Psychiatry Research: Neuroimaging Section 82 Ž1998. 147]160
N-w11 CxMethylspiperone PET, in contrast to w 11 Cxraclopride, fails to detect D 2 receptor occupancy by an atypical neuroleptic Gisela Hagberg a,d,U , Ola Gefvert b , Mats Bergstrom ¨ a,d , Ing-Marie Wieselgren c , b c Leif Lindstrom ¨ , Frits-Axel Wiesel , Bengt Langstrom ˚ ¨ a,d a Uppsala Uni¨ ersity PET Centre, Uni¨ ersity Hospital, S-75185 Uppsala, Sweden Department of Psychiatric Research, Uni¨ ersity of Uppsala, Vasteras ¨ ˚ Central Hospital, S-72189 Vasteras, ¨ ˚ Sweden c Department of Psychiatry, Uni¨ ersity Hospital, Uppsala Uni¨ ersity, S-75017 Uppsala, Sweden d Subfemtomole Biorecognition Project, Japan Science Technology Corporation and Uppsala Uni¨ ersity PET Centre, Uppsala, Sweden b
Received 13 November 1997; received in revised form 24 March 1998; accepted 25 March 1998
Abstract The occupancy of the atypical neuroleptic quetiapine ŽSeroquel. at the D2 dopamine receptor was investigated using the PET tracers w 11 Cxraclopride and N-w 11 Cxmethylspiperone in a group of five schizophrenic patients. A steady-state treatment condition was ensured by dosing the patients with 750 mg quetiapine daily during 3 weeks followed by a period of tapering off the dose. For each patient, PET examinations were performed with both tracers at two of the following doses: 750, 450, 300 andror 150 mg. As control, a group of six healthy untreated volunteers was investigated. The D 2 binding potential in the putamen and the caudate nucleus was determined by using an evaluation method based on the method proposed by Patlak and Blasberg. The receptor occupancy was determined by assuming that the group of healthy volunteers is representative of untreated drug-naive schizophrenic patients. While a significant linear trend of increasing occupancy with increasing quetiapine dose Žreaching 51% " 10% occupancy at the 750 mg dose. was detected with w 11 Cxraclopride Ž P- 0.01., no such trend was apparent for N-w 11 Cxmethylspiperone Ž P) 0.09, maximal occupancy values were 2% "3%, measured for the group of three patients on 450 mg.. The study suggests that N-w 11 Cxmethylspiperone cannot be used for the assessment of D 2 receptor occupancy induced by quetiapine. The result is discussed in terms of endogenous dopamine, tracer kinetics and equilibrium dissociation constants. Q 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Positron emission tomography; w 11 CxRaclopride; N-w 11 CxMethylspiperone; Quetiapine; Schizophrenia; Dopamine receptors
U
Corresponding author. Tel.: q46 18 4713384; fax: q46 18 4713390; e-mail:
[email protected]
0925-4927r98r$19.00 Q 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0925-4927Ž98.00020-1
148
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
1. Introduction Positron emission tomography ŽPET. has allowed in vivo characterization of the binding of neuroleptic drugs to the dopamine D 2 receptor in the striatum ŽWiesel, 1995; Farde, 1996.. The first radiolabelled receptor ligand used for pharmacological studies in humans was N-methylspiperone ŽNMSP.. NMSP not only interacts with D 2 receptors but has a high affinity for serotonergic 5-HT2 receptors. With the advent of atypical neuroleptics, which generally interact with both the serotonergic and dopaminergic receptor systems, it was proposed that 11 C-NMSP may potentially be useful to study treatment effects on both receptor systems simultaneously. This idea was reinforced by the fact that the distribution of the dopaminergic and the serotonergic receptor systems in the human brain is fairly disparate Žsee, for instance, Pazos et al., 1987; Camps et al., 1990.: only approx. 20% of 11 C-NMSP binding in the striatum is due to serotonergic receptors ŽFrost et al., 1987. and the 11 C-NMSP binding to dopaminergic receptors in the frontal cortex is less than 10% ŽFrost et al., 1987. and can be regarded as negligible ŽNordstrom ¨ et al., 1995a.. 11 C-NMSP has been used previously ŽWong et al., 1986. to determine B max of dopaminergic D 2 receptors in drug-naive schizophrenic patients. Briefly, the procedure entailed two PET scans, with and without previous injection of haloperidol. The outcome of the measurements indicated an altered D 2 receptor expression in schizophrenic patients. In contrast, using another D 2 receptor PET ligand, w 11 Cxraclopride ŽFarde et al., 1987, 1990. ŽRAC., no alterations were found. Such contradictory results may, amongst other things ŽAndreasen et al., 1988; Nordstrom ¨ et al., 1995a; Wiesel, 1995., be explained by the different affinity of the two ligands for the D 2 receptor system, which, in the human putamen measured in vitro, is reported to be 0.2 nM for NMSP and 3.9 nM for RAC ŽHall et al., 1990.. Not only under normal physiological conditions but also during treatment with drugs with a low D 2 receptor affinity, it can be assumed that the results might be affected by the choice of PET tracer. Indeed, such effects are suggested by the
discrepant findings obtained when investigating the D 2 receptor occupancy of the atypical neuroleptic clozapine. While Nordstrom ¨ et al. Ž1995b. Ž . and Farde et al. 1992 , working with 11 C-RAC, showed 20]67% D 2 receptor occupancy in a group of schizophrenic patients on 125]600 mg clozapine, Goyer et al. Ž1996., working with 11 C-NMSP, detected occupancies of 15]34% in a group of schizophrenic patients on 300]900 mg clozapine. These findings suggest that 11 C-NMSP might be inadequate to study the occupancy of atypical neuroleptics, which have a low affinity for the D 2 system. The present study was performed to shed further light on these issues. A group of schizophrenic patients was treated with different doses of the putative atypical neuroleptic, quetiapine ŽSeroquel W , Zeneca Pharmaceuticals, UK., and investigated by PET with both 11 C-NMSP and 11 C-RAC. A group of healthy untreated volunteers was investigated in the same way for comparison. 2. Patients and methods 2.1. Appro¨ als This study was approved by the ethics and radiation safety committees of the University of Uppsala. 2.2. Patients and study design This was an open-label, non-randomized trial. Written informed consent was obtained from all participating individuals. Five male Caucasian patients Žmean age: 34 " 8 years, range: 25]45 years. with schizophrenia or schizophreniform psychosis according to DSM-III-R, normal laboratory blood tests, ECG and physical examination were included in the study. Previous neuroleptics were discontinued the day before the first intake of the study drug and depot neuroleptics 12 months before. Each patient received standard escalating doses until 250 mg quetiapine ŽSeroquel. three times daily was attained. The duration of the escalating phase was 1 week. The treatment at this dose level then continued for 3
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
weeks, after which the dose was decreased weekly in a standard stepwise manner so that each subject received, in order, 250, 150, 100 and 50 mg quetiapine three times daily. At two of the four descending doses, each patient underwent PET examinations with both 11 C-NMSP and 11 C-RAC, 2 h after medication: three of the patients had PET scans using 11 C-NMSP and 11 C-RAC while on a daily dose of 750 mg and 450 mg quetiapine, one patient was examined while on 750 mg and 300 mg and one patient while on 300 mg and 150 mg quetiapine. In parallel, six healthy volunteers Žmean age: 32 " 6 years, range: 28]39 years., with normal laboratory blood tests, ECG and physical examination, had PET examinations. Two of them had both 11 C-NMSP and 11 C-RAC PET and further four had either 11 C-NMSP or 11 C-RAC PET. 2.3. PET tracers The tracers were produced according to standard operating procedures at the Uppsala University PET Centre. The concentrations of the compounds were 4.1" 3.4 m grml for 11 C-RAC and 8.3" 5.0 m grml for 11 C-NMSP. The radiochemical purity was 98% for all raclopride batches and 96 " 2% for 11 C-NMSP. The average injection volumes were 1.5" 0.9 ml for 11 C-RAC and 0.6" 0.3 ml for 11 C-NMSP, corresponding to a total dose of 5.0" 3.5 m g for 11 C-RAC and 3.6" 1.6 m g for 11 C-NMSP. The range of injected radioactivity was 154]355 MBq. The estimated tracer concentration in the putamen Žcalculation based on the standard uptake value. was 0.9" 0.4 nM for 11 C-RAC and 0.8" 0.4 nM for 11 C-NMSP. 2.4. Scanning procedure All patients and volunteers had standard fluid and food intake on the day before and on the day of the PET examination. PET scans were performed with a GEMS PC2048-15B camera ŽHolte et al., 1989. which gives 15 contiguous 6.5-mm thick slices with a resolution of approx. 5 mm. For each patient, at each of the two dose levels, 11 C-RAC PET was performed at 09.00 h and 11 C-NMSP PET at 17.00 h, on the same day. Each scan was thus performed 2 h after intake of the drug. In total, 10 PET scans using 11 C-NMSP and
149
10 using 11 C-RAC were performed on five treated patients. In parallel, untreated healthy volunteers were examined by PET. Two of them had both 11 C-NMSP and 11 C-RAC PET and a further four had either 11 C-NMSP or 11 C-RAC PET. All investigated subjects were put in a head fixation system used at both examinations. Before injection of the tracer, a transmission scan was performed. PET scanning was started at the time of injection and continued for 55 min afterwards. Fifteen time frames Žlength: 1, 3, 5, or 10 min. were collected. The individual frames and a summation file of the last 25 min Ž 11 C-RAC. or 48 min Ž 11 C-NMSP. were reconstructed. Before reconstruction, the files were attenuation corrected using the transmission scan, corrected for scattered radiation ŽBergstrom ¨ et al., 1983. and filtered with a 4.2-mm Hanning filter. 2.5. Analysis of PET data For each patient who had a PET examination performed at two different dose levels, the images obtained at the two occasions were realigned by matching to a common intermediate set of orientation parameters, using an automatic realignment program ŽAndersson, 1995.. Average images were created from the realigned summation images obtained at the two occasions and were used to delineate the regions of interest. One cerebellar and four striatal Žtwo for the putamen and two for the caudate nucleus . standardized regions of interest were generated in one slice for the cerebellum and in one slice for the striatum. The cerebellar slice was chosen where the structure could be easily recognised Žouter part of the cerebellum, mainly grey matter. and the striatal slice was chosen where the signal was most intense. The regions of interest in the right and left hemispheres were combined to single regions of interest for the putamen Žtotal size: 2 cm2 . and the caudate nucleus Žtotal size: 0.7 cm2 .. For each of the three resulting regions, decay-corrected time]activity curves were generated. Instead of using the measurement of a plasma time] activity curve, with adequately measured metabolite levels, the cerebellum time]activity curve
150
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
was used as an input function to the striatal Žputamen and caudate nucleus . time]activitycurves. Cerebellum has been used previously as a reference region to detect D 2 receptor occupancy ŽWong et al., 1984; Lammertsma et al., 1996. and has also been shown to be as sensitive for detecting changes as methods that require a metabolite-corrected plasma input function ŽLammertsma et al., 1996.. For the evaluation of the binding potential of 11 C-RAC, a three-compartment model was assumed and the following four evaluation methods were applied: an integration method ŽFarde et al., 1989., the Logan method ŽLogan et al., 1996., the simplified reference tissue model ŽLammertsma and Hume, 1996. and a method based on the Patlak method ŽPatlak and Blasberg, 1985., described in Appendix A. The binding potential determined by this method is proportional to the Žtrue. binding potential in the target region w BP 9 9 A Ž Bmax rK D ., where Bmax is the concentration of free receptors and K D is the equilibrium dissociation constant of the tracerx. The method also takes into account that the flow and the nonspecific binding may be different in the striatum than in the cerebellum. For the evaluation of the binding of 11 C-NMSP, which on the time scale of the PET scan was assumed to be approximately irreversible, two evaluation methods were applied: the Patlak method ŽPatlak and Blasberg, 1985. and the Wong method ŽWong et al., 1986.. The slope of the Patlak curve was determined by linear regression, from 3 min after the start of measurement to 55 min. When the ligand concentration in a reference brain region is used as an input function and no reversible binding occurs, this slope is proportional to the binding potential if it can be assumed that the tissue-to-plasma outflow rate is much larger than the association rate of the tracer]receptor system Ž k 2 4 k 3 . ŽPatlak and Blasberg, 1985; Wong et al., 1986.. For both 11 C-RAC and 11 C-NMSP, the occupancy of quetiapine at the D 2 receptor was estimated by assuming that the mean obtained in the group of healthy volunteers is a valid estimate for binding potential values of drug-naive
schizophrenics ŽNordstrom ¨ et al., 1995a.. The following equation was used: Occs 1 y
ž
BPSchiz, ttx ? 100, BPHealthy
/
where Occ is the percent occupancy, BPSchitz, ttx is the binding potential of the treated schizophrenic patient and BPHealthy is the mean of the group of healthy volunteers. 3. Results In Fig. 1a,b, the time]activity curves for 11 CNMSP and 11 C-RAC in the cerebellum, putamen and caudate nucleus of a healthy volunteer are shown. In Fig. 1c,d, the corresponding Patlak plots are shown. While the 11 C-RAC binding reaches equilibrium, the 11 C-NMSP binding continues to increase during the entire PET examination time. For the evaluation of the binding potential of 11 C-RAC, four different evaluation methods were applied while two different methods were applied to evaluate the 11 C-NMSP data. For each tracer, the different analysis methods correlated well with each other; moreover, no significant difference in occupancy between the values obtained by the different evaluation methods was found. Consequently, the Patlak methods were arbitrarily chosen for presentation of the results. The average binding potential in the group of healthy volunteers was used as a reference value to calculate the occupancy ŽNordstrom ¨ et al., 1995a.. The occupancy values thus determined for the healthy volunteers Žrelating their binding potential to their own average, thus getting an estimate of the error. and for the group of patients on 750 mg or 450 mg clozapine are listed in Table 1. The highest occupancy values, of more than 50%, are detected by 11 C-RAC. For 11 C-NMSP, negative values are obtained for the group of patients on 750 mg quetiapine, possibly due to the fact that drug-naive schizophrenic patients may have a higher Bmax than healthy controls as reported by Wong et al. Ž1986., although this has not been reproduced by others ŽNordstrom ¨ et al., 1995a.. However, since in the group of three patients on
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
Fig. 1. Time]activity curves obtained in a healthy volunteer for Ža. time]activity curve as input function.
11
C-RAC and Žb.
11
C-NMSP; Žc. and Žd. are corresponding Patlak plots, using the cerebellar
151
152
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
Table 1 Occupancy determined by 11 C-RAC and daily quetiapine, 2 h after the last dose
C-NMSP PET in the same groups of treated schizophrenic patients on 450 or 750 mg
11
Occupancy Ž%.
Untreated volunteers Patients, 750 mg Patients, 450 mg
11
ns4 ns4 ns3
11
C-RAC Ž%.
C-NMSP Ž%.
Putamen
Caudate N
Putamen
Caudate N
0 " 11a 51 " 11 43 " 10
0 " 11a 51 " 9 46 " 5
0 " 4a y9 " 8 2"3
0 " 6a y1 " 8 5"1
a
The average binding potential in the group of healthy volunteers was used as a reference value to determine the occupancy in the group of treated patients.
450 mg the occupancy was detectable, 2 " 3% in the putamen and 5 " 1% in the caudate nucleus, it might well be that the error is higher than the one indicated by the standard deviation in the group of healthy volunteers. The ratio of the average binding potential detected at a dose of 750 mg Žfour patients. to the average binding potential at a dose of 300 mg Žtwo patients. was 0.5 in both the putamen and the caudate nucleus when measured by 11 C-RAC while the same ratio for 11 C-NMSP was 0.9 in the putamen and 0.8 in the caudate nucleus. In Fig. 2, the relation between the administered dose and the percent occupancy is illustrated. Values obtained in the same individual were joined together with an eye guideline. As can be seen, the higher occupancy at higher quetiapine doses is clearly demonstrated by 11 C-RAC. In particular, Fig. 2a shows that in all patients except one, a decrease of the quetiapine dose results in a decrease in the detected occupancy. In contrast, the 11 C-NMSP Patlak slope is hardly affected by different quetiapine doses: in all patients except one, the occupancy actually increases at the lower dose. These features are seen both in the putamen and the caudate nucleus. The trend for decreasing 11 C-RAC binding with increasing quetiapine doses was significant in the putamen Ž r 2 s 0.691, Ps 0.003. and in the caudate nucleus Ž r 2 s 0.644, Ps 0.005. while no significant difference was found for 11 C-NMSP Žputamen: r 2 s 0.072, Ps 0.454; caudate nucleus: r 2 s 0.313, Ps 0.092.. When the group of healthy volunteers was included in the regression analysis,
the discrepancy between the two ligands was further reinforced. The regression coefficients obtained with 11 C-RAC became r 2 s 0.705, Ps 0.0002, in the putamen and r 2 s 0.718, Ps 0.0001, in the caudate nucleus, while for 11 C-NMSP the regression coefficient in the putamen was r 2 s 0.037, Ps 0.509, and in the caudate nucleus, r 2 s 0.022, Ps 0.616. 4. Discussion The main issue in the present report was the apparent lack of receptor occupancy when evaluating D 2-receptor binding of the atypical neuroleptic agent quetiapine with 11 C-NMSP. In contrast, a significant and dose-dependent effect was evident when the occupancy was evaluated with PET using 11 C-RAC. The data are described in detail in separate clinical reports ŽGefvert et al., 1998a,b.. The relevance of these discrepant findings is strengthened by the fact that the two assessments, made with 11 C-RAC and 11 C-NMSP, were done on the same day, in the same patient, at the same dose of quetiapine and at the same time after dosing, close to the maximum quetiapine concentration in plasma Žwhich occurs at 1]2 h post-dosing; Wong et al., 1995.. Furthermore, a steady state condition was ensured by treating the patients with quetiapine for 3 weeks, followed by 1 week of tapering off the dose. Finally, all evaluations of drug effects were made in schizophrenic patients, complemented with studies in untreated healthy volunteers. A three-compartment model and the Patlak evaluation method were used to analyse the PET
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
data. The cerebellum was used as a reference region for the analysis of D 2 binding in the putamen and the caudate nucleus. The occupancy was then calculated by relating the binding potential measured in the treated patients to the binding potential measured in the healthy untreated controls. The sensitivity of the two tracers to detect the effect of increased quetiapine doses was investigated both for the binding potential Ždata not shown. and the occupancy. This was evaluated since an elevated central D 2 receptor density has been detected by 11 C-NMSP PET in the putamen of drug-naive schizophrenic patients ŽWong et al., 1986., suggesting that the use of the binding potential determined in the group of healthy volunteers as a reference might only be valid in
153
the case of 11 C-RAC. However, even when evaluating the relationship between the binding potential and the dose, a significant change was detected by 11 C-RAC, but not by 11 C-NMSP. To verify that the results did not depend on the analysis method used, additional methods were applied Ždata not shown.: the integration method ŽFarde et al., 1989., the Logan method ŽLogan et al., 1996. and the simplified reference tissue model ŽLammertsma and Hume, 1996. to evaluate 11 CRAC binding and the Wong method ŽWong et al., 1986. to evaluate 11 C-NMSP. For each tracer, the different analysis methods correlated well with each other; moreover, no significant differences in occupancy between the values obtained by the different evaluation methods was found. The same
154
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
Fig. 2. Relationship between the quetiapine dose given to schizophrenic patients and the occupancy at the D2 receptor. Each patient was examined at two doses, as indicated by the eye guideline. The average binding potential in the group of healthy untreated volunteers was used as a reference value to determine the occupancy. Occupancy detected by 11 C-RAC PET in Ža. the putamen; Žb. the caudate nucleus; and by 11 C-NMSP PET in Žc. the putamen; and Žd. in the caudate nucleus.
result was found independent of the evaluation method used: increased quetiapine doses resulted in an increase of the occupancy when measured by 11 C-RAC. In contrast, 11 C-NMSP failed to detect changes in D 2 receptor occupancy. PET investigations of another atypical antipsychotic, clozapine have yielded similar results. While Nordstrom ¨ et al. Ž1995b., working with 11 C-RAC, have shown 20]67% D 2 receptor occupancy in a group of schizophrenic patients on 125]600 mg clozapine, Goyer et al. Ž1996., working with 11 C-
NMSP, detected occupancies of only 15]34% in a group of schizophrenic patients on higher doses: 300]900 mg clozapine. In contrast, the detection of the D 2 receptor occupancy of classical neuroleptics, such as haloperidol, is well-established. Haloperidol doses of 4]12 mg results in occupancy values of approx. 80% irrespective of which PET tracer is used, 11 C-NMSP ŽWong et al., 1986; Smith et al., 1988; Wolkin et al., 1989. or 11 C-RAC ŽFarde et al., 1992.. For the explanation of the discrepant results
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
between the quetiapine-induced occupancy as observed with the two different tracers, both well-characterized as D 2 receptor antagonists, it is necessary to look in more detail into tracer]receptor interactions in the presence of a competing drug and endogenous dopamine. Several scenarios might explain the discrepant results, namely Ži. quetiapine-induced increase of endogenous dopamine; Žii. different receptor binding sites; Žiii. different receptor subtypes; Živ. differences in tracer kinetics; and Žv. differences in equilibrium dissociation constant.
4.1. Quetiapine-induced increase of endogenous dopamine
It is well-known that raclopride binding, in contrast to N-methylspiperone binding, is affected by endogenous dopamine levels ŽFarde et al., 1990; Inoue et al., 1991a,b; Logan et al., 1991; Ginovart et al., 1997.. Results from combined micro-dialysis of extracellular dopamine and 11 CRAC in anaesthetized rats ŽHume et al., 1992. demonstrated that an amphetamine-induced 20fold increase of the dopamine concentration was necessary to decrease the 11 C-RAC binding by 37%. It was furthermore stated that a four-fold increase of extracellular dopamine only led to a slight decrease of 11 C-RAC binding. In the present study, the 11 C-RAC binding decreased more than two-fold when the quetiapine dose was increased from 300 to 750 mg. Consequently, a more than 20-fold increase of the dopamine release would have been necessary to entirely explain this decrease in 11 C-RAC binding. For both clozapine and quetiapine, a 50% increase of the accumulation of 3-methoxytyramine, indicative of increased dopamine release, has been reported ŽSaller and Salama, 1993.. Although it is unknown how much dopamine release is increased in schizophrenic patients by treatment with quetiapine, it seems improbable that the discrepant results demonstrated in the present study depend entirely on dopamine release, since such massive dopamine increases are required to account for the change in 11 C-RAC binding.
155
4.2. Different receptor-binding sites Another explanation may be that quetiapine binds to a D 2 site which is more similar to the binding site of raclopride than the binding site of N-methylspiperone. Several studies support the hypothesis of dissimilarities in binding sites between the two tracers. For instance, the inhibition constant of 3 H-NMSP was found to be twoto four-fold higher than that of 3 H-RAC when blocking with benzamides ŽHall et al., 1990.. The hypothesis of different binding sites is furthermore supported by the finding that approx. 50% higher Bmax values are detected when using w 3 Hx raclopride than when using w 3 Hxspiperone. In this context it has been hypothesized that a differential binding of the two ligands to D 2 receptor monomers and dimers may be involved ŽSeeman et al., 1992, 1997.. 4.3. Different receptor subtypes While 11 C-RAC also binds to the dopamine D 3 receptor subtype ŽSokoloff et al., 1990., 11 C-NMSP binds to the D4 receptor subtypes ŽSeeman et al., 1993.. Haloperidol, clozapine and quetiapine all have a similar affinity for the D 3 receptor. Due to this and due to the restricted D 3 receptor density in the striatum, it is unlikely that differences in binding to this receptor subtype would have led to any binding discrepancies. On the contrary, the three neuroleptics have large differences in the equilibrium dissociation constant for the D4 receptors. Haloperidol has the highest in vitro affinity, the affinity of clozapine is approx. two- to fourfold lower and the affinity of quetiapine is more than 100-fold lower than that of haloperidol ŽRoth et al., 1995; Schotte et al., 1996; Seeman et al., 1997.. Consequently, had 11 C-NMSP been selective for the D4 receptor only, the occupancy detected by this ligand might have been different for the three neuroleptics. Since 11 C-NMSP has a high affinity for D 2 receptors and since a 50% quetiapine occupancy was measured by 11 C-RAC PET, while the occupancy was virtually non-existent when detected by 11 C-NMSP, binding to the D4 receptor subtype cannot alone account for the discrepant findings presented here.
156
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
4.4. Differences in tracer kinetics The two tracers used have different in vivo kinetics. After a bolus injection, the rapidly equilibrating ŽFarde et al., 1989. 11 C-RAC reaches an equilibrium that is not a true equilibrium, but merely a transient one, due to the clearance of the tracer from the tissue. The fewer binding sites that are available, due to a competing drug, the lower is the level of the transient equilibrium. 11 C-NMSP does not reach any transient equilibrium at all during the PET examination, since the dissociation from the receptor is very slow and the examination time limited. The effect of a competing drug is to slow down the rate of tracer]receptor association ŽHulme and Birdsall, 1992., an effect that is exploited to quantify the in vivo occupancy by 11 C-NMSP PET. The discrepant results presented here seem to suggest that the low-affinity drug quetiapine does not affect the rate of 11 C-NMSP-receptor association enough to be detected during the PET time window, in contrast to in vivo studies of haloperidol ŽWong et al., 1986; Smith et al., 1988; Wolkin et al., 1989; Karbe et al., 1991. and in vitro autoradiographic studies of clozapine ŽHall et al., 1995.. It is known that 11 C-NMSP is more affected by the relation between the tissue to plasma outflux parameter, k 2 , and the rate constant for tracer]receptor association, k 3 , as well as changes in cerebral blood flow ŽPatlak and Blasberg, 1985; Wong et al., 1986. than 11 C-RAC ŽLogan et al., 1994.. We are currently performing simulations of PET experiments by generating time-radioactivity curves, using metabolite-corrected plasma curves as input ŽHagberg et al., 1998a.. Numerical solutions of the three-compartment equations, which take into account the presence of an equilibrating competing drug, are calculated. The simulated data are evaluated in the same way as the experimental data in the present study. The results suggest that for a true occupancy of 50%, the occupancy detected by 11 C-RAC would be 50% while the occupancy detected by 11 C-NMSP would be 35%, assuming a k 2 in the order of k 3 for both tracers ŽWong et al., 1986; Farde et al., 1989.. Furthermore, the simulations show that a treatment-induced alteration of the cerebral blood flow does not affect the occupancy detected by
11
C-NMSP, provided that the change in cerebral blood flow is the same in the striatum as in the reference region. 4.5. Differences in equilibrium dissociation constant
The K D value, the equilibrium dissociation constant, is commonly determined experimentally in vitro: a situation that is quite different from in vivo. In vitro, it is possible to add sufficient amounts of the competing drug, thus accounting for ligand differences in K D , and to wait sufficiently long to allow that equilibrium binding is achieved, thus accounting for differences in the dissociation rate of the ligands. Moreover, in test tube experiments, generally, the radioligand is already present while the drug is added. In contrast, in human experiments, patients are already treated with pharmacological doses of quetiapine and the radioligand is added to it. The in vitro K D value of 3 H-NMSP determined for D 2 receptors in human brain membranes ŽHall et al., 1990. is 0.2 nM, which indicates an almost 20 times higher affinity than the one of 3 H-RAC Ž K D s 3.9 nM.. The affinity of quetiapine Žassessed in vitro by competition with 3 H-spiperone. is in the order of 329 nM and for clozapine 132 nM ŽSaller and Salama, 1993. while the affinity of classical neuroleptics, such as haloperidol, is in the order of 1.4 nM ŽSchotte et al., 1996.. Since the K D values of 11 C-RAC and 11 C-NMSP differ, we hypothesized that this might explain the discrepant findings reported here. However, using the simulation model described above and in vitro K D values, we found that no discrepancies are expected and that the common three-compartment model suggests that 11 C-RAC and 11 C-NMSP have a comparable sensitivity to detect the receptor binding of both low and high affinity drugs ŽHagberg et al., 1998b.. In vivo K D values that are higher than in vitro have been found both for 11 C-RAC ŽFarde et al., 1989. and for another high-affinity D 2 receptor ligand, w 18 Fxfluoropropylepidepride ŽVotaw et al., 1993.. Both the association rate and the dissociation rate affect the K D value; thus an increase of the apparent K D can be explained either in terms of a decreased association rate or an increased
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
dissociation rate. The lipophilicity of the tracer has been proposed to affect either of these two rates. For instance, Votaw et al. Ž1993. found that the dissociation rate of the high-affinity tracer Žin vitro K D s 0.14 nM, in vivo K D s 2.2 nM. w 18 Fxfluoropropylepidepride measured in vivo by displacement could not be predicted by the common three-compartment model. It was hypothesized that a moderate lipophilicity and a reduced diffusion lead to an increase of the apparent dissociation time. Another group ŽDelforges et al., 1996. proposed that the so-called reaction volume is introduced into the common threecompartment model. This parameter admits that the free ligand concentration in the vicinity of the receptor may be different from the mean free ligand concentration in the tissue and results in an apparent decrease of the association rate Žor of the B max .. A significant linear correlation was found between the logarithm of the reaction volume and the lipophilicity. Also in vitro, an effect that can be explained in terms of the lipophilicity of the tracer has been found: Seeman et al. Ž1997. and Seeman and Kapur Ž1997. reported that D 2 receptor equilibrium dissociation constants of neuroleptic drugs depend on the radioligand used. The effect was related to a measure of lipophilicity: the tissue to buffer partition. In the case of quetiapine, K D lines of 160 nM and 650 nM were reported when investigated in competition with w 3 Hxraclopride and w 3 Hxspiperone, respectively ŽSeeman et al., 1997; Seeman, personal communication.. We introduced these values in our simulations and found that a quetiapine concentration that yielded a 50% occupancy measured by 11 C-RAC resulted in a quetiapine occupancy detected by 11 C-NMSP of approx. 16% ŽHagberg et al., 1998b.. These occupancy values are lower than the ones expected if the tissue to buffer partition is not taken into account but still does not fully explain why a close to negligible occupancy was detected by 11 C-NMSP. In contrast, our simulations of 11 CNMSP and 11 C-RAC binding in the presence of haloperidol or clozapine, using the equilibrium dissociation constants reported by Seeman and Kapur Ž1997., yielded occupancies comparable to the ones reported in the literature by others
157
ŽWong et al., 1986; Smith et al., 1988; Wolkin et al., 1989; Farde et al., 1992; Nordstrom ¨ et al., 1995b; Goyer et al., 1996.. 5. Conclusion In the present report, findings were presented supporting that 11 C-NMSP, in contrast to 11 CRAC, fails to detect changes in D 2 receptor occupancy of the atypical neuroleptic quetiapine. The relevance of the discrepant findings is strengthened by the fact that the two assessments, with 11 C-RAC and 11 C-NMSP, were done on the same day, in the same patient, at the same dose of quetiapine and at the same time after dosing, close to the maximum quetiapine concentration in plasma. The finding we have reported is an empirical one; the intent of this article was not to explain why the results obtained by the two PET tracers differ. However, we speculate that it is crucial to choose a PET tracer that is as similar as possible, especially in terms of the Žapparent. dissociation and association rates, to the pharmaceutical agent one wants to study in order to accurately measure receptor occupancy. Acknowledgements Zeneca Pharmaceuticals Alderley Edge, Macclesfield, UK, is gratefully acknowledged for the use of data and the Swedish Medical Research Council for financial support Žgrant 8318.. Appendix 1: The Patlak method The Patlak method ŽPatlak and Blasberg, 1985. is commonly used for evaluating irreversible binding only. Here the method was adapted to reversible binding by fitting the Patlak curve to a simplified empirical expression ŽEq. Ž1... First the Patlak time, C Ž t . was calculated: t
CŽt. s
H0 C
CBL
Ž t . dt
CC B LŽ t .
,
where CC B LŽ t . is the value of the cerebellar
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
158
time]activity curve at timepoint t. Then, the ratio, p¨ al, between the target region Žputamen or caudate nucleus . and the cerebellum was calculated: p¨ al s
Ct o t Ž t . , CC B L Ž t .
where Ct o t Ž t . is the value of the time]activity curve at timepoint t in the putamen or caudate nucleus. The ratio asymptotically approaches the equilibrium value BPPatlak , which is proportional to the binding potential in the target region Ž BP X X rK D , where Bmax is the concentration of A Bmax free receptors and K D is the equilibrium dissociation constant of the tracer.. The equilibrium value BPPatlak was determined for the reversible 11 C-RAC binding according to the following equation: p¨ al Ž C . s a y BPPatlak ? eyk ) C
Ž1.
where a , BPPatlak and k are constants determined by the Nelder]Mead simplex algorithm Žimplemented in Matlab, version 4.0, The Math Works, Natick, MA, USA.. The factor a takes into account that the flow and the non-specific binding may be different in the striatum than in the cerebellum. References Andersson, J.L.R., 1995. A rapid and accurate method to realign PET scans utilizing image edge transformation. Journal of Nuclear Medicine 36, 657]669. Andreasen, N.C., Carson, R., Diksic, M., Evans, A., Farde, L., Gjedde, G.A., Hakim, A., Nair, S., Sedvall, G., Tune, L., Wong, D.F., 1988. Workshop on schizophrenia, PET and dopamine D 2 receptors in the human neostriatum. Schizophrenia Bulletin 14, 471]484. Bergstrom, ¨ M., Eriksson, L., Bohm, C., Blomqvist, G., Litton, J., 1983. Correction for scattered radiation in a ring detector positron camera by integral transformations of the projections. Journal of Computer Assisted Tomography 7, 42]50. Camps, M., Kelly, P.H., Palacios, J.M., 1990. Autoradiographic localization of dopamine D2 receptors in the brain of several mammalians. Journal of Neural Transmission } General Section 80, 105]1127. Delforges, J., Syrota, A., Bendriem, B., 1996. Concept of
reaction volume in the in vivo ligand]receptor model. Journal of Nuclear Medicine 37, 118]125. Farde, L., Wiesel, F.A., Hall, H., Halldin, C., Stone-Elander, S., Sedvall, G., 1987. No D 2 receptor increase in PET study of schizophrenia. Archives of General Psychiatry 44, 671]672. Farde, L., Eriksson, L., Blomqvist, G., Halldin, C., 1989. Kinetic analysis of central w 11 Cx-raclopride binding to D2 receptors studied by PET } a comparison to the equilibrium analysis. Journal of Cerebral Blood Flow and Metabolism 9, 696]708. Farde, L., Wiesel, F.A., Stone-Elander, S., Halldin, C., Nordstrom, ¨ A.L., Hall, H., Sedvall, G., 1990. D2 dopamine receptors in neuroleptic-naive schizophrenic patients. Archives of General Psychiatry 47, 213]219. Farde, L., Nordstrom, ¨ A.L., Wiesel, F.A., Paul, S., Halldin, C., Sedvall, G., 1992. Positron emission tomographic analysis of central D2 and D2 dopamine receptor occupancy in patients treated with classical neuroleptics and clozapine. Relation to extrapyramidal side effects. Archives of General Psychiatry 49, 538]544. Farde, L., 1996. The advantage of using positron emission tomography in drug research. Trends in Neuroscience 19, 211]214. Frost, J.J., Smith, A.C., Kuhar, M.J., Dannals, R.F., Wagner, H.N., 1987. In vivo binding of 3 H-N-methylspiperone to dopamine and serotonin receptors. Life Sciences 40, 987]995. Gefvert, O., Bergstrom, ¨ M., Langstrom, ˚ ¨ B., Lindstrom, ¨ L., Yates, R., 1998a. Time course of central nervous dopamine-D2 and 5-HT2 receptor blockade and plasma drug concentrations after discontinuation of quetiapine ŽSeroquel W . in patients with schizophrenia. Psychopharmacology 135, 119]126. Gefvert, O., Bergstrom, ¨ M., Langstrom, ˚ ¨ B., Lindstrom, ¨ L., Yates, R., 1998b, Dose-dependent dopamine-D2 and 5-HT2 receptor occupancy of quetiapine ŽSeroquel. in patients with schizophrenia, manuscript. Goyer, P.F., Berridge, M.S., Morris, E.D., Semple, W.E., Compton-Toth, B.A., Schultz, S.C., Wong, D.F., Mirald, F., Meltzer, H.Y., 1996. PET measurements of neuroreceptor occupancy by typical and atypical neuroleptics. Journal of Nuclear Medicine 37, 1122]1127. Ginovart, N., Farde, L., Halldin, C., Swahn, C.G., 1997. Effect of reserpine induced depletion of synaptic dopamine on 11 C raclopride binding to D2 dopamine receptors in the monkey brain. Synapse 25, 321]325. Hagberg, G., Gefvert, O., Lindstrom, ¨ L., Bergstrom, ¨ M., Langstrom, ˚ ¨ B., 1998a. D2 receptor occupancy determined by 11 C-raclopride PET and reference region methods. Proceedings of Neuroreceptor Mapping, Neuroimage supplement, p. 20. Hagberg, G., Gefvert, O., Bergstrom, ¨ M., Wieselgren, I.M., Lindstrom, ¨ L., Wiesel, F.A., Langstrom, ˚ ¨ B., 1998b. 11 C-Nmethylspiperone PET detects different levels of dopamine D 2 receptor occupancy of an atypical neuroleptic than11 C-
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160 raclopride PET. ŽAbstract. Schizophrenia Research 29, 165]166. Hall, H., Wedel, I., Halldin, C., Kopp, J., Farde, L., 1990. Comparison of the in vitro receptor binding properties of N-w 3 Hx-methylspiperone and w 3 Hx-raclopride to rat and human brain membranes. Journal of Neurochemistry 55, 2048]2057. Hall, H., Halldin, C., Nyberg, S., Farde, L., Sedvall, G., 1995. Effect of clozapine and risperidone on 5-HT2 and D2 dopamine receptor binding in the post-mortem human brain. An autoradiographic study. European Neuropsychopharmacology 5, 19]25. Holte, S., Eriksson, L., Dahlbom, M., 1989. A preliminary evaluation of the Scanditronix PC2048-15B brain scanner. European Journal of Nuclear Medicine 15, 719]721. Hulme, E.C., Birdsall, N.J.M., 1992. Strategy and tactics in receptor-binding studies. In: Hulme, E.C. ŽEd.., Receptorligand Interactions. A Practical Approach. IRL Press at Oxford University Press, Oxford, New York, Tokyo, pp. 63]176 Hume, S.P., Myers, R., Bloomfield, P.M., Opacka-Juffry, J., Cremer, J.E., Ahier, R.G., Luthra, S.K., Brooks, D.J., Lammertsma, A.A., 1992. Quantitation of carbon-11labeled raclopride in rat striatum using positron emission tomography. Synapse 12, 47]54. Inoue, O., Kobayashi, K., Tsukada, H., Ito, T., Langstrom, ˚ ¨ B., 1991a. Differences in in vivo receptor binding between w 3 Hx N-methylspiperone and w 3 Hxraclopride in reserpine treated mouse brain. Journal of Neural Transmission } General Section 85, 1]10. Inoue, O., Tsukada, H., Kobayashi, K., 1991b. Swim stress alters in vivo binding of w 3 Hx N-methylspiperone. Neuropharmacology 30, 1101]1106. Karbe, H., Wienhard, K., Hamacher, K., Huber, M., Herholz, K., Coenen, H.H., Stocklin, G., Lovenich, A., Heiss, W.D., ¨ ¨ 1991. Positron emission tomography with w 18 Fxmethylspiperone demonstrates D 2 dopamine receptor binding differences of clozapine and haloperidol. Journal of Neural Transmission } General Section 86, 163]173. Lammertsma, A.A., Bench, C.J., Hume, S.P., Osman, S., Gunn, K., Brooks, D.J., Frackowiak, R.S.J., 1996. Comparison of methods for analysis of clinical 11 C raclopride studies. Journal of Cerebral Blood Flow and Metabolism 16, 42]52. Lammertsma, A.A., Hume, S.P., 1996. Simplified reference tissue model for PET receptor studies. Neuroimage 4, 153]158. Logan, J., Dewey, S.L., Wolf, A.P., Fowler, J.S., Brodie, J.D., Angrist, B., Volkow, N.D., Gatley, S.J., 1991. Effects of dopamine on measures of w 18 Fx-N-methylspiroperidol binding in the basal ganglia: comparison of simulations and experimental results from PET studies in baboons. Synapse 9, 195]207. Logan, J., Volkow, N.D., Fowler, J.S., Wang, G.J., Dewey, S.L., MacGregor, R., Schlyer, D., Gatley, S.J., Pappas, N., King, X, 1994. Effects of blood flow on w 11 Cxraclopride binding in the brain: model simulations and kinetic analysis of PET
159
data. Journal of Cerebral Blood Flow and Metabolism 14, 995]1010. Logan, J., Fowler, J.S., Volkow, N.D., Wang, G.J., Ding, Y.S., Alexoff, D.L., 1996. Distribution volume ratios without blood sampling from graphical analysis of PET data. Journal of Cerebral Blood Flow and Metabolism 16, 834]840. Nordstrom, ¨ A.L., Farde, L., Eriksson, L., Halldin, C., 1995a. No elevated D 2 dopamine receptors in neuroleptic-naive schizophrenic patients revealed by positron emission tomography and w 11 Cx N-methylspiperone. Psychiatry Research: Neuroimaging 61, 67]83. Nordstrom, ¨ A.L., Farde, L., Nyberg, S., Karlsson, P., Halldin, C., Sedvall, G., 1995b. D1 , D2 , and 5-HT2 receptor occupancy in relation to clozapine serum concentration: a PET study of schizophrenic patients. American Journal of Psychiatry 152, 1444]1449. Patlak, C.S., Blasberg, R.G., 1985. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. Generalizations. Journal of Cerebral Blood Flow and Metabolism 5, 584]590. Pazos, A., Probst, A., Palacios, J., 1987. Serotonin receptors in the human brain. IV: Autoradiographic mapping of serotonin-2-receptors. Neuroscience 21, 123]139. Roth, B.L., Tandra, S., Burgess, L.H., Sibley, D.R., Meltzer, H.Y., 1995. D4 dopamine receptor binding affinity does not distinguish between typical and atypical antipsychotic drugs. Psychopharmacology 120, 365]368. Saller, C.F., Salama, A.I., 1993. Seroquel: biochemical profile of a potential atypical antipsychotic. Psychopharmacology 112, 285]292. Schotte, A., Janssen, P.F.M., Gommeren, W., Luyten, W.H.M.L., VanGompel, P., Lesage, A.S., De Loore, K., Leysen, J.E., 1996. Risperidone compared with new and reference antipsychotic drugs: in vitro and in vivo receptor binding. Psychopharmacology 124, 57]73. Seeman, P., Guan, H., Civelli, O., Van Tool, H.H.M., Sunahara, R.K., Nyznik, H.B., 1992. The cloned D2 receptor reveals different densities of dopamine receptor antagonist ligands. Implications for human brain positron emission tomography. European Journal of Pharmacology 227, 139]146. Seeman, P., Guan, H., Van Tol, H.H.M., 1993. Dopamine D4 receptors are elevated in schizophrenia. Nature 365, 441]445. Seeman, P., Kapur, S., 1997. Clozapine occupies high levels of dopamine D2 receptors. Life Science 60Ž12., 207]216. Seeman, P., Corbett, R., Van Tol, H.H.M., 1997. Atypical neuroleptics have low affinity for dopamine D 2 receptors or are selective for D4 receptors. Neuropsychopharmacology 16, 93]110. Smith, M., Wolf, A.P., Brodie, J.D., 1988. Serial 18 F-N-methylspiroperidol PET studies to measure changes in antipsychotic drug D 2 receptor occupancy in schizophrenic patients. Biological Psychiatry 23, 653]663. Sokoloff, P., Giros, B., Martres, M.P., Bouthenet, M.L., Schwartz, J.C., 1990. Molecular cloning and characteriza-
160
G. Hagberg et al. r Psychiatry Research: Neuroimaging Section 82 (1998) 147]160
tion of a novel dopamine receptor ŽD3. as a target for neuroleptics. Nature 357, 146]151. Votaw, J.R., Kessler, R.M., dePaulis, T., 1993. Failure of the three compartment model to describe the pharmacokinetics in brain of a high affinity substituted benzamide. Synapse 15, 177]190. Wiesel, F.A., 1995. Positron emission tomography, receptors and schizophrenia. In: Hafner, H., Gattaz, W.F. ŽEds.., ¨ Search for the Causes of Schizophrenia, vol. III. Springer, Berlin, Heidelberg, New York, pp. 252]262. Wolkin, A., Brodie, J.D., Rotrosen, J., 1989. Dopamine receptor occupancy and plasma haloperidol levels. Archives of General Psychiatry 46, 482]483. Wong, D.F., Wagner, H.N., Tune, L.E., Dannals, R.F., Links, J.M., Frost, J.J., Ravert, H.T., Wilson, A.A., Rosenbaum,
A.E., Gjedde, A., Douglas, K.H., Petronis, J.D., Folstein, M.F., Toung, J.K.T., Burns, H.D., Kuhar, M.J., 1984. Effects of age on dopamine and serotonin receptors measured by positron tomography in the living human brain. Science 226, 1393]1396. Wong, D.F., Wagner, H.N., Tune, L.E., Dannals, R.F., Pearlson, G.D., Links, J.M., Tamminga, C.A., Broussolle, E.P., Ravert, H.T., Wilson, A.A., Toung, J.K.T., Malat, J., Williams, J.A., O’Tuama, L.A., Snyder, S.H., Kuhar, M.J., Gjedde, A., 1986. Positron emission tomography reveals elevated dopamine receptors in drug-naive schizophrenics. Science 234, 1558]1563. Wong, J. et al., 1995. Multiple dose pharmacokinetics of quetiapine in elderly schizophrenic patients, Annual Meeting of the American Psychiatric Association, abstract 612.
