INTRODUCTION

Cocaine increases extracellular levels of dopamine (DA), serotonin (5-HT) and norepinephrine (NE) by blocking the neural plasma membrane transporters for those neurotransmitters. Increased extracellular DA (DAex) levels in mesocorticolimbic DA systems have been postulated to mediate cocaine reward (Kuhar et al, 1991; Koob and Nestler, 1997; Bardo, 1998; Kelley and Berridge, 2002). However, homozygous dopamine transporter (DAT) knockout (KO) mice (DAT−/− mice) express intact cocaine reward in conditioned place preference (CPP) (Sora et al, 1998) and drug self-administration paradigms (Rocha et al, 1998). Cocaine reward is eliminated in double-KO mice with no DAT gene copies and either no or one copy of the SERT gene (Sora et al, 2001), but not in double-KO mice with neither DAT nor NET gene copies (Hall et al, 2002). Further, serotonin transporter (SERT) blockade with fluoxetine or norepinephrine transporter (NET) blockade with nisoxetine can yield rewarding effects in DAT-KO mice, which are never seen in wild-type animals (Hall et al, 2002).

We and others have postulated that the retention of cocaine reward in DAT-KO mice may be due to (a) roles for non-DA systems in normal cocaine reward and (b) adaptations to the lifelong loss of DAT found in DAT-KO mice (Kirkpatrick, 2001; Sora et al, 2001; Uhl et al, 2002). Some of these adaptive changes could come from involvement of redundant monoaminergic systems in cocaine reward. Since each transporter displays significant affinities for each monoamine (Faraj et al, 1994; Giros et al, 1994; Gu et al, 1994; Eshleman et al, 1999), the absence of its cognate transporter might allow a monoamine to diffuse further from its site of release and be accumulated by another transporter.

Cocaine and selective norepinephrine transporter (NET) blockers (eg reboxetine) are each reported to increase DAex in NAc of DAT-KO mice, suggesting that NET could act as an alternative uptake site for DA in such animals and that NET blockade might be a mechanism for both the cocaine- and nisoxetine-induced rewards found in DAT-KO mice (Carboni et al, 2001; Hall et al, 2002). However, in vitro data fail to identify cocaine influences on CPu or NAc DA uptake in DAT-KO mice (Budygin et al, 2002; Moron et al, 2002). The simple idea that NET mediates cocaine reward in the absence of DAT is also incompatible with observations that cocaine reward is ablated in DAT/SERT double-KO mice that express normal levels of NET (Sora et al, 2001).

Roles for 5-HT systems in cocaine reward (or aversion) are also less than clear from current data (Cunningham and Callahan, 1991; Kleven et al, 1995; Rocha et al, 1997; Kleven and Koek, 1998; Lee and Kornetsky, 1998; Parsons et al, 1998; Shippenberg et al, 2000; Baker et al, 2001; Sasaki-Adams and Kelley, 2001). Homozygous SERT-KO mice display enhanced cocaine CPP that is increased even more in combined SERT/NET double-KO mice (Sora et al, 1998; Hall et al, 2002). SERT-KO mice, in themselves and in combination with DAT-KOs, thus provide interesting models in which to investigate 5-HT, DA, and 5-HT/DA interactions important for psychostimulant reward.

In this present study, we have therefore examined baselines and drug-induced changes of DAex and 5-HTex in several brain regions implicated in psychostimulant effects, the NAc, CPu and prefrontal cortex (PFc) in DAT-KO, SERT-KO, and both heterozygous and homozygous DAT/SERT double-KO mice. We have studied the effects of both the nonselective blocker cocaine and the selective SERT and DAT blockers fluoxetine and GBR12909. These investigations provide insights into adaptive processes found in these mice and into 5-HT, DA, and 5-HT/DA interactions of the possible importance for cocaine reward.

MATERIALS AND METHODS

Animals

Mutant mice lacking DAT, SERT, and littermate wild-type mice were obtained from heterozygote crosses on 129/C57 mixed genetic backgrounds. DAT/SERT double-KO mice were obtained by intercrossing single KO lines as described previously (Sora et al, 2001). DNA extracted from tail biopsies was genotyped using PCR. Mice were group-housed (two to four per cage) with food and water ad libitum in a room maintained at 22±2°C and 65±5% humidity under a 12 h light–dark cycle. Male and female mice from 10–24 weeks old of each genotype group (n=4–8) were used in each experiment equally. All animal experiments were performed in accordance with the Guidelines for the Care of Laboratory Animals of the Tokyo Institute of Psychiatry.

For the CPu cocaine study, all the nine DAT × SERT genotypes were examined (DAT+/+ SERT+/+, DAT+/+SERT+/−, DAT+/+SERT−/−, DAT+/−SERT+/+, DAT+/−SERT+/−, DAT+/−SERT−/−, DAT−/− SERT+/+, DAT−/−SERT+/−, and DAT−/−SERT−/−). For NAc and PFc cocaine studies and for fluoxetine CPu and NAc studies, the four homozygous genotypes were examined (wildtype, DAT−/−SERT+/+, DAT+/+SERT−/−, DAT−/−SERT−/−). GBR12909 effects on CPu 5-HTex levels were examined in wild-type and DAT+/+SERT−/− mice.

Surgery

Mice were stereotaxically implanted with microdialysis probes under sodium pentobarbital anesthesia (50 mg/kg) in CPu (anterior +0.6 mm, lateral +1.8 mm ventral −4.0 mm from bregma), NAc (anterior +1.2 mm, lateral +1.0 mm ventral −5.0 mm from bregma) or PFc (anterior +2.0 mm, lateral +0.5 mm ventral −3.0 mm from bregma) according to the atlas of Franklin and Paxinos (1997). Probe tips were constructed with regenerated cellulose membranes that provided 50 kDa molecular weight cutoffs, outer diameters of 0.22 mm, and membrane lengths of either 1 mm (NAc) or 2 mm (CPu and PFc) (Eicom, Kyoto, Japan). Dialysis probe placements were verified histologically at the ends of each experiment (Figure 1), and experimental data were excluded if the membrane portions of the dialysis probes lay outside the central CPu, medial PFc or NAc core or shell regions, respectively.

Figure 1
figure 1

Location of dialysis probes in coronal sections of PFc (a), NAc (b), and CPu (c). The arrows illustrate the implantation sites of dialysis probes.

Microdialysis and Analytical Procedure

At 24 h after implantation, probes in freely moving mice were perfused with Ringer's solution (147 mM Na+, 4 mM K+, 1.26 mM Ca2+, 1 mM Mg2+, and 152.5 mM Cl−, pH 6.5) at 1 μl/min for 180 min. DAex and 5-HTex baselines were obtained from average concentrations of three consecutive 10 min, 10 μl samples. These and subsequent 10 min, 10 μl dialysate fractions were analyzed using an AS-10 autoinjector (Eicom), high-performance liquid chromatography (HPLC), with a PPS-ODS reverse-phase column (Eicom) and a ECD-100 graphite electrode detector (Eicom). The mobile phase consisted of 0.1 M phosphate buffer (pH 5.5) containing sodium decanesulfonate (500 mg/l), EDTA (50 mg/l), and 1% methanol. Detection limits for DA and 5-HT were 1 fmol/sample with signal-to-noise ratios of at least 2. In vitro recoveries from the 1- and 2-mm membrane length probes were 10 and 15%, respectively.

Drugs

Test drugs were dissolved in saline for systemic administration or in Ringer's solution for local infusion via microdialysis probes. After establishment of stable baselines, cocaine HCl (10 mg/kg for subcutaneous injection or 100 μM for local infusion; Dainippon, Osaka, Japan), fluoxetine (20 mg/kg, Sigma, Tokyo, Japan), GBR12909 (10 mg/kg, Sigma) or saline (10 ml/kg) was administered subcutaneously (s.c.) and dialysates collected for 3 or 2 h, respectively.

Statistics

Baselines of DAex and 5-HTex were compared across genotype groups using two-way ANOVAs (DAT genotype, and SERT genotype). DA and 5-HT responses to drugs were expressed as percentages of baselines. Effects of drugs on DAex and 5-HTex were assessed by calculating the areas under time–response curves (AUC) for the first 120 or 180 min after drug administration. AUCs were analyzed using two-way ANOVAs (Drug, Genotype). Least significant difference tests were applied for multiple comparisons and P-values less than 0.05 were considered statistically significant. Statistical analyses used STATISTICA (StatSoft Inc., Tulsa, OK).

RESULTS

Baselines of DAex and 5-HTex in CPu, NAc, and PFc

The mean (±SEM) baselines of DAex and 5-HTex in dialysates from the CPu, NAc, and PFc in mice, who were subsequently treated with either vehicle or test drugs, are shown in Table 1. Two-way ANOVA of DAex baselines confirmed that DAT-KO had significant effects on DAex baselines in CPu (F(1, 91)=299.77, P<0.00001) and NAc (F(1, 55)=101.49, P<0.00001), but not PFc (F(1, 33)=0.07, P=0.79). Dialysate DA in homozygous DAT-KO mouse CPu and NAc was approximately 10-fold higher than that in mice with either one or two copies of the DAT gene. 5-HTex baselines were unaffected by DAT-KO in any region.

Table 1 The Baselines (fmol/10 min) of DAex and 5-HTex in CPu, NAc and PFc

SERT-KO exerted significant effects on 5-HTex baselines in each of these three regions (F(2, 91)=87.06, P<0.00001; F(1, 55)=29.95, P<0.00001; F(1, 33)=80.37, P<0.00001, respectively). In CPu, NAc, and PFc, 5-HTex baselines in mice with no SERT gene were six to ten times as large as that found in mice with one or two copies of SERT gene. DAex baselines were unaffected by SERT-KO in any region.

Interestingly, there was a significant interaction between DAT and SERT genotype effects on basal NAc dialysate DA levels (F(12, 55)=4.33, P<0.05). DAex levels in NAc of mice with no DAT or SERT genes (DAT−/−SERT−/−) were higher than those of mice with no DAT genes but two SERT genes (DAT−/−SERT+/+).

Systemic Cocaine Effects on DAex in CPu, NAc, and PFc

DAex level changes in CPu, NAc, and PFc following systemic cocaine administration are shown in Figure 2a, c, and e. DA responses to cocaine in the CPu of wild-type and DAT+/− mice peak at 40–60 min (Figure 2a). Cocaine also induces a slower DA response curve in the CPu of homozygous DAT-KO mice (DAT−/−SERT+/+), peaking at about 90 min (Figure 2a). This pattern is not observed in NAc, where DAT−/−SERT+/+ mice do not exhibit any larger increments in DAex levels (Figure 2c). In further contrast, wild-type, DAT−/−SERT+/+, and DAT−/−SERT−/− mice each exhibit indistinguishable cocaine-induced DA responses in PFc (Figure 2e).

Figure 2
figure 2

(a, c, and e) Temporal pattern of DA response to cocaine (10 mg/kg, s.c.) in CPu, NAc, and PFc, respectively. Each point represents the mean (±SEM) of the percentage of DAex baselines. The time of injections is indicated with an arrow. (b, d, and f) Histogram represents the mean AUC (±SEM) of DA response to saline or cocaine in CPu, NAc, and PFc during 180 min interval after injection. **P<0.01, ***P<0.001 compared to the saline group of the same genotype; #P<0.05, ###P<0.001 compared to the cocaine-treated wild-type group.

Drug effects on DAex levels can be assessed by studying AUCs (Figure 2b, d, and f). ANOVAs of mean AUC (±SEM) for drug effects on DAex levels reveal that drugs have significant effects on DA AUC in CPu (F(1, 62)=132.32, P<0.0001), NAc (F(1, 34)=80.60, P<0.0001), and PFc (F(1, 28)=67.59, P<0.0001). Genotype and drug × genotype interactions were significant for DA AUC in CPu (F(8, 62)=5.45, P<0.0001; F(8, 62)=3.41, P<0.01; respectively) and NAc (F(3, 34)=23.82, P<0.0001; F(3, 34)=36.09, P<0.0001; respectively), but not in PFc (F(3, 28)= 0.89, P=0.46; F(3, 28)=0.94, P=0.43; respectively). In CPu (Figure 2b), DAT-KO mice exhibit statistically significant cocaine-induced increments in DAex levels, although these increases are less than those found in wild-type mice. By contrast, in DAT−/−SERT+/− and DAT−/−SERT−/− mice, the same genotypes that do not exhibit rewarding effects of cocaine also do not exhibit cocaine-induced increases in DAex in CPu. No significant differences are observed in cocaine-induced DA AUC increases in CPu between the DAT+/+ and DAT+/− mice. In NAc (Figure 2d), cocaine fails to increase DAex in DAT−/− SERT+/+ or in DAT−/−SERT−/− mice. There are no significant differences in cocaine-induced DA increases in NAc between wild-type and DAT+/+SERT−/− mice. In PFc (Figure 2f), cocaine produces significant increases in DAex in all genotypes.

Systemic Cocaine Effects on 5-HTex in CPu, NAc, and PFc

The temporal patterns of 5HT responses to cocaine in CPu, NAc, and PFc are shown in Figure 3a, c, and e. DAT+/+ SERT−/− and DAT+/−SERT−/− mice show gradual 5HT responses to cocaine in CPu (Figure 3a) and NAc (Figure 3c), but not in PFc (Figure 3e). 5-HT response curves produced by cocaine are observed in CPu (Figure 3a) and NAc (Figure 3c) in all genotypes except DAT−/− SERT−/− mice. The peak of 5-HT response are smaller for SERT−/− mice than for either SERT+/+ or SERT+/− mice. SERT−/− mice exhibit no 5-HT response to cocaine in PFc (Figure 3c), while wild-type mice exhibit robust increases.

Figure 3
figure 3

(a, c, and e) Temporal pattern of 5-HT response to cocaine (10 mg/kg, s.c.) in CPu, NAc, and PFc, respectively. The time of injections is indicated with an arrow. (b) Histogram represents the mean AUC (±SEM) of 5-HT response to saline or cocaine in CPu, NAc, and PFc during the 180 min interval after injection. *P<0.05, ***P<0.001 compared to the saline group of the same genotype; #P<0.05, ###P<0.001 compared to the cocaine-treated wild-type group.

Drug effects on 5-HTex levels are expressed as mean AUC (±SEM) in Figure 3b, d, and f. Two-way ANOVAs of the AUC for 5-HT responses to cocaine show significant effects of Drug, Genotype, and Drug × Genotype interactions in CPu (F(1, 62)=181.49, P<0.0001; F(8, 62)=5.01, P<0.0001; F(8, 62)=4.88, P<0.0001; respectively), NAc (F(1, 34)=31.57, P<0.0001; F(3, 34)=6.44, P<0.0001; F(3, 34)=8.41, P<0.0001; respectively) and PFc (F(1, 28)=57.74, P<0.0001; F(3, 28)=11.55, P<0.0001; F(3, 28)=15.59, P<0.0001; respectively). In CPu (Figure 3b) and NAc (Figure 3d), multiple AUC comparisons reveal that cocaine significantly increases 5-HTex in DAT+/+SERT−/− and DAT+/−SERT−/− mice, but not in DAT−/−SERT−/− mice. SERT+/− mice display cocaine-induced increases in 5-HTex in CPu that are similar to those found in wild-type values. PFc 5-HTex levels are not altered significantly by cocaine in SERT−/− mice (Figure 3f).

Systemic Fluoxetine Effects on DAex in CPu and NAc

The temporal patterns of DA response to fluoxetine in CPu and NAc of DAT+/+ SERT+/+, DAT−/−SERT+/+, DAT+/+SERT−/−, and DAT−/−SERT−/− mice are shown in Figure 4a and c. In CPu (Figure 4a), DAT−/− SERT+/+ mice exhibit gradual DA responses to fluoxetine that display time courses similar to those of cocaine and persist for at least 3 h (Figure 4a). Two-way ANOVAs of DA AUC responses show significant effects of Drug (F(1, 33)=9.62, P<0.01) and Drug × Genotype interactions (F(1, 33)=4.94, P<0.01). Multiple comparisons reveal that fluoxetine significantly increases DA AUC only in the CPu of DAT−/−SERT+/+ mice (Figure 4b). In NAc, DA responses to fluoxetine display no significant effects of either Drug (F(1, 29)=0.0076, P=0.93), Genotype (F(1, 29)=0.49, P=0.69), genotype (F(1, 29)=0.69, P=0.41) or Drug × Genotype interaction (F(3, 29)=1.55, P=0.22) (Figure 4d).

Figure 4
figure 4

(a and c) Temporal pattern of DA response to fluoxetine (20 mg/kg, s.c.) in CPu and NAc, respectively. The time of injections is indicated with an arrow. (b and d) The histogram represents the mean AUC (±SEM) of DA response to saline or fluoxetine in CPu and NAc during the 180 min interval after injection. *P<0.05 compared to the saline group of the same genotype; #P<0.05 compared to the fluoxetine-treated wild-type group.

Systemic GBR12909 Effects on 5-HTex in CPu

The temporal pattern of CPu 5-HT response to GBR12909 is shown in Figure 5a. DAT+/+SERT−/− mice exhibit remarkable 5-HTex increases after administration of GBR12909, which are not seen in WT mice. These SERT-KO mice continue to display elevated CPu 5-HTex levels for at least 3 h. Two-way ANOVA of the AUC of the DA response to GBR 12909 shows significant effects of Drug (F(1, 13)=14.43, P<0.01), Genotype (F(1, 13)=7.63, P<0.05), and Drug × Genotype interactions (F(1, 13)= 5.74, P<0.05). Multiple comparisons show that GBR12909 administration significantly increases CPu 5-HTex in DAT+/+SERT−/−, but not in wild-type mice (Figure 5b).

Figure 5
figure 5

(a) Temporal pattern of 5-HT response to GBR12909 (10 mg/kg, s.c.) in CPu. The time of injections is indicated with an arrow. (b) The histogram represents the mean AUC (±SEM) of 5-HT response to saline or GBR12909 in CPu during 180 min interval after injection. **P<0.01 compared to the saline group of the same genotype; ##P<0.01 compared to the GBR12909-treated wild-type group.

Local Cocaine Effects on DAex and 5-HTex in CPu

DAex and 5-HTex level changes in CPu following local cocaine infusion are shown in Figure 6a and c. Local cocaine cannot induce DA response curve in CPu of DAT−/− SERT+/+ and DAT−/−SERT−/− mice, but produces gradual 5-HT response curve in DAT+/+SERT−/− mice.

Figure 6
figure 6

(a and c) Temporal pattern of DA and 5-HT response to local cocaine infusion (100 μM) in CPu, respectively. Horizontal bar indicates the time of infusions. (b and d) The histogram represents the mean AUC (±SEM) of DA and 5-HT response to saline or cocaine in CPu during 120 min interval after injection. *P<0.05, ***P<0.001 compared to the saline group of the same genotype; ###P<0.001 compared to the cocaine-treated wild-type group.

ANOVAs of mean AUC (±SEM) for DA responses reveal significant effects of Drug, Genotype, and Drug × Genotype interactions in CPu (F(1, 24)=161.46, P<0.0001; F(3, 24)=48.20, P<0.0001; F(3, 24)=47.30, P<0.0001; respectively). Multiple AUC comparisons show that local cocaine fails to increase DAex in CPu of DAT−/− SERT+/+ or in DAT−/−SERT−/− mice (Figure 6b). ANOVAs of mean AUC (±SEM) for 5-HT responses also reveal significant effects of Drug, Genotype, and Drug × Genotype interactions in CPu (F(1, 24)=43.26, P<0.0001; F(3, 24)=9.55, P<0.0001; F(3, 24)=9.70, P<0.0001; respectively). Multiple comparisons reveal that local cocaine significantly increases 5-HTex in wild-type, DAT+/+SERT−/− and DAT−/−SERT+/+ mice, but not in DAT−/−SERT−/− mice (Figure 6d). Moreover, there were no significant changes in NAc DAex in DAT−/−SERT+/+ and DAT−/−SERT−/− mice after local cocaine infusions (data not shown).

DISCUSSION

These microdialysis results reveal parallels with and differences from the patterns of KO effects on reward elicited by cocaine and fluoxetine that we have previously reported in these mouse strains. We can thus evaluate hypotheses about the pharmacological profiles and brain localization of processes hypothesized to mediate cocaine reward with regard to their convergence or divergence with this microdialysis data.

Differential DA Responses in to Cocaine in CPu, NAc, and PFc and Correlations with Assessments of Cocaine Reward

The current data do not provide simple correlations with models that postulate that enhanced NAc DAex levels alone are necessary and sufficient for cocaine reward. Although this hypothesis has been supported by data from microinjection and lesion studies (Kuhar et al, 1991; Koob and Nestler, 1997; Bardo, 1998; Kelley and Berridge, 2002), many results from gene KO studies fail to support the simple hypothesis that DA alone mediates the rewarding effects of cocaine. Our current observations that cocaine does not increase DAex in NAc of homozygous DAT-KO mice contrasts with the nearly-intact cocaine reward found in these animals (Rocha et al, 1998; Sora et al, 1998). These in vivo microdialysis data are also consistent with studies which document failure of cocaine to block DA uptake in NAc samples taken from DAT homozygous mice in in vitro experiments (Budygin et al, 2002; Moron et al, 2002).

The current data also fail to provide simple correlations with models that postulate that enhanced PFc DA levels are necessary and sufficient for cocaine reward. This hypothesis has also been supported by a substantial body of lesion and microinjection data (Goeders and Smith, 1983; Goeders et al, 1986; Bardo, 1998; Tzschentke, 2001). Cocaine increases DAex in PFc of both wild-type and homozygous DAT-KO mice that exhibit cocaine reward and DAT/SERT double homozygous KO mice that do not display cocaine reward.

Intriguingly, the current results for DA in CPu appear to provide the best fit with studies of cocaine-induced place preferences. Although intra-CPu cocaine does not affect DAex levels in DAT-KO mice, systemic cocaine causes about 1.5-fold increase in peak DAex concentrations in CPu dialysate from DAT-KO mice that are rewarded by cocaine, but not from DAT/SERT double homozygous KO mice that lack cocaine CPP. Systemic fluoxetine also increases CPu DAex levels in homozygous DAT-KO mice in which this compound is rewarding, but not in wild-type mice or homozygous SERT-KO mice in which fluoxetine does not produce a place preference.

Differential 5-HT Responses to Cocaine in CPu, NAc, and PFc and Correlations with Assessments of Cocaine Reward

Although cocaine-induced increases in CPu and NAc 5-HTex are found in SERT-KO mice that exhibit enhanced cocaine CPP, the magnitude of the increases in 5-HTex after cocaine administration is attenuated when it is compared with wild-type mice. Interestingly, chronic SERT blockade with fluoxetine can also potentiate cocaine reward (Cunningham and Callahan, 1991; Kleven and Koek, 1998). It is conceivable that the attenuation of cocaine-induced 5-HTex rise may lead mice more sensitive to the reward effect of cocaine. These sorts of data, and the current results, continue to point to possible roles for 5-HT in cocaine reward, especially in light of the more complex hypotheses of the basis of cocaine reward discussed below.

5-HTex Clearance by DAT, DAex Clearance by NET, and opportunities for ‘Promiscuous Uptake’

Removal of a transporter that usually provides inactivation, re-accumulation, and recycling of a released monoamine neurotransmitter provides opportunities for greater diffusion of the monoamine, documented by higher extracellular dialysate concentrations noted here. Removal of a cognate transporter also enhances the opportunities for transmitter uptake by a transporter that normally recognizes another monoamine. The presence of the same vesicular transporter in DAT-, SERT-, and NET-expressing neurons provides the opportunity for the monoamine that has been taken up by a non-cognate plasma membrane transporter to be accumulated into vesicles, and to be re-released as a ‘false transmitter’ (Liu and Edwards, 1997; Uhl et al, 2000). DA accumulation by NET-expressing neurons also provides the opportunity for DA to be subjected to β-hydroxylation to produce norepinephrine, providing a ‘true’ transmitter for noradrenergic neurons. It is interesting to note that elimination of monoamine transporters has different effects on basal monoamine levels in different brain regions, supporting ideas that factors that mediate DA and 5-HT clearance from synaptic clefts may differ substantially from one terminal field to another.

Many of the present and previously reported results appear to provide evidence for uptake by non-cognate transporters, and even for possible ‘false transmission’ in these transporter-KO mice. Cocaine and the selective DAT blocker GBR12909 produces a substantial increase in dialysate 5-HT in SERT-KO mice that is not found in wild-type animals. These findings were supported by previous reports that have documented 5-HT uptake by cultured neurons from SERT-KO mice that could be blocked by selective DAT blockers (Pan et al, 2001), and 5-HT-like immunoreactivity in substantia nigra and ventral tegmental area dopaminergic neurons (Zhou et al, 2002). False transmission may be region-dependent, with differences in the relative densities of DAT- SERT- and NET-expressing neural elements providing differential opportunities for such processes.

Moreover, our observations of virtually identical PFc DAex baselines in each of these KO strains appear to support a relatively reduced prominence of DAT-mediated DA uptake in this region even in wild-type mice. These observations are compatible with the relatively sparse distribution of PFc DAT in several species (Freed et al, 1995; Sesack et al, 1998), in contrast with more prominent NET and SERT expression. They are also in accord with pharmacological and other evidence for significant NET-mediated DA uptake in rodent PFc (Di Chiara et al, 1992; Tanda et al, 1997; Yamamoto and Novotney, 1998). DA may thus be accumulated by NET in PFc of both wild-type and DAT-KO mice.

The current observations in DA response to cocaine and fluoxetine in CPu of DAT-KO mice may provide a different picture. Although systemic cocaine and fluoxetine increase significantly CPu DAex in DAT-KO mice, local cocaine fails to change it. These results demonstrate that SERT does nt play a role of ‘promiscuous uptake’ in DA clearance. Systemic cocaine- or fluoxetine-induced DA increase in CPu of DAT-KO mice may result from DA release from activated DA neuron rather than local clearance by SERT.

Comparisons with Other Results

Observations that CPu dialysate monoamine levels apparently provide the best parallel with the loss of cocaine CPP found in current results could be consistent with a previously underappreciated role for CPu structures in mediating some of the ‘learned’ features of cocaine reward that are manifest in conditioned place preference testing (White and McDonald, 2002). These structures can be critical for stimulus-response ‘habit’ learning, including that related to reward (Jog et al, 1999; Reynolds et al, 2001). It is conceivable that this structure may play an even greater role in DAT-KO mice that lack cocaine-induced DAex elevations in NAc.

The failure of dialysis results for DA alone in NAc or PFc to parallel cocaine reward effects of various KOs and the apparent parallel in CPu should not prevent further consideration of: (a) multiple compensating contributions of monoamines to the rewarding effects of cocaine; (b) contributions of cocaine effects on monoamines in other brain regions, for example, ventral pallidum (Gong et al, 1996, 1997), ventral tegmental area (Roberts and Koob, 1982; Ranaldi and Wise, 2001) for cocaine reward; (c) effects of nonmonoaminergic adaptations to the retained cocaine reward in the transporter KO mouse strains that retain such reward. Monoamine actions in brain regions such as the ventral tegmental area have been postulated to be central to the rewarding actions of major drug classes, such as opiates (Wise, 1989; Garzon and Pickel, 2001) and stimulants. It is quite conceivable that monoamine actions in areas not sampled in the current studies could play roles in normal cocaine reward mechanisms, and in adaptations that may underlie the retention of cocaine reward in DAT- and in SERT-KO mice. Mice with single or multiple transporter deletions display many adaptive alterations, as assessed through behavioral, neurochemical, per- or post-synaptic receptor binding, gene expression, and other analytical approaches. None of the current data should hinder attempts to add more explanatory power for the remarkable behavioral pharmacological profiles displayed by these KO mice through use of any or all of these alternative approaches.

The current results in NAc and CPu DA response to cocaine in DAT-KO mice produced in our laboratory, while highly reproducible in our hands, differ from those obtained in reports from another line of DAT-KO mice that which showed that systemic cocaine and reboxetine (NET blocker) increased DAex remarkably in NAc of DAT-KO mice (Carboni et al, 2001). The different DA response to cocaine in NAc and CPu between Carboni's and our DAT KOs may be due to the different DNA construction which was used to disrupt DAT gene. Moreover, our findings are consistent with other reports which demonstrated that cocaine could not affect DA clearance in NAc of DAT-KO mice via in vitro experiments. It is noteworthy that (1) DAex baseline in NAc of DAT-KO mice is about 10 times greater than that in wild-type mice, and that (2) the capacity for DA uptake of NET is far weaker than that of DAT (Giros et al, 1994; Gu et al, 1994). These may be the reasons why NET cannot show redundancy for DAT in NAc.

In summary, the present work adds to previous data concerning the behavioral consequences of DAT and SERT deletion, by suggesting that cocaine CPP does not necessarily correlate with simple elevations of DA the NAc or PFc. It points out unanticipated correlations with DAex elevations in CPu. It is interesting that the CPu findings parallel behavioral observations of the rewarding profiles of not only cocaine but also of fluoxetine in these varying mouse strains. While these correlations do not prove causation, the data support careful re-examination of CPu roles in psychostimulant reward (or reward learning) in both wild-type and DAT-KO mice, including both the dorsal and ventral CPu regions likely to be sampled with our microdialysis approaches. Another view of the current results is that the double homozygous DAT/SERT combined KO mice that failed to display either cocaine-induced DAex or 5-HTex elevations in NAc also failed to exhibit cocaine CPP, suggesting perhaps that either DAex or 5-HTex elevation can mediate cocaine reward and that the absence of both effects is required to eliminates the cocaine CPP. The current data also add to the growing body of evidence that may indicate uptake of released monoamines by non-cognate transporters when their cognate transporters are deleted, and provide evidence for the brain-region specificity of these processes in wild-type and in transporter KO mice. Each of these findings adds pieces to the complex puzzle of the mediation of cocaine reward by monoaminergic brain systems.