Increased Caudate Dopamine D2 Receptor Availability as a Genetic Marker for Schizophrenia

Context  Schizophrenia has a heritability of about 80%, but the detailed molecular genetic basis of the disorder has remained elusive. Relative hyperfunction of the subcortical dopamine system has been previously suggested to be one of the key pathophysiologic mechanisms in schizophrenic psychosis.

Objective  To examine the contributions of genetic vulnerability for schizophrenia to the dopamine system in the human brain.

Design  Population-based twin cohort study.

Setting  Finland.

Participants  Six monozygotic and 5 dizygotic unaffected co-twins of patients with schizophrenia were ascertained, along with 4 monozygotic and 3 dizygotic healthy control twins with no family history of psychotic disorders.

Main Outcome Measures  Striatal dopamine D₂ receptor availability estimated with positron emission tomographic imaging and carbon 11 (¹¹C)–labeled raclopride, and performance on neuropsychological tests sensitive to frontal lobe functioning and to schizophrenia vulnerability.

Results  Unaffected monozygotic co-twins had increased caudate D₂ density compared with unaffected dizygotic co-twins and healthy control twins. Higher D₂ receptor binding in caudate was associated with a poor performance on cognitive tasks related to schizophrenia vulnerability in the whole sample.

Conclusions  The caudate dopamine D₂ receptor up-regulation is related to genetic risk for schizophrenia. Higher dopamine D₂ receptor density in caudate is also associated with poorer performance on cognitive tasks involving corticostriatal pathways. This finding suggests that caudate dopamine dysregulation is also a trait phenomenon related to psychosis vulnerability. This pattern of results provides a theoretical rationale for early pharmacologic intervention approaches using dopamine D₂ receptor blocking drugs.

Schizophrenia has a heritability of about 80%,¹ but the detailed molecular genetic basis of the disorder has remained elusive.² Modern neuroimaging techniques have made it feasible to study brain function at a molecular level in vivo. Among the brain neurotransmitter systems, the dopamine (DA) system has been most intensively studied in schizophrenic patients. Previous neuroimaging studies suggest that there is dysregulation of the subcortical DA system, as indicated by increased responsivity to pharmacologic challenges of subcortical DA neurons.³ In vivo receptor imaging studies indicate increased density of striatal postsynaptic D₂ receptors in patients with schizophrenia.^4,5 Thus far, it is unknown whether this increased activity of the subcortical DA system represents a common final pathway in the schizophrenic psychosis as a state-related phenomenon, or whether it could be a genetically mediated trait associated with vulnerability to the disease. The search for endophenotypes, measurable mediators between predisposing genes and clinical phenotypic features,⁶ may facilitate identification of the genetic underpinnings of this multifactorial neuropsychiatric disease with heterogeneous clinical manifestations.

In the present study, we examined unaffected co-twins of schizophrenic patients as well as healthy control twins by positron emission tomography (PET) and carbon 11 (¹¹C)–labeled raclopride, a DA D₂ receptor antagonist ligand. This study paradigm has proven to be feasible in studying the striatal DA system in vivo.⁷ Group comparisons were performed across 3 levels of genetic loading (control twins [n = 14], dizygotic [DZ] co-twins [n = 5], and monozygotic [MZ] co-twins [n = 6]). The MZ co-twins share 100% and the DZ co-twins on average 50% of their segregating genes with the proband. All subjects also underwent comprehensive neuropsychological testing for functions attributed to the frontal lobe, a region that has a regulatory influence on the subcortical DA system.⁸ This study was set out to test the hypothesis that the unaffected co-twins of schizophrenic patients will show striatal DA receptor disturbances similar to those seen in patients with the disease. We further hypothesized that this aberration would correlate with performance on cognitive tests related to frontal lobe functioning, which we have previously shown to be associated with genetic vulnerability to schizophrenia.⁹

The study protocol was reviewed and approved by the institutional review boards or ethical committees of the University of California (Los Angeles), the University of Turku (Turku, Finland), the University of Pennsylvania (Philadelphia), and the National Public Health Institute of Finland (Helsinki), and all participants signed institutional review board– or ethical committee–approved informed consent forms. The study was performed in accordance with the Declaration of Helsinki.

Subjects were drawn from a twin cohort consisting of virtually all same-sex twin pairs born in Finland from 1940 to 1957 (N = 9562 pairs). Questionnaire-based classification identified 2495 MZ pairs, 5378 dizygotic pairs, and 1689 pairs of unknown zygosity.¹⁰ This cohort was screened, for the period of 1969 through 1991, for a history of hospitalization, medicine prescription, and/or work disability due to psychiatric indication, in 3 national computerized databases: the hospital discharge register, the register for fully reimbursed medications, and the pension register.⁹ These searches identified 348 index twin pairs, with at least 1 co-twin with a diagnosis of schizophrenia or schizoaffective disorder, and 9214 healthy pairs, with no schizophrenia diagnosis in either co-twin according to any of the 3 sources. After exclusion due to death or emigration, a total of 130 twin pairs consisting of 60 (27 MZ and 33 DZ) index pairs were chosen randomly from the available index pairs (N = 229: 50 MZ, 121 DZ, and 58 unknown zygosity), along with 70 (34 MZ and 36 DZ) demographically balanced healthy pairs. Index pairs in which, on direct interview, either the proband had a diagnosis of schizoaffective disorder–affective type or the co-twin had a psychotic disorder diagnosis were excluded (n = 1 concordant MZ pair, 1 discordant MZ pair, 1 discordant DZ pair, 1 unpaired DZ proband, and 1 unpaired MZ proband). Healthy pairs were excluded if there was a history of psychosis-related treatment or work disability in any of their first-degree relatives or if either co-twin was found, on direct interview, to meet diagnostic criteria for a psychotic disorder or schizotypal, paranoid, or schizoid personality disorder (n = 15 pairs: 6 MZ and 9 DZ). The selected sample of 111 pairs consisted of 8 pairs (7 MZ and 1 DZ) concordant and 48 pairs (18 MZ and 30 DZ) discordant for schizophrenia, and 55 healthy pairs (28 MZ and 27 DZ). Of this sample, 25 subjects took part in the PET study. Our final sample of subjects who underwent PET consisted of 11 healthy co-twins from pairs discordant for schizophrenia (6 MZ and 5 DZ) and 14 twins from 7 healthy pairs (4 MZ and 3 DZ). One pair of healthy control DZ twins was excluded from the analysis because one of the twins was diagnosed as having a cluster A personality disorder. The zygosity of the studied twins was confirmed by genetic markers as described previously.⁹ No subjects were taking medication affecting the central nervous system.

The PET experiments were performed as described previously.⁷ In brief, each subject received an intravenous rapid bolus dose of 5.1 to 7.0 mCi (190.2-257.6 MBq) of [¹¹C]raclopride (specific radioactivities, 1.01 ± 0.2 mCi/nmol [37.3 ± 8.3 MBq/nmol]; amount of radiotracer, 2.08 ± 0.51 μg; mean ± SD). No differences between groups of genetic liability in any of these parameters were detected. The uptake of [¹¹C]raclopride was measured by means of a whole-body PET scanner (GE Advance; General Electric Co, Milwaukee, Wis) in 3-dimensional mode with 35 sections of 4.25-mm thickness covering the whole brain for 51 minutes after injection using 13 time frames. [¹¹C]raclopride was prepared as previously reported.⁷

Regions of interest (ROIs) were drawn for the calculation of regional time-activity curves.⁷ The ROIs for caudate, putamen, thalamus, and cerebellum were manually delineated on the magnetic resonance images coregistered and resliced according to PET images by means of a mutual information method as implemented in SPM99 (Statistical Parametric Mapping; Institute of Neurology, University College of London, London, England).¹¹ All ROIs were drawn on 3 planes. No partial volume effect correction was done. The ROI analyses were performed with Imadeus software (release 1.0; Forima Inc, Turku, Finland). The simplified reference tissue model was applied to the time-activity data of each ROI to calculate binding potential (BP) values. Here, BP represents the ratio of the rate constants for transit between nondisplaceable and specific binding compartment (BP = k₃/k₄) in terms of kinetic modeling and in terms of receptor parameters, the product of receptor density (B_max), apparent affinity (1/K_d), and nonspecific binding in the brain (f₂) (BP = f₂B_max/K_d) (see Slifstein and Laruelle¹² for discussion of the nomenclature). The applicability of the simplified reference tissue model to [¹¹C]raclopride has been previously validated for ROI-based analysis in striatum with cerebellum as a reference region.¹³

All subjects were tested with a comprehensive neuropsychological battery out of which a canonical liability variable was constructed. This variable reflects performance on tests of spatial working memory (immediate recall of sequences of spatial locations, sum of forward and backward orders of presentation; Visual Span subtest of Wechsler Memory Scale–Revised), divided attention (simultaneously counting backward and performing spatial cancellation, percentage decrement from single-task performance in a Brown-Petersen dual-task paradigm), intrusions during recall of a word list (“recall” of items not on the learning list; California Verbal Learning Test), and choice reaction times to visual targets (pressing the appropriate button when a target appears on either the left or the right side of fixation; average time to respond in a Posner paradigm) (see Table 1 from Cannon et al⁹). Deficits on these tests have been shown to be heritable in twins discordant for schizophrenia, increasing in severity with increasing genetic proximity to the proband, and each of these tests contributed uniquely to the discrimination of degree of genetic liability for schizophrenia in a canonical discriminant analysis.⁹ The canonical liability variable represents combination of these 4 items. Findings on the neuropsychological data have already been published,⁹ and the present study sample represents a subgroup of the original sample.

Data were analyzed by means of the general linear mixed model with repeated measures (SAS, version 6.12; SAS Institute Inc, Cary, NC), correcting for dependency (ie, correlation) among the healthy co-twins, by treating twin pair as a random variable, and adjusting the model error terms accordingly (Satterthwaite option). Differences in mean BP between groups were tested by modeling risk group (healthy MZ co-twins from discordant pairs, healthy DZ co-twins from discordant pairs, and healthy twin pairs) as a fixed-effect predictor, while covarying for age at scanning, since striatal D₂ receptors decline with age.¹⁴ To test for possible differences in laterality, hemisphere entered the model as a within-subject repeated-measures factor, and a group × hemisphere interaction entered the model to test for possible differences in laterality between the groups. Whenever one of these terms contributed significantly to the prediction of [¹¹C]raclopride BP, contrast analyses were performed comparing conditions within the term collapsing over nonsignificant terms in the model. This approach maintains the hypothesis-wise type I error rate at .05 because a predictor’s contribution to particular dependent measures is evaluated only if its effect is found to vary at the multivariate level. The significance of each predictor was tested while accounting for all other model terms simultaneously. Since the SAS proc mixed procedure did not allow for the use of multiple repeated-measures factors and a random factor within the same model, a separate model was used for each ROI (caudate, putamen, and thalamus), and a Bonferroni-corrected P value of .017 was used to determine the significance of the predictors.

The relationship between D₂ BP in the caudate and our canonical liability variable⁹ was examined by means of mixed-model regression analysis, predicting our canonical liability variable with D₂ BP in the caudate, while covarying for age, and correcting for dependency (ie, correlation) among the healthy co-twins, as in the linear mixed model.

Intraclass correlation coefficients (ICCs) and their confidence intervals for healthy control MZ twin pairs were calculated for the BP in each of the regions studied. The ICC calculation for healthy control DZ twin pairs was not attempted because of the small number of DZ twin pairs (n = 2). Differences in ICCs between brain regions were compared by F tests.

Differences between groups in subject demographic characteristics were tested by analysis of variance for continuous data and the χ² test for nominal data with a significance P value of .05. For handedness, alcoholism, and other Axis I disorders, analyses were performed while accounting for correlated measurements.

Independent confirmatory voxel-based statistical image analyses were performed to confirm the results achieved with the conventional ROI analysis and to enable detailed visualization of the results. The analysis procedure was described earlier in detail.^15,16 Briefly, parametric images were calculated by receptor parametric mapping based on the simplified reference tissue model.¹⁷ Preprocessing and statistical analysis of parametric images were performed with SPM99¹¹ and Matlab 6.5 for Windows (MathWorks, Natick, Mass). The voxels in which the BP of [¹¹C]raclopride differed between groups were determined using subtraction analysis with T contrast. The correlation between the BP of [¹¹C]raclopride and canonical liability variable was analyzed with a multiple regression model where the effect of age was modeled as a nuisance covariate. Both analyses were performed with volume of interest analysis confined to the same regions as for conventional ROI analysis. A P value of .05 corrected for multiple comparisons was considered the criterion of significance in the voxel-based analyses.

The demographic characteristics of the subjects are presented in Table 1. There were no statistically significant differences between groups on any variable.

The [¹¹C]raclopride BP values for each study group are given in Table 2. We found a significant group effect (F_2,43 = 4.72, P = .01) and a significant age effect (F_1,43 = 5.83, P = .02) for [¹¹C]raclopride BP in the caudate. No hemispheric lateralization of this effect was observed. When t contrasts were analyzed, MZ co-twins exhibited significantly higher BP values than controls (8.4%, t₄₃ = 2.96, P = .005) (Figure 1). The independent voxel-based receptor parametric mapping analysis confirmed this finding and showed that the most significant difference was located in the left caudate nucleus (Figure 2). The MZ co-twins had also significantly higher BP values than DZ co-twins (8.2%; t₄₃ = 2.39, P = .02) (Figure 1). We failed to detect differences between DZ co-twins and controls, a finding that would fit in the model of greater degree of deficit with increasing genetic vulnerability. No significant group effects were detected in putamen or thalamus. Caudate D₂ receptor BP was negatively correlated with a canonical genetically related liability variable in the whole study sample (N = 25) (F_1,21 = 6.22, r = −0.34, P = .02), but not in groups studied separately. The association is illustrated in Figure 3, which plots D₂ BP and the model-predicted canonical variability scores (adjusted for twinship and age). High caudate D₂ BP predicted lower scores (ie, poor performance) on these neurocognitive tasks. The receptor parametric mapping analysis also confirmed this finding, showing negative correlations bilaterally between D₂ BP and cognitive test scores in caudate nucleus (Figure 4).

The ICCs for healthy MZ twin pairs were 0.60 for caudate (95% confidence interval, −0.29 to 0.95), 0.37 for putamen (95% confidence interval, −0.54 to 0.91), and 0.23 for thalamus (95% confidence interval, −0.65 to 0.87). No significant differences in ICC values across brain regions were found.

The results of the present study using PET and [¹¹C]raclopride in MZ and DZ twin pairs discordant for schizophrenia indicate that caudate DA D₂ receptor up-regulation is related to genetic risk for schizophrenia. Correlation analysis showed that caudate D₂ up-regulation was associated with poor performance in cognitive tasks previously shown to be related to psychosis vulnerability. To our knowledge, this is the first evidence of hyperactive subcortical DA system in unaffected co-twins at genetic risk for schizophrenia.

We interpret this finding of increased [¹¹C]raclopride BP as an increase in postsynaptic DA D₂ receptor density. However, some limitations of our study must be considered. First, the signal of caudate [¹¹C]raclopride binding may have a component of binding to D₃ receptors, since [¹¹C]raclopride binds to both D₂ and D₃ receptors with almost similar affinity.^18,19 Dopamine D₃ receptors are located in the human brain mainly in the nucleus accumbens, ventral parts of putamen and caudate, and the islands of Calleja.^20,21 The D₃ receptor density in the ventral striatum has been reported to be elevated in drug-free patients with schizophrenia,²² although the D₃ receptor messenger RNA levels were found to be normal in another study.²³ The caudate ROI in the present study consists mainly of dorsal caudate,^24,25 where D₃ receptor binding constitutes about 10% of the D₂-like (D₂ and D₃) receptor binding.²¹ Therefore, and in the absence of direct evidence of D₃ receptor density abnormalities in relatives of patients with schizophrenia, increased density of caudate D₃ receptors in unaffected co-twins is not likely to explain our findings. Second, partial volume effects in small structures like the head of caudate may affect the observed ROI-based BP estimates due to the limited spatial resolution of the PET scanner. Although caudate volume abnormalities have been suggested in schizophrenia²⁶ and schizotypal personality disorder,^27,28 no such changes have been found in healthy siblings^29,30 or unaffected MZ co-twins³¹ of patients with schizophrenia. The volumes of the manually delineated caudate ROIs did not differ between the study groups. Moreover, we were able to confirm our finding with a voxel-based receptor mapping analysis, an independent and objective analysis method free of operator-derived error in defining brain regions. Thus, it seems unlikely that the present findings are explained by group differences in the caudate volume. Third, factors other than receptor density may affect the BP of [¹¹C]raclopride, since it represents the product of receptor density and affinity (for discussion, see Slifstein and Laruelle¹²). In this setting, it is not possible to exclude group differences in levels of baseline caudate DA concentration. Interestingly, patients with schizophrenia have been shown to exhibit increased baseline striatal D₂ receptor occupancy by endogenous DA.⁵ However, it is not known whether this result is related to vulnerability or to the outbreak of schizophrenia. In healthy volunteers, most of the interindividual variance in BP values is more reflective of variations in receptor density than affinity.^32,33 As alterations in endogenous DA levels would lead to a different interpretation of the data presented herein, future studies should consider using a DA depletion paradigm in subjects with high genetic risk for schizophrenia. Also, other phenomena related to receptor-ligand interaction, such as receptor internalization and synaptic vs extrasynaptic D₂ receptors,³⁴ can also be involved in these findings. Cigarette smoking and alcohol use are not likely confounders of our results, since there were no differences across groups in smoking habits or major alcohol use (Table 1). Fourth, the small sample size in the present study limits the interpretation of findings, and the present results must be considered preliminary in that sense. In addition, we failed to demonstrate differences in D₂ BP between DZ co-twins and healthy controls, a finding that would fit in the model of greater degree of abnormality with increasing genetic loading for schizophrenia. In contrast, such a linear scaling has been observed for neuropsychological dysfunction.⁹ The number of DZ co-twins in the present study may have been too small to disclose subtle differences in D₂ BP, as, according to the model described above, the DZ co-twins should exhibit endophenotypic abnormalities intermediate in magnitude between those of MZ co-twins and healthy controls (ie, D₂ BP increase of 0%-9% in the present study). It is also possible that there is a threshold for this endophenotypic phenomenon, but the sample size is too small to make definitive inferences. Future studies with larger samples are warranted.

Since the formation of the DA hypothesis of schizophrenia, extensive research efforts have been made to evaluate the role of striatal D₂ receptors in the pathophysiology of schizophrenia.^35,36 Results from in vivo imaging studies using PET and single-photon emission computed tomography have yielded conflicting results. Initial findings of increased striatal D₂ receptor density in neuroleptic drug-naive or drug-free patients with schizophrenia were achieved with the use of butyrophenone-class radioligands, such as N-methylspiperone labeled with carbon 11 or bromospiperone labeled with bromine 77.^37-39 These findings have not generally been replicated with benzamide-class radioligands, such as [¹¹C]raclopride or iodobenzamide labeled with iodine 123.^40-42 Many factors, such as differences in susceptibility to fluctuations in intrasynaptic DA levels, differential binding to D₄ receptors, and differential binding to monomer vs dimer forms of the D₂ receptor, have been proposed to account for this discrepancy.^4,36,43 A recent meta-analysis of all published in vivo imaging studies evaluating striatal D₂ receptor density in patients with schizophrenia showed that there may be a small but significant increase of 12% in striatal D₂ receptor density (effect size of 0.51).⁴ Moreover, effect size for studies performed with butyrophenones was significantly higher than that for studies performed with other ligands.⁴ The meta-analysis demonstrates that so far all studies have been underpowered to detect such a small increase in D₂ receptor density in a clinically and etiologically heterogeneous population of patients with schizophrenia. A recent study suggests that D₂ receptor binding is increased in patients with poor prognosis, whereas patients with good prognosis show no alterations in D₂ receptor binding compared with healthy volunteers,⁴⁴ pointing out the need for more homogeneous groups of patients for future studies. We were able to show a significant elevation in caudate D₂ receptor binding with [¹¹C]raclopride, a benzamide radioligand, in this small sample of unaffected MZ co-twins of patients with schizophrenia, a phenomenon that has not been shown previously in small samples of patients with schizophrenia using benzamides. The increase in unaffected MZ co-twins of patients with schizophrenia in the present study is in the same range as the increase in patients with schizophrenia according to the meta-analysis.⁴ It may be that factors associated with overt psychosis, such as phasic DA bursts, could affect [¹¹C]raclopride binding and partly mask a more marked D₂ receptor up-regulation in patients with schizophrenia. It has been suggested that increased endogenous DA levels in actively psychotic patients with schizophrenia, a phenomenon recently noticed in vivo with PET,⁵ would decrease [¹¹C]raclopride binding and underestimate the proposed D₂ receptor up-regulation.^45,46 Also, studies in nonhuman primates and in humans have demonstrated striatal D₂ receptor down-regulation after long-term exposure to DA-releasing agents, such as cocaine and amphetamine.^47-50 These aspects may partly explain why we observed a D₂ receptor increase in such a small sample with the use of [¹¹C]raclopride and PET in nonpsychotic subjects with genetic predisposition for schizophrenia. Our findings suggest that the D₂ up-regulation includes a genetic liability component for schizophrenia rather than being purely a psychosis-related pathophysiologic mechanism.

Increased caudate D₂ receptor binding could reflect a hyperactive subcortical DA system, a pathophysiologic mechanism strongly implicated in schizophrenia.³ A hyperactive subcortical DA system is considered to result from a failure of prefrontal cortical glutamatergic efferents to regulate midbrain DA neurons^8,51 and to be involved in the disruption of the prefrontal-striatal-pallidal-thalamic-prefrontal loop crucial for controlling behavior in response to environmental stimuli.⁸

Major efferent connections from the prefrontal cortex project especially to the head of caudate nucleus,⁵² which has been proposed to represent associative striatum as opposed to striatal regions having primarily motor or limbic projections. More specifically, caudate is a part of the associative loop receiving glutamatergic innervations from the dorsolateral prefrontal cortex, whereas putamen is more involved in the sensorimotor neural loop receiving projections mainly from primary motor and premotor cortex as well as the supplementary motor area.²⁴ There is evidence that responsivity of the DAergic system is lower in associative striatum than sensorimotor or limbic striatum in healthy volunteers, perhaps because of asymmetric negative feedback circuits between subdivisions of striatum and the midbrain.²⁵ It may be speculated that in schizophrenia, this intricate pattern of nonreciprocal feedback loops is disrupted, leading to overstimulation of DA receptors in the associative striatum. Our findings are also compatible with the concept of tonic and phasic DA release⁵¹ and the reduction in tonic DA release in schizophrenia, which, in turn, leads to disinhibition of the phasic DA release and subsequent functional deterioration of the neural circuitry outlined previously. The decreased tonic DA tone could up-regulate postsynaptic D₂ receptors as a compensatory process, and thereby the system might become sensitized to phasic DA bursts. In light of the present results, the dysregulated subcortical DA system may represent a genetically inherited risk factor for schizophrenia and, in interaction with environmental and psychological stressors, could be activated and bring about overt symptoms of psychosis. The increase in D₂ receptor density seen in this study as a genetically determined phenomenon might be secondary to prefrontal DAergic hypofunction, which has, in turn, been shown to be a heritable risk factor for schizophrenia in terms of a polymorphism of DA catabolizing enzyme.⁵³ The high ICC value (0.60) for caudate BP in healthy MZ control twins in the present sample is in line with the hypothesis of marked genetic influences on caudate D₂ receptor binding.

Possible associations between the known functional variants of the D₂ receptor gene and schizophrenia have been intensively studied during the past decade, but the evidence has been inconclusive.^54,55 In addition, there are no consistent effects of these gene variants on D₂ receptor density. Thus, it seems unlikely that the up-regulation of caudate D₂ receptors as a heritable risk factor for schizophrenia would be mediated by direct variations in the D₂ receptor gene. Instead, we propose that the biological mediator of this genetic predisposition lies upstream along the prefrontal-striatal-pallidal-thalamic-prefrontal loop, with most evidence suggesting the prefrontal cortex. In favor of this hypothesis, earlier data suggest an association of markers of neuronal pathology specifically in the dorsolateral prefrontal cortex with increased amphetamine-induced striatal DA release in schizophrenic patients.⁵⁶ This adds to the proposal that hyperactive subcortical DA system is secondary to dysfunction of dorsolateral prefrontal cortex in schizophrenia. Clearly, other cortical areas, such as the temporal cortex, have influence on the prefrontal-striatal pathway. Recently, it has been shown in primates that neonatal medial temporal lesions lead to exaggerated striatal DA release,⁵⁷ implying that neurodevelopmental disruption of corticocortical connectivity affects the function of the subcortical DA system.

Dissection of an illness into more discretely determined subcomponents has become a recommended strategy in diseases with complex genetics such as schizophrenia. We have previously demonstrated a heritable influence on cortical gray matter reduction in schizophrenia, with unaffected co-twins showing reductions in polar and dorsolateral prefrontal cortex to a degree commensurate with their genetic proximity to an affected proband.⁵⁸ We also found evidence that deficits on neurocognitive tests sensitive to the functioning of the frontal lobe, including spatial working memory, divided attention, executive aspects of verbal memory retrieval, and choice reaction time, varied in a dose-dependent fashion with increasing genetic risk among the MZ and DZ co-twins of schizophrenic patients.⁹ Animal models have shown that spatial working memory in particular is critically dependent on the cortical DA system as mediated by the D₁ receptor,⁵⁹ and that there is a reciprocal relationship between activity in the mesolimbic and mesocortical DA systems, suggesting that increased subcortical DA activity may be accompanied by reduced cortical (prefrontal) DA activity.^51,60 Recently, a tight relationship between reduced activity in prefrontal cortex during a working memory task and increased subcortical DA function has been presented in schizophrenic patients.⁶¹ In our study, a significant negative correlation was found between caudate D₂ receptor density and the composite measure of liability-related cognitive functioning reflecting performance on a set of frontally mediated tests, suggesting a common inherited substrate extending across the structural-anatomic, cellular-receptor, and behavioral levels of analysis.

We have demonstrated an increased caudate D₂ receptor density in unaffected co-twins of schizophrenic patients. There may be a heritable preexisting change in neural DAergic machinery increasing susceptibility toward schizophrenia. External or internal stressors may further activate the DA system and lead to overt expression of psychosis. These findings encourage further search for neurochemical endophenotypes as a possible future tool for early detection of individuals at risk for schizophrenia⁶² and also lend theoretical support to early pharmacologic intervention approaches using antipsychotic drug treatment of the DA D₂ blocker class.⁶³ The present results should be independently replicated in a larger sample to establish the role of D₂ receptors in the genetic background of schizophrenia.

Correspondence: Jarmo Hietala, MD, PhD, Department of Psychiatry, Central Hospital, University of Turku, Kiinamyllynkatu, Turku 20520, Finland (jahi@utu.fi).

Submitted for Publication: March 26, 2004; final revision received August 27, 2004; accepted September 9, 2004.

Funding/Support: This study was supported by grant MH52857-05 from the National Institute of Mental Health, Bethesda, Md; grant 13649 from the Turku University Central Hospital; and research grants from Oy H. Lundbeck Ab, Turku; Lilly Foundation, Vantaa, Finland; Research Foundation of Orion Corporation, Espoo, Finland; the Instrumentarium Foundation, Helsinki; The Finnish Medical Foundation, Helsinki; and the Varsinais-Suomi Regional Fund of the Finnish Cultural Foundation, Turku (Dr Hirvonen).

Previous Presentation: This study was presented in part at the 45th Annual and Fourth Mediterranean Meeting of the Scandinavian College of Neuro-Psychopharmacology; April 23, 2004; Juan-les-Pins, France.

Acknowledgment: The assistance of Ulla Mustonen (National Public Health Institute of Finland, Helsinki), the staff in the Turku PET Laboratory, and the magnetic resonance imaging unit of Helsinki University Hospital is greatly appreciated.