|Home | About | Journals | Submit | Contact Us | Français|
Methylphenidate (MP) is widely used to treat Attention-Deficit Hyperactivity Disorder (ADHD). Variable number of tandem repeats (VNTR) polymorphisms in the dopamine D4 receptor (D4) gene have been implicated in vulnerability to ADHD and on the response MP. Here we examined the contribution of D4 receptors (D4Rs) to baseline brain glucose metabolism (BGluM) and to the regional metabolic responses to MP. We compared BGluM (measured with microPET and [18F]2-fluoro-2-deoxy-D-glucose) at baseline and after MP (10 mg/kg ip) administration in mice with genetic deletion of the D4. Images were analyzed using a novel automated image registration procedure. Baseline D4−/− mice had lower metabolism in prefrontal cortex (PFC) and greater metabolism in the cerebellar vermis (CBV) than D4+/+ and D4+/− mice; when given MP, D4−/− mice increased metabolism in PFC and decreased it in CBV; whereas in D4+/+ and D4+/− mice MP decreased metabolism in PFC and increased it in CBV. These findings provide evidence that D4Rs modulate not only the PFC, which may reflect activation by dopamine of D4Rs located in this region, but also the CBV, which may reflect an indirect modulation since D4Rs are minimally expressed in this region. Since individuals with ADHD show structural and/or functional abnormalities in these brain regions the association of ADHD with D4Rs may reflect its modulation of these brain regions. The differential response to MP as a function of genotype could explain differences in brain functional responses to MP between patients with ADHD and healthy controls and between ADHD patients with different D4 polymorphisms.
Attention Deficit Hyperactivity Disorder (ADHD) is the most frequent psychiatric disorder of childhood and it is increasingly being recognized in adults (Fergason, 2000; Volkow et al., 2001), (Pary et al., 2002). The stimulant methylphenidate (MP) is widely used to treat symptoms of ADHD (inattention, hyperactivity, impulsivity). MP increases extracellular levels of dopamine (DA) and norepinephrine (NE) by blocking their respective transporters (Kuczenski & Segal, 1997). Imaging studies have shown that MP increases cerebellar glucose metabolism in normal individuals (Volkow et al., 1997a), and that it attenuates task-induced increases in brain glucose metabolic activity during exposure to a cognitive task in proportion to the baseline metabolic measures (Volkow et al., 2008). These findings suggest that MP’s effects on brain metabolic activity depend on the context (task performed) as well as the individual (baseline brain metabolism) and that the therapeutic effects of MP may reside in part on its ability to increase the efficiency of the brain when performing a cognitive task (Volkow et al., 2008).
ADHD has been associated with a polymorphism in the dopamine D4 receptor (D4); a 7-repeat variant of the 48 base-pair (bp) VNTR located in exon 3, which results in decreased D4 efficacy (Terwilliger & Ott, 1992; LaHoste et al., 1996; Swanson et al., 1998; Himelstein et al., 2000; Sunohara et al., 2000; Tahir et al., 2000; Cheon et al., 2007). This polymorphism has also been associated with poor response to MP treatment in ADHD (Seeger et al., 2001; Hamarman et al., 2004; Cheon et al., 2007) and with volumetric changes in cerebellum (Monuteaux et al., 2008) and prefrontal cortex (Durston et al., 2005).
In the brain, the D4 is highly expressed in the frontal cortex and in the hypothalamus (Ariano et al., 1997; Tarazi & Baldessarini, 1999; Oak et al., 2000) and with very low expression in cerebellum (CB) (predominantly in white matter) (Barili et al., 2000). Deletion of D4 in knockout (KO) mice results in lower basal extracellular DA levels in the striatum as well as decreased KCl-evoked overflow of DA in the striatum and nucleus accumbens core (Thomas et al., 2007). These mice also show reduced exploration of novel stimuli, decreased spontaneous locomotor activity (Rubinstein et al., 1997; Dulawa et al., 1999) and abnormal behavioral responses (locomotor activity and conditioned place preference) to stimulant drugs (amphetamine, MP and cocaine) (Thanos et al., 2009).
Here we test the hypothesis that dopamine D4 receptors (D4Rs) influence baseline activity in prefrontal cortex and cerebellum, which are brain regions implicated in ADHD, and that D4Rs also modulate the response to stimulant medication in these brain regions. For this purpose we compared regional brain glucose metabolism (marker of brain activity) at baseline and in response to MP in D4 KO (D4−/−) with that in heterozygous (D4+/−) and in wild-type (D4+/+) mice. Specifically we hypothesized that since D4Rs are highly expressed in the prefrontal cortex we hypothesized that D4−/− mice would differ in baseline and in MP-induced changes in prefrontal metabolism. Because MP-induced increases seen in earlier human studies have been postulated to reflect prefrontal regulation of cerebellar activity we also hypothesized differences in baseline and in MP-induced changes in cerebellar metabolism between D4−/− and wild-type mice.
All mice were produced as previously described (Rubinstein et al., 1997). We studied male D4 wild-type (n=8), heterozygous (n=8) and KO mice (n=8) that had been bred at the Brookhaven National Laboratory (BNL) Animal Facility. At the time of the experiments, the mice were approximately 3–4 months old and weighed 36.2g±2.1 (D4+/+), 36.5g±1.5 (D4+/−), and 33.7g±2.4 (D4−/−). They were single-housed under a 12 hour light/dark cycle in clear acrylic cages with wire-mesh tops and food and water was available ad-libitum. All experiments were conducted in conformity with the National Academy of Sciences Guide for the Care and Use of Laboratory Animals (NAS & NRC, 1996) and BNL Institutional Animal Care and Use Committee protocols.
MP hydrochloride (Sigma, St. Louis, MO) was dissolved in saline to produce a concentration of 10 mg/kg. 2-[18F]-flouro-2-deoxy-D-glucose (FDG) was synthesized at the BNL Cyclotron.
All animals were scanned with FDG twice and each scan was performed 1 week apart. Animals were fasted overnight prior to the scan. The Baseline-Scan was conducted first and used as a control scan during which each animal received an intraperitoneal (IP) injection of approximately 0.2 mCi FDG and was immediately placed in its home cage. Each mouse was awake for a period of 40 minutes during the FDG uptake. Mice were then anesthetized with a mixture of ketamine (100 mg/kg) and xylazine (10 mg/kg) and placed in a stereotaxic head holder (David Kopf Instruments; Tujunga, CA, USA) in a prone position on the bed of the scanner 5–7 minutes after the anesthesia administration. The MP-scan was the same as Baseline except that it was administered 1 week later and each animal was given 10 mg/kg (IP) MP 5 min prior to FDG.
An R4 small animal positron emission tomography (μPET) scanner (Concorde CTI Siemens, Knoxville, TN) was used for FDG μPET imaging. The R4 μPET has a transaxial resolution of 2.0 mm full width at half maximum, with a field of view (FOV) of 11.5 cm. Animals were placed in the center of the FOV and were scanned under a static imaging protocol for 80 minutes using a ramp filter with cutoff at Nyquist frequency. After scanning, all images were corrected for photon scatter and reconstructed using the ordered subset expectation maximization algorithm provided by Concorde CTI. Attenuation correction was not carried out due to factors previously described (Alexoff et al., 2003; Schiffer et al., 2007).
MicroPET image analysis was performed as described previously (Pascau et al., 2009) with slight modification especially for mouse brain. To obtain statistical parametric maps from different animals, these studies must be properly co-registered. For this purpose, all the images in the dataset were co-registered to a reference image (manually selected by the user). The registration algorithm makes use of normalized mutual information to find the rigid transformation that co-registers both images, and works in two multiresolution steps: At the first step (lower resolution), the whole reference image is used, whereas at the second step (higher resolution) the reference image is masked in such a way that only those pixels inside the brain are used to compute the cost function (which is used as a quantitative measure of the quality of the alignment). To minimize registration errors, this process was repeated three times, selecting different reference images in every repetition. These three reference images are also registered against each other. Finally, by combining all these geometrical transformations, we can automatically detect any incorrect registration making use of consistency measures. Once the whole dataset is properly registered, an FDG μPET template image was created by averaging all co-registered images. In order to circumvent potential confounds associated with differences in metabolism that could underlie glucose uptake values, we assessed the regional metabolic change of each animal relative to its global activity (whole-brain). This prevented the characterization of metabolic changes being attributed to differences in animal metabolism. It also prevented misleading glucose uptake effects due to injected dose, weight differences between animals and variable absorption from the IP cavity. Images were analyzed using the statistical parametric mapping (SPM2) software package using a method previously described (Thanos et al., 2008b). In order to get a more accurate anatomical representation of potential areas of activation, a high-resolution magnetic resonance imaging (MRI) brain scan of an age and body weight matched C57/BL6 wild-type mouse was acquired on a 21.1 Tesla magnet (50μm isotropic, TR/TE 250/5 ms) and was subsequently used for μPET-MRI coregistration using previously described procedures (Thanos et al., 2008b).
An analysis of variance (ANOVA) model was used that defined six different groups that corresponded to D4+/+, D4+/− and D4−/− mice with and without MP. Images were subtracted after intensity normalization to 100 by the proportional scaling method. After estimation of the statistical model, an F contrast was applied to reveal the effects of interest. These effects were overlaid on the previously generated high-resolution MRI template to get a more accurate representation of the areas of activation as previously described (Thanos et al., 2008a). An uncorrected P-value of 0.001 was used as threshold to determine statistical significance for the model.
The original F contrasts for the comparisons between D4+/+ vs. D4+/−, D4+/+ vs. D4−/− and D4+/− vs. D4−/− did not reveal any significant cluster differences at baseline (α=.05). In contrast the F contrast that assessed the differences between the combined scans of D4+/+ and D4+/− vs. those of D4−/− mice yielded three significant clusters: right prefrontal cortex (PFC) (KE=13; F=24.5; Zscore=4.14; p<0.001), left PFC (KE=6; F=26.8; Zscore=4.3; p<0.001), cerebellar vermis (CBV) (KE=23; F=26.3; Zscore=4.26; p<0.001) (Figure 1 and Table 1).
The SPM analysis yielded three significant interaction effects (Table 1). In the first interaction, and most significant, the D4+/+ and D4+/− mice showed greater relative baseline metabolism in two regions located in the right (KE=13; F=24.5; Zscore=4.14; p<0.001) and left PFC (KE=6; F=26.8; Zscore=4.3; p<0.001), compared to D4−/− mice whereas the D4−/− mice had greater relative baseline metabolism in a region located in the CBV (KE=23; F=26.3; Zscore=4.26; p<0.001) than D4+/− and D4+/+ mice (Figure 2).
MP significantly increased metabolic activity in the CBV in D4+/+ and D4+/− mice but decreased it in D4−/− mice (KE=23; F=26.3; Zscore=4.26; p<0.001) (Table 1; Figure 3) and it decreased metabolism in left (KE=6; F=26.8; Zscore=4.3; p<0.001), and right PFC (KE=13; F=24.5; Zscore=4.14; p<0.001) in D4+/+ and D4+/− mice but increased it in D4−/− mice (Table 1 and Figure 3).
The linear regression analysis performed between MP-induced changes in metabolism in the left and right PFC and those in the CBV showed that this was significant for a negative correlation between the left PFC and CBV in D4−/− mice but not in D4+/+ and D4+/− mice (r=0.88; p=0.007; Figure 4).
In this study, we document significant differences between D4 genotypes in baseline and in MP-induced changes in metabolism in PFC and in CBV. D4 KO mice when studied at baseline showed lower metabolism in PFC regions and greater metabolism in CBV metabolism when compared with D4+/− and D4−/− mice. Also MP in D4 KO mice increased PFC metabolism and decreased metabolism in CBV whereas in D4+/− and D4−/− mice MP elicited the opposite pattern of metabolic changes in these brain regions. These findings support our hypothesis that D4Rs modulate baseline activity in prefrontal regions and in CBV. Since studies have shown both volumetric and functional differences in the PFC and CB of ADHD patients (for review see (Giedd et al., 2001; Brennan & Arnsten, 2008)) and the D4 is associated with ADHD our findings suggest that differences in D4 expression or function may underlie the reported dysfunction of these brain regions in subjects with ADHD. Studies comparing prefrontal and CBV activity between ADHD patients with the D4 VNTR 7 repeat polymorphism and those without it are necessary to test this hypothesis.
In the wildtype and heterozygous mice acute treatment with MP evoked a relative increase in metabolism in CBV, which is consistent with results from imaging studies reporting MP-induced increases in cerebellar metabolism, which is most accentuated in the CBV (Volkow et al., 1997b). Moreover studies measuring cerebral blood flow or BOLD responses with functional MRI (fMRI) have also consistently identified the cerebellum and the prefrontal cortex as targets for MP effects (Giedd et al., 2001; Schweitzer et al., 2003; Volkow et al., 2005).
The lower basal brain glucose metabolism (BGluM) in the PFC (bilaterally) of D4 KO compared to wild-type and heterozygote counterparts suggests that D4 modulate activity in the PFC and thus are likely to influence PFC-related behaviors. Studies in humans have shown that the PFC is involved in cognitive inhibition, impulse control, organizational planning, working memory, sensory gating and attention (Godefroy et al., 1996; Goldman-Rakic, 1996; Itami & Uno, 2002; Aron et al., 2004). Similarly, animal lesion studies have shown that the PFC is associated with attentional impairment in primates and rodents (Bartus & Levere, 1977; Birrell & Brown, 2000). Interestingly, individuals with ADHD show deficits in the above PFC-related behaviors as well as decreased PFC volume (Castellanos et al., 1996; Casey et al., 1997; Bush et al., 1999; Rubia et al., 1999; Sowell et al., 2003) and activity (Zametkin et al., 1990; Vaidya et al., 1998; Bush et al., 1999; Rubia et al., 1999; Schweitzer et al., 2000; Schweitzer et al., 2003).
D4Rs are expressed preferentially in the PFC (Tarazi & Baldessarini, 1999) and are localized on glutamatergic and GABAergic neurons (Mrzljak et al., 1996; Wedzony et al., 2000) where they have been shown to regulate cognitive functions (Fuster, 2001). Similarly mice lacking D4Rs show impaired inhibitory and/or excitatory activation in the prefrontal cortex (Rubinstein et al., 2001), a trait that may underlie their behavioral phenotype. D4−/− mice show less locomotor activity in novel and familiar environments, greater vigilance in approach avoidance paradigms and increased excitability in prefrontal cortical neurons when compared to D4+/+ mice (Rubinstein et al., 1997; Rubinstein et al., 2001; Falzone et al., 2002). Furthermore, D4−/− mice show reduced exploration of novel stimuli (Dulawa et al., 1999; Tan et al., 2003). Evidence for the role for prefrontal D4Rs in attention is given by a recent study showing that the Spontaneously-Hyperactive Rat, a rat which exhibits increased hyperactivity and impaired attentional, show lower D4 levels in PFC when compared with control strains that don’t show attentional deficits (Li et al., 2007).
Pharmacological studies also support the role of D4Rs in cognitive tasks linked with prefrontal cortex (working memory and attention). Specifically, the selective D4 agonist A-412997 was reported to improve performance in the novel object recognition task and to increase extracellular DA concentration in the PFC in rats (Woolley et al., 2009). On the other hand, D4 antagonism (L745,870) has been reported to decrease cognitive-related behavior in rats (Braszko, 2009a; b) including decreases in working-memory performance in rats with good baseline performance but increases in rats with poor baseline performance (Zhang et al., 2004).
We found a differential effect of MP based on D4 genotype; MP increased metabolism in PFC in D4−/− mice but decreased it in D4+/+ and D4+/− mice. This finding further supports our hypothesis that D4Rs regulate PFC activity. It is also consistent with findings form imaging studies in non-ADHD humans that showed MP-induced decreases in PFC cerebral blood flow (CBF) in controls (Mehta et al., 2000), and with CBF increases in PFC in children with ADHD (Teicher et al., 1996; Vaidya et al., 1998). PET imaging studies using FDG observed both increases and decreases in metabolism in frontal brain regions after intravenous MP in normal adults (Volkow et al., 1997a) but no changes after oral MP in adults with ADHD tested under no stimulation conditions (Matochik et al., 1993; Matochik et al., 1994). The PFC effects of MP are likely to reflect in part its DAergic effects in the PFC since MP increases synaptic DA in the PFC of human subjects (Montgomery et al., 2007).
Thus, our results support the hypothesis that increased PFC activation in ADHD patients (in response to MP as observed in fMRI studies), may reflect dysfunctional D4Rs on GABA and glutamate neurons in PFC, a hypothesis that would have to be tested by further investigation. Though some caution the interpretation of behavioral, physiological and biochemical studies of the PFC in D4 KO mice (Wang et al., 2009).
We found greater basal metabolism in D4−/− compared to D4+/+ and D4+/− mice in the CBV, which is a brain region that modulates DA neurotransmission in caudate and accumbens (Nieoullon et al., 1978) via itsprojections to the ventral tegmental area (Snider et al., 1976), and has furthermore been implicated both in the pathophysiology of ADHD and on the therapeutic effects of MP (Anderson et al., 2002). Thus, D4 regulation of CBV activity may also underlie the association between D4Rs and ADHD. The opposite pattern of relative metabolism between the CBV and the PFC across genotype suggests that these two brain regions may work in concert in D4−/− mice. Indeed, projections in the form of parallel, closed-loop circuits, from the PFC to the striatum, CB and back, have been previously described (for review see (Brennan & Arnsten, 2008). Finally, the CB and striatum, where MP has significant effects on DA release, are connected via a polysynaptic pathway involving the intralaminar nuclei of the thalamus (Ichinohe et al., 2000; Hoshi et al., 2005). Based on these findings, it has been suggested that such circuits may be involved in impaired regulation of higher-order cognitive functions as well as impaired motor control (Brennan & Arnsten, 2008), both of which are observed in ADHD patients.
D4+/+ and D4+/− mice showed increased metabolic activity in response to MP in the CBV whereas D4−/− mice showed a decrease. The differential responses to MP in the CBV as a function of D4 are also consistent with the differential response of the CBV to MP as a function of the ADHD phenotype; specifically MP increased T(2) relaxation time in the ADHD children with high score on hyperactivity whereas it reduced it in ADHD children without symptoms of hyperactivity (Anderson et al., 2002). The results in the CBV are opposite to the differential activation observed in the PFC with MP as described above. Apart from the traditionally defined role of the CB in motor control and coordination, findings have shown that the CB plays a role in cognitive and emotional processes including memory, learning, and attention processes (Leiner et al., 1989; Barkley et al., 1992; Andreasen et al., 1995; Desmond et al., 1998; Rapoport et al., 2000; Schmahmann, 2004) that are disrupted in ADHD (Barkley et al., 1992) and have been shown to be improved by MP (Wilens & Biederman, 1992). The involvement of the CB in ADHD has further been suggested by structural imaging studies that showed smaller CBV in children with ADHD than in controls (Berquin et al., 1998; Mostofsky et al., 1998; Castellanos et al., 2001). fMRI studies have also showed a specific activation profile of the CB in response to MP in normal subjects (Anderson et al., 2006). PET imaging studies with FDG and [15O]H2Ohave also shown MP-specific increases in cerebellar energy metabolism and CBF respectively (Matochik et al., 1993; Volkow et al., 1997a; Schweitzer et al., 2003; Udo de Haes et al., 2007).
In addition to the NE transporter (NET) (Kung et al., 2004), the DA transporter (DAT) is also expressed in the CBV (albeit at low levels) (Melchitzky & Lewis, 2000). Thus MP effects in the CBV are likely to reflect both its blockade of the NET as well as that of DAT (Patrick et al., 1987; Volkow et al., 1992; Houk & Wise, 1995; Pontieri et al., 1995; Gatley et al., 1996; Strazielle et al., 1999; Middleton & Strick, 2001; Glaser et al., 2006). However, we contend that the increases in cerebellar metabolism with MP are also likely to reflect striatal cerebellar networks as evidenced by imaging studies that showed that striatal D2 receptor levels predicted MP-induced increases in cerebellar metabolism in healthy controls (Volkow et al., 1997a).
We found a significant negative correlation between MP-induced metabolic changes in the left PFC and those in the CBV in D4−/− mice. A similar but not significant effect (p=.09) was observed in D4+/− mice. This finding suggests a functional interaction between these two brain regions that is mediated by the presence of D4, a conjecture that should be further investigated.
Our interpretation of the present FDG findings is limited by the fact that we cannot account for compensatory changes in D4 KO mice. This potential interference might be circumvented in future studies through the use of conditional KO, gene-therapy or D4 vectors. Another limitation is that the baseline scan always preceded the MP scan and thus we cannot rule out the possibility that differences in metabolism observed with MP may be confounded by an order effect and/or a differential effect of the anesthetic used in the first measurement. Nevertheless, this possibility seems unlikely, given the excellent within-subject reproducibility reported in previous μPET studies (Alexoff et al., 2003; Marsteller et al., 2006). Finally, this is the first study to report the use of SPM in mice and thus further work will be required to determine the sensitivity of this analytical method, especially as it relates to our use of global normalization (Borghammer et al., 2009).
In summary this study identifies the PFC and the CBV as regions that distinguish D4−/− and D4+/+ mice providing evidence of the regulation of the activity of these brain regions by D4Rs.
This work was supported by the NIAAA (AA 11034 & AA07574, AA07611).