Borsini Et Al-2002-CNS Drug Reviews

8/18/2019 Borsini Et Al-2002-CNS Drug Reviews

1/26

Pharmacology of Flibanserin

Franco Borsini1, Kennett Evans2, Kathryn Jason3, Frank Rohde1,

Barbara Alexander 1, and Stephan Pollentier 1

1 Boehringer Ingelheim Pharma KG, Biberach an der Riss, Germany;2 Boehringer Ingelheim Pharmaceuticals, Burlington, Ontario, Canada;

3 Boehringer Ingelheim Pharmaceuticals, Ridgfield, CT, USA

Key Words: Flibanserin—Serotonin—Dopamine—Receptors—Firing rate—Adenylyl

cyclase—Animal models—Antidepressants.

ABSTRACT

Flibanserin has preferential affinity for serotonin 5-HT1A, dopamine D4, and serotonin

5-HT2A receptors. In vitro and in microiontophoresis, flibanserin behaves as a 5-HT1A

agonist, a very weak partial agonist on dopamine D4 receptors, and a 5-HT2A antagonist. In vivo flibanserin binds equally to 5-HT1A and 5-HT2A receptors. However, under higher

levels of brain 5-HT (i.e., under stress), flibanserin may occupy 5-HT2A receptors in

higher proportion than 5-HT1A receptors. The effects of flibanserin on adenylyl cyclase

are different from those of buspirone and 8-OH-DPAT, two other purported 5-HT1A re-

ceptor agonists. Flibanserin reduces neuronal firing rate in cells of the dorsal raphe, hip-

pocampus, and cortex with the CA1 region being the most sensitive in the brain.

Flibanserin-induced reduction in firing rate in the cortex seems to be mediated through

stimulation of postsynaptic 5-HT1A receptors, whereas the reduction of the number of

active cells seems to be mediated through dopamine D4 receptor stimulation. Flibanserin

quickly desensitizes somatic 5-HT autoreceptors in the dorsal raphe and enhances tonic

activation of postsynaptic 5-HT1A receptors in the CA3 region. Flibanserin preferentially

reduces synthesis and extracellular levels of 5-HT in the cortex, where it enhances

extracellular levels of NE and DA. Flibanserin displays antidepressant-like activity in

most animal models sensitive to antidepressants. Such activity, however, seems qualita-

tively different from that exerted by other antidepressants. Flibanserin seems to act via

direct or indirect stimulation of 5-HT1A, DA, and opioid receptors in those animal models.

Flibanserin does not display consistent effects in animal models of anxiety and seems to

exert potential antipsychotic effects. Flibanserin may induce some sedation but does not

induce observable toxic effects at pharmacologically relevant doses.

117

CNS Drug ReviewsVol. 8, No.2, pp. 117– 142© 2002 Neva Press, Branford, Connecticut

Address correspondence and reprint requests to: Franco Borsini, Boehringer Ingelheim Pharma KG,Birkendorfer Strasse 65, 88397 Biberach an der Riss, Germany.

Tel.: +49 (7351) 54-7297; Fax: +49 (7351) 54-98647; E-mail: [email protected]

mailto:[email protected]:[email protected]:[email protected]


2/26

BINDING STUDIES

Flibanserin shows the highest affinity for cloned human serotonin 5-HT1A receptors

compared with cloned human D4 (isoforms 4.2, 4.4, and 4.7) and 5-HT2A receptors(Table 1). This rank of affinity values is still maintained when the affinity for 5-HT1A and

5-HT2A receptors was evaluated in brain tissue. However, flibanserin displays higher af -

finity for 5-HT1A and 5-HT2A receptors in cloned cells than in cerebral tissue (the affinity

for D4 receptors was not tested in tissues).

Flibanserin also shows some affinity for human D2L and D3 receptors and rat NE-

alpha1 and 5-HT7 receptors (12). Flibanserin has different affinity for rat (> 10,000 nM)

and human (305–785 nM) D2 receptors (12). The affinity for all other receptors, including

the 5-HT transporter (12), varies from low to very low (Table 1).

Flibanserin does not inhibit monoamine oxidase activity (Borsini et al., 1998).

RECEPTOR ACTIVITY

The activity of flibanserin on 5-HT1A receptors was assessed by using biochemical

(forskolin-stimulated adenylyl cyclase in vitro) and electrophysiological (firing rate after

CNS Drug Reviews, Vol. 8, No. 2, 2002

118 BORSINI ET AL.

TABLE 1. Binding profile of flibanserin

Receptor Species Tissue K i (nM)

5-HT1A Human CHO 1

D4 Human CHO 4–245-HT2A Human CHO 49

5-HT1A Humanrat Cortex, hippocampus, dorsal raphe 15–50

5-HT2A Humanrat Cortex 115–133

D2L Human HEK293 305–785

D3 Human CHO 364–479

NE-alpha1 Rat Cortex 523

5-HT7 Rat COS 990

> 1,000 nM

5-HT1B, 5-HT1DE , 5-HT2C , 5-HT6, 5-HTT, D1, H1, opioid-mu, opioid-kappa, sigma, Na+ channelsite 2

> 10,000 nM5-HT3, 5-HT4, NE-alpha2, NE-beta1, NE-beta2, H2, M1, M2, M3, D5, BDZ, NMDAMK-801, NMDAglycine, NMDA phencyclidine, nicotinic, A1, Na

+ channelsite 1, Ca2+ chaneltype L, estrogen, proges-terone, testosterone, GABA-Aagonist site, GABA-ACL – channel, insulin, NK-1, opiod-delta, rat D2

CHO, Chinese Hamster Ovary cells; COS, CVI origin CV 40 cells. Where two K i values are reported,

they represent the lowest and highest K i values obtained in two or more experiments. See ref. 19 and 50

for methods in human 5-HT2A and 5-HT1A receptors in cells, respectively. These experiments were

conducted by G. B. Schiavi at Boehringer Ingelheim, Italy, and Cerep, France. See references 33, 48,

and 67 for methods in human dopamine receptors. These experiments were conducted by J. Mierau

and G. B. Schiavi of Boehringer Ingelheim. See references 43 and 45 for methods on 5-HT1A and

5-HT2A in human tissues. These experiments were conducted by D. Marazziti and L. Palego of Uni-

versity of Pisa, Italy. See references 52 and 62 for methods on NK 1 and glucocorticoid receptors, re-

spectively. These experiments were conducted by Cerep, France. See reference 12 for methods for all

other receptor studies.


3/26

microiontophoretic application) indices. Activity at dopamine D4 receptors was investi-

gated using various biochemical parameters (adenylyl cyclase, microphysiometry, and

GTP-shift).

Activity on 5-HT2A receptors was assessed biochemically by determining the phospha-tidylinositol turnover.

Activity at 5-HT1A Receptors

Flibanserin behaves as a 5-HT1A agonist in cloned cells and the cortex and hippo-

campus of human and rat brain (Table 2). In the dorsal raphe, flibanserin behaves as an ag-

onist when firing rate was taken as an index of activity in rats (58) but was devoid of ag-

onist activity when forskolin-stimulated cAMP was measured in human tissue (44).

In contrast to the dorsal raphe nucleus of rats (24), 5-HT1A receptors in the dorsal raphe

nucleus of humans appears to be linked with adenylyl cyclase (44). In human tissues, the

EC50 of flibanserin in reducing forskolin-stimulated cAMP formation is similar to the af-

finity values for 5-HT1A receptors (Table 1 and 2). In contrast, the EC50 of flibanserin in

reducing forskolin-stimulated cAMP formation in cells or rat tissues is different from the

affinity values for 5-HT1A receptors. The coupling system between the receptor and the

G-proteins is important to demonstrate activity for agonists (55). Thus, it may be that the

relative amount of various G-proteins present in human tissue is different from that

present in CHO cells or rat tissue. Flibanserin binds to only one site in human 5-HT1A re-

ceptors (44), whereas it binds to high- and low-affinity sites in rodent 5-HT1A receptors

(12). Thus, a different proportion of high-low-affinity sites might explain the different

values between 5-HT1A affinity and EC50 on cAMP. Nevertheless, flibanserin is more

potent in reducing cAMP in the hippocampus than in the cortex, both in rats and humans.

In contrast, when electrophysiology is considered, flibanserin is more potent in the cortex

than the hippocampus (Table 2).


FLIBANSERIN 119

TABLE 2. Activity of flibanserin on cAMP in vitro and on firing rates

after microiontophoretical application in vivo

Tissue Parameter Species Effect EC50 (nM) IT50 (nC)

CHO cells cAMP Human Full agonism 457Cortex cAMP Human Agonism 28

Hippocampus cAMP Human Agonism 4

Dorsal raphe cAMP Human No agonism –

Cortex cAMP Rat Full agonism 913

Hippocampus cAMP Rat Full agonism 317

Cortex Firing rate Rat Full agonism 1,260

Hippocampus Firing rate Rat Partial agonism 1,365

Dorsal raphe Firing rate Rat Full agonism 260

EC50 indicates the concentration that reduces forskolin-induced cAMP formation by 50% in rat tissue

(12), and in CHO cells (experiment was conducted by E. Giraldo of Boehringer Ingelheim, Italy). See

reference 44 for methods in CHO cells and reference 55 for methods in human tissue. IT50 indicates

the current x seconds that suppresses firing rate by 50% (58).


4/26

The effect of flibanserin in 5-HT1A cloned cells, in human and rat cortex seems to be

mediated by stimulating 5-HT1A receptors (Table 3), as shown by the ability of all the

5-HT1A antagonists we used to reduce the effect of flibanserin. However, it is less clear

whether 5-HT1A receptors mediate the effects of flibanserin in the hippocampus (Table 3),as suggested by the observation that two out of five 5-HT1A antagonists failed to anta-

gonize flibanserin (17,44,58). Other purported 5-HT1A agonists, such as 8-OH-DPAT and

buspirone, failed to inhibit cAMP in the rat cortex (12) or human hippocampus (44), re-

spectively, so 8-OH-DPAT and buspirone were used as antagonists in these two regions in

our studies.

Activity on D4 Receptors

As far as the effects of flibanserin on dopamine D4 receptors are concerned, these were

only tested in cloned cells (Table 4). Flibanserin behaved as an antagonist of forskolin-

stimulated cAMP when dopamine D4 receptor density was < 450 fmolmg protein or as afull agonist when flibanserin concentrations were very high (> 30 ìM). Flibanserin be-

haved as a partial agonist (microphysiometry and GTP shift) when dopamine D4 receptor

density was > 750 fmolmg protein, which is much higher than that has been reported in

the human brain (0.2–50 fmol protein; 49,63). In an attempt to evaluate the affinityac-tivity of flibanserin for dopamine D4 receptors directly in human brain, the binding of the

suggested dopamine D4 ligand nemonapride was investigated in human postmortem

cortex and striatum. However, no dopamine D4 receptors were detected in the cortex and

caudatum of nonpsychiatric subjects (D. Marazziti, University of Pisa, Italy, personal

communication, January 2001).

Activity at 5-HT2A Receptors

Flibanserin behaves as a 5-HT2A-receptor antagonist. In fact, it does not enhance PI

turnover per se and antagonized 5-HT-stimulated turnover in mouse cortex (12). The af-

finity values for tissue 5-HT2A receptors (115–133 nM) and the K i in reducing 5-HT-

stimulated turnover (113 nM) are in good agreement.

Activity at 5-HT7 Receptors

Flibanserin (up to 1 ìM) did not modify relaxation induced by the 5-HT7 agonist

5-carboxamidotryptamine in ileum that was contracted by administration of substance P(41). At higher concentrations, flibanserin reduced substance P contractions (41).

IN VIVO BINDING TO 5-HT1A AND 5-HT2A RECEPTORS

Information from in vitro binding studies is limited by the fact that the endogenous

neurotransmitter is not present. The presence in vivo of 5-HT, which has higher affinity for

5-HT1A than 5-HT2A receptors (27), may differently influence the binding of flibanserin to

these receptors. Thus, an experiment was conducted to evaluate the in vivo binding of

flibanserin (Table 5). Flibanserin, in spite of different in vitro affinity values, occupies


120 BORSINI ET AL.


5/26

C N S Dr u

gR

e v i e w s , V o l . 8 , N o .2

,2

0 0 2

TABLE 3. Effects of various 5-HT 1A antagonists on activity of flibanserin on cAMP i

or firing rate after microiontophoretic application in vivo

Tissue Parameter Species

Effect of 5-HT1A antagonists

WAY100135 WAY100635 Tertatolol Pindobind BMY

CHO cells cAMP Human Antagonism

Cortex cAMP Human Antagonism Antagonism

Cortex cAMP Rat Antagonism

Cortex Firing rate Rat Antag

Hippocampus cAMP Human No antagonism

Hippocampus cAMP Rat Antagonism

Hippocampus Firing rate Rat No antagonism Antag

Data are from references 15, 17,44, and58. Thedata for WAY100135 in CHOcellswere obtainedby E. Giraldo andR.

a commercially available enzyme immunoassay kit. WAY100635, cyclohexanecarboxamide, N-[2-[4-(2-methoxyphe

trihdrochloride; BMY 7378, 8-azaspiro[4, 5]decane-7,9-dione, 8 [2-[4-(2-methoxyphenyl]-1-piperazinyl]ethyl]-,dihydroch

(4-indolyloxy)-2-hydroxy-propyl-[Z]-1,8-diamino-p-menthane.


6/26

5-HT1A and 5-HT2A receptors in vivo in similar percentages (60). In addition, flibanserin

preferentially occupies 5-HT1A and 5-HT2A receptors in the cortex at doses as low as

1 mgkg. It has been reported that a full agonist needs to occupy about 20% of receptors to

induce an effect (47,66). In contrast, antagonism is evident with at least 70% of receptor occupancy (28,38). Thus, flibanserin should be capable of triggering 5-HT1A-mediated


122 BORSINI ET AL.

TABLE 4. Effect of flibanserin on cells cloned with human dopamine D4 receptors

Test

Receptor density

(cell–fmolmg protein) Effect

Adenylyl

cyclase

CHO-446 Flibanserin behaved as an antagonist ( K i = 339 nM) against 1 ìM do-

pamine-induced suppression of 30 ìM forskolin (FSK)-stimulatedcAMP concentration. At the same experimental conditions, clozapine

antagonized dopamine effects ( K i = 115 nM). However, flibanserin be-

haved as an agonist (about 80% inhibition of FSK-induced cAMP) at

concentrations > 30 ìM. Dopamine (EC50 = 346 nM) and pramipexole

(EC50 = 34 nM) reduced FSK-induced cAMP by about 80%

Microphy-

siometry

CHO-780 Flibanserin behaved as a partial agonist ( K i = 54 nM), displaying only

39% of the efficacy of pramipexole. Flibanserin behaved as an anta-

gonist (IC50 = 83 nM) against pramipexole

GTP shift CHO-780 Flibanserin behaved as partial agonist; its K i shifted 6.1 times in the

presence of Gpp(NH)p. At the same experimental conditions, the K i of

pramipexole shifted 96 times and that of haloperidol 0.6 timesGTP shift Sf9-860900 Flibanserin behaved as partial agonist; its K i shifted 8.5 times in the

presence of Gpp(NH)p. The K i of pramipexol shifted 25.5 times and that

of haloperidol 0.9 times

The data on cAMP were obtained by E. Giraldo and R. Targa, Boehringer Ingelheimm, Italy (see ref. 2

for methods). Other data were provided by J. Mierau and H. Ensinger of Boehringer Ingelheim,

Germany (see ref. 48 for methods).

TABLE 5. 5-HT 1A and 5-HT 2A receptor occupancy by flibanserin in rats,

at 30 min after i.p. administration of the drug

Brain region

Dose of flibanserin

(mgkg)

Receptor occupancy (%)

5-HT1A 5-HT2A

Cortex 1 20 13

10 53 42

30 77 69

Hippocampus 1 No significant occupancy No significant occupancy

10 60 50

30 79 73

Brainstem 1 No significant occupancy No significant occupancy

10 44 53

30 50 72

Data, modified from ref. 60.


7/26

mechanisms only in the cortex at a dose of 1 mgkg, while flibanserin doses as high as

30 mgkg should be necessary to exert antagonist properties on 5-HT2A receptors.

Table 5 refers to an experiment performed in normal rats. In case of stress, 5-HT levels

may increase several fold (39,64), and thus 5-HT may differently compete with flibanserinfor binding to 5-HT1A and 5-HT2A receptors. In order to understand this, a computer simu-

lation was performed in conditions where the concentration of 5-HT varied from 0 to

1 ìM (Table 6). When 5-HT is present in concentrations between 1 and 10 nM, flibanserin

has higher affinity for 5-HT1A than for 5-HT2A receptors. However, when 5-HT concentra-

tions are between 100 and 1,000 nM, flibanserin should significantly occupy 5-HT2A re-


FLIBANSERIN 123

TABLE 6. Computer simulations describing the expected behavior of flibanserin in vivo

at various concentrations of 5-HT

5-HT concentration (nM)

Apparent flibanserin affinity values (nM)

K i for 5-HT1A K i for 5-HT2A

0 10 100

1 13 100

10 43 104

100 343 140

1000 3343 500

These simulations with 5-HT1A and5-HT2A receptors, for which 5-HT hasmedianaffinityvalues ( K d )

of 3 and 250 nM, respectively. Flibanserin was considered to have affinity values ( K i ) of about 10

and 100 nM for 5-HT1A and 5-HT2A receptors, respectively. The apparent K i values of the drug for the

receptors was calculated by simulating the presence of different 5-HT concentrations (1–1,000 nM),according to the following equation (valid for a competitive interaction for the same binding site):

K i(app) = K i(1+ [5-HT] K d ). This simulation was performed by T. Mennini and M. Gobbi, Mario Negri Institute, Milan, Italy.

TABLE 7. Plasma concentrations of flibanserin after intraperitoneal administration in rats

Dose (mgkg)

Concentration (ìM) at various times

1 hour 3 hours 8 hours

4 1.5 0.8 0.26

16 9.1 6.9 3.0

32 10 7.2 3.5

64 17.6 14.2 10.5

Plasma protein binding of flibanserin in rats is 97%. Thus, the calculated free concentration

of flibanserin is as follows:

Dose (mgkg)

Free concentration (nM) at various times

1 hour 3 hours 8 hours

4 44 23 8

16 274 206 91

32 300 216 105

64 528 426 314

Values are means of 6 experiments, whose variation coefficient was 31–86%.


8/26

ceptors at lower concentrations than those that are necessary to occupy 5-HT1A receptors.

According to the hypothesis that 5-HT2A antagonism should be present to favor 5-HT1Aagonism in the cortex (8), such a condition should occur for flibanserin when brain 5-HT

levels exceed 100 nM and if flibanserin concentration is between 104 and 343 nM. Asshown in Table 7, flibanserin can reach sufficient concentrations at doses from 16 mgkg.

ELECTROPHYSIOLOGICAL EFFECTS

AFTER SYSTEMIC ADMINISTRATION

Electrophysiological data on firing rate were obtained by giving flibanserin intrave-

nously (Table 8). The most striking activity of flibanserin was on firing rate of CA1 hippo-

campal neurons, where it reduced firing rate by 50% at a dose as low as 3 ìgkg (A. Ceci,Boehringer Ingelheim, data on file). To date, the effects of 5-HT1A antagonists on fliban-

serin action on CA1 neurons have not been evaluated. Flibanserin reduced firing rate by

50% in the dorsal raphe at 200 ìgkg, in CA3 hippocampal neurons at 1.4 mgkg, and in

cortical neurons at 3 mgkg (58). The potency difference of flibanserin in reducing thefiring rate of neurons in the CA1 and CA3 regions may not be surprising. In fact, it has

been suggested that the 5-HT1A receptor recognition site may be different in the CA3 and

CA1 areas (3,5). It is interesting to compare these values with those necessary for 50%

5-HT1A receptor occupancy after intraperitoneal (i.p.) administration (Table 5). For this

comparison, one must consider that the bioavailability of flibanserin in rats is 35%. Thus,

the ED50 values (1.4 to 3 mgkg, i.v.) of flibanserin in reducing the firing rate of neuronsin the CA3 region and cortex may be in line with ED50 values for receptor occupancy

(about 10 mgkg, i.p., with higher occupancy in hippocampus). In contrast, flibanserin re-

duced the firing rate of the dorsal raphe neurons by 50% at 0.2 mgkg, i.v., whereas it oc-

cupied 50% of 5-HT1A receptors in the midbrain only at a dose as high as 30 mgkg, i.p.However, receptor occupancy in the midbrain may not reflect receptor occupancy in the

dorsal raphe. Nevertheless, flibanserin behaved as a full agonist in the dorsal raphe nu-

cleus and its effect was blocked by 5-HT1A antagonists (Table 9) (58; E. Esposito, “Mario

Negri Sud,” Italy, personal communication, 1992). The effect of flibanserin in CA3 was


124 BORSINI ET AL.

TABLE 8. Effects of intravenous flibanserin on firing rate of neurons

in different brain regions in the rat

Dose, mgkg

Tissue

Cortex Hippocampus Dorsal raphe

0.003 CA1: 50% reduction


0.2 50% reduction

0.6 100% reduction


3 50% reduction CA3: 80% reduction10 80% reduction


9/26

also found to be antagonized by 5-HT1A antagonists (58). However, the effect of fliban-

serin on dorsal raphe and hippocampus is different. The effect on the firing rate of the

dorsal raphe is short lasting. In fact, flibanserin 5 mgkg, s.c., reduced dorsal firing rate by

only 25% after 2 days and did not reduce it after 7 days, a consequence of receptor desen-sitization (57). In contrast, the same dose of flibanserin enhanced tonic activation of

postsynaptic 5-HT1A receptors in the CA3 region after 2 days administration as much as

after 7 days (57).

As far as the firing rate of cortical neurons is concerned, flibanserin-induced inhibition

seems to be mediated by 5-HT1A receptors, even if one out of three 5-HT1A antagonists

used in the cortex did not antagonize flibanserin (Table 9). The effect of flibanserin in the

cortex was not blocked by the chemical destruction of 5-HT-containing neurons by

5,7-DHT (15), which significantly depleted brain 5-HT by 80% (NE was reduced by 27%

[not significant]). Thus, it appears that flibanserin does not require 5-HT neurons to exert

its inhibitory effects on the cortex. Once more, the other 5-HT1A agonists 8-OH-DPAT and buspirone, which were used as comparators, behaved differently from flibanserin. In fact,

in contrast with flibanserin, buspirone excited cortical neurons, and 8-OH-DPAT excited

them at low doses and inhibited them at high doses (15). Additionally, the inhibitory ef-

fects of 8-OH-DPAT were completely blocked by 5,7-DHT (15).

In addition to reducing firing rate, flibanserin also decreased the number of spontane-

ously active cells in the cortex (12). Flibanserin exerted a significant effect already at

2 mgkg, i.p., and induced the maximal inhibition (about 80%) at 8 mgkg (Fig. 1). Themaximum effect was at 6 h and a significant reduction was still observed after 24 h

(Fig. 2). When a 5-HT1A antagonist was used to evaluate whether this effect of flibanserin

was mediated by 5-HT1A

receptors, the results were unclear. In fact, the 5-HT1A

antagonist

N-(1,1-dimethylethyl)-4-(2-methoxyphenyl)-alpha-phenyl-[(±)WAY100135] (30 mgkg,s.c.) antagonized flibanserin only in one of the two experiments performed. The effect of

flibanserin was also antagonized by the D4 antagonist 1H-pyrrolo[2,3-b]pyridine,3-[[4-

(4-chlorophenyl)piperazin-1-yl]methyl] (L-745, 870) (at 30 mgkg, i.p., but not at 10 mgkg).Thus, the effect of firing rate appears to be mediated by 5-HT1A receptors, whereas the

effect on reduction of spontaneously active cells appear to be mediated by dopamine D4receptors and probably 5-HT1A receptors (Fig. 3).

Flibanserin, 5 mgkg, administered s.c. for 2 days, reduced the electrophysiological ef-fects of the 5-HT2 agonist (±)-2,5-dimethoxy-4-iodoamphetamine [DOI] by 33% when ap-

plied microiontophoretically into the cortex (57). This antagonism towards DOI disappeared

after 7 days. However, it is questionable whether the antagonistic activity of flibanserin


FLIBANSERIN 125

TABLE 9. Effect of various 5-HT 1A antagonists, given intravenously, on the reduction

of firing rates induced by intravenous flibanserin

Tissue

5-HT1A antagonists

WAY100135 WAY100635 Tertatolol

Cortex Antagonism Potentiation Antagonism

Hippocampus CA3 Antagonism

Dorsal raphe Antagonism Antagonism

Data from references15and 58and from E.Esposito, “MarioNegriSud,” Italy, personalcommunication.


10/26

versus DOI is mediated by blockade of 5-HT2A receptors. In fact, consistent 5-HT2A re-

ceptor occupancy by flibanserin is obtained after only 30 mgkg. Furthermore, 5-HT1A re-

ceptor stimulation has been reported to reduce 5-HT2-dependent events (26,65). Thus, the an-

tagonism of flibanserin versus DOI may reflect the 5-HT1A agonist activity of flibanserin.


126 BORSINI ET AL.

0

2

4

6

8

10

12

0.5 h 3 h 6 h 24 h 48 h

Vehicle

Flibanserin

**

**

*

*

N u m

b e r

o f

s p o n

t a n e o u s

l y

a c

t i v e c e

l l s

i n t h e c o r t e x

Time

Fig. 2. Time-course of the effect of flibanserin on the number of spontaneously active cells in the cortex of rats.

Columns represent mean S.E.M. from 4 rats. The method was according to Ceci et al. (22). Flibanserin was

given i.p. at 8 mgkg. * P < 0.05; ** P < 0.01. This experiment was performed by A. Ceci in BI, Italy.

0

2

4

6

8

10

12

0 mg/kg 0.5 mg/kg 2 mg/kg 8 mg/kg 16 mg/kg

Dose N u m b e r o f s p o n t a n e o u s

l y a c t i v e

c e l l s i n t h e c o r t e

x

*

** **

Fig. 1. Effect of flibanserin on the number of spontaneously active cells in the cortex of rats. Columns represent

mean S.E.M. from 4 rats. The method was according to Ceci et al. (22). The effect of flibanserin was observed

6 h after i.p. administration. * P < 0.05; ** P < 0.01. The experiment was performed by A. Ceci, Boehringer

Ingelheim, Italy.


11/26

EFFECTS ON TURNOVER

AND EXTRACELLULAR LEVELS OF 5-HT

Flibanserin exerts a preferential action in the cortex (Table 10). In fact, flibanserin,

2 mgkg, reduces 5-HT turnover in the cortex without altering it in hippocampus and brainstem (18). This preferential activity on the cortex distinguishes flibanserin from the

other 5-HT1A receptor agonists, buspirone and 8-OH-DPAT, which reduce 5-HT turnover

to a similar extent in the cortex, hippocampus, and brainstem (18). The effect of 8-OH-DPAT

and buspirone was interpreted to result from somatic presynaptic 5-HT1A receptor acti-

vation, which inhibits 5-HT turnover in all the 5-HT-containing axons departing from the

raphe nuclei. Flibanserin begins to inhibit 5-HT turnover in the hippocampus at 16 mgkg,this dose is 8 times higher than the dose necessary to begin to reduce 5-HT turnover in the

cortex. In contrast, flibanserin did not consistently affect 5-HT turnover in the brainstem,where raphe nuclei are located. Thus, it is not clear why flibanserin reduces 5-HT turnover

in the cortex at doses lower than those required to induce the same effect in the hippo -

campus. Flibanserin reduced 5-HT turnover in the cortex by 50% at 8 mgkg i.p., whichfits well with its occupancy of cortical 5-HT1A receptors. 5-HT1A receptors are not located

on cortical or hippocampal nerve endings (35). Thus, it may be that flibanserin stimulates

5-HT1A receptors in the cortex, which in turn trigger a negative feedback on 5-HT release.

Such negative feedback has been shown to occur in the cortex (20,21,34). This may ex-

plain the preferential effect of flibanserin only on those neurons that project to the cortex.

Such feedback mechanisms seem to work less efficiently in the hippocampus.

A preferential effect of flibanserin on the cortex was observed also in microdialysis ex-

periments. A reduction in cortical extracellular 5-HT levels was observed already with fli-

banserin, 3 mgkg, i.p. (Table 10). Cortical extracellular 5-HT levels were reduced by


FLIBANSERIN 127

0

2

4

6

8

10

12

Vehicle

Flibanserin

N u m

b e r

o f

c e

l l s p e r t

r a c

k

Vehicle WAY100135 L-745,870

#

* *

Fig. 3. Effects of WAY100135 and L745870 on the flibanserin-induced reduction of spontaneously active cells

in the cortex of rats. Columns represent mean S.E.M. from 4 rats. The method was according to Ceci et al. (22).

Flibanserin was administered i.p. 3 h before recording. WAY 100135 (30 mgkg) and L645870 (30 mkg) weregiven i.p. 30 min before recording. # P < .01 vs. respective vehicle. * P < .01 two-way ANOVA. This experiment

was performed by A. Ceci, Boehringer Ingelheim, Italy.


12/26

about 50% at a dose of 10 mgkg, i.p., which fits well with the occupancy of cortical5-HT1A receptors by flibanserin. Extracellular levels of 5-HT were not changed in the hip-

pocampus by flibanserin at doses that reduced extracellular 5-HT levels in the cortex.

Presynaptic 5-HT1BD receptors have been reported to inhibit 5-HT release (40). However,flibanserin has low affinity for these receptors. Thus, the microdialysis experiments also

support the hypothesis that flibanserin may trigger a feedback mechanism on 5-HT cor-

tical neurons through stimulation of cortical postsynaptic 5-HT1A receptors. Electro-

physiological experiments showed that very close systemic doses of flibanserin stimulate

5-HT1A receptors in the cortex and the CA3 to a similar extent. Thus, experiments with

microdialysis also support the notion that the negative feedback does not work well in hip-

pocampus. However, another possible hypothesis might be that flibanserin mainly acts on

those 5-HT1A somatic autoreceptors located on dorsal raphe neurons that specifically

project to the cortex. In this case, the occupancy of 5-HT1A receptors in the midbrain (44%

at 10 mgkg, i.p.) might not reflect the real receptor occupancy in the dorsal raphe. Thesame holds for the data on 5-HT synthesis that were obtained from all the brainstem.

When the microdialytic probe was inserted into the dorsal raphe, however, a reduction of

5-HT extracellular concentration was not observed at the dose of 3 mgkg, which induced5-HT extracellular reduction in the cortex. Flibanserin significantly reduced 5-HT

extracellular levels in the dorsal raphe only at 10 mgkg, i.p. Buspirone and 8-OH-DPAT,which have been reported to be active on somatic 5-HT1A autoreceptors, behave differ-

ently from flibanserin. In fact, both buspirone and 8-OH-DPAT reduce 5-HT synthesis and


128 BORSINI ET AL.

TABLE 10. Effects on 5-HT turnover and extracellular levels of 5-HT, NE, DA, and GABA

in various brain regions at 45 min after i.p. administration of flibanserin to rats

Dose

(mgkg)

Brain regions

Cortex Hippocampus Brainstemdorsal raphe

5-HT turnover

1 No significant reduction No significant reduction No significant reduction

2 20% reduction No significant reduction No significant reduction



16 69% reduction 18% reduction 38% reduction

32 68% reduction 24% reduction No significant reductionMicrodialysis

3 5-HT levels 30% reduction No effect No significant reduction

NE levels 30% increase

DA levels No significant increase

GABA levels No effect

10 5HT levels 50% reduction No effect 70% reduction

NE levels 90% increase

DA levels 100% increase

GABA levels No effect

Data on 5-HT turnover are from ref. 18. Data on microdialysis are from R. Invernizzi, “Mario Negri”

Institute, Milan, Italy, personal communication.Brainstemwasused for the5-HT turnover, anddorsal

raphe was used for extracellular 5-HT concentrations.


13/26

extracellular levels in the terminal regions and dorsal raphe at the same doses (18,37).

Thus, the hypothesis that flibanserin may primarily interact with dorsal raphe neurons to

exert its neurochemical effects is not so clearly supported by experimental findings.

Flibanserin, at 3 mgkg, increased the extracellular concentration of norepinephrine(NE). Since a direct action of flibanserin on noradrenergic receptors may be excluded

(Table 1), it may be that this elevation of NE (and DA, see below) levels is secondary to

the reduced release of 5-HT. In fact, it has been reported that 5-HT may inhibit the release

of NE and DA in the cortex (53,59,61).

The effects of flibanserin at 10 mgkg in increasing cortical NE and DA are very clear,as this increase reaches about 100%. However, the increase in DA lasts for a shorter time

than that of NE (Table 11). This increase in DA might explain the dopamine D4-mediated

effects of flibanserin in reducing the number of spontaneously active cells. In fact, these

effects were also exerted to the same extent by the full dopamine D 4 agonist pramipexole

(A. Ceci, Boehringer Ingelheim, personal communication, 1998), but the biochemical data

indicated that flibanserin is a partial agonist on dopamine D4 receptors. Thus, it seems

conceivable that flibanserin reduces the number of spontaneously active cells via increas-

ed release of DA, which in turn activates dopamine D4 receptors.

SUMMARY OF FUNCTIONAL ASPECTS

The first interaction of flibanserin with brain structures occurs in the hippocampus and

dorsal raphe, as revealed by electrophysiological experiments. The mechanism of this in-teraction in CA1 has not been explored. In the CA3 and dorsal raphe, flibanserin inhibits

firing rate through stimulation of 5-HT1A receptors. All these events happened at doses

below and around 1 mgkg. From 1 up to 10 mgkg, flibanserin interacts with the cortex.This cortical interaction is evident electrophysiologically and biochemically, and it seems

to be mediated by 5-HT1A receptors, even if dopaminergic D4-mediated effects were also

observed. The interaction with 5-HT2A receptor is not very clear. Above 10 mgkg, thereis also a biochemical interaction of flibanserin with the hippocampus and dorsal raphe.


FLIBANSERIN 129

TABLE 11. Duration of action of flibanserin

Test

Dose

(mgkg, i.p.)Peak time(minutes)

Duration of action(hours)

Microdialysis: 5-HT in dorsal raphe 10 60 < 1.5

Microdialysis: DA in cortex 10 60 < 1.5

Microdialysis: 5-HT in cortex 10 60 > 2

Microdialysis: NE in cortex 10 30 > 2

Ex vivo binding in cortex 10 30 < 3

Ex vivo binding in midbrain 10 30 < 3

Ex vivo binding in hippocampus 10 30 > 3

5-HT turnover in cortex 16 45 > 6 Number of active cellstrack 8 360 > 24


14/26

C N S Dr u

gR

e v i e

w s , V o l . 8 , N o .2

,2

0 0 2

TABLE 12. Summary of the electrophysiological and biochemical effects of intraperitonea

Dose (mgkg) Cortex Hippocampus

32 68% inhibition of 5-HT synthesis 24% inhibition of 5-HT synthesis

30 69% 5-HT2A receptor occupancy 73% 5-HT2A receptor occupancy 72% 577% 5-HT1A receptor occupancy 79% 5-HT1A receptor occupancy 50% 5

16 69% inhibition in 5-HT synthesis 18% inhibition in 5-HT synthesis 38% in

10 42% 5-HT2A receptor occupancy 50% 5-HT2A receptor occupancy 53% 5

53% 5-HT1A receptor occupancy 60% 5-HT1A receptor occupancy 44% 5

50% inhibition of 5-HT extracellular levels 70%

90% increase in NE extracellular levels

100% increase in DA extracellular levels

8 50% inhibition in 5-HT synthesis

75% reduction of active cells

4 28% inhibition in 5-HT synthesis3 30% increase in NE extracellular levels

30% inhibition of 5-HT extracellular levels

50% inhibition of firing rate

2 20% inhibition in 5-HT synthesis

22% reduction of active cells

1.4 50% inhibition of firing rate in CA3

1 20% 5-HT2A receptor occupancy

13% 5-HT1A receptor occupancy

0.2 50% in

0.003 50% inhibition of firing rate in CA1


15/26

Thus, flibanserin exerts two different levels of effects in hippocampus and dorsal

raphe. The interaction of flibanserin with these two brain structures is observed electro-

physiologically at low doses and biochemically at higher doses (Table 12). In contrast,

flibanserin exerts a consistent activity profile in the cortex, since both electrophysiological

and biochemical techniques reveal effects at similar doses (Table 12). This heterogeneous

picture may reflect the well-known heterogeneous pharmacology of 5-HT1A receptors

(6,7,9,56). In fact, in the dorsal raphe, flibanserin behaves as a full agonist when firing rate

is considered as an index of activity, but it is devoid of stimulant properties when adenylyl

cyclase is taken as a measure of activity. Likewise, in hippocampus, flibanserin behaves as

partial agonist when firing rate is considered, but as a full agonist when adenylyl cyclaseis measured.

Nevertheless, flibanserin seems to exert unique pharmacological properties. In fact, its

effects are different from those of two purported 5-HT1A agonists, buspirone and 8-OH-DPAT

(Table 13). However, it is difficult to understand whether this difference solely depends on

the different drug effects at 5-HT1A receptors or on other factors. In fact, buspirone and

8-OH-DPAT can bind also to dopamine D2 and 5-HT7 receptors, respectively, and fliban-

serin to dopamine D4 and 5-HT2A receptors.

Finally, the terminal half-life values of flibanserin in plasma were 0.9 h (i.v.) and 1.9 h

(p.o.) in rats. The half-life values are in agreement with the duration of some effects in-

duced by flibanserin, but not with others (Table 11), such as 5-HT turnover in the cortex or

the number of active cells in the cortex. So far, it is unknown whether the metabolites of

flibanserin play a pharmacological role.

BEHAVIORAL EFFECTS

Rats seems more susceptible than mice to the inhibitory effects of flibanserin on motor

activity (14). In fact, reduced motor activity was observed at doses of 8 to 16 mgkg, i.p.,

in rats, and no reduction of motor activity was seen with doses up to 32 mgkg, i.p., in mice.Flibanserin induced failure to grasp the wire in the traction test (R. Cesana, Boehringer

Ingelheim, data on file) in 50% of mice at a dose of about 45 mgkg, i.p., or 95 mgkg,


FLIBANSERIN 131

TABLE 13. Comparison of functional effects of flibanserin, buspirone and 8-OH-DPAT

Flibanserin Buspirone 8-OH-DPAT

cAMP in rat cortex Inhibits Inactive Inactive

cAMP in human cortex Inhibits Inactive InhibitsFiring rate in rat cortex Inhibits Excites Excitesinhibits

Active cells in rat cortex Inhibits Inhibits

cAMP in rat hippocampus Inhibits Inhibits Inhibits

cAMP in human hippocampus Inhibits Inactive

Firing rate in hippocampus Inhibition in CA1

higher than in CA3

Inhibition in CA3

higher than in CA1

cAMP in human dorsal raphe Inactive Inhibits

Firing rate in rat dorsal raphe Inhibits Inhibits Inhibits


16/26

p.o. At 64 mgkg, p.o., mice also displayed hypomotility and hypothermia. These effects

were more pronounced at 128 mgkg. At this high dose, the following additional alter-ations were also observed: reduced grip strength, ptosis and head weaving, and occasional

catalepsy. Some signs of a weak serotonergic syndrome were observed at 64 mgkg, i.p.,in rats (13).

Animal Models Sensitive to Antidepressants

Flibanserin was tested in 13 animal models sensitive to antidepressants. It was active in

the following 10 models (Table 14): learned helplessness in rats (16), bulbectomized rats

(16), chronic mild stress in rats and mice (25), muricidal rats (10), amphetamine with-

drawal in rats (31), REM sleep latency reduction (W. Gaida, Boehringer Ingelheim, data

on file), forced swimming test in mice (23), distress calls in chicks (10), and fixed ratio

(FI) schedule in mice (10). In contrast, flibanserin did not show antidepressant-like be-

havior in the forced swimming test (10) and differential-reinforcement-of-low rate 72 sec[DRL-72] (16) in rats and tail suspension in mice (10). Active doses ranged between 10

and 32 mgkg, i.p., except in the chronic mild stress model in mice where activity was

seen after 2.5 and 5 mgkg, i.p., and in REM sleep latency in rats, where an effect was ob-

served after 1 mgkg, p.o. In rats, the antidepressant-like effects of flibanserin occur at thesame doses that induce sedation. Thus, sedation may be a possible explanation for the

antimuricidal activity of flibanserin or the effect of flibanserin in bulbectomized rats.

However, there are some tests where flibanserin did not alter rat motor behavior, e.g.,

swimming in the Morris maze (14) or escape behavior in the learned helplessness test

(16). Flibanserin can even increase rat motor activity in amphetamine withdrawal (31).

Thus, it is difficult to ascertain to what extent sedation may be responsible for the effect of flibanserin in muricidal or bulbectomized rats. However, flibanserin seems to induce se-


132 BORSINI ET AL.

TABLE 14. Effects of flibanserin in animal models of depression

Test Antidepressant-like effects

Rat

Learned helplessness yes

Bulbectomy yes

Chronic mild stress yes

Muricidal behavior yes

Amphetamine withdrawal yes

REM sleep latency yes

Forced swimming test no

Differential reinforcement 72 sec no

Mouse

Chronic mild stress yes

Fixed ratio yes

Forced swimming test yes

Tail suspension no

Chick Distress-call yes


17/26

dation in resting animals or in animals whose motivation to explore is not high (as in the

actimeter); at the same time, it increases the drive of animals toward goal-oriented be-

havior, such as in the learned helplessness test (16) or the fixed ratio operant behavior

(10). Thus, one may understand why there is no motor reduction in swimming to searchfor the hidden platform in the Morris maze, or why there is a higher escaping activity in

the learned helplessness test. This might also explain the contrasting finding of “de-

pressive-like” effects in the forced swimming test in rats and antidepressant-like effects in

mice. Rats have lower motivation to explore the environment, since they are pretested

24 h earlier. Mice, however, are tested in the forced swimming test without pretest and

have, therefore, no familiarity with the environment. Likewise, flibanserin reduced calls in

rat pups that were pretested (54) but increased calls in chicks that were not pretested (10).

The reduction of movement power induced by flibanserin in the tail suspension test may

reflect some effects of flibanserin on spinal 5-HT1A receptors (46), since mice are sus-

pended with the head down in the tail suspension test. However, all these thoughts are

highly speculative, and the failure of flibanserin in the forced swimming test and DRL-72in rats, as well as in tail suspension in mice, remains difficult to explain. In fact, fliban-

serin is a 5-HT1A agonist and a 5-HT2 antagonist and one or both these two mechanisms

have been reported to play a role in those tests (see 10,16). This points out, once more, the

difference between flibanserin and the other 5-HT1A agonists.

In summary, flibanserin did not show antidepressant-like activity in all of the tests, and

induced different behavioral changes than those observed with imipramine (Table 15).

Only clinical trials will tell whether flibanserin will be an antidepressant. However, if

flibanserin is an antidepressant, it may be different from the classical antidepressants.

Flibanserin’s mechanism of action in animal models of depression was investigated

in the forced swimming test in mice (23) and learned helplessness in rats (11). In bothmodels, the effects of flibanserin were not reduced by brain 5-HT depletion, suggesting

that the effect of flibanserin does not require the integrity of 5-HT containing neurons.

However, whereas the effect of flibanserin seems to be mediated by 5-HT1A receptor stim-

ulation in mice, this does not seem to occur in rats. In both models, dopamine D1 or D2 an-

tagonists reduced the effects of flibanserin. Active doses of flibanserin in these models

ranged from 16 to 32 mgkg. At this dose range, 5-HT1A and 5-HT2A receptors should beoccupied by approximately 50 to 70% (Table 5). Since flibanserin has lower affinity for

both, rat D1 and D2 receptors (Table 1), it is possible that flibanserin activates dopamine

receptors indirectly, through release of DA. At 10 mgkg flibanserin has been shown to

increase dopamine release (Table 10). The effect of flibanserin in the learned helplessness


FLIBANSERIN 133

TABLE 15. Behavioral differences between flibanserin and imipramine

Test Flibanserin ImipramineFluoxetine

Learned helplessness

in rats

Active after acute administration Active after repeated administration

Chronic mild stress

in mice

Active after acute administration Active after repeated administration

Chick-distress call Increases calls in the “protest”-phase Increases calls in the “frustration”-

phaseDRL-72 Inactive Active after acute administration


18/26

test was also blocked by naloxone (11). Interestingly, like many other antidepressants,flibanserin activates ì-opioid receptors without inducing rewarding properties (11).

By repeated daily administration flibanserin did not produce tolerance to its antide-

pressant-like effects in the chronic mild stress procedure in mice (25), in bulbectomized

rats (16), and in the forced swimming test in mice (Fig. 4). In contrast to imipramine, a

subeffective dose of flibanserin remains inactive even after repeated administrations

(Fig. 5). Thus, it seems that, in contrast to imipramine or fluoxetine, flibanserin has anti-

depressant-like effects after a single administration and that these effects are maintained

over time. With repeated administration of flibanserin to animals, the effect of the 5-HT1Aagonist 8-OH-DPAT on body temperature is attenuated, whereas the effect on forced

swimming test is still present (Fig. 6). It has been suggested that 8-OH-DPAT induces its

effects in the forced swimming test and on body temperature by acting on 5-HT1A post-

and presynaptic receptors, respectively (42). Thus, it appears that flibanserin exert its anti-

depressant-like effect independently from involvement at 5-HT1A presynaptic receptors.

This contrasts with classical antidepressants, the 5-HT postsynaptic activity of which ap-

pears to be secondary to the downregulation of 5-HT1A presynaptic receptors (1,4).

Animal Models Sensitive to Anxiolytics

Flibanserin, at 2 to 50 mgkg, was tested in the elevated plus maze (16), conflict test

(Table 16), and ultrasonic model (54) in rats and in the darklight exploratory test (16) and

stress-induced hyperthermia (16) in mice. At doses ranging from 5 to 50 mgkg s.c.,


134 BORSINI ET AL.

0

60

120

180

240

Vehicle - acute

Flibanserin - acute

I m m o

b i l i t y

t i m e

( s e

c )

Vehicle - Chronic Flibanserin - chronic

*

*

Fig. 4. Force swimming test in mice. Effects of an acute dose of flibanserin, 16 mgkg, or vehicle in mice re-

peatedly treated with either flibanserin, 32 mgkg, or vehicle. Values represent median with interquartile range

from 10 mice. Flibanserin or vehicle was given i.p. Vehicle or flibanserin 32 mg kg was given from day 1 to 5

and, after 2 days of pause, from day 8 to 10. On day 11, the animals received vehicle or flibanserin 32 mgkg

only in the morning. The challenging dose of 16 mgkg was given on day 12. The test, as described in Cesana,et al. (23), was performed 30 min after the challenging dose of flibanserin. This experiment was conducted by

R. Cesana, Boehringer Ingelheim, Italy. Wilcoxon test: * P < 0.01.


19/26

flibanserin reduced ultrasonic vocalization in rat pups, without altering body temperature

or motor activity. At 8 to 16 mgkg i.p., flibanserin exerted anxiolytic-like effects in the

two tests in mice. It is worth noting that flibanserin induced strong sedative effects at

16 mgkg in mice in the dark light test, which was carried out during the dark phase of

the lightdark cycle. This contrasts with the absence of sedative effects on motor activity

in the actimer when registered during the light phase of the lightdark period. In rats,flibanserin was devoid of any effect in the elevated plus maze, and exerted anxiogenic-like

effects in the conflict test (Table 16). The possibility that flibanserin can increase pain sen-

sitivity (as the conflict test is based on a weak electrical shock) was evaluated in mice.


FLIBANSERIN 135

0

60

120

180

240

V E H +

V E H

V E H

+ I M I

I M I +

V E H

I M I +

V E H

V E H +

V E H

V E H

+ I M I

I M I +

V E H

I M I +

V E H

I m m o b i l i t y t i m e ( s e c )

0

60

120

180

240

V E H +

V E H

V E H

+ F L I

F L I +

F L I

F L I +

V E H

V E H +

V E H

V E H

+ F L I

F L I +

V E H

F L I +

V E H

I m

m o b i l i t y t i m e ( s e c )

After 5 day treatment After 10 day treatment

After 10 day treatmentAfter 5 day treatment

*

Fig. 5. Effect of a repeated administration of an acute inactive dose of flibanserin and imipramine in the forced

swimming test in mice. Values are medians with interquartile range from 10 mice. Drugs were given i.p. Mice

were given flibanserin 8 mgkg, imipramine 8 mgkg or vehicle twice daily from days 1 to 4. On day 5, theforced swimming test, as described by Cesana et al. (23), was performed 30 min after drug or vehicle adminis-

tration. The animals were again treated in the afternoon. After 2 days’ pause, the same procedure was followed

from days 8 to 12. This test was performed by R. Cesana, Boehringer Ingelheim, Italy. Wilcoxon test: * P < 0.05.


20/26


136 BORSINI ET AL.

34

34.5

35

35.5

36

36.5

3737.5

38

38.5

39

Chronic Vehicle Chronic Flibanserin

Chronic Vehicle Chronic Flibanserin

Vehicle

8-OH-DPAT

Vehicle

8-OH-DPAT

100

120

140

160

180

200

220

240

R e c t a l

t e m p e r a t u r e ( ° C )

I m m o b i l i t y t i m e ( s e c )

*

*

#

Fig. 6. Effects of 8-OH-DPAT on hypothermia and immobility test in mice repeatedly treated with 32 mg kgflibanserin. Values are medians with interquartile range from 10 mice. Mice were given vehicle or flibanserin

32 mgkg i.p. twice daily from days 1 to 5 and, after 2 days’ pause, from days 8 to 10. On day 11, the animals re-

ceived vehicle or flibanserin 32 mgkg only in the morning. On day 12, 8-OH-DPAT 3 mgkg was administereds.c., and 30 min later body temperature was measured and, thereafter, the forced swimming test was performed.

This test was performed by R. Cesana, Boehringer Ingelheim, Italy. Wilcoxon test: * P < 0.05 vs. respective

vehicle; # P < 0.01 vs. 8-OH-DPAT in chronic vehicle group.

TABLE 16. Effects of flibanserin in the conflict test in rats

Drug Dose (mgkg p.o.)

Number of lever presses in 3 min

Predrug session Postdrug session

Flibanserin 3.7 8.5 (2–32.8) 8.8 (0.8–24)

Flibanserin 7.4 19.3 (1–32.8) 1 (0.3–20)*

Flibanserin 29.6 20 (4–34) 0.5 (0.3–1.5)*

Diazepam 10 17.5 (0.8–32.5) 37.5 (16.3–47)*

Values are median and interquartile range (in bracket) of 17 rats. The operant conflict procedure was

closely related to the original one (30). Predrug session was conducted 1 h after oral vehicle adminis-

tration and 5 h before drug administration. Postdrug session was conducted 1 h after oral flibanserin.

Sedative effects were observed after 29.6 mgkg flibanserin. This experiment was performed byE. Lehr, Boehringer Ingelheim, Germany. Wilcoxon test: * P < 0.01.


21/26

Flibanserin exerted no algesic effects. The conflict procedure was based on food pellet de-

livery, thus a possible effect of flibanserin on food intake cannot be ruled out since it re-

duces food intake in rats at doses as low as 5 mgkg (32). The mechanism of action of

flibanserin in animal models of anxiety has not yet been investigated.

Animal Models Sensitive to Antipsychotics

Flibanserin was studied in tests of psychostimulant-induced hypermotility (14). It re-

duced hypermotility induced by the agonist (+)SKF-100147 or by d-amphetamine at doses

of 4 and 8 mgkg, i.p. in mice and at a dose of 8 mgkg, i.p., in rats. In these experiments,where the animals were habituated to the actimer, flibanserin did not significantly change

motor activity per se. Flibanserin also reduced d-amphetamine-induced stereotypy at doses

of 8 and 16 mgkg in rats and apomorphine-induced stereotypy at 16 mgkg, i.p. (14).

Other Behavioral Effects

5-HT 2A antagonism. Flibanserin, 4.1 mgkg, antagonized head twitches induced by the5-HT2 agonist DOI in mice (12). However, it is questionable whether this effect depends

on 5-HT2A antagonism or rather on 5-HT1A receptor activation (60).

Analgesic affect . Flibanserin antagonized 50% of phenylquinone-induced writhings in

mice at an oral dose of 10.4 (confidence limit = 2.5–42.9), indicating that it has a weak an-

algesic activity (36). Flibanserin displayed antinociceptive effects in the hot plate in mice

only at a dose as high as 128 mgkg, p.o. At this dose, a clear reduction of motor activitywas observed, which could have affected the motor response of animals in the hot plate

test. Nevertheless, this activity was reduced by naloxone and potentiated by morphine.

Interaction with sedative compounds. At 8 mgkg i.p., flibanserin potentiated pentobarbital-induced loss of righting reflex in mice (mean duration of the loss of righting reflex

after 40 mgkg, i.p., pentobarbital: vehicle = 41.3 5 min; flibanserin 8 mgkg = 69.4 6.4 min,

P < 0.05; flibanserin 32 mgkg = 119.6 6.9 min, P < 0.01). At 16 mgkg i.p., but not atlower doses, flibanserin also potentiated barbital-induced loss of righting reflex in mice

(mean duration of the loss of righting reflex after 180 mgkg, i.p., barbital: ve-

hicle = 85.1 21.5 min; flibanserin 16 mgkg = 202.3 28.4 min; P < 0.05; flibanserin

32 mgkg = 289.6 35.3 min; P < 0.01). At 16 mgkg, i.p., flibanserin also potentiateddiazepam-induced loss of righting reflex in mice (number of mice out of 10 with loss of

righting reflex after 4 or 8 mgkg i.p. diazepam: vehicle = 2 and 4, respectively; fliban-

serin 16 mgkg = 6 and 10, respectively; P < 0.05). At 100 mgkg, p.o., but not at lower doses, flibanserin potentiated ethanol-induced loss of righting reflex in mice (number of

mice out of 10 with loss of righting reflex after 1.6 mgkg i.p. ethanol: vehicle = 2; fliban-

serin 100 mgkg = 8; P < 0.01). However, at this dose, flibanserin did not alter ethanol-induced aerial righting reflex (see ref. 29 for methods).

The fact that flibanserin potentiated either barbital (mainly excreted as unchanged by

kidney) and pentobarbital (mainly metabolized by the liver) seems to exclude a potential

metabolic interference between flibanserin and these two barbiturates. This suggests that

flibanserin may interact with the macromolecular complex of GABA where the two barbi-

turates act. This hypothesis is supported also by the finding that flibanserin potentiated the


FLIBANSERIN 137


22/26

effects of other drugs acting as agonists on this GABA macromolecular complex, such as

diazepam and ethanol. However, the interaction of flibanserin with this receptor complex

does not seem to be direct, as suggested by the low affinity of flibanserin for the benzodia-

zepine receptors, the GABA-A agonist site, and GABA-A Cl –

channel (Table 1). Fliban-serin failed to increase extracellular concentrations of GABA (Table 4). Thus, flibanserin

seems to interact with GABAergic transmission indirectly.

In addition, in rats at single doses up to 32 mgkg i.p., flibanserin does not impair

learning in a water maze (14). At doses up to 64 mgkg i.p. (maximal dose tested) fliban-serin did not produce any relevant changes in blood pressure or heart rate of conscious

rats. At doses up to 100 mgkg p.o., flibanserin did not injure gastric mucosa or alter gas-

trointestinal transit in rats. In rats, at doses up to 30 mgkg i.p., flibanserin did not alter

basal- or cholinergic stimulation-induced salivation; at 30 mgkg p.o. it had no effect onurinary volume or electrolytes.

CONCLUSIONS

Flibanserin is capable of modifying animal behavior at two different dosage levels

(Table 17). At doses below 10 mgkg, it exerts prolongation of REM latency, anti-anhe-

donic effects, mild analgesia, facilitates exploration in the lightdark test, reduces ultra-sound vocalization, antagonizes exploratory activity in bulbectomized rats, reduces psycho-

stimulant-induced hypermotility, and potentiates barbiturate-induced loss of righting

reflex. At these doses, flibanserin seems to exert effects solely by activating 5-HT1A re-ceptors. At higher doses (16 to 40 mgkg) flibanserin had effects in other tests involvinginteractions with other receptors.

The mechanism of action of flibanserin is not the same in all tests. It may directly or in-

directly activate 5-HT1A, D1, D2, D4, and ì-opioid receptors. Flibanserin also increases the

function of GABA multicomplex receptors. The antagonistic activity of flibanserin at

5-HT2A receptors has been observed in biochemical and electrophysiological tests; it is

less evident in behavioral models. In most behavioral tests flibanserin exerts its maximal

activity for a short period of time, which is in agreement with its half-life. This contrasts

with its longer duration of action observed in some biochemical and electrophysiological

parameters. The only exception in behavioral tests is represented by the activity of

flibanserin in the chronic mild test model in mice, where its activity lasted 18 h (25). The

contribution of potential active metabolite(s) cannot be excluded, and this issue needs to

be elucidated.

The effects on motor activity seem to be test-related at low doses, but sedation is

clearly evident at doses above 32 mgkg. Additionally, flibanserin does not appear to in-terfere with potential side effects induced by antidepressants, such as serotonergic syn-

drome (13) and cardiovascular effects (J. Van Ryn, Boehringer Ingelheim, data on file).

The fact that flibanserin, in contrast to other 5-HT1A agonists, reduces the activity of

cortical pyramidal cells makes this compound a unique tool to study the physiopathology

of cortical glutamatergic pyramidal cells. This fact, combined with the findings that fli-

banserin displays antidepressant-, antipsychotic- and probably anxiolytic-like properties,

indicates that this compound may be an interesting psychotropic drug.


138 BORSINI ET AL.


23/26

REFERENCES

1. Artigas F, Romero L, de Montigny C, Blier P. Acceleration of the effect of selected antidepressant drugs in

major depression by 5-HT1A antagonists. Trends Neurosci 1996;19:378–383.

2. Asghari V, Sanyal S, Buchwaldt S, Paterson A, Javanovic V, Van Tol HHM. Modulation of intracellular

cyclic AMP levels by different human dopamine D4 receptor variants. J Neurochem 1995;65:1157–1165.

3. Beck SG, Choi KC, List TJ. Comparison of 5-hydroxytryptamine1A-mediated hyperpolarization in CA1 and

CA3 hippocampal pyramidal cells. J Pharmacol Exp Ther 1992;263:350–335.

4. Blier P, de Montigny C. Current advances and trends in the treatment of depression. Trends Pharmacol Sci

1994;15:220–226.5. Blier P, Lista A, de Montigny C. Differential properties of pre- and postsynaptic 5-hydroxytryptamine

1A re-

ceptors in the dorsal raphe and hippocampus. I. Effect of spiperone. J Pharmacol Exp Ther 1993;265:7–15.


FLIBANSERIN 139

TABLE 17. The lowest doses of flibanserin that have behavioral effects or the highest doses

free of any behavioral changes

Dose Route

Activity Inactivity

Rats Mice Rats or mice

3 p.o. REM latency

2.5 i.p. Chronic mild stress Chronic mild stress

4 i.p. Antagonism of amphetamine

hypermotility

Antagonism of SKF hypermotility

5 s.c. Ultrasound vocalization

8 i.p. Reduction in motor activity Lightdark exploratory test FSTreverse

Antagonism of amph. hypermotility

Antagonism of SKF hypermotility

Pentobarbital-induced loss of RR

s.c. Forced swimming test

10 i.p. Bulbectomized rats

Hypomotility after amph. withdrawal

p.o. Antagonism of writhings

16 i.p. Killer rats Forced swimming test Elevated

plus maze

Stress-induced hyperthermia

Diazepam-induced loss of RR

Barbital-induced loss of LRR

20 i.p. DRL-7224 i.p. Learned helplessness

30 i.p. Tail sus-

pension test

32 i.p. Morris

maze

p.o. FST

40 p.o. FI

100 p.o. Ethanol-induced loss of RR

128 p.o. Antinociception in the hot-plate

FST, forced swimming test; SKF, SKF100147; amph., amphetamine; RR, righting reflex.


24/26

6. Blier P, Lista A, de Montigny C. Differential properties of pre- and postsynaptic 5-hydroxytryptamine1A re-

ceptors in the dorsal raphe and hippocampus. II. Effect of pertussis and cholera toxin. J Pharmacol Exp Ther

1993;265:15–23.

7. Bohmaker K, Bordi F, Meller E. The effect of pertussis toxin on dopamine D2 and serotonin 5-HT1A

autoreceptor-mediated inhibition of neurotransmitter synthesis: Relation to receptor reserve. Neuropharma-cology 1992;31:451–459.

8. Borsini F. Balance between cortical 5-HT1A and 5-HT2 receptor function: Hypothesis for a faster antide-

pressant action. Pharmacol Res 1994;30:1–11.

9. Borsini F. Pharmacology of 5-HT1A

receptors: Critical aspects. CNS Spectrum 1998;3:17–38.

10. Borsini F, Cesana R. Further characterization of potential antidepressant action of flibanserin. Psycho-

pharmacology 2001;159:64–69.

11. Borsini F, Cesana R. Mechanism of action of flibanserin in the learned helplessness paradigm in rats. Eur J

Pharmacol 2001;433:81–89.

12. Borsini F, Brambilla A, Ceci A, et al. Flibanserin. Drugs Fut 1998;23:9–16.

13. Borsini F, Brambilla A, Cesana R, Grippa N. Lack of interaction between flibanserin and antidepressants in

inducing serotonergic syndrome in rats. Int J Neuropsychopharmacol 2001;4:9–15.

14. Borsini F, Brambilla A, Grippa N, Pitsikas N. Behavioral effects of flibanserin (BIMT 17). Pharmacol

Biochem Behav 1994;64:137–146.15. Borsini F, Ceci A, Bietti G, Donetti A. BIMT 17, a 5-HT1A receptor agonist5-HT2A antagonist, directly ac-

tivates postsynaptic 5-HT inhibitory responses in the rat cerebral cortex. Naunyn-Schmiedeberg’s Arch

Pharmacol 1995;352:283–290.

16. Borsini F, Cesana R, Kelly J, et al. BIMT 17: A putative antidepressant with a fast onset of action? Psycho-


17. Borsini F, Giraldo E, Monferini E, et al. BIMT 17, a 5-HT2A

receptor antagonist and 5-HT1A

receptor full

agonist in rat cerebral cortex. Naunyn-Schmiedeberg’s Arch Pharmacol 1995;352:276–282.

18. Brambilla A, Baschirotto A, Grippa N, Borsini F. Effect of flibanserin (BIMT 17), fluoxetine, 8-OH-DPAT

and buspirone on serotonin synthesis in rat brain. Eur Neuropsychopharmacol 1999;10:63–67.

19. Bryant HU, Nelson DL, Button D, et al. A novel class of 5-HT2A receptor antagonist: Aryl aminoguanidines.

Life Sci 1996;15:1259–1268.

20. Casanovas JM, Hervas I, Artigas F. Postsynaptic 5-HT1A

receptors control 5-HT release in the rat medial

prefrontal cortex. Neuroreport 1999;10:1441–1445.21. Ceci A, Baschirotto A, Borsini F. The inhibitory effect of 8-OH-DPAT on the firing activity of dorsal raphe

serotonergic neurons in rats is attenuated by lesion of the frontal cortex. Neuropharmacology 1994;33:

709–713

22. Ceci A, Fodritto F, Borsini F. Repeated treatment with fluoxetine decreases the number of spontaneously

active cells per track in frontal cortex. Eur J Pharmacol 1994;271:231–234.

23. Cesana R, Ciprandi C, Borsini F. The effect of BIMT 17, a new potential antidepressant, in the forced swim-

ming test in mice. Behav Pharmacol 1995;6:688–694.

24. Clarke WP, Yocca FD, Maayani S. Lack of 5-hydroxytryptamine1A-mediated inhibition of adenylyl cyclase

in dorsal raphe of male and female rats. J Pharmacol Exp Ther 1996;277:1259–1266.

25. D’Aquila P, Monleon S, Borsini F, Brain P, Willner P. Anti-anhedonic actions of the novel serotonergic agent

flibanserin, a potential rapidly acting antidepressant. Eur J Pharmacol 1997;340:121–132

26. Darmani NA, Martin BR, Pandy U, Glennon RA. Does functional relationship exist between 5-HT1A

and

5-HT2 receptors? Pharmacol Biochem Behav 1990;36:901–906.27. Engel G, Göthert M, Hoyer D, Schlicker E, Hillenbrand K. Identity of inhibitory presynaptic 5-hydro-

xytryptamine (5-HT) autoreceptors in the rat brain cortex with 5-HT1B

binding sites. Naunyn-Schmiede-

berg’s Arch Pharmacol 1986;332:1–7.

28. Ericson H, Radesäter AC, Servin E, Magnusson O, Mohringe B. Effects of intermittent and continuous

subchronic administration of raclopride on motor activity, dopamine turnover and receptor occupancy in the

rat. Pharmacol Toxicol 1996;79:277–286.

29. Frey G, Breese GR, Mailman RB, Vogel RA, Ondrusek MG, Mueller RA. An evaluation of the selectivity of

fenmetazole (DH-524) reversal of ethanol-induced changes in central nervous system function. Psycho-


30. Geller I. Relative potencies of benzodiazepines as measured by their effects on conflict behavior. Arch Int

Pharmacodyn Ther 1964;159:243–247.

31. Gilbert Evans SE, DeSouza NJ, Vaccarino FJ. The effects of flibanserin on amphetamine withdrawal-

induced hypolocomotion in rats. Soc Neurosci Abs 1998;24:2133.32. Gilbert Evans SE, Vaccarino FJ. Flibanserin, a 5-HT

1A agonist5-HT2A antagonist, decreases sucrose intake

in operant and non-operant paradigm. Soc Neurosci 2000;26:394.


140 BORSINI ET AL.


25/26

33. Grandy DK, Zang Y, Bouvier C, et al. Multiple human D5 dopamine receptor genes: A functional receptor

and two pseudogenes. Proc Natl Acad Sci USA 1991;88:9175–9179.

34. Hajos M, Hajos-Korcsok E, Sharp T. Role of the medial prefrontal cortex in 5-HT1A receptor-induced inhi-

bition of 5-HT neuronal activity in the rat. Br J Pharmacol 1999;126:1741–1750.

35. Hall MD, El Mestikawy, Emerit MB, Pichat L, Hamon M, Gozlan H. [3

H]8-hydroxy-2-(di-n-propylamino)-tetralin binding to pre- and postsynaptic 5-hydroxytryptamine sites in various regions of the rat brain.

J Neurochem 1985;44:1685–1696.

36. Hendershot LC, Forsaith J. Antagonism of the frequency of phenylquinone-induced writhing in the mouse

by weak analgesics and nonanalgesics. J Pharmacol Exp Ther 1959;125:237–240.

37. Hjorth S, Sharp T. Effect of the 5-HT1A

receptor agonist 8-OH-DPAT on the release of 5-HT in dorsal and

median raphe-innervated rat brain regions as measured by in vivo microdialysis. Life Sci 1991;48:

1779–1786.

38. Kapur S, Remington GR, Zipursky RB, Wilson AA, Houle S. The D2 dopamine receptor occupancy of

risperidone and its relationship to extrapyramidal symptoms: A PET study. Life Sci 1995;10:PL103–PL107.

39. Kirby LG, Chou-Green JM, Davis K, Lucki I. The effects of different stressors on extracellular 5-hydro-

xytryptamine and 5-hydroxyindoleacetic acid. Brain Res 1997;760:218–230.

40. Limberger N, Deicher R, Starke K. Species differences in presynaptic serotonin autoreceptors: Mainly

5-HT1B but possibly in addition 5-HT1D in the rat, 5-HT1D in the rabbit and guinea-pig brain cortex. Naunyn-Schmiedeberg’s Arch Pharmacol 1991;343:353–364.

41. Lucchelli A, Santagostino-Barbone MG, D’Agostino G, Masoero E, Tonini M. The interaction of antide-

pressant drugs with enteric 5-HT7 receptors. Naunyn-Schmiedeberg’s Arch Pharmacol 2000;362:284–289.

42. Luscombe GP, Martin KF, Hutchins LJ, Gosden J, Heal DJ. Mediation of the antidepressant-like effect of

8-OH-DPAT in mice by postsynaptic 5-HT1A

receptors. Br J Pharmacol 1993;108:669–677.

43. Marazziti D, Marracci S, Palego L, et al. Localization and gene expression of serotonin1A

receptors in

human brain postmortem. Brain Res 1994;658:55–59.

44. Marazziti D, Palego L, Giromella A, et al. Pre- versus post-synaptic effects of flibanserin and buspirone on

adenylyl cyclase activity in human brain. Int J Neuropsychopharmacol 2002. In press.

45. Marazziti D, Rossi, A, Palego L, et al. [3H]Ketanserin binding in human brain postmortem. Neurochem Res

1997;22:753–757.

46. Marlier L, Teillhac JR, Cerruti C, Privat A. Autoradiographic mapping of 5-HT1A

, 5-HT1B

and 5-HT2 re-

ceptors in the rat spinal cord. Brain Res 1991;550:15–23.47. Meller E, Goldstein M, Bohmaker K. Receptor reserve for 5-hydroxytryptamine

1A-mediated inhibition of

serotonin synthesis: Possible relationship to anxiolytic properties of 5-hydroxytryptamine1A agonists. Mol

Pharmacol 1990;37:231–237.

48. Mierau J, Schneider FJ, Ensinger HA, Chio CL, Lajiness ME, Huff RM. Pramipexole binding and activation

of cloned and expressed dopamine D2, D3 and D4 receptors. Eur J Pharmacol 1995;290:29–36.

49. Murray AM, Hyde TM, Knable MB, et al. Distribution of putative D4 dopamine receptors in postmortem

striatum from patients with schizophrenia. J Neurosci 1995;15:2186–2191.

50. Newman-Tancredi A, Wootton R, Strange PG. High-level stable expression of recombinant 5-HT1A 5-hydro-

xytryptamine receptors in Chinese hamster ovary cells. Biochem J 1992;285:933–938.

51. Ohno M, Yamamoto T, Watanabe S. Blockade of 5-HT2 receptors protects against impairment of working

memory following transient forebrain ischemia in the rat. Neurosci Lett 1991;129:185–188.

52. Patacchini R, Maggi CA. Tachykinin receptors and receptor subtypes. Arch Int Pharmacodyn 1985;329:

161–184.

53. Pehek EA, Yanming B. Ritanserin administration potentiates amphetamine-stimulated dopamine release in

the rat prefrontal cortex. Prog Neuro-Psychopharmacol Biol Psychiat 1997;21:671–682.

54. Podhorna J, Brown RE. Flibanserin has anxiolytic effects without locomotor side effects in the infant ultra-

sonic vocalization model of anxiety. Br J Pharmacol 2000;130:739–746.

55. Raymond JR, Albers FJ, Middleton JP. Functional expression of human 5-HT1A receptors and differential

coupling to second messengers in CHO cells. Naunyn-Schmiedeberg’s Arch Pharmacol 1992;346:127–137.

56. Romero L, Celada P, Artigas F. Reduction of in vivo striatal 5-hydroxytryptamine release by 8-OH-DPAT

after inactivation of GiGo proteins in dorsal raphe nucleus. Eur J Pharmacol 1994;265:103–106.

57. Rueter LE, Blier P. Electrophysiological examination of the effects of sustained flibanserin administration

on serotonin receptors in rat brain. Br J Pharmacol 1999;126:627–638.

58. Rueter LE, de Montigny C, Blier P. In vivo electrophysiological assessment of the agonistic properties of

flibanserin at pre- and postsynaptic 5-HT1A receptors in the rat brain. Synapse 1998;29:392–404.59. Saito H, Matsumoto M, Togashi H, Yoshioka M. Functional interaction between serotonin and other

neuronal systems: Focus on in vivo microdialysis studies. Jpn J Pharmacol 1996;70:203–225.


FLIBANSERIN 141


26/26

60. Scandroglio A, Monferini E, Borsini F. Ex vivo binding of flibanserin to serotonin 5-HT1A and 5-HT2A re-

ceptors. Pharmacol Res 2001;43:179–183.

61. Schmidt CJ, Fadayel GM. The selective 5-HT2A receptor antagonist, MDL 100,907, increases dopamine

efflux in the prefrontal cortex of the rat. Eur J Pharmacol 1995;273:273–279.

62. Stelner R, Wittliff JL. A whole cell assay for glucocorticoid binding sites in normal human lymphocytes.

J Clin Chem 1985;31:1855–1860.

63. Sumiyoshi T, Stockmeier CA, Overholser JC, Thimpson PA, Meltzer HY. Dopamine D4 binding and effects

of guanine nucleotides on [3H] raclopride binding in postmortem caudate nucleus of subjects with schizo-

phrenia or major depression. Brain Res 1995;681:109–116.

64. Thorre K, Chaouloff F, Sarre S, Meeusen R, Ebinger G, Michotte Y. Differential effects of restraint stress on

hippocampal 5-HT metabolism and extracellular levels of 5-HT in streptozotocin-diabetic rats. Brain Res

1997;772:209–216.

65. Willins DL, Meltzer HY. Direct injection of 5-HT2A receptor agonists into the medial prefrontal cortex pro-

duces a head-twitch response in rats. J Pharmacol Exp Ther 1997;282:699–706.

66. Yocca FK, Iben L, Meller E. Lack of apparent receptor reserve at postsynaptic 5-hydroxytryptamine1A

re-

ceptors negatively coupled to adenylyl cyclase activity in rat hippocampal membranes. Mol Pharmacol

1992;41:1066–1072.

67. Zhou QY, Grandy DK, Thambi L, et al. Cloning and expression of human and rat D1 dopamine receptors.

Nature 1990;347:76–80.

142 BORSINI ET AL.

Borsini Et Al-2002-CNS Drug Reviews

Documents

Transcript of Borsini Et Al-2002-CNS Drug Reviews