A New Framework for Investigating Antipsychotic Action in Humans: Lessons from PET Imaging S Kapur

PERSPECTIVE A new framework for investigating antipsychotic action in humans: lessons from PET imaging S Kapur

Schizophrenia Division and the PET Centre, The Clarke Institute of Psychiatry, Department of Psychiatry, University of Toronto; Rotman Research Institute, Baycrest Centre, University of Toronto

With a decade of neuroreceptor imaging of antipsychotics behind us, this article attempts to synthesise what has been learnt about the mechanism of action of antipsychotics using these techniques. The data show that: (i) the ‘typical’ antipsychotics bind mainly to the dopamine

D2 receptor, and that 60–80% D2 occupancy may provide optimal antipsychotic response with little extrapyramidal side effects; (ii) all the clinically available ‘atypical’ antipsychotics show

a higher occupancy of the 5-HT2 than D2 receptors; (iii) however, these ‘atypical’ antipsychotics differ in their D2 occupancy. The D2 occupancy of risperidone is within the typical range (ie Ͼ 60%) while that of clozapine is clearly lower (Ͻ60%); (iv) antipsychotics with com-

bined 5-HT2/D2 antagonism lose some of their ‘atypical’ properties if used in doses where Ͼ their D2 occupancy is too high ( 80%). Based on these data a framework is suggested wherein antipsychotics may be classiﬁed on the basis of their D2 and 5-HT2 occupancy in patients at steady state while taking clinically relevant doses. Within this framework typical antipsy-

chotics are classiﬁed as ‘high-D2’, risperidone as ‘high-D2 high-5HT2’ and clozapine as a ‘low- D2 high-5HT2’ antipsychotic. The justiﬁcation, limitations and the value of this framework in understanding and investigating newer antipsychotics is discussed. Keywords: schizophrenia; antipsychotics; dopamine receptors; serotonin receptors; positron emission tomography

Introduction D2 receptors and found that typical antipsychotics occupied 70–89% D receptors at conventional clinical It has been a decade since the first studies1,2 of the 2 doses. Moreover, patients with extrapyramidal side- effects of antipsychotics on brain receptors were effects (EPS) had, on average, higher D occupancy reported. Much has been learnt since. The purpose of 2 than those who did not exhibit EPS (82% vs 72%). This this article is to synthesize this information. This is not relationship was further supported by Nordstrom et al6 a comprehensive review of PET data on this subject. who found that those with greater than 60% D occu- Instead it represents a selective analysis of the extant 2 pancy showed much better antipsychotic response data with a view to develop a theoretical perspective than those with occupancies below 60%. These data and a clinical understanding of the role of the dopa- suggest that there may be a therapeutic window, per- mine D and serotinin 5-HT receptors in the treatment 2 2 haps between 60% and 80% D occupancy, which may of schizophrenia. 2 yield adequate antipsychotic response with low or minimal EPS. Typical antipsychotics, D2 occupancy and clinical In a clinical context, it has been shown that low correlates doses of oral haloperidol, 1–5 mg day−1 (plasma levels of 0.5–5.8 ng ml−1) lead to occupancies ranging from 50 It has been known for almost two decades that antipsy- − to 88%; and that 2 mg day 1 of haloperidol may be suf- chotic effects and extrapyramidal side effects (as well ficient to get most patients into the putatively thera- as catalepsy in animals) are related to the dopamine D 2 peutic 60–80% occupancy range.7 To test the clinical blocking properties of antipsychotics.3,4 However, validity of this, seven patients, all suffering from a first prior to the advent of PET imaging it was not possible episode of schizophrenic illness, were treated with hal- to observe any of these relationships at a receptor level − operidol 2 mg day 1 in an open clinical-PET investi- in vivo in humans. Farde et al5 reported the first sys- gation.8 The average D receptor occupancy was 67% tematic study of the action of antipsychotics on in vivo 2 and significant improvement in positive as well as 8 negative symptoms was observed. This clinical-D2 occupancy correlation fits rather nicely with McEvoy Correspondence: Dr S Kapur, PET Centre, The Clarke Institute of 9 Psychiatry, 250 College Street, Toronto, ON, Canada M5T 1R8. et al who applied a clinical ‘neuroleptic threshold’ E-mail: kapurȰclarke-inst.on.ca technique to determine the minimal effective dose. The Received 28 April 1997; revised and accepted 24 July 1997 median threshold dose for the first episode patients PET imaging of antipsychotics S Kapur

136 Ͼ was 2.1 mg of haloperidol—and the addition of an 80% D2 occupancy using SPECT). Thus at effective extra 5–10 mg day−1 of haloperidol did not signiﬁcantly doses (between 200–400 mg day−1)22 clozapine occu-

augment their clinical response, but did increase the pies signiﬁcantly fewer D2 receptors than are occupied incidence of EPS. Several recent studies by Stone et by typical antipsychotics and certainly fewer than al,10 Janicak et al11 as well as Ortega-Soto et al12,13 all those associated with EPS. point to the effectiveness of low dose antipsychotics Risperidone, like clozapine, exhibits high levels of −1 (2–5 mg day of haloperidol or equivalent) and uni- 5-HT2 occupancy. However, risperidone’s D2 occu- formly show that higher doses (10–50 mg day−1) pancy is clearly higher than that of clozapine.23–25 increase the incidence of EPS-related side-effects with- Kapur et al25 have shown that, on average, 2 mg day−1

out an appreciable change in clinical response. of risperidone occupied 66% of D2 receptors, Even with optimal dosing, a signiﬁcant number of 4 mg day−1 occupied 73% while 6 mg day−1 occupied patients (estimated between 20–50%) do not respond 79%. While most patients with risperidone did not

satisfactorily to treatment with D2 blockers. The cause experience EPS, those who did experience EPS had a of this non-response is not known, but non-responders much higher D2 occupancy than those who did not 14,15 ± ± show no difference in their D2 occupancy profiles. (79% 4% vs 68% 5%). Thus risperidone’s high lev- This has led to the suggestion that non-responders to els of 5-HT2 occupancy (whatever its other benefits) D2 occupancy may reflect a biologically distinct subset cannot provide absolute protection from EPS once D2 of the illness.14,15 In fact Pilowsky et al16 show that blockade is near complete.25,26

some patients who do not respond despite high and The putatively negative effects to too much D2 block- adequate D2 occupancy, respond to much lower D2 ade can also be observed in the multi-centre studies occupancy with clozapine (which affects several recep- which compared 2–16 mg day−1 of risperidone with −1 27,28 −1 tor systems in addition to D2). Even those who do 10–20 mg day of haloperidol. Six mg day of ris- respond to the D2-mediated mechanism of response peridone was numerically and statistically superior to usually show less than satisfactory improvement in haloperidol in terms of positive, negative and EPS syn- negative symptoms and cognitive impairment. These dromes. However, beyond 6 mg day−1, the positive and limitations of the typical antipsychotics (some patients negative symptom response decreased and EPS are refractory, many get EPS, and negative and positive increased, such that risperidone 10–16 mg day−1 was symptom improvement is limited) necessitated the statistically indistinguishable from haloperidol. Doses development of ‘atypical’ antipsychotics. of risperidone beyond 6 mg day−1 give greater than 25,29 80% D2 occupancy. Thus, it would seem that push- ing D occupancy to saturation, in atypical drugs like Atypical antipsychotic, 5-HT and D occupancy, 2 2 2 risperidone, may not only diminish their EPS superior- and clinical correlates ity, but may also decrease their superiority on positive While there is no consensus on how to deﬁne an atypi- and negative symptoms. Therefore, it is not sufﬁcient

cal antipsychotic, most authorities agree that clozapine that a drug display the ‘right 5-HT2/D2 afﬁnity ratio’ in (and more recently risperidone, olanzapine and the test tube.17,18 It is imperative that the drug be used

sertindole) are ‘atypical’. All four of the above agents in humans in doses where D2 occupancy is not satu- are distinguished from typical antipsychotics by one rated.25 pharmacological commonality: a higher afﬁnity for the

5-HT2 as opposed to D2 dopamine receptor in vitro as Beyond the typical/atypical dichotomy: towards a well as ex vivo in animals.17,18 Since reliable neuroim- new framework aging data in humans are available only for clozapine Until recently, clozapine was the only ‘atypical’ anti- and risperidone, this section focuses on these two psychotic but several new antipsychotics are available atypical antipsychotics. now. They all differ from the typical antipsychotics, A comprehensive PET study of the receptor profile but it is increasingly becoming evident that these of clozapine shows that the D2 occupancy in patients newer antipsychotics may not replicate all of cloza- −1 treated with 125–600 mg day of clozapine varied pine’s features. One can classify these drugs on the from 20–67%, while at these doses the 5-HT2 occu- basis of their in vitro pharmacological character- 19 pancy is always greater than 80%. No patient on cloz- istics,30–32 their neurophysiological actions,33 their apine demonstrated D2 occupancy in the range that is gene-induction profiles34 or their actions in animal likely to produce EPS (ie 80% or greater). We find simi- models.35 However, all these schemes are based on pre- lar results in a series of nine patients whose doses clinical characteristics of these medications. Suggested −1 ranged from 75–900 mg day (and plasma levels from here is an alternative framework for classifying anti- −1 107 to 859 ng ml of clozapine). All nine patients exhi- psychotics, derived from and based on the in vivo 5- Ͼ bit very high 5-HT2 occupancy ( 90% in all patients, HT and D occupancies of the antipsychotics at steady 18 2 2 as measured using F-setoperone) and low to modest state at clinically relevant doses. D2 dopamine occupancy (24–66% as measured using 11 C-raclopride; preliminary analysis communicated by Classifying antipsychotics using the D2/5-HT2 Dr Robert Zipursky). Similar findings have also been framework reported by other groups using PET20 and SPECT16,21 Figure 1 displays the criteria for classifying antipsy- (see Pickar et al21 for the only clozapine patient with chotics in this framework. Antipsychotics which show PET imaging of antipsychotics S Kapur

Ͻ 137 an average D2 occupancy of 60%, at clinically rel- only one pharmacological property—a higher affinity evant doses, are classified as ‘low D2’ while those with for 5-HT2 as compared to D2 receptors. Clinical evi- a higher D2 occupancy are ‘high-D2’. This watershed of dence is now accumulating that the combined 5HT2/D2 60% is based on current evidence that at least 60% D2 antagonism may provide an additive benefit on nega- occupancy is required for antipsychotic response if tive symptoms and may modify the expression of extra- that is the only mechanism invoked.6 The next division pyramidal symptoms.36,39 This, combined with the fact of the drugs is along their 5HT2 properties. Based on that it is possible to measure 5HT2 occupancy in a the review of evidence presented above,25,31,36 drugs reliable manner19,23,40,41 makes it the second axis for Ͼ which show: (i) 5HT2 occupancy 80%; and (ii) 5-HT2 this framework. occupancy greater than D2 occupancy are classified as It is not the purpose of this framework to detract 42–44 ␣ ‘high-5HT2’, all others as ‘low-5HT2.’ Therefore a drug from the possibly important roles of the D4, 1 and ␣ 45–48 49–51 52 which has a 80% 5HT2 occupancy, but also has a sim- 2 adrenergic, cholinergic, glutamatergic and ultaneous 80% D2 occupancy would not be a ‘high- other systems. It is the author’s view that the complete 5HT2’ drug in this framework since it would violate the answer will encompass more than just two receptors. second clause. However, till we are there, we need simple frameworks Every antipsychotic medication can be unambigu- which facilitate investigation. Thus, this framework is ously classified into one of the four cells provided one useful not only for those who are interested in D2 and has data regarding D2 and 5-HT2 occupancy at clini- 5-HT2. But, it also provides a basis for controlling the cally relevant doses. This issue of clinically relevant effects of the D2 and 5-HT2 systems as one investigates doses is crucial. In this framework, haloperidol and the role of the D4 or adrenergic or glutamatergic system perhaps every ‘typical’ antipsychotic belongs in the in antipsychotic response. high-D2 low-5HT2 group. Risperidone is a high-D2 −1 high-5HT2 antipsychotic since (in the 4–6 mg day Why does one need in vivo human data at clinically Ͼ Ͼ range) it occupies 60% D2 receptors and 80% 5-HT2 relevant doses? receptors.37 Clozapine (in the 200–400 mg day−1 range) Previous efforts at pharmacological classification have occupies less than 60% D2 receptors and greater than relied on in vitro indices. However, measures such as 80% 5-HT2 receptors and therefore is a low-D2 high- inhibition constant (Ki) or the dissociation constant 5HT2 antipsychotic. At present, there is no known (Kd) disregard a number of factors which are clearly effective antipsychotic in the low-D2 low-5HT2 group. relevant in the clinical situation. First, the measures of Early results suggest that olanzapine and sertindole are affinity obtained in vitro can differ remarkedly from 31 clearly high-5-HT2 drugs, though their D2 character- those obtained in vivo (see Schotte et al for a demon- istics are not unequivocally ascertained. We have stration of this in the context of atypical recently scanned three patients who were being treated antipsychotics). Second, the in vitro numbers do not with higher than recommended doses (ie 30 and account for the action of metabolites, which are often 40 mg day−1) of olanzapine and found greater than 80% an important issue in the clinical situation. Third, the

D2 occupancy in all three cases. This would suggest in vitro analyses usually do not account for compe- that olanzapine, unlike clozapine, is in all likelihood tition with the endogenous agonist, a facet which a high-D2 high-5-HT2 antipsychotic. Though, as stated affects all drug-receptor interactions in vivo. Finally, above, accurate delineation of the status of this drug the in vitro afﬁnity of a drug only tells of its ‘potential’ would require studies done in the clinically rec- for occupying receptors. What matters clinically is not ommended (10–20 mg day−1) dose range. the potential, but the actual number of receptors occupied. For these reasons more meaningful comparisons

Why choose D2 and 5-HT2 receptor systems as the can be made based on the effects of these drugs in vivo basis? in humans (which factor in in vivo afﬁnity, dose,

D2 blockade is a necessary condition for antipsychotic metabolism, as well as human pharmacokinetics) at response. There are (in 1997) no known antipsychotics clinically relevant doses. which are devoid of D2 occupancy. For many patients, D2 blockade may even be a sufficient condition for anti- Useful attributes of this framework psychotic response. Even clozapine occupies a modest The distinguishing attribute of this scheme is that it 19 number of the dopamine D2 receptors. While the emanates from pharmacological and clinical obser- degree of D2 occupancy may differ across the antipsy- vations in humans and is clearly operationalised and chotics, some level of D2 dopamine receptors occu- therefore testable. This framework readily accommo- pancy is required for antipsychotic response. For this dates the observed pre-clinical and clinical differences reason, D2 receptor occupancy forms the first axis in between the currently available antipsychotics. Anti- this framework. psychotics in the high-D2 low-5HT2 cell all exhibit The second axis is that of 5HT2 occupancy. Meltzer catalepsy in animals, show prolactin elevation in et al17,18,38 have documented in in vitro and ex vivo humans, exhibit EPS in the clinically relevant doses studies that the high ratio of 5HT2 to D2 dopamine and are associated with a high risk of TD over pro- affinity distinguishes atypical antipsychotics. In fact all longed usage. The low-D2 high-5HT2 group shows no the currently available atypical antipsychotics (or minimal) catalepsy in animals, does not cause sig- (clozapine, risperidone, olanzapine, sertindole) share nificant prolactin elevation and does not induce the PET imaging of antipsychotics S Kapur 138

Figure 1 A framework for the classiﬁcation of antipsychotic drugs into four groups based on their in vivo D2 and 5-HT2 occupancy at clinically relevant doses. It is predicted that drugs such as olanzapine, sertindole, quetiapine and ziprasidone

will fall in the high-5HT2 category. What needs to be ascertained is whether they will be high-D2 drugs like risperidone or low D2 drugs like clozapine.

Figure 2 The preclinical and clinical proﬁles that distinguish medications which are clustered in each of the four groups corresponding to Figure 1.

classical Parkinsonian EPS and is not associated with peutic efﬁcacy (perhaps reﬂecting the modulation of

TD. It is contended that these differences obtain these attributes by the presence of high-5HT2). directly from the different D2 properties of these two Another value of this framework is in generating a groups. Furthermore, the low-D2 high-5HT2 group is hypothesis regarding the mechanism of action of anti- known to be effective against negative symptoms and psychotic drugs. A classiﬁcation is of value only if may be effective in a subset of refractory patients. members within a group share properties which make

Medications in the high-D2 high-5HT2 category do them distinct from members in other groups. It is pre- exhibit catalepsy in animals, prolactin elevation in dicted that all the drugs in a given cell in Figure 1

humans and EPS (reflecting their high-D2 status)—but would exhibit similar properties with respect to the in animals as well as in humans seem to exhibit a much following attributes: (i) efficacy vs positive symptoms; wider margin between these side-effects and thera- (ii) efficacy vs negative symptoms; (iii) potential for PET imaging of antipsychotics S Kapur 139 causing catalepsy in animals, prolactin elevation in of the present findings, the real test lies in whether this humans, Parkinsonian EPS in humans and TD with framework can predict and accommodate new findings long-term use. If this prediction is wrong, that is, if we in this area. Have the PET data told us how antipsy- do find a drug which is in the same cell as clozapine, chotics work? Not really. Have they told us how and but substantially differs from clozapine along the above where to look for the answer? Perhaps yes. cited attributes, it would suggest that our current thoughts regarding 5-HT2/D2 require modification. Pharmacological contrast of such a drug with clozapine Acknowledgements would lead us to the source of clozapine’s differential The author would like to thank Drs Robert Zipursky, efficacy. Similarly, if a drug is found in the high-D2 Gary Remington and Philip Seeman for the several high-5HT2 group which never gives prolactin elevation spirited discussions which fostered the ideas or EPS at all the clinical doses, that would defy this expressed here. The author is supported by a Clinician framework. But, pharmacological contrast of that drug Scientist Development Award from the Medical with risperidone would point us to the neuro- Research Council of Canada. transmitter systems, beyond just D2 and 5-HT2, which may be involved in protecting one from EPS and prolactin elevation. References Finally, at present, there are no antipsychotic treat- 1 Cambon H, Baron JC, Boulenger JP et al. In vivo assay for neurolep- ments in the low-D2 low-5HT cell. However, this is a tic receptor binding in the striatum: positron tomography in very important cell. If we were to find a medication humans. Br J Psychiatry 1987; 151: 824–830. which shows antipsychotic benefits while in this cell 2 Farde L, Wiesel FA, Halldin C, Sedvall G, PET-determination of central D1- and D2- dopamine receptor occupancy in neuroleptic (for example a predominant D4 antagonist would fall treated schizophrenics. In: Heiss W-D, Pawlick G, Herholz K, Wien- in this cell), it would be a radical departure from the hard K (eds). Clinical Efficacy of Positron Emission Tomography currently available treatments and may bring with it a Martinus Nijhoff Publishers: Dordrecht, 1987, pp 213–219. new definition of antipsychotic treatment. 3 Sanberg PR. Haloperidol-induced catalepsy is mediated by post- synaptic dopamine receptors. Nature 1980; 284: 472–473. Limitations of this framework 4 Seeman P, Lee T, Chau-Wong M, Wong K. Antipsychotic drug doses and neuroleptic/dopamine receptors. Nature 1976; 261: Biological systems tend to be continuous rather than 717–719. quantum, therefore, any effort to encapsulate nature 5 Farde L, Nordstrom AL, Wiesel FA, Pauli S, Halldin C, Sedvall into discrete divisions (such as Figure 1) is clearly an G. Positron emission tomographic analysis of central D1 and D2 oversimplification. Nonetheless, these divisions are dopamine receptor occupancy in patients treated with classical neuroleptics and clozapine: relation to extrapyramidal side effects. provided because they explain the current data, they Arch Gen Psychiatry 1992; 49: 538–544. are heuristically useful and provide a framework for 6 Nordstrom AL, Farde L, Wiesel FA et al. Central D2-dopamine generating testable hypotheses. receptor occupancy in relation to antipsychotic drug effects—a The major limitation of this scheme is that it only double-blind PET study of schizophrenic patients. Biol Psychiatry accommodates receptors which can be measured in 1993; 33: 227–235. 7 Kapur S, Remington G, Jones C et al. Relationship between D2 humans. This is just a technical limitation. It is hoped receptor occupancy and plasma haloperidol: a PET study. Eur Neu- that as our knowledge and interest in a particular ropsychopharmacol 1996; 6: 73. receptor system evolves, the ability to measure them in 8 Kapur S, Remington G, Jones C et al. High levels of dopamine D2 vivo will develop in concert. A more salient limitation receptor occupancy with low dose haloperidol treatment: a PET study. Am J Psychiatry 1996; 153: 948–950. is that this framework deals with phenomena only at 9 McEvoy JP, Hogarty GE, Steingard S. Optimal dose of neuroleptic the cell-surface receptor level. Antipsychotic response in acute schizophrenia: a controlled study of neuroleptic threshold lags behind the occupation of receptors. It is only logi- and higher haloperidol dose. Arch Gen Psychiatry 1991; 48: 739– cal to assume that biological changes at the post-recep- 745. tor system are involved. At present little is known 10 Stone CK, Garver DL, Griffith J, Hirschowitz J, Bennett J. Further evidence of a dose-response threshold for haloperidol in psychosis. about which post-receptor changes are clinically rel- Am J Psychiatry 1995; 152: 1210–1212. evant and there are few techniques to measure these 11 Janicak PG, Javaid JI, Sharma RP, Leach A, Dowd S, Davis JM. A in the living human brain. It is the author’s view that double-blind randomized study of three haloperidol plasma levels receptor and post-receptor changes are complementary for acute psychosis. Acta Psychiatrica Scand 1997; 95: 343–350. 12 Ortega-Soto HC, Brunner E, Apiquian R, de la Torre P. Therapeutic views of a cascade that effects antipsychotic response. minimum dose of haloperidol (HLP) in schizophrenia. Neuropsych- Therefore, the four groups currently presented in opharmacology 1994; 10: 140S. Figure 1 would be expected to have distinct intracellu- 13 Ortega-Soto HA, Brunner E, Apiquian R, de la Torre MP, Ulloa RE. lar finger-prints. And our understanding of antipsy- Typical antipsychotics: the threshold doses strategy. Eur Neuropsy- chotic action would be advanced by mapping differ- chopharmacology 1996; 6: 171. 14 Wolkin A, Barouche F, Wolf AP. Dopamine blockade and clinical ences at the receptor levels (low-D2 vs high-D2) to their response: evidence for two biological subgroups of schizophrenia. different consequences at the post-receptor level. Am J Psychiatry 1989; 146: 905–908. In summary, the first decade of in vivo human neuro- 15 Pilowksy LS, Costa DC, Ell PJ, Murray RM, Verhoeff NPLG, Kerwin receptor imaging has produced interesting empirical RW. Antipsychotic medication, D{2} dopamine receptor blockade and clinical response: a {+123}I IBZM SPET (single photon emis- regularities. This article has attempted to review these sion tomography) study. Psychol Med 1993; 23: 791–797. findings and impose on them a theoretical framework. 16 Pilowsky LS, Costa DC, Ell PJ, Murray RM, Verhoeff NPLG, Kerwin While the proposed framework can accommodate most RW. Clozapine, single photon emission tomography, and the D2 PET imaging of antipsychotics S Kapur 140 dopamine receptor blockade hypothesis of schizophrenia. Lancet 34 Robertson GS, Matsumura H, Fibiger HC. Induction patterns of Fos- 1992; 340: 199–202. like immunoreactivity in the forebrain as predictors of atypical 17 Meltzer HY, Matsubara S, Lee JC. Classification of typical and antipsychotic activity. J Pharamcol Exp Therapeut 1994; 271: atypical antipsychotic drugs on the basis of dopamine D-1, D-2 and 1058–1066. serotonin-2 pKi values. J Pharmacol Exp Therapeut 1989; 251: 35 Swerdlow NR, Braff DL, Bakshii VP, Geyer MA. An animal model 238–246. of sensorimotor gating deficits in schizophrenia predicts antipsy- 18 Meltzer HY, Matsubara S, Lee JC. The ratios of serotonin-2 and chotic drug action. In: Csernansky JG (ed). Antipsychotics. Hand- dopamine-2 affinities differentiate atypical and typical antipsy- book of Experimental Pharmacology. Heidelberger Platz 3/W-1000 chotic drugs. Psychopharmacol Bull 1989; 25: 390–392. Berlin 33/Germany: Springer-Verlag Berlin, 1996, pp 289–312. 19 Nordstrom A-L, Farde L, Nyberg S, Karlsson P, Halldin C, Sedvall 36 Kapur S, Remington G. Serotonin-dopamine interaction and its rel-

G. D1,D2, and 5-HT2 receptor occupancy in relation to clozapine evance to schizophrenia. Am J Psychiatry 1996; 153: 466–476. serum concentration: a PET study of schizophrenic patients. Am J 37 Nyberg S, Farde L, Eriksson L, Halldin C, Eriksson B. 5-HT2 and Psychiatry 1995; 152: 1444–1449. D2 dopamine receptor occupancy in the living human brain. A PET 20 Karbe H, Wienhard K, Hamacher K et al. Positron emission tom- study with risperidone. Psychopharmacology 1993; 110: 265–272. ography with (18F)methylspiperone demonstrates D2 dopamine 38 Meltzer HY. Clinical studies on the mechanism of action of cloza- receptor binding differences of clozapine and haloperidol. J Neural pine: the dopamine-serotonin hypothesis of schizophrenia Trans–Gen Sect 1991; 86: 163–173. [Review]. Psychopharmacology 1989; 99: S18–S27. 21 Pickar D, Su TP, Weinberger DR et al. Individual variation in D-2 39 Meltzer HY. The role of serotonin in schizophrenia and the place dopamine receptor occupancy in clozapine-treated patients. Am J of serotonin-dopamine antagonist antipsychotics. J Clin Psycho- Psychiatry 1996; 153: 1571–1578. pharmacol 1995; 15: 2S–3S. 22 Pollack S, Lieberman JA, Fleischhacker WW et al. A comparison 40 Kapur S, Remington G, Jones C et al. Is loxapine an ‘atypical’ anti- of European and American dosing regimens of schizophrenic psychotic: a PET study of the in vivo 5-HT2 and D2 occupancy. patients on clozapine: efﬁcacy and side effects. Psychopharmacol Eur Neuropsychopharmacol 1996; 6: 72. Bull 1995; 31: 315–320. 41 Kapur S, Jones C, DaSilva J, Wilson A, Houle S. Reliability of a

23 Farde L, Nyberg S, Oxenstierna G, Nakashima Y, Halldin C, Erics- simple non-invasive method for the evaluation of 5-HT2 receptors son B. Positron emission tomography studies on D-2 and 5-HT2 using [18F]-setoperone PET imaging. Nucl Med Commun 1997; 18: receptor binding in risperidone-treated schizophrenic patients. J 395–399. Clin Psychopharmacol 1995; 15: S19–S23. 42 Seeman P, Guan H-C, Van Tol HHM. Dopamine D4 receptors elev- 24 Busatto GF, Pilowsky LS, Costa DC, Ell PJ, Verhoeff NPLG, Kerwin ated in schizophrenia. Nature 1993; 365: 441–445. RW. Dopamine D-2 receptor blockade in vivo with the novel anti- 43 Seeman P, Guan HC, Vantol HHM. Schizophrenia: elevation of psychotics risperidone and remoxipride—an I-123-IBZM single dopamine D-4-like sites, using [H-3]nemonapride and [I-125]epide- photon emission tomography (SPET) study. Psychopharmacology pride. Eur J Pharmacol 1995; 286: R3–R5. 1995; 117: 55–61. 44 Seeman P, Corbett R, Van Tol H. Atypical neuroleptics have low

25 Kapur S, Remington G, Zipursky RB, Wilson AA, Houle S. The D2 affinity for dopamine D2 receptors or are selective for D4. Neuropsy- dopamine receptor occupancy of risperidone and its relationship chopharmacology 1997; 16: 93–110. to extrapyramidal symptoms: a PET study. Life Sci 1995; 57: 45 Svensson TH, Mathe JM, Andersson JL, Nomikos GG, Hildebrand PL103–Pl107. BE, Marcus M. Mode of action of atypical neuroleptics in relation 26 Kapur S. 5-HT2 antagonism and EPS benefits: is there a causal con- to the phenyclidine model of schizophrenia: role of 5-HT2 receptor nection? Psychopharmacology 1996; 124: 35–39. and alpha(1)-adrenoreceptor antagonism. J Clin Pyschopharmacol 27 Chouinard G, Jones B, Remington G et al. A Canadian multicenter 1995; 15: S11–S18. placebo-controlled study of fixed doses of risperidone and haloper- 46 Litman RE, Hong WW, Weissman EM, Su TP, Potter WZ, Pickar D. idol in the treatment of chronic schizophrenic patients. J Clin Psy- Idazoxan, an alpha 2 antagonist, augments fluphenazine in schizo- chopharmacol 1993; 13: 25–40. phrenic patients: a pilot study. J Clin Psychopharmacol 1993; 13: 28 Marder SR, Meibach RC. Risperidone in the treatment of schizo- 264–267. phrenia. Am J Psychiatry 1994; 151: 825–835. 47 Grenhoff J, Svensson TH. Prazosin modulates the firing pattern of 29 Farde L, Nyberg S, Oxenstierna G, Nakashima Y, Halldin C, Erics- dopamine neurons in rat ventral tegmental area. Eur J Pharmacol son B. Positron emission tomography studies on D2 and 5-HT2 1993; 233: 79–84. receptor binding in risperidone-treated schizophrenic patients: J 48 Tassin JP. NE/DA interactions in prefrontal cortex and their poss- Clin Psychopharmacol 1995; 15: 19S–23S. ible roles as neuromodulators in schizophrenia. J Neural Transm 30 Schotte A, Janssen PF, Megens AA, Leysen JE. Occupancy of cen- Suppl 1992; 36: 135–162. tral neurotransmitter receptors by risperidone, clozapine and halo- 49 Freedman R, Hall M, Adler LE, Leonard S. Evidence in postmortem peridol, measured ex vivo by quantitative autoradiography. Brain brain tissue for decreased numbers of hippocampal nicotinic recep- Res 1993; 631: 191–202. tors in schizophrenia. Biol Psychiatry 1995; 38: 22–33. 31 Schotte A, Janssen PFM, Gommeren W et al. Risperidone compared 50 O’Keane V, Abel K, Murray RM. Growth hormone responses to pyr- with new and reference antipsychotic drugs: in vitro and in vivo idostigmine in schizophrenia: evidence for cholinergic dysfunc- receptor binding. Psychopharmacology 1996; 124: 57–73. tion. Biol Psychiatry 1994; 36: 582–588. 32 Gerlach J, Peacock L. New antipsychotics: the present status. Int 51 Tandon R, De Quardo JR, Goodson J, Mann NA, Greden JF. Effect Clin Psychopharmacol 1995; 10: 39–48. of anticholinergics on positive and negative symptoms in schizo- 33 Grace AA, Bunney BS, Moore H, Todd CL. Dopamine-cell depolar- phrenia. Psychopharmacol Bull 1992; 28: 297–302. ization block as a model for the therapeutic actions of antipsychotic 52 Javitt DC. Glutamate receptors and schizophrenia: opportunities drugs. Trends Neurosci 1997; 20: 31–37. and caveats. Mol Psychiatry 1996; 1: 16–17.