Psychiatry Research 113(2002) 41–47

0165-1781/02/$ - see front matter� 2002 Elsevier Science Ireland Ltd. All rights reserved.PII: S0165-1781Ž02.00224-X

D-Fenfluramine-evoked serotonergic responses inolanzapine-treated schizophrenic patients

Hugh Jones , Vivienne A. Curtis , Padraig Wright , Lyn S. Pilowsky , James V. Lucey *a a a a b,

Institute of Psychiatry, De Crespigny Park, London SE5 8AF, UKa

Department of Psychiatry, Royal College of Surgeons in Ireland, Dublin, Irelandb

Received 12 December 2001; received in revised form 28 June 2002; accepted 19 September 2002

Abstract

Antagonist activity at the 5-HT receptor may contribute to the therapeutic efficacy of atypical antipsychotics in2

schizophrenia. This neuroendocrine study examined the in vivo functional serotonergic(5-HT) activity of the atypicalantipsychotic olanzapine. We examined central 5-HT responses by measuring the serum prolactin(PRL) over 5 h in2

response to 30 mg ofD-fenfluramine(DFEN) in two groups of male schizophrenic patients. Blunted PRL responsesto DFEN indicate functional 5-HT receptor antagonism. Seven patients treated with olanzapine at a mean(S.D.)2

dose of 13.1(4.6) for a mean of 28 weeks were compared with a matched group of eight patients who had receivedno antipsychotic treatment for at least 2 weeks. Baseline PRL levels did not differ significantly in the two patientgroups and were within the normal range. The olanzapine-treated patients showed a significantly lower maximalDFEN-evoked PRL response and a significantly lower group=time overall PRL release compared with the untreatedpatient group. We have previously demonstrated a similar degree of functional in vivo 5-HT antagonism with the2

atypical antipsychotic clozapine. This study thus suggests that this activity may not contribute to the unique clinicalefficacy of clozapine.� 2002 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Serotonin; Neuroendocrine; Schizophrenia;D-Fenfluramine; Antipsychotic

1. Introduction

The neurochemical basis for the therapeuticprofile of olanzapine and other atypical antipsy-chotics remains uncertain. Clozapine, the prototyp-ical atypical antipsychotic, has relatively low invitro and in vivo affinity at dopamine D receptors2

(Nordstrom et al., 1995; Schotte et al., 1996). This¨

*Corresponding author. Academic Centre, James ConnollyMemorial Hospital, Dublin 15, Ireland. Tel.:q353-1-821-3844x5635; fax:q353-1-886-4051.

E-mail address: jvlucey@rcsi.ie(J.V. Lucey).

led to renewed interest in other receptor systems,in particular serotonin 5-HT receptors, that might2a

contribute to the mechanism of antipsychotic drugaction. Many atypical antipsychotics show high invitro affinity at this receptor(Bymaster et al.,1996; Schotte et al., 1996). Olanzapine has beenshown to produce significant in vivo occupancy at5-HT receptors using in vivo receptor imaging2

(Kapur et al., 1998). A relatively high ratio ofaffinity at 5-HT yD receptors has been used to2a 2

differentiate atypical from typical antipsychoticdrugs(Meltzer, 1999). The novel 5-HT antagonist2

42 H. Jones et al. / Psychiatry Research 113 (2002) 41–47

M-100907, which is devoid of activity at dopaminereceptors, may have significant activity, at leastagainst positive psychotic symptoms(Carlsson,2000). In vivo receptor imaging studies haveexamined 5-HT receptor blockade by a number2a

of atypical antipsychotics. Many of these studiesexamining atypical antipsychotics show receptorblockade at variance with their in vitro data(Kapuret al., 1999). Thus, there is a need to establish thein vivo 5-HT activity of these drugs in clinically2a

treated patient samples.An early neuroendocrine study with olanzapine

demonstrated functional serotonergic antagonismin rats using inhibition of pergolide-evoked corti-costerone release as a measure of serotonergicantagonism(Fuller and Snoddy, 1992). The agentmeta-chlorophenylpiperazine(m-CPP) has beenused in in vivo studies to assess drug effects inschizophrenic patients. This has, however, relative-ly non-specific actions at 5-HT receptors thatprobably include activity at 5-HT in addition to1a

5-HT receptors(Meltzer and Maes, 1995).2ay2c

Blunting of 5-HT-mediated neuroendocrineresponses usingm-CPP is observed following treat-ment with the atypical antipsychotic agent cloza-pine, but not the conventional antipsychoticshaloperidol or fluphenazine(Kahn et al., 1993;Owen et al., 1993). A recent study has examinedthis probe in olanzapine-treated schizophrenicpatients(Scheepers et al., 2000). They reportedsignificant serotonergic antagonism with olanza-pine treatment, although the degree of antagonismwas reported as less than that in a clozapine-treated patient sample. Differences in the function-al activity at 5-HT receptors may contribute to thedifferential therapeutic efficacy of olanzapine andclozapine.The prolactin(PRL) response toD-fenfluramine

(DFEN) is a specific test of serotonergic activity(Invernizzi et al., 1986). DFEN is a specific 5-HTreleasing agent and reuptake inhibitor(Garattini etal., 1987). Its neuroendocrine effects in humanshave been shown to be mediated by postsynaptic5-HT and 5-HT receptors(Goodall et al.,2c 2a

1993). Manufacture of this compound has nowstopped because of a reported association betweenthe use of DFEN in combination with phentermine

in the development of valvular heart defects(Con-nolly et al., 1997).Drug-free chronic schizophrenic subjects show

enhanced DFEN-evoked PRL responses comparedto healthy control subjects(Monteleone et al.,1999). We have previously demonstrated thatsymptomatic improvement following clozapinetreatment is associated with alterations in DFEN-evoked PRL responses(Curtis et al., 1995). Thisreduction in DFEN-evoked PRL release persistsduring chronic clozapine treatment and shows asignificant positive correlation with specific posi-tive schizophrenic symptoms, namely hallucina-tions and delusions(Jones et al., 1998a).Risperidone treatment is also associated with asignificant reduction in DFEN-evoked PRL release(Jones et al., 1998b).We examined central 5-HT receptor actions of

the atypical antipsychotic olanzapine. We used theDFEN-evoked PRL response as a measure ofcentral 5-HT antagonism. We hypothesised that2

olanzapine-induced blunting of serotonergic activ-ity would be evident through a reduction in theDFEN-evoked PRL response. We wished furtherto examine the hypothesis that the atypical anti-psychotics clozapine and olanzapine differ in theiractivity at 5-HT receptors as measured by the2

DFEN-evoked PRL response.

2. Methods

Ethical approval was obtained for the studyfrom local ethical committees. All patients gavewritten informed consent before entry.

2.1. Patient sample

Two groups of male schizophrenic patients wereincluded in the study. Seven patients with a DSM-IV diagnosis of schizophrenia had received olan-zapine for a mean(S.D.) of 28 (17) weeks. Thesepatients were receiving a mean(S.D.) daily olan-zapine dose of 13.1(4.6) mg (range 10–20 mgdaily). The patients in this group had a low levelof positive and negative symptoms, as shown bytheir mean symptom scores(Table 1). All patientshad a history of previous antipsychotic treatment.Four patients had been treated with at least three

43H. Jones et al. / Psychiatry Research 113 (2002) 41–47

Table 1Clinical symptom ratings in olanzapine-treated and untreatedpatient groups

Untreated patients Olanzapine-treated(ns8) patients(ns7)

SAPS score 51(32) 19 (18)SANS score 49(13) 16 (14)

Results are given as mean(S.D.).

different antipsychotic drugs during a period of atleast 10 years. Two patients had been treated withat least two different antipsychotic drugs during aperiod of at least 5 years. One patient had beentreated previously with a single typical antipsy-chotic for a period of 1 year. None of the patientshad been classified as antipsychotic treatmentresistant and they had not been considered fortreatment with the atypical antipsychotic clozapine.These were compared with a similar group of

eight patients with a DSM-IV diagnosis of schiz-ophrenia. These comparison subjects had previous-ly received unsuccessful treatment with at leastthree different typical antipsychotic drugs duringthe preceding 2 years and all were tested after amean(S.D.) neuroleptic-free ‘washout’ period of3.87 (1.89) weeks. This group had never receivedtreatment with any atypical antipsychotic.Mean (S.D.) age in the olanzapine group of

39.6 (6) years and in the untreated patient groupof 37 (13.7) years were not significantly different(tsy0.7, d.f.s13,Ps0.94). Mean(S.D.) weightin the olanzapine group of 73.5(3.8) kg and inthe untreated patient group of 72.8(6.8) kg werenot significantly different (tsy0.24, d.f.s13,Ps0.8).All patients were assessed at the time of the

neuroendocrine challenge using the Scale for theAssessment of Positive Symptoms(SAPS)(Andreasen, 1981a) and the Scale for the Assess-ment of Negative Symptoms(SANS) (Andreasen,1981b) by two investigators(HJ and VC). Clinicalratings on the two patient groups are shown inTable 1.

2.2. Neuroendocrine challenge

Patients presented to hospital at 09:00 h afteran overnight fast. An in-dwelling intravenous cath-

eter was inserted and patients were allowed to restfor 30 min. Patients were then given 30 mg ofD-fenfluramine(DFEN) and baseline blood sampleswere taken. Further blood samples were taken at60, 120, 180, 240 and 300 min. Serum wasseparated by centrifuge at 3000 rev.ymin for 15min and serum frozen prior to blind batch analysisfor PRL estimation.Patients were encouraged to remain supine and

awake throughout the study. Patients did not eatand drank only water during the study. The IMxsystem microparticle enzyme immunoassay wasused, with a coefficient of variation of 5% overthe range 506.2–620 mUyl for PRL measurement.

2.3. Data analysis

Baseline hormone values for each DFEN chal-lenge test were calculated by taking the mean ofvalues at timey15 and 0 min. For each patientPRL release was calculated by an area-under-the-curve analysis on all time points. The curvebetween points was estimated by a straight lineand the area calculated by simple algebra. PRLrelease in the two patient groups was comparedusing independent two-samplet-tests. PRL releasewas also compared over all time points using arepeated-measures analysis of variance to test forsignificant effects of patient group and time onPRL release. In this analysis, patient age, baselinePRL levels and patient weight were incorporatedas covariates into the model.

3. Results

3.1. Baseline PRL measures

Mean PRL release over all time points in thetwo patient groups is shown in Fig. 1. Baselinemean(S.D.) PRL levels of 212(90) mUyl in theolanzapine-treated sample and 280(136) mUyl inthe untreated sample were not significantly differ-ent (ts1.1, d.f.s13, Ps0.3). There was no sig-nificant correlation between baseline PRL levelsand overall PRL release as measured by the areaunder the curve(AUC) (Rs0.09,Ps0.74).

44 H. Jones et al. / Psychiatry Research 113 (2002) 41–47

Fig. 1. DFEN-evoked PRL release in the olanzapine and untreated schizophrenic patient samples. Graph shows mean PRL releasein the two subject groups at each time point. Error bars represent 1 S.D. at each time point.

3.2. AUC analysis

Mean (S.D.) AUC values in the olanzapine-treated group of 59.7(61.7) and in the untreatedpatient group of 379.7(331.3) were significantlydifferent (ts2.5, d.f.s13, Ps0.02). In the olan-zapine-treated group, AUC was not significantlycorrelated with olanzapine dose(rs0.4, Ps0.3)or olanzapine treatment duration(rs0.3,Ps0.4).In the untreated group, AUC was not significantlycorrelated with previous antipsychotic treatment(rs0.1, Ps0.8). For all the patients consideredtogether, AUC was not significantly correlatedwith patient age(rsy0.2,Ps0.47), illness dura-tion (tsy0.047, Ps0.87) or weight (tsy0.2,Ps0.46).

3.3. Repeated-measures ANOVA

Patient age, weight and illness duration wereused as covariates in the analysis.

3.3.1. Within-subject effectsRepeated-measures ANOVA revealed a signifi-

cant group=time interaction (Fs3.0, d.f.s1,5,Ps0.02). There was no significant effect on PRLrelease by time(Fs1.4, d.f.s1,5, Ps0.22) orthe interaction of group=time (Fs3.1, d.f.s1,5,

Ps0.01). There were no significant interactionsbetween time and any of the covariates.

3.3.2. Between-subject effectsNo factor produced a significant between-sub-

ject effect. The only factor for which the effectapproached significance was patient group(Fs2.4, d.f.s1,10, Ps0.1). However, prolactinrelease was being compared across all time points,and the nature of the action of DFEN is that noeffect would be expected before 120 min, hencethe need for the long duration of the study.

4. Discussion

4.1. 5-HT antagonism of olanzapine

The study reports significant attenuation of thePRL response to DFEN in patients chronicallytreated with olanzapine. We believe this result isindicative of significant attenuation of 5-HT res-2

ponsivity in olanzapine-treated patients. However,that is not the only possible explanation for theresult. Olanzapine has significant D antagonism2

and such receptors importantly control PRLrelease. However, in all the olanzapine-treatedpatients, baseline PRL levels were within normallimits compared with levels in drug-free subjects.This suggests that the D antagonism of olanzapine2

45H. Jones et al. / Psychiatry Research 113 (2002) 41–47

did not contribute significantly to its effect onPRL release during this study.It seems unlikely that olanzapine treatment

attenuated PRL release by involving mechanismsother than the 5-HT system. Typical antipsychoticdrugs enhance baseline PRL values consistent withtheir potent D antagonism, but do not attenuate2

PRL release during a 5-HT challenge(Kahn et al.,1993; Owen et al., 1993). These results supportthe hypothesis that the attenuation of DFEN-evoked PRL release in olanzapine-treated patientsis mediated by the 5-HT antagonism of olanzapine.

4.2. Study limitations

The neuroendocrine challenge was not per-formed using a placebo-controlled design; howev-er, it is practically and ethically difficult to keeppatients drug-free merely to allow such a protocol.Although baseline PRL neuroendocrine status didnot differ significantly between the two patientgroups, there may be some concern at the increasedmean PRL level in the untreated group. Thisdifference may reflect increased anxiety in themore symptomatic untreated patients. Indeed, by60 min when no specific effects of DFEN onperipheral PRL levels would be expected, the meanPRL levels were more similar in the two groups.As far as possible, the groups were closely

matched. In particular, age, illness duration andbody weight did not differ significantly betweenthe two patient groups. All patients had receivedexposure to previous typical antipsychotic treat-ment and most had a previous duration of suchtreatment for more than 5 years. It is, however,true that the olanzapine group was not treatmentresistant as strictly defined, although they mostlyhad a history of poor response to at least onetypical antipsychotic. While the untreated groupwas described as treatment resistant, none of thesepatients had been treated with an atypical anti-psychotic. The samples are small, yet such is thesize of the effect that the significant difference isunlikely to represent a type 1 error. The use of abetween-group design may be considered inferiorto a within-group design, but in so far as possiblewe have overcome this limitation by close match-ing of samples.

It may be argued that the D antagonism of2

olanzapine contributed to blunting of the DFEN-evoked PRL release in these patients. However,baseline PRL levels in this group were not signif-icantly elevated. Furthermore, 5-HT-mediated neu-roendocrine responses are not reduced in patientsreceiving typical antipsychotics, despite the potentD antagonism of these drugs(Kahn et al., 1993;2

Owen et al., 1993).

4.3. 5-HT antagonism in novel antipsychotics2

Previous studies by ourselves and otherresearchers demonstrate significant in vivo func-tional 5-HT antagonism by the atypical antipsy-2

chotic clozapine(Kahn et al., 1993; Jones et al.1998a) and the novel antipsychotic risperidone(Jones et al., 1998b). It is not possible to comparedirectly patients in our studies treated with olan-zapine and clozapine. However, the profoundblockade of functional 5-HT activity demonstrat-2

ed by olanzapine in this study suggests that clo-zapine treatment is most unlikely to exhibit greaterfunctional in vivo 5-HT antagonism than olanza-2

pine treatment. This is consistent with evidencefor similar activity at this receptor of these twodrugs from both in vitro neurochemical(Bymasteret al., 1996; Schotte et al., 1996) and in vivoreceptor imaging data(Kapur et al., 1999). How-ever, it is at variance with the recently reportedneuroendocrine study using a different 5-HT probe(mCPP) (Scheepers et al., 2000). This differencemay reflect differences in the receptor selectivityof the two neuroendocrine probesmCPP andDFEN. DFEN mediates its effects mainly at 5-HT receptors, whereasmCPP shows greater2ay2c

selectivity at the 5-HT and 5-HT receptors.1 2c

(Kahn and Wetzler, 1991; Meltzer and Maes, 1995;Sabbe et al., 2001). Our study adds to existingevidence that the 5-HT receptor site may be a2

common site of action for olanzapine and otheratypical antipsychotic drugs.

4.4. 5-HT antagonism and treatment-resistant2

schizophrenia

Drugs with little 5-HT activity, such as halo-2

peridol, produce clear therapeutic benefit in up to

46 H. Jones et al. / Psychiatry Research 113 (2002) 41–47

70% of patients with schizophrenia. 5-HT antag-2

onism may, however, be important in the treatmentof the patient subpopulation with a poor thera-peutic response to typical antipsychotic drugs.Exaggerated serotonergic neuroendocrineresponses have been demonstrated in unmedicatedschizophrenic patients with a history of poorresponse to typical antipsychotics(Monteleone etal., 1999). In drug-naıve patients, exaggerated¨serotonergic neuroendocrine responses areassociated with poorer therapeutic response tosubsequent typical antipsychotic treatment(Mohret al., 1998). Persistent positive symptoms inclozapine-treated schizophrenic patients are signif-icantly associated with exaggerated serotonergicneuroendocrine responses(Jones et al., 1998a).Olanzapine may not share the same clinical

efficacy as clozapine in ‘treatment-resistant’ schiz-ophrenia(Chakos et al., 2000). This study suggeststhat the functional 5-HT antagonism of clozapine2

as measured by DFEN-evoked PRL release isunlikely to be greater than that achieved by olan-zapine. Hence the differential therapeutic activityof these drugs may not depend crucially on theiractivity at 5-HT receptors. Nevertheless, this2a

study demonstrates that olanzapine exhibits signif-icant in vivo 5-HT antagonism. Such activity may2

contribute to either the clinical efficacy or side-effect profile of this antipsychotic.

Acknowledgments

The hormonal assays were funded by an unres-tricted educational grant from Eli Lilly.

References

Andreasen, N.C., 1981. The Scale for the Assessment ofPositive Symptoms(SAPS). University of Iowa, Iowa City,Iowa.

Andreasen, N.C., 1981. The Scale for the Assessment ofNegative Symptoms(SANS). University of Iowa, IowaCity, Iowa.

Bymaster, F.P., Calligaro, D.O., Falcone, J.F., et al., 1996.Radioceptor binding profiles of the atypical antipsychoticolanzapine. Neuropsychopharmacology 14, 87–96.

Carlsson, A., 2000. Focussing on dopaminergic stabilisers and5-HT receptor antagonists(editorial). Current Opinion in2A

CPNS Investigational Drugs 2, 22–24.

Chakos, M., Lieberman, J., Hoffman, E., et al., 2000. Effect-iveness of second generation antipsychotics in patients withtreatment resistant schizophrenia: a review and meta analysisof randomized trials. American Journal of Psychiatry 158,518–526.

Curtis, V.A., Wright, P., Reveley, A., Kerwin, R., Lucey, J.V.,1995. Effect of clozapine onD-fenfluramine-evoked neu-roendocrine responses in schizophrenia and its relation toclinical improvement. British Journal of Psychiatry 166,642–646.

Fuller, R.W., Snoddy, H.D., 1992. Neuroendocrine evidencefor antagonism of serotonin and dopamine receptors byolanzapine(LY 170053) an antipsychotic drug candidate.Research Communications in Chemical Pathology and Phar-macology 77, 87–93.

Garattini, S., Mennini, T., Samanin, R., 1987. From fenflura-mine racemate toD-fenfluramine. Specificity and potencyof the effects on the serotoninergic system and food intake.Annals of the New York Academy of Sciences 499,156–166.

Goodall, E.M., Cowen, P.J., Franklin, M., Silverstone, T.,1993. Ritanserin attenuates anorexic responses toD-fenflur-amine in human volunteers. Psychopharmacology 499,156–166.

Invernizzi, R., Berrettera, C., Garattini, S., Samanin, R., 1986.D- and L-isomers of fenfluramine differ markedly in theirinteraction with brain serotonin and catecholamines in therat. European Journal of Pharmacology 20, 9–15.

Jones, H.M., Curtis, V.A., Wright, P., Lucey, J.V., 1998.Neuroendocrine evidence that clozapine’s serotonergicantagonism is relevant to its efficacy in treating hallucina-tions and other positive schizophrenic symptoms. AmericanJournal of Psychiatry 155, 838–840.

Jones, H.M., Curtis, V.A., Wright, P., Lucey, J.V., 1998.Risperidone is associated with blunting ofD-fenfluramineevoked serotonergic responses in schizophrenia. Internation-al Clinical Psychopharmacology 13, 199–203.

Kahn, R.S., Siever, L., Davidson, M., Greenwald, C., Moore,C., 1993. Haloperidol and clozapine treatment and theireffect on m-CPP-mediated responses in schizophrenia:implications for the mechanism of action of clozapine.Psychopharmacology 112, S90–S94.

Kahn, R.S., Wetzler, S., 1991.mCPP as a probe of serotoninreceptors: a review. Biological Psychiatry 30, 1139–1166.

Kapur, S., Zipursky, R.B., Remington, G., 1999. Clinical andtheoretical implications of 5-HT and D receptor occupancy2 2

of clozapine, risperidone, and olanzapine in schizophrenia.American Journal of Psychiatry 156, 286–293.

Kapur, S., Zipursky, R.B., Remington, G., et al., 1998. 5-HT2

and D receptor occupancy of olanzapine in schizophrenia:2

PET investigation. American Journal of Psychiatry 155,921–928.

Meltzer, H.Y., 1999. The role of serotonin in antipsychoticdrug action. Neuropsychopharmacology 21, 106S–111S.

Meltzer, H.Y., Maes, M., 1995. Pindolol pretreatment blocksstimulation bymeta-chlorophenylpiperazine of prolactin butnot cortisol secretion in normal men. Psychiatry Research58, 89–98.

47H. Jones et al. / Psychiatry Research 113 (2002) 41–47

Mohr, P., Horacek, J., Motlova, L., Libiger, J., Czobor, P.,1998. Prolactin response toD-fenfluramine challenge test asa predictor of treatment response to haloperidol in acuteschizophrenia. Schizophrenia Research 30, 91–99.

Monteleone, P., Tortorella, A., Borriello, R., Cassandro, P.,Maj, M., 1999. Prolactin hyperresponsiveness toD-fenflur-amine in drug-free schizophrenic patients: a placebo con-trolled study. Bioligical Psychiatry 45, 1606–1611.

Nordstrom, A.L., Farde, L., Nyberg, S., Karlsson, P., Halldin,¨C., Farde, L., 1995. D , D and 5HT receptor occupancy1 2 2

in relation to clozapine serum concentration: a PET studyof schizophrenic patients. American Journal of Psychiatry152, 1444–1449.

Owen, R.R., Guhewez-Estinou, R., Hsiao, J., et al., 1993.Effects of clozapine and fluphenazine treatment on responses

to m-chlorophenylpiperazine infusions in schizophrenia.Archives of General Psychiatry 50, 636–644.

Sabbe, B., Hulstijn, W., Maes, M., et al., 2001. Psychomotorslowing, neuroendocrine responses and behavioural changesafter oral administration ofmeta-chlorophenylpiperazine innormal volunteers. Psychiatry Research 105, 151–163.

Scheepers, F.E., Gispen-de-Wied, C.C., Kahn, R.S., 2000. Theeffects of olanzapine on neuroendocrine responses inducedby the serotonin agonistm-chlorophenylpiperazine(m-CPP).Schizophrenia Research 41, 244.

Schotte, A., Janssen, P.F.M., Gommeren, W., et al., 1996.Risperidone compared with new and reference antipsychoticdrugs: in vitro and in vivo receptor binding. Psychophar-macology 124, 57–73.