Main

Glutamate is the primary excitatory neurotransmitter in the brain, and there is an increasing amount of evidence for the role of glutamine receptors in neuropsychiatric disease (e.g., Javitt and Zukin 1991; Coyle 1987; Choi 1988; Ellison 1995). To further elucidate the role of the glutamate receptor system in behavior and neurochemical function, particularly in human subjects, the development of a pharmacologic probe is extremely important. The importance of a pharmacologic probe for glutamatergic function is underscored by the lack of a suitable glutamate receptor radiotracer for in vivo imaging (e.g., Blin et al. 1991). This would permit the development of methods to measure glutamate modulation of other neurotransmitter systems that may provide greater insight into the role of glutamate in normal function and in disease.

The “dissociative anesthetics” such as phencyclidine (PCP) and its derivative ketamine have been administered as a pharmacologic challenge to evaluate the role of glutamate receptor antagonism. These agents bind to a site within the N-methyl-D-aspartate (NMDA) receptor complex, a glutamate receptor subtype (Corssen and Domino 1966; Zukin and Zukin 1976; Vincent et al. 1979). At anesthetic doses (1.0 mg/kg and higher), ketamine has multiple neurochemical effects including the inhibition of monoamine transport and acetylcholinesterase and the blockade of sigma receptors (e.g., Smith et al. 1981; Oye et al. 1991; Cohen et al. 1974). The administration of subanesthetic doses of ketamine (0.1–0.5 mg/kg) has demonstrated a relatively selective effect in binding to the NMDA site (Javitt and Zukin 1991; Thomson et al. 1985), as there is a several-fold difference in the affinities of monoamine transporter and sigma binding sites compared with the NMDA binding site (e.g., Smith et al. 1981; Oye et al. 1991).

Administration of anesthetic doses of PCP to patients and of subanesthetic doses of PCP to normal volunteers altered in behavior in a manner that has been compared with psychotic symptoms (e.g.,Javitt and Zukin 1991; Krystal et al. 1994). Both PCP and ketamine administration exacerbated symptomatology in schizophrenic patients (e.g., Lahti et al. 1995a). Given the hypothesis that hyperdopaminergic function underlies psychotic symptoms, the psychotic-like behaviors observed after PCP and ketamine administration may be attributable to secondary effects on the dopamine system. This is consistent with the preclinical, neurochemical observations that NMDA receptor antagonists such as ketamine alter the firing rate of mesocortical and mesolimbic dopamine neurons and increase concentrations of extracellular dopamine in the striatum and prefrontal cortex (e.b., French 1994; McGeer et al. 1977; Moghaddam et al 1990; Verma and Moghaddam 1996).

Recently, dopamine concentrations and dopamine modulation has been measured in vivo using positron emission tomography (PET) and single photon emission computed tomography (SPECT) imaging modalities, radiotracers for the dopamine (D2) receptor and pharmacologic challenges that alter dopamine concentrations directly or indirectly through other neurotransmitter systems. Studies in non-human primates and human subjects have demonstrated that a pharmacologic increase in dopamine concentrations by administration of dopamine agonists and reuptake inhibitors such as d-amphetamine, GBR 12909, and methylphenidate decreased the striatal binding of D2 radiotracers (18F-N-methylspiroperidol. 123I-IBZM, 11C-raclopride), due to the increased competition between dopamine and for binding to the striatal D2 receptor (Dewey et al. 1991, 1993a; Innis et al. 1992; Volkow et al, 1994). Subsequent studies demonstrated that pharmacologic alterations of other systems known to modulate striatal dopamine (e.g., GABA, acetylcholine, serotonin) altered binding consistent with the preclinical data demonstrating that dopamine concentrations is modulated by these neurotransmitter systems (e.g., Dewey et al. 1992, 1993a, b, 1995; Smith et al. 1997). It is important to note these studies measured alterations in striatal D2 binding as the low densities of D2 receptors in extrastriatal regions cannot be detected with PET and the D2 radiotracers used in these studies. Given this limitation, it is of interest to assess cortical input to the striatum within the context of this PET experimental design. As the primary cortical input to the striatum is glutamatergic, this study was designed to develop a method to assess in vivo the functional integrity of cortical input to the striatum. The present study was undertaken to evaluate whether administration of subanesthetic doses of ketamine would decrease striatal 11C-raclopride binding in normal control subjects, consistent with an increase in dopamine concentrations.

MATERIALS AND METHODS

Subject Screening and Selection

Subjects underwent medical (physical and laboratory testing, toxicology screening, physical examination) and psychiatric evaluation and a magnetic resonance imaging (MRI) scan (Siemens 1.5T Magnatom Vision). Subjects were excluded based upon a history of or current significant medical, psychiatric, or neurologic disease, substance abuse, family history (first- or second-degree relatives) of psychiatric illness or substance abuse or use of prescription or over-the-counter medications with central nervous system (CNS) effects (e.g., antihistamines, cold medications) within the past month. Seven, right-handed male subjects were enrolled in the study, with a mean age of 31.0 ± 5.3 years (range 26–39 years). Subjects fasted after midnight and were given a standard breakfast 3 h before scanning. After a complete description of the study to the subjects, written informed consent was obtained. The experimental protocol and consent forms were approved by the Institutional Board of Research Associates at the New York University School of Medicine and the Human Studies Research Committee at Brookhaven National Laboratories.

PET Procedures

For each subject, two 11C-raclopride studies were performed on the same day. After performing a baseline 11C-raclopride scan, ketamine was administered (0.5 mg/kg, IV, infused over 20 min). Two pretreatment intervals were evaluated to determine when the maximal effects of ketamine on 11C-raclopride binding would be observed. In four subjects, the second 11C-raclopride scan was performed immediately after the ketamine infusion ended. In three subjects (subjects 3,4,5), 11C-raclopride was injected 10 min into the ketamine infusion, and the ketamine infusion continued for the initial 10 min of the 11C-raclopride study. As ketamine has a relatively short half-life in plasma, injecting the 11C-raclopride during the infusion could prolong potentially the pharmacologic effect. In the case of both time points, the timing of the second scan corresponded to the interval during which the maximal plasma concentration and the maximal behavioral effects of ketamine have been observed. The ketamine dose and administration method was consistent with that of the study reported by Krystal et al. (1994).

The 11C-raclopride studies were performed on the Computer Technology Inc. (CTI) 931-08/12 tomograph. Prior to scanning, catheters were placed in an antecubital vein for radiotracer injection and in a radial artery for blood sampling. The subject was positioned in the scanner using external lasers wet parallel to the cantho-meatal line and to the mid-line as marked on the head. Repositioning accuracy was insured by the use of the headholder and the positioning pins that secure the headholder to the cradle and the laser markers. The sensitivity of the machine was tested with a germanium phantom. After a blank scan, a 10-min transmission scan was obtained with a 68G source, to be used for attenuation correction of the emission scan.

11C-raclopride was synthesized by the reaction of carbon-11–labeled methyl iodide with nor-raclopride (Farde et al. 1986). 11C-raclopride was injected intravenously (5.09–13.41 mCi, Specific Activity 0.5–1.5 Ci/μmol). Scanning began immediately and lasted for 60 min. The scanning protocol involved 10 scans of 1 min, followed by 10 scans of 5 min. Continuous blood samples were obtained by an automated device (Ole Dich, Denmark) for the first 2 min after injection, at the peak of radioactivity in blood. For the remainder of the study, manual sampling was used. Selected plasma samples (1, 5, 10, 20, 30, 45, and 60 min) were analyzed for the presence of unchanged 11C-raclopride, as described previously (Volkow et al. 1993). Subject were scanned in a quiet, dimly lit room, with eyes open and ears unoccluded. The subjects were instructed that they would receive either ketamine or placebo before each study.

Plasma samples were obtained for analyses of the dopamine metabolite homovanillic acid (HVA), prolactin, and cortisol. The analytical methods and their coefficients of variation have been reported previously (Hollander et al. 1992). Plasma samples were obtained before and immediately after each of the two 11C-raclopride scans. The initial plasma sample (before the first 11C-raclopride study) was obtained approximately 30 min after placement of the arterial and venous catheters.

Data Analysis

Regions of interest (ROIs) were identified for the striatum (specific binding) and cerebellum (nonspecific binding). The ROIs for the striatum were drawn on four slices beginning at the head of the caudate and encompassing the caudate and putamen on the other three slices. Receptor availability was measured by a graphical analysis method designed for reversible systems (Logan et al. 1990). The date will be reported in terms of the distribution volume (DV), which is a linear function of the free receptor concentration. This method has been described in detail in earlier publications (e.g., Logan et al. 1990, 1995). The DV parameter is less sensitive to noise than the individual kinetic parameters, which often have large standard errors associated with their determination. In addition, the DV as derived is flow independent, since the flow terms appear in the numerator and denominator and thus, cancel out. The insensitivity of the DV analysis to changes in cerebral blood flow has been addressed experimentally in the analysis of two reversible tracers for two different neurotransmitter systems (11C-flumazenil, 11C-raclopride; Logan et al. 1995; Holthoff et al. 1991). To take into account changes in nonspecific binding, the data will be presented as the cerebellar DV (average of the two studies for each subject) subtracted from the striatal DV. As the magnitude of change of the cerebellar DV for each subject is within the range of the test-retest variability for the cerebellar DV (Volkow et al. 1993), the two cerebellar DV values were averaged to reduce the potential noise contributed by the cerebellar DV. In subtracting the cerebellum from the striatum, the striatal data is weighted more and the differences in receptor availability observed are due to a greater extent to changes in striatal binding. The use of the subtraction approach is supported by the observation that the striatum and cerebellum have comparable amounts of nonspecific binding (Farde et al. 1989) and that the administration of ketamine did not alter the cerebellar DV in a systematic manner (either increase or decrease) in excess of the test-retest variability for the cerebellum (Volkow et al. 1993). In considering the effect of ketamine on rCBF that might contribute to the changes in nonspecific binding, ketamine administration has been shown to alter rCBF primarily in limbic areas, but not in the striatum and cerebellum, the regions of interest for the 11C-raclopride studies (Lahti et al. 1995b). Thus, regionally selective changes in rCBF in the striatum and cerebellum are not likely to represent a major confounding factor.

Statistical Analysis

Repeated measures analysis of variance (ANOVA) was used to investigate the effects of ketamine on plasma levels of HVA, cortisol, prolactin, the rate of metabolism of 11C-raclopride and DV measurements for 11C-raclopride. If a significant repeated measures effect was obtained, post-hoc ANOVAs were performed to compare the individual time points.

RESULTS

As the subjects were being scanned during the maximal behavioral effects of the drug, they were questioned about their response to the drug after scanning. The subjects consistently reported feelings of anxiety, a sensation that they were floating and a general feeling of derealization that could not be easily described. All subjects reported that the maximal effect of the drug occurred within the first half hour of scanning.

The levels of cortisol, prolactin, and HVA in plasma are shown in Table 1. For the HVA and prolactin measures, the main effects of condition (placebo/drug) and time (pre/post-scan) and the condition by time interaction were not-significant (HVA: f[1,6] = 1.70, p > .1, f[1,6] = 0.19, p < .1 and f[1,6] = 0.00, p < .1, respectively: prolactin: (f[1,6] = 5.47, p > .05, f[1,6] = 3.35, p > .1 and f[1,6] = 1.95, p > .1, respectively). For cortisol, there was a significant effect of drug, but a nonsignicant effect of time and a drug by time interaction (f[1,6] = 8.96, p < .05, f[1,6] = 5.61, p > .05 and f[1,6] = 0.00, p < .1, respectively). To follow-up the significant effect of condition for cortisol, the sum of the baseline measures was compared with the two post-ketamine measures and a significant effect was obtained for the pre-scan 2 measure, but not for the post-scan 2 measure (f[1,6] = 12.43, p < .05, f[1,6] = 4.25, p < .1). Table 2 shows the percent of unchanged 11C-raclopride in arterial plasma over time for the two scans. Ketamine administration did not significantly alter the rate of metabolism of 11C-raclopride as the repeated measures ANOVA demonstrated a significant main effect of time, as expected, but not a significant effect of condition or a time by condition interaction (f[5,25] = 39.16, p < .001, f[1,6] = 0.29, p < .1 and f[1,5] = 0.24, p < .1, respectively).

Table 1 Plasma Homovanillic Acid (HVA) Levels (ng/mL), Prolactin Levels (ng/mL), Cortisol Levels (μg/dL)
Full size table
Table 2 Percentage of Total Carbon-11 as 11C-raclopride in Plasma
Full size table

The DVs were decreased significantly in the striatum, but not in the cerebellum (f[1,6] = 11.86, p < .05, f[1,6] = 0.03, p > .1, respectively). A significant decrease in the subtracted DVs was observed. The cerebellar DVs subtracted from the striatal DVs (striatum − cerebellum) for the baseline and ketamine pretreatment condition were decreased significantly by ketamine administration (f[1,6] = 11.86, p < .05). The values for the subtracted DVs for the individual subjects are shown in Table 3 and Figure 1 . For the subtracted DVs, the magnitude of decrease varied across subjects. A separate set of analyses were performed to determine whether the magnitude of response differed as a function of timing of ketamine infusion (the subjects who received the second 11C-raclopride injection at the end of the ketamine infusion versus those who received the second 11C-raclopride injection in the middle of the ketamine infusion). No significant group differences were observed in the change in 11C-raclopride binding, cortisol, prolactin, or HVA. Finally, significant correlations were not obtained between the change in 11C-raclopride binding and the change in cortisol, prolactin, or HVA.

Table 3 Effects of Ketamine on the DV of the Striatum-Cerebellum for 11C-raclopride Binding
Full size table
Figure 1
figure 1

The effect of ketamine administration on 11C-raclopride binding. Ketamine administration resulted in a decrease in 11C-raclopride binding consistent with an increase in dopamine activity.

Full size image

DISCUSSION

In the present study, ketamine administration decreased 11C-raclopride binding, consistent with an increase in dopamine concentrations. An alternative explanation is that the results obtained could be due to a shift of D2 receptors from a low to a high affinity state for dopamine. In addition, either a presynaptic mechanism (for example, increased dopamine release or decreased reuptake) or a postsynaptic mechanism (for example, increased affinity of receptors for synaptic dopamine) could account for the present findings.

For the subjects as a group, the change in 11C-raclopride binding was greater in magnitude than the test-retest variability of 11C-raclopride binding in studies performed on 2 consecutive days at the same time of day (Volkow et al. 1995). Whereas all of the subjects showed a decrease in binding, the magnitude of decrease varied across subjects. The variability was not accounted for by differences between subjects in age, baseline 11C-raclopride binding, behavioral response to ketamine infusion, baseline levels, or the magnitude of change in levels of cortisol, prolactin, or HVA. The differential timing of the 11C-raclopride injection relative to the ketamine infusion did not result in a systematic difference in response within the two groups of subjects. As the maximal effects of ketamine on behavior occurred during the PET scan, it was possible only to interview the subject after the PET scan. The behavioral effects of ketamine in these subjects were consistent with the effects reported in other studies designed specifically to measure behavioral effects. Whether the variability observed is inherent to the subjects in this study remains to be evaluated in a larger sample of subjects.

An increase in cortisol was observed after ketamine administration. Cortisol levels are also increased by pharmacologic challenges that increase dopamine concentrations directly and indirectly, such as d-amphetamine, methylphenidate, and fenfluramine (e.g., Smith et al. 1994, 1997; McBride et al. 1990). However, the neuroendocrine effects in the present study could also occur by a direct blockade of hypothalamic NMDA receptors (Gribkoff 1991). The lack of correlation between the change in 11C-raclopride binding and the change in the neuroendocrine measures was also observed in a previous 11C-raclopride study using fenfluramine challenge to measure serotonin modulation of dopamine concentrations (Smith et al. 1997). The lack of correlation may be due to the differential time course of the effects of ketamine on neuroendocrine function and dopamine concentrations.

The results of the present study are consistent with the findings of in vivo microdialysis studies that ketamine administration increased extracellular dopamine concentrations in the prefrontal cortex and the striatum, to a greater extent in prefrontal cortex (e.g., Verma and Moghaddam 1996). With regard to the mechanism underlying this effect, there is evidence that glutamate modulation of striatal dopamine occurs both at the level of the dopaminergic cell bodies of the ventral tegmental area (VTA) and substantia nigra (SN) and the level of the glutamatergic afferents of the prefrontal cortex that modulate concentrations in the VTA, SN, or in the striatum by providing excitatory input to the GABA interneurons (Monaghan and Cotman 1985; Carter 1982; Sesack and Pickel 1992). In evaluating these different mechanisms, a recent in vivo microdialysis study demonstrated that glutamatergic projections from the prefrontal cortex to the VTA were critical in controlling striatal dopamine release (Karreman and Moghaddam 1996). These microdialysis results, when considered with the effects of PCP or ketamine administration on cerebral metabolism and blood flow described below, support the hypothesis that the alterations in striatal dopamine concentrations observed in the present study may be a secondary consequence of altered cortical function. Radiotracers are currently in development to measure extrastriatal D1 and D2 receptors that may be sensitive to alterations in dopamine concentrations (e.g., 18F-desmethoxyfallypride, 11C-NNC 01-0112; Mukherjee et al. 1995; Halldin et al. 1994). When a suitable radiotracer is developed, it will be possible to measure glutamate modulation of dopamine in cortical and limbic areas in addition to the striatum.

The effects of acute ketamine or PCP administration differ from that of amphetamine administration (as reviewed by Javitt and Zukin 1991; Ellison 1995; Krystal et al. 1994); however, both drugs mimic different aspects of symptomatology in schizophrenia (negative versus positive symptoms). Both classes of drugs and have different regional effects on cerebral metabolism that resemble the metabolic alterations observed in schizophrenia. PCP administration to rats and ketamine administration to schizophrenic patients increased metabolism in sensory cortices, motor cortex, thalamus, and limbic areas (e.g., Gao et al. 1993; Weissman et al. 1987; Lahti et al. 1995b). D-amphetamine administration to schizophrenic patients decreased metabolism in frontal and temporal cortical areas (e.g., Wolkin et al. 1987). It is possible that deficits in the dopamine and glutamate systems may relate to different aspects of symptomatology and metabolic alterations in schizophrenia. The variability in symptomatology observed between patients may be related to different degrees of glutamate-dopamine imbalance.

In contrast, there are similarities between chronic use of PCP and ketamine and the stimulant agents (e.g., cocaine, amphetamine), and both classes of drugs are known for their abuse liability (for a review, see Siegel 1978). There are similarities in the pattern of abuse of both classes of agents in that “binge” intake patterns are observed, followed by a rebound depressive phase and in some individuals, persistent psychotic symptoms (Burns and Lerner 1981; Ellison 1995). As the release of dopamine is the primary neurochemical effect of cocaine and amphetamine and a secondary effect of PCP and ketamine, the addictive properties and associated symptoms of chronic use of these agents may be related to a common dopaminergic mechanism (e.g., Nomikos et al. 1990).

In conclusion, the administration of a subanesthetic dose of ketamine decreased 11C-raclopride binding, consistent with an increase in dopamine concentrations. These findings are consistent with the ability of NMDA receptor blockade to alter dopamine concentrations demonstrated by neurophysiologic and neurochemical studies. This experimental paradigm for measuring glutamate modulation of dopamine concentrations in vivo may be used to elucidate etiologic and therapeutic mechanisms in the psychotic and addictive disorders.