk intake in non-deprived rats (Experiment 5.1). These

equated doses included: FLU (3.6 mg/kg), DMI (5.5 mg/kg),

Ratio 1 = FLU (1.25 mg/kg) + DMI (2.5 mg/kg) and Ratio 2 =

FLU (2.35 mg/kg) + DMI (1.16 mg/kg).

Experiment 5.3: Chronic administration of FLU. DMI and
FLU+DMI

Baseline bodyweight and 24 h intake were collected for

3 days prior to drug administration. Rats were assigned to

one of four injection groups (n=6 per group) matched for

baseline intake and body weight. Injection groups included:

vehicle, FLU (5 mg/kg), DMI (5 mg/kg) and FLU (5 mg/kg) +

DMI (5 mg/kg). Rats were injected once daily and intake of

powdered chow (corrected for spillage) and bodyweight were

recorded.














Results


Experiment 5.1: Isobolographic analysis of FLU + DMI


Data from the effects of the individual drugs on non-

deprived milk intake were analyzed in an identical manner to

the previous experiment:(1) 30 min milk intakes were

transformed by dividing test intake by baseline intake, (2)

ANOVA and post-hoc tests were performed on these scores, and

(3) non-linear regression was used to fit curves to the

dose-response data to determine one-half maximum asymptote

(ED0o), the asymptotic standard error and the 95% confidence

intervals (CI). The only differences for analysis of the

fixed dose ratios were that: (1) ANOVA and nonlinear

regression were performed on the total dose of the mixture

(e.g., 1.5, 3.0, 6.0 and 12.0 mg/kg) and (2) the EDso

predicted off of this curve was then divided into the

appropriate ratio. For example, an ED50 of 6.0 mg/kg (total

dose) predicted from the dose response curve for DMI:FLU

administered in a 1:2 ratio would reflect a mixture of 2.0

mg/kg FLU and 4.0 mg/kg DMI.

FLU, DMI and the FLU + DMI combination significantly

suppressed milk intake in nondeprived-rats in a dose

dependent manner. The dose response curves are shown in












79

Figure 5.1 and the results are summarized in Table 5.1. The

isobolographic analysis (shown in Figure 5.2) revealed that

when co-administered, FLU and DMI interact in a dose-

additive manner; the ED,0 values for both combinations fall

within the 95% confidence intervals of the dose additive

line.



Table 5.1. Effects of FLU and DMI alone and administered in
a 1:2 and 2:1 ratio on 30 min sweetened milk intake in non-
deprived rats.

Drug and doses ANOVA ED50 (95% CI)
(mg/kg) mg/kg

DMI (2.0-16.0) F(4,31)=10.5, 5.5 (3.8-7.3)
P<0.01

FLU (2.0-16.0) F(4,31)=13.2, 3.6 (1.5-5.7)
P<0.01

DMI:FLU 1:2 ratio F(4,25)=17.5, 3.8 (2.5-5.1)
P<0.01

DMI:FLU 2:1 ratio F(4,25)=26.1, 3.5 (2.0-5.0)
P<0.01



Experiment 5.2: Effects of FLU. DMI and FLU + DMI on water
intake in water deprived rats

Figure 5.3 shows the effects of FLU, DMI and the two

combinations of FLU/DMI on 1 h water intake in 24 h water

deprived rats. ANOVA revealed a significant effect of drug

[F(4,25)=4.1, P<.05)]. Post-hoc tests revealed that DMI,















1.6

1.4

I 1.2

1.0

0.8

0.6

0.4

0.2


0.0 2.0 4.0 8.0 16.0 0.0 2.0 4.0 8.0 16.0

Fluoxetine (mg/kg) Desipramine (mg/kg)





S1.6 -

1.4 -

r 1.2 -

1.0 -

0.8

0.6-

0.4

0.2


0.0 1.5 3.0 6.0 12.0 0.0 1.5 3.0 6.0 12.0

Fluoxetine/Desipramine 1:2 Fluoxetine/Desipramine 2:1



Figure 5.1. Dose response curves for the anorectic
effects of FLU, DMI, FLU/DMI 1:2 and FLU/DMI 2:1
ratios in non-deprived rats drinking sweetened milk.
ED50 values for all curves were estimated using
non-linear regression.









81







8


7
FLU 1: DMI 2

S 6 /


5 -


4-

e / FLU 2: DMI 1
.4 3


2
C) /



-




0 1 2 3 4 5 6 7 8

Fluoxetine (mg/kg)


Figure 5.2. Isobolographic plot for the anorectic
effects of the FLU/DMI combination in non-deprived
rats drinking sweetened milk. Points along the
dotted lines projecting out from the origin reflect
all possible combinations of the drugs at a given
ratio. Notice that the ED,, values for both fixed-
dose ratios clearly fall within the dose-additive
region.









82









18


15


S12
2 *


, 9


6


3 L




VEH FLU DMI Ratiol Ratio2






Figure 5.3. Effects of FLU, DMI and 2 ratios of
FLU:DMI on water intake in 24 h water-deprived
rats. Only FLU was not significantly reduced
relative to control rats. These doses of FLU, DMI
and both ratios were estimated to suppress milk
intake by 50% in Experiment 5.1.














Ratio 1 and Ratio 2-treated rats were the only groups that

differed significantly relative to the vehicle-treated

group.

Experiment 5.3: Daily Administration of FLU. DM and FLU/DMI

Food intake

Figure 5.4 (panel A) shows percent baseline 24 h food

intake in rats treated with vehicle, FLU, DMI or FLU + DMI

for 13 days. A group by day ANOVA revealed significant main

effects for group [F(3,20) = 45.6, P<0.01] and day

[F(12,240) = 19.5, P<0.01]. One way ANOVAs conducted for

each test day revealed a significant effect of drug on food

intake for each test day. Post-hoc Bonferroni tests showed

that FLU- and DMI-treated rats significantly differed

relative to the vehicle group on all days except on days 8-

10 and day 13. Rats treated with FLU + DMI differed from all

groups on all test days, except on days 12 and 13 when they

did not differ relative to DMI-treated rats.

Body weight

Figure 5.4 (panel B) shows percent baseline bodyweight

in rats treated with vehicle, FLU, DMI or FLU + DMI for 13

days. A Group by Day ANOVA revealed significant main effects

for Group [F(3,20) = 11.8, P<0.01] and Day [F(12,240) =

19.7, P<0.01]. One way ANOVAs conducted on each test day

revealed a significant effect of drug on bodyweight for each
















140

120

100 -

80 -

60 -

40

20 -


105 -
-r-
a)
100 -

95
0)

S90
-

S85

80
.2 9-


PRE1 2 3 4 5 6 7 8 9 101112


-0-

-U-
-A-


Control
DMI
FLU
FLU/DMI


Days





Figure 5.4. Percent baseline 24 h food intake (panel A)
and bodyweight (panel B) in rats treated with FLU
(5 mg/kg/day), DMI (5 mg/kg/day) and FLU (5 mg/kg/day)
+ DMI (5 mg/kg/day).


4: w:
'^-H-lh














test day. Post-hoc Bonferroni tests showed that: (1) DMI-

treated rats differed significantly relative to the vehicle

group from day 4 and on (all P's<0.05), (2) FLU-treated rats

differed from the vehicle group from day 6 and on (all

P's<0.05), (3) FLU or DMI-treated rats did not differ

relative to each other on any test day and (4) rats treated

with the combination differed relative to all groups from

day 2 and on, except on day 7 when they did not differ from

rats treated with the individual drugs and on days 12 and 13

when they did not differ relative to DMI-treated rats.

Discussion


The present experiments used an isobolographic analysis

to quantify to what extent FLU, a reuptake inhibitor for 5-

HT, would interact with DMI, a reuptake inhibitor for NE, to

suppress intake. The analysis revealed that when FLU and DMI

were administered in a 1:2 or a 2:1 ratio, their anorectic

effects were dose-additive. The EDs values for both FLU/DMI

ratios fell within the 95% CI of the dose additive line.

How do these findings compare to our previous results

with DFEN/PHEN, where milk intake was suppressed in non-

deprived rats, in a synergistic manner? A consideration of

the pharmacological profile of each of these agents may

provide a possible explanation. First, while DFEN and FLU

have similarities (both inhibit the reuptake of 5-HT), DFEN














apparently has additional 5-HT-ergic properties. For

example, while 5-HT depleting lesions abolish the anorectic

effects of FLU, the hypophagic actions of DFEN are only

slightly affected (Rowland, Patel, Roth & Cespedes,

submitted). This suggests that in contrast to FLU, DFEN

(and/or its active metabolite dexnorfenfluramine; DNFEN) may

be exerting direct postsynaptic effects.

PHEN has not been as well researched as other diet

drugs (like FEN and DFEN), perhaps because it received FDA

approval (1959) before many modern neurochemical methods

were available. Based on a review of the (limited)

literature, it appears that PHEN can exert effects on NE

(Samanin et al., 1975), DA (Samanin et al., 1975; Hoebel,

Hernandez, Schwartz, Mark & Hunter, 1987; Shoaib et al.,

1997; Balcioglu & Wurtman, 1998) and 5-HT (Shoaib et al.,

1997). To what extent PHEN's effects are due to reuptake

inhibition, releasing properties, and/or direct postsynaptic

effects has not been conclusively determined. However,

PHEN's ability to enhance brain 5-HT may contribute to the

synergistic anorectic effects of the DFEN/PHEN combination.

Two recent studies (using microdialysis) have shown that the

mixture of PHEN and FEN can have additive effects on

increasing brain 5-HT levels (Shoaib et al., 1997; Balcioglu

& Wurtman, 1998).














In contrast to PHEN, DMI is a tricyclic with

predominantly NE uptake inhibiting properties (Hytell,

1982). DMI is not entirely selective and may weakly inhibit

uptake of 5-HT (Sangdee & Franz, 1979). DMI also exerts some

anti-cholinergic effects (Sulser, 1964; Brimblecombe &

Green, 1967) that have been attributed to muscarinic

receptor antagonism at post-ganglionic parasympathetic

receptor sites; this proposed mechanism of action, however,

has been challenged by Pendleton, Miller and Ridley (1980).

Overall, it appears that the DFEN/PHEN combination may

be exerting more "diverse" pharmacological effects than the

FLU/DMI combination in terms of sites of action (presynaptic

and postsynaptic) and additive effects on brain 5-HT. These

properties may be accounting for DFEN/PHEN's synergistic

suppression of acute intake. It is noteworthy, that even

additive interactions can have important clinical

implications. Given the adverse effects that emerge at

higher doses of FLU (Wong et al., 1995; Goldstein et al.,

1995), the possibility of reducing the dose of FLU by co-

administering another agent that interacted in a dose-

additive manner would still be of therapeutic significance.

To what extent are the intake reducing effects of the

FLU/DMI combination specific? As a basic screening test for

behavioral specificity, doses of FLU, DMI and the













combinations that reduce non-deprived milk intake by 50%

were administered to 24 h water deprived rats given a 1 h

water intake test. When administered at these doses, water

intake was significantly suppressed by DMI (-60-70%), Ratio

1 and Ratio 2 (both ~40%). In contrast, intake was not

significantly suppressed by FLU (~20%). These findings argue

that (1) FLU is more "selective" for suppressing milk intake

in non-deprived rats relative to water-intake in water

deprived rats and (2) the combinations are no less selective

than DMI alone. In fact, DMI appears to be somewhat more

potent at suppressing water intake in water-deprived rats

versus milk intake in non-deprived rats. DMI's potent

effects on water intake can potentially be explained by: (1)

DMI's anticholinergic activity; elevated levels of

acetylcholine have been implicated in (especially cellular)

thirst (Grossman, 1990) or (2) the NE reuptake inhibition

induced by DMI; elevated central NE has been shown to have

pronounced inhibitory effects on drinking (Montgomery,

Singer & Purcell, 1971).

The results obtained from the effects of chronic

administration of FLU (5 mg/kg/day), DMI (5 mg/kg/day) and

FLU (5 mg/kg/day) + DMI (5 mg/kg/day) on 24 h intake and

body weight paralleled their effects on acute intake. The

combination yielded anorectic and weight reducing effects














that appeared to be dose-additive. After 7 days of

injections rats treated with the individual drugs had

reduced food intake by around 25-35% of their baseline

intake and incurred about a 5% reduction in baseline

bodyweight. Rats treated with the combination of FLU/DES had

reduced their food intake by around 70% of their baseline

intake and incurred about a 10% reduction in baseline

bodyweight.

Interestingly, chronic TCA administration in humans is

typically correlated with weight gain (Gottfries, 1981;

Berken, Weinstein & Stern, 1984). However, when the effects

of these drugs are assessed in different animals models of

ingestive behavior they consistently yield anorectic and

weight reducing effects (Blavet & Defeudis, 1982; Towell,

Willner & Booth, 1985; Nobrega & Coscina, 1987; Durcan,

McWilliam, Campbell, Neale & Dunn, 1988). Whether or not

administration of the combination of FLU + DMI (Aranow, et

al., 1989; Bergstrom, et al., 1991) prevented DMI-induced

weight gain in individuals suffering from depression was not

reported.

Because these drugs were administered once daily and

have relatively long half-lives, it is important to consider

the potential impact of FLU on DMI's pharmacokinetics.

CYP2D6 is the cytochrome P450 enzyme that is believed to be

responsible for metabolizing several drugs including TCA's,













some neuroleptics, some cardiovascular drugs and codeine

(Baumann, 1996; Richelson, 1998). FLU and other SSRI's have

been shown to have inhibitory effects on this enzyme

(Baumann, 1996). As a consequence, plasma levels of DMI may

have been elevated to a greater extent and for a longer

period of time in rats that received the combination of FLU

+ DMI versus rats that received only the individual agents.

The potential for FLU to enhance plasma DMI is substantiated

by research in vitro (liver microsomes in rats; Fuller &

Perry, 1989), in vivo (blood and brain concentrations of DMI

in rats; Fuller & Perry, 1989) and in several clinical

reports (plasma concentrations; Preskorn et al., 1994;

Bergstrom, et al., 1991; Aranow, et al., 1989). While we did

not measure plasma drug levels, rats in our chronic study

received one-half the dose of the combination relative to

those in previous reports (10 versus 20 mg/kg total dose;

Fuller & Perry, 1989). To what extent the anorectic and

weight reducing effects of the combination may be

attributable to FLU's impact on DMI's pharmacokinetics

remains to be determined.















CHAPTER 6
GENERAL DISCUSSION


Presented within this dissertation are the only

existing experiments that have attempted to formally examine

the interaction of PHEN and DFEN in animal models of

feeding. Based on these studies two main themes are included

in the general discussion.


How Effective Are Anorectic Combinations?



The present research effort was designed around the

premise that because so many neurotransmitter and

neuropeptide systems are implicated in the control of

ingestive behavior, administering an appetite suppressant

that acts on an individual system may not be the most

effective approach for long-term reduction in intake and

weight loss. The other remaining systems may be sufficient

to maintain intake and/or compensate for the affected

system. An alternative approach would be to administer

combinations of drugs that target multiple neurotransmitter

systems. The mixture of FEN and PHEN was selected as a model

to explore this premise because these drugs act via














different systems (5-HT and DA/NA, respectively) and their

increasing "off-label" use had not been substantiated by any

preclinical work in non-humans.

Acute and chronic administration of these drugs had

profound effects on ingestive behavior in rats. When a dose

of PHEN (5 mg/kg) that was ineffective at suppressing milk

intake was combined with a dose of DFEN (2 mg/kg) that was

modestly effective at suppressing intake, milk intake was

completely eliminated for several test days. Similar effects

were observed when 24 h intake and bodyweight was compared

in ad-lib fed rats receiving DFEN, PHEN and DFEN/PHEN via

osmotic minipump. When administered alone, these drugs were

modestly effective at suppressing intake and reducing

bodyweight (-6 %), however, when given in combination (at

full doses) they were extremely effective. Daily intake was

practically eliminated for several test days and initial

bodyweight was reduced by 16%, effects that do not appear to

be attributable to an increased metabolic influence. These

results are particularly impressive given the expected

increase in "hunger" that likely accompanies several days of

virtually no food intake in rats with ad-lib access to food.

One could argue however, that the above findings are

not especially surprising. The expected outcome of a




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EPKTJB89E_V0RHC7 INGEST_TIME 2013-09-28T02:08:26Z PACKAGE AA00014293_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES