A comparison of morphine-induced locomotor activity and mesolimbic dopamine release in C57BL6, 129Sv and DBA2 mice


Address correspondence and reprint requests to N. P. Murphy, Neural Circuit Mechanisms Research Group, RIKEN Brain Science Institute, 2–1 Hirosawa, Wakoshi, Saitama, 351–0198, Japan. E-mail: nmurphy@brain.riken.go.jp


Inbred mouse strains show marked variations in morphine-induced locomotion and reward behaviors. As increases in mesolimbic dopamine release and locomotion have been implicated as being critical aspects of drug-seeking and reward-related behaviors, the present study sought to determine the relationship between morphine-induced changes in locomotion and mesolimbic dopamine release. Freely moving microdialysis of the ventral striatum was performed in mouse strains chosen on the basis of their documented differences in locomotor and reward response to morphine (C57BL6 and DBA2) and use in the production of genetically modified mice (129Sv). Both C57BL6 and 129Sv mice showed significant increases in locomotion and ventral striatal extracellular dopamine levels following subcutaneous morphine administration (3 mg/kg), with the former strain showing the largest increase in both parameters. Ventral striatal extracellular DA levels increased in DBA2 mice to a similar extent as 129Sv mice following morphine administration, despite this strain showing no locomotor response. Intra-strain analysis found no correlation between morphine-induced locomotion and mesolimbic dopamine release in any of the strains studied. Thus, no universal relationship between morphine-induced mesolimbic dopamine release and locomotion exists between, and particularly within, inbred mouse strains. Furthermore, morphine-induced increases in mesolimbic activity correlate negatively with the rewarding potential of morphine described in previously reported conditioned place preference studies.

Abbreviations used


The mesolimbic dopamine (DA) system has long been postulated as a fundamental component of mammalian reward circuitry (Koob and Le Moal 1997; Wise 1998; Di Chiara 1999; Ikemoto and Panksepp 1999; Robinson and Berridge 2000). The relationship between mesolimbic DA release, locomotion and drug reward remains a hotly debated issue. In particular, can knowledge of the drug-induced response of one of these parameters predict the response of the others?

The complexity of this issue is reflected in the fact that several decades of research in this area have been unable to categorically assign a function to mesolimbic DA in motivated behaviors. However, this is not to say that little has been achieved in this time. The initial view of the mesolimbic DA pathway as a ‘pleasure circuit’ has been refined to the current view of it as a critical link between motivation and motivated behaviors. The ‘psychomotor stimulant theory of addiction’ devised by Wise and Bozarth (1987) posits that forward locomotion is a necessary manifestation of these goal-directed behaviors. In addition, activation of the mesolimbic DA pathway is critical in this event. Thus, drug-induced activation of the mesolimbic DA system and locomotion ought to some extent correlate with each other, as well as to the incentive properties of that particular drug.

Inbred mice strains have served as valuable models of drug abuse susceptibility for several decades due to large inter-strain variability in responses to commonly abused drugs. For instance, C57BL6 mice exhibit strong morphine-induced locomotor responses (Oliverio and Castellano 1974a,b; Brase et al. 1977; Belknap et al. 1989, 1998; Wenger 1989) yet show relatively weak morphine-induced place preferences (Cunningham et al. 1992; Semenova et al. 1995). This contrasts with DBA2 mice, which show no morphine-induced locomotor response (Oliverio and Castellano 1974a,b; Brase et al. 1977; Belknap et al. 1989, 1998; Wenger 1989) despite displaying stronger conditioned place preferences than C57BL6 mice (Cunningham et al. 1992; Semenova et al. 1995). As has been pointed out by others (Cunningham et al. 1992), this conflicts with the psychomotor stimulant theory of addiction insofar as the mouse strains showing the lowest locomotor response ought also to show the lowest reward response.

Presently, limited information regarding the neurobiology of these differences between inbred mice strains is available owing to the fact that the majority of studies addressing the neurochemistry of reward behaviors have been conducted in the rat. Investigators have been reluctant to address neurochemical questions in the mouse, probably because of the relatively small size of the species and the associated technical difficulties. However, the recent explosion in the use of genetically modified animals to probe highly specific genotype effects on animal behavior and physiology has encouraged investigators to adapt neurochemical techniques, such as microdialysis, to the mouse. Furthermore, the necessity for a complete understanding of the influence of genetic background on animal behavior and neurochemistry has become decidedly more pressing as deliberate genetic modifications (such as gene ‘knockout’) have often been made on mixed genetic backgrounds (e.g. C57BL6 × 129Sv hybrids). Ironically, it is these very studies that have often turned investigators to studying pure inbred mice strains in an attempt to identify the source of variability in data produced by crossing them.

With these matters in mind, the current study sought to address a number of critical issues surrounding the relationship between drug-induced mesolimbic DA release and locomotion by attempting to answer the following questions. First, do different inbred mice strains show similar mesolimbic DA responses to morphine? Second, how does this neurochemical response correlate with the locomotor response? Finally, are morphine-induced mesolimbic DA and/or locomotor responses able to predict the inter-strain variations in morphine reward identified in previous studies?

In order to address these questions, microdialysis of the ventral striatum (the terminal region of the mesolimbic DA neurons) was performed in the freely moving mouse during morphine administration. Spontaneous locomotor activity in the horizontal plane was measured in synchrony. The C57BL6 and DBA2 mice strains were initially selected for study because of their well-characterized and contrasting behavioral responses to morphine. In addition to these two strains, the lesser investigated 129Sv strain was included in the study, as this strain is often selected as a progenitor for the production of transgenic and knockout mice. Thus, characterization of the 129Sv strain may prove useful for future studies.

Materials and methods


Experimental protocols used throughout the study were approved by the institutional review committee and were in accord with NIH ethics guidelines. Age-matched male mice of the C57BL/6 J, 129 × 1/SvJ and DBA2/J strains (mean ages at commencement of surgery 71.4 ± 1.8, 72.2 ± 2.0 and 73.3 ± 1.9 days, respectively) were purchased from Jackson Laboratories (Bar Harbor, ME, USA). Animals were housed four animals per cage in a temperature-controlled colony room with a 12-h light/dark cycle (lights on 7.00 h) where they received food and water ad libitum.


Mice were maintained under halothane anesthesia (2–3%, vaporized in equal parts oxygen and nitrous oxide) and mounted in a stereotaxic frame. Prior to microdialysis probe implantation, the size of the skull was measured by determining the distance between bregma and the intersection of the lambdoid and sagittal sutures.

Microdialysis probes of a concentric design were constructed as previously described (Maidment et al. 1989) using a 2-mm long, 300-µm diameter polyacrylonitrile membrane (AN69, Hospal). Each animal was implanted with a microdialysis probe in the ventral striatum (coordinates relative to bregma in millimeters, A: 1.1–1.4, L: − 0.8, V: 5.0) according to the atlas of Franklin and Paxinos (1997). Microdialysis probes were continually perfused using Instech Laboratories (Plymouth Meeting, PA, USA) syringe pumps at a rate of 2 µL/min with an artificial cerebrospinal fluid containing 125 mm NaCl, 2.5 mm KCl, 0.9 mm NaH2PO4, 5 mm Na2HPO4, 1.2 mm CaCl2, 1 mm MgCl2 (pH 7.4). Three miniature self-tapping screws were placed in the skull to serve as anchors. The whole assembly was fixed with dental cement prior to attaching the animal to a tether leading to a single-channel liquid swivel (Instech Laboratories). A single piece of photoelectric tape (15 mm × 10 mm) was attached to the back of each animal using cyanocrylate adhesive in order to allow monitoring of locomotion (see below). Each animal was placed into a cylindrical plexiglass microdialysis cage (18 cm diameter × 21 cm height) and left overnight to recover. During this time and the remainder of the experiment, animals continued to receive food and water ad libitum.

Microdialysis procedure

Microdialysis sample collection started within 3–5 h of lights on. One hour prior to this, a low intensity lamp fixed 94 cm above the dialysis cages was illuminated to allow simultaneous monitoring of locomotor activity by reflection using a Videomex V system (Columbus Instruments, Columbus, OH, USA). Dialysate samples were collected manually at 15-min intervals into tubes pre-loaded with 3 µL of 12.5 mm perchloric acid/250 µm EDTA at room temperature (22°C). Tubes were immediately cooled to 4°C after collection until completion of the experiment, after which they were stored at −80°C until analysis. Locomotor activity was recorded in 15-min sessions in synchrony with dialysate sample collection.

After 1 h of sample collection, animals were subcutaneously administered morphine sulfate (3 mg/kg, corrected for water content, Mallinckrodt, St Louis, MO, USA) or vehicle (sterile 0.9% NaCl, in a volume of 10 mL/kg body weight). A further 6 h of dialysate sample collection and locomotor monitoring was performed.

Neurochemical analysis

High performance liquid chromatography with electrochemical detection (Antec Leyden, Leiden, the Netherlands) was used for determination of DA in dialysates using a mobile phase consisting of sodium acetate (75 mm), sodium dodecane sulfonate (0.75 mm), EDTA (10 µm), triethylamine (0.01%), acetonitrile (12%), methanol (12%), tetrahydrofuran (1%), pH 5.5, pumped at a rate of 200 µL/min (Shimadzu model LC-10AD, Columbia, MD, USA) through a 100 × 2 mm column (3 µm, Hypersil C18, Keystone Scientific, Bellefonte, PA, USA). The system was calibrated at regular intervals and provided a limit of detection of 0.25 fmol for a 5-µL injection of sample. Data were collected and analyzed using ChromPerfect software (Justice Innovations, Mountain View, CA, USA).

Histological verification of probe placement

At the end of microdialysis experiments, animals were anesthetized by intraperitoneal injection of pentobarbitone (60 mg/kg) and transcardially perfused with approximately 30 mL of heparinized (20 mg/L) isotonic saline followed by approximately 100 mL of 10% buffered formalin. Brains were sectioned at 50-µm thickness and alternate sections were mounted on gelatinized slides prior to staining with cresyl violet. Sections were analyzed blind in order to determine the precise location of probe implantation. Only microdialysis probes located within the ventral striatum between the anterior and posterior coordinate of + 0.74 mm and 1.78 mm relative to bregma were included in subsequent analysis.

Data analysis and statistical testing

Data were analyzed to determine differences between strains in basal and morphine-induced effects on ventral striatal extracellular DA and locomotor activity.

Absolute mean basal concentrations (i.e. the mean of four consecutive pre-drug samples) of DA in dialysates were analyzed by one-way anova. For comparison of morphine-induced effects on DA overflow, dialysate DA concentrations were normalized to the respective mean basal concentration. Statistically significant differences were identified by three-way univariate repeated measures anova (strain × treatment × normalized dialysate DA content) over the 3-h post-morphine injection period.

Locomotion data were analyzed only in absolute form. Significant differences in mean basal locomotor activity (i.e. the mean of four consecutive pre-drug recordings) were tested for by one-way anova. Significant differences within and between strains in morphine-induced locomotor response were analyzed by univariate three-way repeated measures anova (strain × treatment × locomotion) over the 3-h post-morphine injection period. Following anova, contrast analyses were used to identify specific differences between treatments and groups. Correlations between locomotor and mesolimbic DA release were tested using Pearson's correlation. Statistical significance was taken at p-values less than 0.05.


Characteristics of mice studied

Despite being identical in age, the three strains of mice studied showed conspicuous differences in body mass and skull size (Table 1). In particular, DBA2 mice were significantly heavier (F2,60 = 9.04), followed by C57BL6, then 129Sv mice. C57BL6 mice had significantly smaller skulls than 129Sv and DBA2 mice (F2,57 = 9.75).

Table 1.  Physical characteristics of mice strains studied
StrainBody weight (g)Skull size (mm)
  1. Data represent mean ± SEM. Numbers in parentheses indicate n.

C57BL625.4 ± 0.5 (21)3.93 ± 0.07 (19)
129Sv24.0 ± 0.6 (20)4.27 ± 0.05 (19)
DBA227.3 ± 0.6 (22)4.20 ± 0.05 (22)

Histological analysis of probe position

Histological analysis of the placement of microdialysis probes (Fig. 1) revealed increased incidences of probes placed in anterior regions of the ventral striatum of 129Sv mice, particularly in comparison to C57BL6 mice. As a consequence, significant differences in the mean anterior–posterior probe locations were observed between the three strains (F2,60 = 3.4). Despite this difference in probe distribution, no correlation was found between either basal levels of DA in ventral striatal dialysates and probe position within strains, or magnitude of morphine-induced DA release and probe position within strains (data not shown).

Figure 1.

Box plots showing distribution of microdialysis probes in the anterior–posterior coordinate across the three strains studied. Vertical box edges mark 25th and 75th percentiles. Vertical line within box marks median (50th percentile). Whiskers extend to farthest observation not farther than 1.5 times the distance between the quartiles. More extreme data values are plotted with individual markers. (C57BL6, n = 21; 129Sv, n = 22; DBA2, n = 22).

Morphine-induced changes in locomotion

Mean locomotion over the hour prior to drug administration was 3.0 ± 0.8, 1.4 ± 0.3 and 1.6 ± 0.3 m/h for C57BL6, 129Sv and DBA2 strains, respectively. Thus, C57BL6 mice tended to show a higher basal locomotion though this narrowly failed to reach statistical significance (F2,69 = 2.819, p = 0.067).

Three-way univariate repeated measure anova of the 3-h period following morphine or saline administration revealed a significant interaction between strain, treatment and time (F22,698 = 5.494). Saline injection induced transient increases in locomotion in all three strains (Fig. 2). No significant differences were found between any of the strains studied following saline injection.

Figure 2.

Effect of subcutaneous vehicle (upper panel) or morphine administration (3 mg/kg; lower panel) on locomotor activity of C57BL6 (●), 129Sv (○) and DBA2 (●) mice. Data are expressed as a mean ± SEM (saline treatment: C57BL6, n = 12; 129Sv, n = 11; DBA2, n = 10; morphine treatment: C57BL6, n = 15; 129Sv, n = 12; DBA2, n = 12).

Subcutaneous injection of morphine induced significant increases in locomotion (vs. saline injection) in C57BL6 and 129Sv strains only (Fig. 2). The morphine-induced locomotion observed in C57BL6 mice was significantly higher than that seen in 129Sv (and DBA2) mice, approximating a threefold difference. Although DBA2 mice showed no significant change in locomotion following morphine administration (vs. saline-treated animals) when considered over the 3-h period following injection, a clear suppression of the saline-induced locomotion was observed followed by a small increase in locomotion approximately 1 h after injection.

Morphine-induced changes in ventral striatal extracellular DA

No significant differences in basal dialysate DA concentration were evident between the three strains studied (p = 0.287). Basal dialysate DA concentrations were as follows: C57BL6 = 0.70 ± 0.06 nm, 129Sv = 0.78 ± 0.07 nm, DBA2 = 0.65 ± 0.05 nm.

Three-way univariate repeated measures anova of normalized ventral striatal extracellular DA over the 3-h post-injection period revealed a significant interaction between strains, drug treatment and ventral striatal extracellular DA (F22,598 = 2.170). Similar to morphine-induced locomotor changes, effects on ventral striatal extracellular DA occurred mostly within 3 h of injection (Fig. 3).

Figure 3.

Effect of subcutaneous vehicle (upper panel) or morphine administration (3 mg/kg; lower panel) on DA release in the ventral striatum of C57BL6 (●), 129Sv (○) and DBA2 (●) mice. Data are expressed as a percentage (mean ± SEM) of the mean DA concentration of four pre-drug dialysate samples (saline treatment: C57BL6, n = 9; 129Sv, n = 8; DBA2, n = 10; morphine treatment: C57BL6, n = 12; 129Sv, n = 12; DBA2, n = 12).

Saline administration induced an immediate but transient increase in ventral striatal extracellular DA in all three strains, this being most prolonged in DBA2 mice (Fig. 3). Following the short-lived immediate increase, saline-injected C57BL6 mice showed continually decreasing ventral striatal DA levels though no significant statistical difference was found between these and other vehicle injected strains (p = 0.544 vs. 129Sv, p = 0.129 vs. DBA2).

Subcutaneous injection of morphine induced increases in ventral striatal extracellular DA levels in all animals, the amplitude of which varied according to strain (Fig. 3). C57BL6 mice showed the greatest increase in extracellular DA. This peaked 1.5 h after morphine injection. DBA2 mice showed a similar temporal profile, but the amplitude of the response was approximately half that observed in C57BL6 mice. Although DA levels peaked at approximately the same time following injection in all three strains, 129Sv mice showed a more irregular increase. It is notable that DA levels were approaching pre-injection levels within 3 h of morphine administration in all three strains. When compared with the respective saline-injected control over the same 3-h post-injection period, morphine-induced increases in ventral striatal extracellular DA levels proved to be statistically significant for all three strains studied with C57BL6 mice, inducing a significantly larger increase than that seen in 129Sv and DBA2 mice. No significant differences existed between morphine-induced ventral striatal extracellular DA levels in 129Sv and DBA2 mice (p = 0.366).

Relationship between morphine-induced locomotion and ventral striatal extracellular DA

In addition to showing the greatest locomotor and mesolimbic DA responses following morphine administration, C57BL6 mice also showed the strongest temporal relationship between changes in these two parameters as is apparent when the profiles are displayed together (Fig. 4). Thus, the rate of rise and peak response in the locomotor and mesolimbic DA in C57BL6 mice occurred almost concurrently. In contrast, DBA2 and 129Sv mice appeared to exist at opposite extremes. Whereas the peak locomotor response preceded the peak mesolimbic response in 129Sv mice, the inverse was true for DBA2 mice.

Figure 4.

Comparison of the profile of locomotor activity and DA release in the ventral striatum following subcutaneous morphine administration (3 mg/kg). Locomotor activity is expressed as a mean ± SEM Ventral striatal DA is expressed as a percentage (mean ± SEM) of the mean DA content of four pre-drug dialysate samples (locomotion: C57BL6, n = 15; 129Sv, n = 12; DBA2, n = 12; ventral striatal DA: C57BL6, n = 12; 129Sv, n = 12; DBA2, n = 12).

Intra-strain analysis of the mean morphine-induced locomotor and mesolimbic DA responses during the 3-h period following injection found no significant correlation between these parameters in any of the strains studied when considered as either peak responses in the 3-h period following morphine administration (Fig. 5), or as mean responses in the same period (data not shown).

Figure 5.

Intra-strain correlation between subcutaneous morphine-induced (3 mg/kg) ventral striatal DA release and locomotor activity. Ventral striatal DA is expressed as the peak increase over basal levels in the 3 h following morphine administration. Locomotor data is expressed as the peak activity in the same period of time. Broken lines represent linear fits.

Inter-strain analysis also failed to show a significant correlation between morphine-induced locomotion and mesolimbic DA release whether considered as peak mean morphine induced responses (r2 = 0.952, p = 0.140), mean (over 3-h post-drug period) morphine induced responses (r2 = 0.917, p = 0.186) or mean (over 3-h post-drug period) morphine induced responses corrected for saline effects (r2 = 0.937, p = 0.160).


Despite several decades of intensive research, the precise role of the mesolimbic DA system in drug-seeking behavior and reward remains unclear. However, most investigators concur that activation of this system is an important (if not always necessary) event in the neurobiology underlying drug-seeking behavior (Koob and Le Moal 1997; Wise 1998; Di Chiara 1999; Ikemoto and Panksepp 1999; Robinson and Berridge 2000). The relationship between drug reward and locomotion is at times contentious. Drug-induced locomotion has often been considered to be a readily measurable index of drug reinforcement potential, due much to the postulate that forward locomotion is a fundamental aspect of the ‘approach behavior’ characteristic of drug-seeking. The neural substrates, in particular the mesolimbic DA system, that underlie these processes are believed to be mutual to some extent. Indeed, this mutuality was considered inextricable enough for Wise and Bozarth (1987) to propose the ‘psychomotor stimulant theory of addiction’. This theory set forth the hypothesis that locomotor activation is the ‘common denominator’ in the behavioral response to many addictive substances, with the mesolimbic DA system implicated as the probable neural substrate of these two behaviors. Indeed, there is no doubt that a large body of evidence supports this theory. For example, strong correlations between locomotion and mesolimbic DA release induced by amphetamine (Sharp et al. 1987) and food (McCullough and Salamone 1992) have been noted in the rat. Studies comparing outbred rats find that the strains requiring the highest doses of morphine to induce a conditioned place preference also require the highest dose of morphine to activate the mesolimbic DA system (Shoaib et al. 1995). However, there is also substantial evidence that mesolimbic DA may not be involved in drug-seeking. For example, neurotoxic lesions of the ventral striatum variably produce reductions in morphine-induced place preference (Kelsey et al. 1989; Shippenberg et al. 1993; Olmstead and Franklin 1996) and self-administration (Pettit et al. 1984). Furthermore, animals appear to be quite capable of responding to natural rewards in the absence of a functioning mesolimbic system (Caine and Koob 1994; see Salamone 1994 for review).

The C57BL6 and DBA2 mice strains have been employed in the research of abused drugs for several decades (Crawley et al. 1997). Prior to the development of transgenic and knockout technologies, inbred strains served as (and arguably continue to serve as) some of the best genetic models of drug abuse susceptibility available. The wide variability that inbred strains show in their locomotor response to morphine has been noted countless times. In this respect, the data presented in the current study concur well with others showing firstly that differences in basal locomotion exist between strains (Castellano and Oliverio 1975), and secondly that DBA2 (Oliverio and Castellano 1974a,b; Brase et al. 1977; Belknap et al. 1989, 1998; Wenger 1989) and 129Sv (Belknap et al. 1998) mice exhibit weaker locomotor responses to morphine than C57BL6 mice.

The possibility of inter-strain variations in the bioavailability, potency, and efficacy of morphine at opioid receptors are worthy of consideration before any conclusion on the source of differences locomotor and mesolimbic activating effects are drawn. Although reports vary, up to 70% higher concentrations of morphine have been found in brains of C57BL6 compared with DBA2 mice following peripheral administration (Brase et al. 1977; Gwynn and Domino 1984; Belknap et al. 1989). Yet studies covering large numbers of varied strains draw no correlations between brain or plasma morphine concentration and morphine sensitivity (Belknap et al. 1998). However, differences in the availability of morphine metabolites may also be worthy of consideration (Wahlstrom et al. 1986).

Consideration must also be given to the dose of morphine administered. A dose of 3 mg/kg was chosen in the current study based on the fact that this dose is both effective and submaximal in conditioned place preference studies of morphine reward (Cunningham et al. 1992; Semenova et al. 1995; Maldonado et al. 1997). However, an analysis of a wide range of morphine doses would be optimal and this must be borne in mind when drawing conclusions from the data presented here.

Numerous investigators have sought neurochemical explanations for the inter-strain variability in basal and morphine-induced locomotion in the mouse. The current study reflects on these investigations in order to compare the action of morphine on mesolimbic activity across select inbred mouse strains. The reader is referred elsewhere for a complete review of the early literature (Oliverio and Castellano 1981). Evidence for a role of the mesolimbic system itself in opiate-induced locomotion transpires from a number of sources. For example, locomotion can be stimulated by direct administration of opiates into the ventral tegmental area (Kelley et al. 1980; Stinus et al. 1980; Joyce et al. 1981). Furthermore, such opiate-induced locomotion is attenuated by either 6-hydroxydopamine lesions of the mesolimbic system (Kelley et al. 1980; Stinus et al. 1980) or administration of a DA receptor antagonist (Joyce and Iversen 1979). DA receptor antagonists also block the hyperlocomotory effects of systemically administered morphine (Longoni et al. 1987). On the other hand, some investigators have reported that 6-hydroxydopamine or electrolytic lesions of the mesolimbic system fail to attenuate systemic opiate-induced locomotion (Stevens et al. 1986; Vaccarino et al. 1986). One explanation may be the existence of a DA-dependent and a DA-independent mechanism via which opiates induce hyperlocomotion (Kalivas et al. 1983).

As far as we are aware, the current study is the first to make a direct comparison between morphine-induced mesolimbic DA release and locomotion in inbred mice strains. The principal issue that we wished to address was if the variability seen in morphine-induced locomotor responses (in the horizontal plane only) across inbred strains is reflected in the activation of the mesolimbic DA system. By addressing this question, we sought evidence for mesolimbic dopamine as a determinant of morphine-induced locomotion. As an extrapolation of this, we wished to make inferences on the validity of the psychostimulant theory of reward.

No differences in basal dopamine levels were found in the ventral striatum of the three mouse strains studied. A no-net-flux determination of the basal extraceullular dopamine levels was not made in the current study. However, previous studies in the same three strains suggest dialysate dopamine concentrations reliably reflect extracellular dopamine levels in the striatum (a closely related structure) when determined by no-net-flux (He and Shippenberg 2000).

The data obtained in the current study ostensibly suggest a relationship exists between the magnitude of morphine-induced locomotion and morphine-induced mesolimbic DA activation across inbred strains. This is true inasmuch as C57BL6 mice show the largest increase in both locomotion and mesolimbic activity following morphine administration. However, two observations made in the current study may argue against mesolimbic DA as underlying morphine-induced locomotion.

First, no correlation between the magnitude of the mean morphine-induced locomotor and mesolimbic DA activation in the 3 h following administration is evident within strains. (In addition, no correlation was found between basal locomotion and basal dialysate DA concentrations, data not shown.) Thus, those animals within any particular strain that showed the largest locomotor response to morphine are not necessarily the same ones that show the largest mesolimbic DA activation. However, several factors may affect the determination of this correlation. In particular, the precise placement of the microdialysis probe in not just the anterior and posterior direction, but also mediolateral and dorsoventral coordinate may be critical. Studies by Di Chiara (1999) suggest a regional difference in the sensitivity of DA output between the core and shell of the nucleus accumbens following administration of abused drugs. Thus, intravenous administration of doses of morphine to rats that normally maintain self-administrative behaviors induce a stronger activation of mesolimbic DA activity in the shell than the core region (Pontieri et al. 1995). Distinction between these two subdivisions was not possible to make in the current study as the 2 mm dialysis membrane employed is capable of spanning both these neuroanatomical regions.

Second, the relative timing of the profiles of locomotor and mesolimbic DA activation varies between strains. Whereas the peak mesolimbic response to morphine is seen prior to the peak locomotor response in DBA2 mice (although it is arguable whether DBA2 mice show a locomotor response at all), the opposite is true for 129Sv mice. Only C57BL6 mice show synchronous profiles in the activation of these two parameters. If mesolimbic DA underlies morphine-induced locomotion, then all the strains studied would be likely to show similar levels of synchrony between morphine-induced locomotion and mesolimbic DA activity. However, conflicting effects of morphine on distinct neural systems subserving locomotion may be possible. For example, increases in mesolimbic DA may stimulate forward locomotion in DBA2 mice, yet some suppressive effect of morphine on a downstream system may prevent expression of this effect. A similar argument may be applicable to the sensitivity of pathways efferent to mesolimbic DA neurons to DA in all of the strains studied (although the lack of correlation between morphine-induced mesolimbic DA release and locomotion within any particular strain would argue against this). Thus, mesolimbic DA may underlie morphine-induced locomotion, but differences between strains in the sensitivity of the efferent systems to mesolimbic DA prevents expression of increases in DA release as locomotion in some strains. However, if this were true then one ought to see corresponding differences between strains in hyper-locomotion induced by psychostimulant drugs. However, this is not the case. For example, cocaine induces higher levels of locomotion in DBA2 than C57BL6 mice (Rocha et al. 1998), though it is currently unknown if DBA2 and C57BL6 show the same mesolimbic response to this drug.

Clearly, if locomotor activation is chosen as an index of mesolimbic DA activity, then establishing a recognizable correlation between these two parameters is a fundamental pre-requisite. This proposal holds true not just for opiates, but all stimuli and drugs capable of activating the mesolimbic DA system.

Unfortunately, the ‘rewarding’ properties of any given drug (in inbred mice strains at least) appear to depend heavily on the paradigm used in their determination. For instance, DBA2 mice show stronger morphine conditioned place preferences (Cunningham et al. 1992; Semenova et al. 1995) and more robust intravenous morphine self-administration than C57BL6 mice (Semenova et al. 1995). However, C57BL6 mice willfully drink much larger quantities of morphine solutions than DBA2 mice, which will barely drink any (Horowitz et al. 1977; Belknap et al. 1993; Berrettini et al. 1994). Clearly, multiple factors may affect the detection of reward behaviors, i.e. assay used, route of administration, voluntary versus involuntary administration, etc. There is little doubt that the method and route of drug administration used in the present study is most akin to that used in a conditioned place preference paradigm inasmuch as mice receive the drug involuntarily. Thus, the data presented here suggest that morphine-induced elevations in mesolimbic DA release are inversely correlated with the magnitude of the reward experienced when assessed by the conditioned place preference paradigm. That is, C57BL6 mice showed the largest increase in mesolimbic DA release following morphine administration yet these animals are reported to show weaker morphine-induced conditioned places preferences (Cunningham et al. 1992; Semenova et al. 1995). Likewise, DBA2 mice, which showed small mesolimbic DA responses, are known to display stronger morphine-induced place preferences than C57BL6 mice (Cunningham et al. 1992; Semenova et al. 1995). However, crucial differences between investigator- and self-administered drugs on mesolimbic DA release must be considered. For instance, mesolimbic DA release is induced in the rat by doses of heroin that are ineffective when self-administered, suggesting that the mesolimbic system is more sensitive to involuntarily administered drugs (Hemby et al. 1995). Furthermore, only a single drug administration was made in the current study, whereas typically multiple drug administrations are made in conditioned place preference studies.

A further consideration may be the role that the size of the arena plays in determining drug responsiveness. Previous studies suggest that morphine experienced in small areas induces a stronger place preference and higher levels of activity, possibly by preserving the novelty of the testing environment (Vezina and Stewart 1987). Furthermore, phenomena such as locomotor sensitization are known only to occur in arenas above a critical size (Kuribara 1997). However, the environment presented to the animals in the present study is unlikely to be considered particularly novel at the time of drug administration, as they were acclimated to it during the previous night.

As far as the authors are aware, the rewarding properties of morphine in 129Sv mice are yet to be reported. 129Sv mice are known to acquire cocaine conditioned place preference less readily than C57BL6 mice (Miner 1997). Furthermore, 129/OlaHsd mice are unwilling to nose-poke for cocaine (Kuzmin and Johansson 2000). Clearly this area of investigation warrants detailed examination in light of the data presented here, and the wide use of this strain in the generation of knockout animals.

In summary, we have shown that variations in morphine-induced locomotion between, and particularly within, different inbred mouse strains do not necessarily correlate with variations in mesolimbic DA response. Thus, drug-induced activation of the mesolimbic DA system may not always be synonymous with increases in locomotor activity (e.g. DBA2 mice). In addition, differences in morphine-stimulated mesolimbic activity do not correlate positively with previously reported differences in morphine reward potential between inbred strains as assessed in the conditioned place preference paradigm. This disagrees somewhat with the psychomotor stimulant theory of addiction (Wise and Bozarth 1987). We caution against the use of locomotor activity as an index of mesolimbic activity in response to morphine, unless the relationship between these two parameters is carefully established for that particular species and strain. Furthermore, the variability seen in these parameters between strains, e.g. C57BL6 versus 129Sv should be considered when analyzing data obtained by crossing them.


This work was supported by NIDA grants #DA05010 and #DA09359.