INTRODUCTION

It is well established that the dopamine D2/3 receptors are the target of most antipsychotic drugs, though it remains unclear whether the D2/3 receptors are involved in the pathophysiology of psychosis. An important recent development has been the evidence from preclinical models indicating that the ‘high-affinity state’ of the dopamine D2 receptors (D2high) is elevated in psychosis (Seeman et al, 2005b). The dopamine D2 and D3 receptors are members of the G-protein-coupled receptors (GPCR) family. The D2 receptor exists in two inter-convertible states: a G-protein-coupled state which has a high affinity for agonist binding and is responsible for the functional effects of dopamine; and a G-protein-uncoupled state which is a functionally inert state with a low-affinity for dopamine. On the other hand, the existence of the G-protein-coupled and uncoupled states of the D3 receptors is controversial (Freedman et al, 1994; Seeman et al, 2005a; Sokoloff et al, 1992), due to the inability of the G-protein analogs (eg guanilylimidodiphosphate, GppNHp) to shift it from the high- to the low-affinity state (Vanhauwe et al, 2000). Nevertheless, the D3 receptor has at least 20-fold higher affinity than D2 for dopamine (Freedman et al, 1994; Seeman et al, 2005a; Sokoloff et al, 1992).

Data from over a dozen preclinical animal models of psychosis (eg amphetamine sensitization, phencyclidine sensitization, ethanol withdrawal, hippocampal lesion) (Seeman et al, 2005b) suggests that there is a selective elevation of high-affinity states in psychosis. Though, the method of distinguishing high- from low-affinity states in vitro involves the addition of high concentrations of G-protein analogs, a method that is not feasible in humans. Molecular imaging techniques such as positron emission tomography (PET) allow for the in vivo study of brain receptors; but, until recently, the available radioligands for D2/3 receptors (eg [11C]-raclopride, [18F]-fallypride) were all ‘antagonist’ radioligands which would not distinguish between the high- and low-affinity states of the receptors. Studying the high-affinity state of the receptor directly in patients requires the use of an ‘agonist’ radioligand.

Recently, our group has developed [11C]-(+)-PHNO ([11C]-(+)-4-propyl-9-hydroxynaphthoxazine) (Wilson et al, 2005), a D2/3 agonist radiotracer for use in humans. [11C]-(+)-PHNO binds with nanomolar affinity to D2 and D3 receptors allowing for a prominent signal from the striatum, globus pallidus (GP), substantia nigra (SN), and anterior thalamus (Freedman et al, 1994; Graff-Guerrero et al, 2008; Narendran et al, 2006; Seeman et al, 1993; Willeit et al, 2006). It shows at least three critical differences compared to antagonist radiotracers. First, unlike antagonist radiotracers which do not distinguish between the low- and high-affinity states, [11C]-(+)-PHNO is assumed to bind preferentially to the high-affinity states of the receptors (Ginovart et al, 2006; Seeman et al, 2007).

Second, (+)-PHNO has a higher in vitro and in vivo affinity for D3 than for D2 receptors. In vitro, Freedman et al (1994) and Parker et al (2006) reported that (+)-PHNO has more than 10-fold higher affinity for D3 than for D2 using conventional in vitro techniques, though Seeman et al (2005a) reported that (+)-PHNO has a higher affinity for the D2high than D3high) using in vitro analysis in the presence of GppNHp. In vivo, Narendran et al (2006) suggested that [11C]-(+)-PHNO has fourfold higher preference for D3 than D2high—a finding supported by studies in nonhuman primates showing that the [11C]-(+)-PHNO signal in the GP and midbrain binding is selectively displaceable with the D3 preferential agonist (BP897) and antagonist (SB-277011) (Narendran et al, 2006; Rabiner et al, 2007b, 2008). Finally, studies in healthy humans show that the [11C]-(+)-PHNO signal in the GP is preferentially displaceable with the D3 antagonist (ABT925) (Abi-Saab et al, 2008; Graff Guerrero et al, 2008). Thus, across species, methods and techniques, [11C]-(+)-PHNO shows a modest higher affinity for D3 over D2.

Finally, [11C]-(+)-PHNO has shown higher sensitivity for amphetamine displacement than [11C]-raclopride in anesthetized cats (Ginovart et al, 2006). However, this higher sensitivity was not found in an ex vivo dissection study in awake rodents (McCormick et al, 2008), though it is suggested by the study in awake and healthy humans (Willeit et al, 2008). The cause of this discrepancy between studies is still elusive, though the cat and human studies suggest that [11C]-(+)-PHNO binds preferentially to the high-affinity states of the dopamine receptor.

The development of [11C]-(+)-PHNO has finally made it possible to test the hypothesis with regards to elevated high-affinity states in schizophrenia. Moreover, as [11C]-(+)-PHNO provides a robust binding from the D3-rich areas in the brain (GP and SN), for the first time it has provided an opportunity to explore directly the suggested involvement of the dopamine D3 receptors in schizophrenia (Griffon et al, 1995; Gurevich et al, 1997; Joyce and Gurevich, 1999).

The current study was aimed to compare the [11C]-(+)-PHNO binding in drug-free patients with schizophrenia and appropriate controls to test the D2 high-affinity states hypothesis and to compare the dopamine D3 binding between patients with acute psychosis and controls.

MATERIALS AND METHODS

Clinical Sample

This study has been approved by the local Research Ethics Board and by Health Canada. Twenty-six subjects were included: 13 drug-free patients and 13 healthy controls. The participants provided written informed consents after the study procedures and risks were explained. They were excluded if they had a current diagnosis of substance abuse or dependence at screening, positive result in urine drug screen at enrollment or before any of the PET scans, history of clinically significant physical illness, pregnant or lactating women at screening or positive urine pregnancy test before PET scans, or metal implants that would preclude the MRI scan.

The patients were evaluated before PET scan with the clinical global impression (CGI) and positive and negative syndrome scale (PANSS).

The patients had the diagnosis of schizophrenia or schizophreniform disorder corroborated with the MINI-Plus structured interview (Sheehan et al, 1998). The age range for inclusion was from 18 to 50 years old. The patients were drug free for at least 2 weeks and were excluded if they had been exposed to depot antipsychotic in the past 2 years; had a minimum PANSS total score of 55, score of >3 on at least two PANSS psychosis items (P1, P2, P3, P5, or P6) or >4 on one psychosis item; and CGI severity score ⩾4 (moderately ill).

The healthy controls were recruited through local advertisement. Psychiatric disorders were excluded using the MINI-Plus structured interview (Sheehan et al, 1998). The age range for inclusion was from 18 to 50 years old. Subjects with any medical or neurological conditions or with axis I psychiatric diagnoses were excluded from the study. Similarly, subjects with substance abuse (other than smoking) within 6 months before their baseline visit were not included. Participants were asked to consume no more than their usual amount of coffee (and if smokers, cigarettes) on the day of PET examination, and to abstain from alcohol intake 24 h before PET scans. Standard urine tests for psychotropic substances were performed at inclusion and immediately before PET scan. Pregnancy was excluded using serum analysis at inclusion and urine pregnancy tests before each scan.

[11C]-(+)-PHNO Synthesis

The radiosynthesis of [11C]-(+)-PHNO has been described in detail elsewhere (Wilson et al, 2005). Briefly, [11C]-propionyl chloride was reacted with 9-hydroxynaphthoxazine to generate a [11C]-amide which is subsequently reduced by lithium aluminum hydride. Purification by HPLC and formulation gave radiochemically pure [11C]-(+)-PHNO as a sterile, pyrogen-free solution suitable for human studies.

Positron Emission Tomography Imaging

Studies were performed using a high-resolution brain PET camera system, HRRT (Siemens Molecular Imaging, Knoxville, TN), which measures radioactivity in 207 brain slices with an interslice distance of 1.2 mm. The in-plane resolution of the scanner is approximately 2.8 mm full-width-at-half-maximum. Transmission scans were acquired using a 137Cesium single photon point source to provide attenuation correction. The emission data were acquired with a head fixation system during PET scans to avoid movement during the acquisition. After being placed on the scanning table, a total of 329±59 MBq (8.9±1.6 mCi) with a specific activity of 955.9±295 mCi/μmol (range: 407–1790 mCi/μmol) and a mass of 2.33±0.39 μg (range: 1.2–3.3 μg) of [11C]-(+)-PHNO was injected as a bolus followed by a flush with 2 ml saline into an intravenous line placed in an antecubital vein. Scanning data were acquired for 90 min after the injection. Once scanning was completed, the data were reframed into 30 frames (1–15 of 1 min duration and 16–30 of 5 min duration). From the 26 participants in this study, one healthy control experienced transient mild nausea. This was presented few minutes (∼10 min) after [11C]-(+)-PHNO injections and lasted less than 5 min. No interruption of the PET procedures was required due to this adverse effect of the radiotracer.

MRI Imaging

Subjects undertook a proton density image (TE=17, TR=6000, FOV=22 cm 2D, 256 × 256, slice thickness of 2 mm, NEX=2) acquired on a 1.5T Signa scanner (General Electric Medical Systems, Milwaukee, WI). These images were used for the analysis of the PET scans.

Image Analysis

The time activity curves (TACs) from the regions of interest (ROIs) caudate, putamen, ventral striatum (VS), GP, SN, and cerebellar cortex were obtained from the dynamic [11C]-(+)-PHNO PET images in native space with reference to co-registered MRI image as previously described (Graff-Guerrero et al, 2008; Rusjan et al, 2006; Willeit et al, 2008). The ROI in the anterior thalamus was drawn manually using the Analyze Software Version 6.0 (AnalyzeDirect, Lenexa, Kansas) in the anterior part of the thalamus as observed in the first axial projection where putamen and thalamus can be shown. The co-registration of the MRI to the PET space image was carried out using the normalized mutual information algorithm as implemented in SPM2 (SPM2, Welcome Department of Cognitive Neurology, London; http://www.fil.ion.ucl.ac.uk/spm).

The TACs were obtained using a in-house software for semi-automated generation of ROIs (Rusjan et al, 2006). The quantitative estimate of binding was estimated using the simplified reference tissue method (Lammertsma and Hume, 1996) with the cerebellar cortex as reference region. This method has been validated to reliable estimate the binding potential no-displaceable (BPND), which compares the concentration of radioligand in the receptor-rich region to the receptor-free region (Innis et al, 2007), and has been validated for use with [11C]-(+)-PHNO (Ginovart et al, 2007). The BPNDs were estimated using the PMOD v2.7 software (PMOD Technologies Ltd, Zurich, Switzerland).

Parametric voxel-wise BPND maps from the dynamic images in native space were generated according to the method of Gunn et al (1997) with the cerebellum as reference region and as implemented in PMOD v2.7 software (PMOD Technologies Ltd). The BPND map images were spatially normalized into the Montreal Neurological Institute brain space by nearest neighbor interpolation and with a voxel size fixed in 2 × 2 × 2 mm using the SPM2 software (Friston, 1995). The normalized images were smoothed with a Gaussian filter in each coordinate direction with a kernel of 4 mm.

Statistical Analysis

The analysis was made using the Statistical Program for the Social Sciences (version 15.0; SPSS, Chicago, IL). Variables were presented as mean±standard deviation (SD). Demographic and clinical characteristics were compared between patients and controls by using Mann–Whitney U-test (gender) on categorical data and independent sample t-tests on continuous data (age). Analysis of variance with Bonferroni correction for multiple comparisons as a post hoc was performed to compare the BPNDs between drug-free patients and sex- and age-matched healthy controls per ROI (caudate, putamen, VS, GP, SN, and anterior thalamus). Pearson's product moment correlations between PANSS (total and subscales) and BPNDs were estimated. Repeated measures analysis of variance was performed to compare the standardized radioactivity in the cerebellum (within subjects factor) between drug-free patients and sex- and age-matched healthy controls (between subjects factor). The significance was always assumed at p<0.05.

The voxel-wise comparison between BPND maps, of drug-free patients and sex- and age-matched healthy controls, was performed with a two sample t-tests as implemented in SPM2. We reported areas only if they met the joint criteria of (a) p(uncorrected)<0.001; (b) an extent ⩾10 voxels; and (c) in the caudate, putamen, GP, VS, midbrain, and anterior thalamus. Additionally, the statistically significant a priori regions would be corrected for multiple comparisons using the false discovery rate (FDR) approach.

RESULTS

The characteristics of the patients and controls are summarized in Table 1. There were no differences between groups in age (drug-free patients mean=25.8±SD=5.9 years; healthy controls=26.8±6.4 years, t=−0.40, p=0.69) and in gender (drug-free patients=9 men; healthy controls=9 men; Mann–Whitney U-test=84.5, p=1.0). No differences emerged between drug-free patients and age- and sex-matched healthy controls in the BPND in any of the ROI (F(6,19)=0.671, p=0.674). The BPND (mean±SD) in the caudate for controls was 2.1±0.2 and for patients was 2.2±0.4 (p=0.52, pair-wise Bonferroni corrected); putamen: controls=2.6±0.3 and patients=2.6±0.2 (p=0.50); VS: controls=3.1±0.4 and patients=3.2±0.7 (p=0.63); GP: controls=2.9±0.6 and patients=2.7±0.5 (p=0.51); SN: controls=1.5±0.6 and patients=1.3±0.4 (p=0.33); thalamus: controls=1.2±0.2 and patients=1.1±0.3 (p=0.43) (Figure 1). This result was corroborated by the voxel-wise comparison of the BPND maps between drug-free patients and controls which did not show any differences (Figure 2). Moreover, the BPND did not correlate with the severity of the illness as measured by the PANSS even before correction for multiple comparisons (Table 2).

Table 1 Characteristics of the Drug-Free Patients Group and Age- and Sex-Matched Healthy Controls
Full size table
Figure 1
figure 1

Comparison of the [11C]-(+)-PHNO binding potential no-displaceables (BPNDs) of every region of interests (ROI) between drug-free patients and age- and sex-matched healthy controls. Black symbols indicate drug-naive patients, grey symbols indicate drug-free patients, and white symbols indicate healthy controls. Horizontal lines indicate mean BPND from each ROI. VS, ventral striatum; GP, globus pallidus; SN, substantia nigra.

Full size image
Figure 2
figure 2

[11C]-(+)-PHNO mean and standard deviation binding potential no-displaceable (BPND) maps of drug-free patients with schizophrenia (n=13) and age- and sex-matched healthy controls (n=13) illustrating the similitude in binding between controls and patients. The BPND maps images correspond to axial projections overlay in a T1 template in the Montreal Neurological Institute (MNI) space. The images correspond to 13 patients and 13 controls. Z corresponds to the millimeters above (+) or below (−) the anterior commissure in the AC–PC plane.

Full size image
Table 2 Pearson's Product Moment Correlations Between [11C]-(+)-PHNO BPND and Positive and Negative Syndrome Scale
Full size table

DISCUSSION

This study represents the first comprehensive investigation of the dopamine D2 receptors in the high-affinity state (D2high) and the D3 receptors in patients. Contrary to our expectation based on preclinical data, we did not find evidence for elevated D2high in untreated patients with schizophrenia or schizophreniform disorder during acute psychosis.

The high-affinity state hypothesis has been sustained in many preclinical models of psychosis (Seeman et al, 2005b) and has been indirectly suggested by dopamine depletion studies in patients (Abi-Dargham et al, 2000). Our results in drug-free patients failed to support this hypothesis. Although the reason for this discrepancy between predictions from in vitro data and findings in patients is not clear, several issues need to be considered.

First, the methods of measuring the high-affinity state in vitro and in patients are very different. The in vitro studies include competition assays between antagonist radiotracers and agonist ligands in the presence of non-hydrolysable GTP analogs (GppNHp), the latter converting the receptor from its high-affinity state to its low-affinity state (De Lean et al, 1980; Seeman et al, 2002, 2005b). This differs from the approach in humans which involves a direct study of the high-affinity states with an agonist radioligand.

Second, there is ongoing debate as to the measurability of high-affinity state in vivo: eg Sibley et al (1983) identified both high- and low-affinity binding sites in membrane cells, but only low-affinity sites on the same intact cells. However, Seeman (2008) identified the high-affinity binding sites on intact cells with [3H]domperidone, but not with [3H]spiperone or [3H]raclopride. Moreover, one PET study in nonhuman primates with [11C]-NPA (D2 agonist) and [11C]-raclopride (D2/3 antagonist) through a Scatchard plot analysis was able to differentiate the density of receptors (Bmax) in the high- vs high+low-affinity state (Narendran et al, 2005); although the same approach in cats with [11C]-(+)-PHNO and [11C]-raclopride did not find any difference in the Bmax between the two radiotracers (Ginovart et al, 2006). On the other hand, studies in anesthetized cats (Ginovart et al, 2006), anesthetized nonhuman primates (Narendran et al, 2004; Seneca et al, 2006), and awake humans (Willeit et al, 2008) showed greater displacement with an agonist radiotracer ([11C]-NPA, [11C]-MNPA, [11C]-(+)-PHNO) than with an antagonist radiotracer in response to an amphetamine challenge. But one ex vivo study in awake rodents did not find any difference in amphetamine displacement between [11C]-(+)-PHNO and [11C]-raclopride (McCormick et al, 2008). These studies suggest that although agonist radiotracers do show a distinct binding profile in vivo, the exact difference varies as a function of the ligand and the tissue/animal characteristics.

Finally, it is theoretically possible that an increase in the D2high and/or D3 is indeed present in acute psychotic state in schizophrenia, but this may be masked by the presence of abnormal high levels of dopamine in the synaptic cleft. The abnormal high levels of dopamine in psychosis has been indirectly shown after an amphetamine challenge (Abi-Dargham et al, 1998), but the dopamine striatal synthesis capacity has reliably been found higher than in controls (Dao-Castellana et al, 1997; Hietala et al, 1999; Huttunen et al, 2007; Reith et al, 1994).

Moreover, if the density of receptors configured in the high-affinity state is extremely high as suggested in normal conditions (∼80%) (Narendran et al, 2005), the difference between patients and controls may be below the resolution of the PET technique. One approach to uncover the differences and to test this hypothesis would require depleting dopamine in patients before imaging studies (Abi-Dargham et al, 2000). We did not deplete dopamine levels and thus cannot definitively rule out a change in D2high/D3 receptors which may have been cancelled by the ambient increase in endogenous dopamine. This is not just a theoretical concern, because studies by Abi-Dargham et al (2000) have shown that patients not only have higher baseline occupancy of dopamine receptors by dopamine, but also that when they are depleted of dopamine, they reveal a higher number of absolute D2 receptors (Abi-Dargham et al, 2000). Thus, a study depleting dopamine would be very critical in definitively resolving this issue.

The ROI and voxel-wise analyses did not reveal any difference in the D3-rich areas (GP and SN) between patients and controls. These results disagree with a previous postmortem study (Gurevich et al, 1997), including drug-free patients with schizophrenia, describing twofold increase in the total number of D3 receptors. The reason of this discrepancy with our results is unclear. But, differences in the sample characteristics between postmortem studies and our data such as age (old vs young), length of illness (years vs months), and previous exposure to antipsychotic medications (Graff-Guerrero et al, 2007) may help to explain the discrepancy on the results.

The current study also provides the first insight to the status of the D3 receptors in the schizophrenia, as the [11C]-(+)-PHNO signal in the SN is predominantly related to D3 (Graff Guerrero et al, 2008; Rabiner et al, 2007a, 2007b). The SN (midbrain) is important because it is the origin of the nigrostriatal dopaminergic system and the D3 receptors in the SN are autoreceptors (Diaz et al, 2000). Thus alterations in these receptors could contribute to the hyperactivity of this midbrain dopamine system which is believed to play an important role in the genesis of schizophrenic symptoms (Murray et al, 2008). However, we do not find support for an alteration of the D3 in the SN.

Another issue to consider is that the [11C]-(+)-PHNO signal reflects its binding to both D2high and D3 receptor populations. The contribution of each receptor to the regional binding varies according to the numbers of each receptor sub-type in the region and to the affinity of the radioligand for that receptor sub-type. Although the precise contribution of D2high and D3 to each ROI is not fully mapped, studies in rodents, baboons, and humans estimate that the vast majority of the [11C]-(+)-PHNO binding in the GP (68–89%) and in the midbrain SN (91–100%) corresponds to D3. On the other hand, the estimation of [11C]-(+)-PHNO binding in the dorsal striatum that corresponds to D3 is between 8 and 50% (Graff Guerrero et al, 2008; Rabiner et al, 2007a, 2007b). These proportions suggest that [11C]-(+)-PHNO binding in the GP and SN is a reasonable proxy for the estimation of D3 receptors in those regions, although the dorsal striatum should be understood as a mixture of predominantly D2high with some contributions from D3. However, this limitation in radioligand selectivity should not take away our empirical finding that we find no difference in the [11C]-(+)-PHNO BPND in drug-free patients with schizophrenia vs controls. Further, our inference that there is no substantial difference in either D2high or D3 is also secure because a change in either of these should have led to an increase in the overall BPND and we do not observe that. There is of course the hypothetical possibility that a simultaneous increase in D2high coupled with a decrease in D3 (or vice versa) could hide a true finding. But, there is no data in the literature that suggest such possibility.

We did not record the smoking status of the healthy controls. This limitation may be important in the light of the effect of cigarette consumption, craving and its mood or hedonic effect on the dopamine system (Barrett et al, 2004; Brody et al, 2008; Montgomery et al, 2007), and the fact that patients with schizophrenia are more likely to be smokers than general population (de Leon and Diaz, 2005). However, given that we found no effect of smoking within our patients with schizophrenia (three smokers vs 10 nonsmokers) (data not shown), we think that it is unlikely that smoking alone could have obscured differences between patients and controls.

The lack of a full kinetic analysis is another potential limitation. This approach would have allowed for the direct estimation of the BP in the ROIs, without the assumption of equivalent nonspecific and free fraction in the reference region. However, our data showed that the cerebellar uptake in the drug-free patients and sex- and age-matched healthy controls was not different (Figure 3; F(1,24)=0.34, p=0.56) and, therefore, it is unlikely that a true finding could have been obscured by differences in the reference region.

Figure 3
figure 3

Time–activity curves (mean±SD) from the cerebellum of 13 drug-free patients with schizophrenia and 13 age- and sex-matched healthy controls. y axis represents standardized uptake values (SUV; calculated as: regional radioactivity concentration/(injected radioactivity/body weight) for [11C]-(+)-PHNO. The time–activity curves from the cerebellum illustrate that there was no difference between groups on free and nonspecific [11C]-(+)-PHNO binding. Error bars correspond to standard deviation (repeated measure ANOVA, F(1,24)=0.34, p=0.56).

Full size image

The current study represents the first effort to measure the D2high and D3 receptors in drug-free patients with psychotic disorders (schizophrenia-spectrum disorders). Our results indicate that patients in an acute psychotic episode do not exhibit an increase of the D2high or D3 receptors in comparison to sex- and age-matched healthy controls. We cannot rule out the possibility that any difference of the D2high or D3 were occluded by endogenous dopamine. The next step to definitively rule out such a possibility would require studies using [11C]-(+)-PHNO but in patients whose dopamine has been depleted. Until such a finding is observed, accounts should not assume a change in the D2high or D3 in the pathophysiology of psychosis.