Introduction

Abstract Introduction Methods Results Discussion References

Cocaine (Crow, 1970; Esposito et al., 1978; Frank et al., 1992; Kokkinidis and McCarter, 1990) and several other habit-forming drugs (Gardner 1992; Wise 1980, 1996; Wise and Rompré, 1989) are not only rewarding² in their own right; they also increase lever-pressing that is maintained by the rewarding effects of lateral hypothalamic electrical stimulation. When tested in the "curve-shift" paradigm (Frank et al., 1987; Gallistel, 1987; Maldonado-Irizarry et al., 1994; Miliaressis et al., 1986) these drugs increase the potency of rewarding stimulation, shifting to the left the functions relating response rate to the frequency, intensity or train length (Frank et al., 1987; Gallistel and Karras, 1984; Gallistel and Freyd, 1987). The curve-shift paradigm offers what is essentially a traditional dose-response analysis of brain stimulation reward (Liebman, 1983) in which it is inferred that leftward or rightward shifts reflect treatments that increase or decrease, respectively, the rewarding potency of the stimulation (Edmonds and Gallistel, 1974). Upward or downward shifts, on the other hand, are caused by changes in the response demands of the task (Edmonds and Gallistel, 1974; Stellar and Neeley, 1982); such shifts, when seen after drug or other experimental treatments and in the absence of changes in response demands, are interpreted as reflecting treatment-induced changes in the response capacity of the animal (Edmonds and Gallistel, 1974; Stellar and Neeley, 1982).

Psychophysicists of brain stimulation reward (Edmonds et al., 1974; Gallistel 1978; Gallistel et al., 1981; Miliaressis et al., 1986) have argued that the most useful metric of drug-induced changes in the rewarding potency of brain stimulation is the amount of stimulation that must be given to offset a given drug treatment and restore the level of an animal's performance to baseline. The emerging consensus (Gallistel 1987; Gallistel and Freyd 1987; Miliaressis et al., 1986; Wise and Rompré, 1989) is that stimulation frequency (the number of stimulation pulses per reward) is the most useful stimulation parameter in drug studies. Rate-frequency determinations have been used to quantify the reward-enhancing effects of systemic amphetamine (Gallistel and Freyd, 1987; Gallistel and Karras, 1984; Wise and Munn, 1993), morphine (Carlezon and Wise 1993b), nicotine (Bauco and Wise, 1994) and phencyclidine (Carlezon and Wise, 1993a), and of central injections of amphetamine (Colle and Wise, 1988) and morphine (Bauco et al., 1993). In the present investigation this approach was used to assess the effects of cocaine across the full range of its effective doses.

In addition, the effects of repeated dosing were assessed by use of this metric. Although it has been widely believed that there is tolerance to the habit-forming effects of addictive drugs, sensitization has been shown to develop to some effects of amphetamine (Piazza et al., 1990) and cocaine (Horger et al., 1990, 1992). Although it is difficult to quantify progressive changes in the rewarding efficacy of drugs in drug self-administration (see, e.g., Gerber and Wise, 1989; Winger et al., 1989) and conditioned place-preference (Bozarth, 1987) paradigms, several drugs augment the rewarding effects of brain stimulation and this effect usually predicts the drug's rewarding effects (Wise, 1996). Studies of the effects of cocaine by use of rate (Kokkinidis and McCarter, 1990) or rate-duration measures (Frank et al., 1988) suggest that there is no sensitization and little, if any, tolerance to the reward-enhancing effects of medial forebrain bundle electrical stimulation. The present study was designed to examine these questions with the rate-frequency curve-shift paradigm.

	Methods

Abstract Introduction Methods Results Discussion References

Subjects. Forty-five 300- to 350-g male Long-Evans rats (Harlan Sprague Dawley, Indianapolis, IN) were used. They were housed individually in polyethylene cages with wood chip bedding and free access to food and water. Lighting was maintained on a normal 12-h light/dark cycle; the animals were tested at the end of the dark phase.

Surgery. Each animal was implanted bilaterally with monopolar stainless steel electrodes (0.25-mm diameter) insulated with varnish (Formvar), except for the rounded tip, under sodium pentobarbital anesthesia (65 mg/kg i.p.). Atropine (0.6 mg/kg i.p.) was administered 20 min before the anesthetic to minimize bronchial secretions,. Flat-skull coordinates were 2.5 mm posterior to bregma, 1.7 mm lateral to the midline and 8.0 mm ventral to the skull surface. Four stainless steel screws were used to anchor each electrode; the screws were wrapped with uninsulated wire that served as an anode. The electrode and screw assembly was embedded in dental cement.

Apparatus. Stimulation was controlled by a microprocessor-based system that controlled the delivery of stimulation via a constant current generator (Mundl, 1980) connected to each animal by flexible leads through a mercury commutator that allowed the animal free movement. Each animal was tested in a 26 × 26-cm cage with an operant lever protruding 2.5 cm from the rear wall at a height of 7.5 cm from the floor. The operant lever controlled a microswitch connected to the current generator. Each test cage was enclosed within a larger wooden chamber that reduced external noise and visual distractions.

Procedure. Self-stimulation testing began 7 days after surgery. The animals were placed in test boxes and allowed to explore without experimenter-administered stimulation. Each initially investigatory depression of the operant lever resulted in delivery of a 200-msec train of 0.1-msec rectangular cathodal pulses to the selected electrode. Each stimulation train was followed by an 800-msec "time-out" period when responding was not reinforced; the purpose of the time-out was to allow significant decay (Black et al., 1985) of proactive post-pulse "priming" effects of earned stimulation (Gallistel et al., 1974), which thus increased the sensitivity of the paradigm to changes in drug-induced response-reinforcement per se (see "Discussion"). Initial stimulation frequency was 72 Hz, and initial current intensity was 100 µA (a low level). If the animal did not learn to lever-press for stimulation at this intensity the current was increased daily in 50-µA increments until the animal learned to lever press reliably or until the current intensity reached an upper limit of 800 µA. Once a current level was reached that established steady responding at a minimum rate of 30 lever presses per minute, the current was fixed and the animal was allowed to lever press freely for 1 h/day on 3 consecutive days. If the animal did not continue lever-pressing or if the stimulation produced aversive side effects (e.g., gross head or body movements to one side, spinning, retreating to a corner of the test cage, vocalization or jumping) the animal was retested with use of the contralateral electrode. One or two test sessions were sufficient to produce reliable lever-pressing in each of the animals reported in the present experiments; no animals were dropped from the present experiment because of failure to learn the task or to meet this criterion for stability.

Confirmation of electrode placements. At the end of testing each animal was anesthetized and a 1.5 mA anodal current was passed for 15 sec through the stimulating electrode. The animals were then perfused with physiological saline followed by 10% formaldehyde. Next, the brains were immersed in a formalin-cyanide solution (10% formaldehyde, 3% potassium ferrocyanide, 3% potassium ferricyanide and 0.5% trichloroacetic acid) for 15 min. This solution forms a blue reaction product with iron particles that have been expelled by the anodal current into the tissue around the electrode tip. After storage for at least 1 week in 10% formaldehyde the brains were frozen, sliced into 40-µm sections and stained with thionin for determination of electrode placements relative to the drawings and coordinates of Swanson (1992).

Estimation of self-stimulation threshold. Self-stimulation frequency thresholds were defined as the minimum stimulation frequency required to sustain lever pressing at greater than chance rates. Because of the greater variability of responding at stimulation frequencies near threshold levels, threshold estimates involved extrapolation and curve fitting based on the more reliable range of stimulation frequencies. A regression line was fitted to the data points for the frequencies estimated, by interpolation, to sustain responding at 20%, 30%, 40%, 50% and 60% of maximum, and threshold was estimated as the point at which this regression line crossed the abscissa (Miliaressis et al., 1986).

Drug. Cocaine hydrochloride was dissolved in sterile physiological saline and administered i.p.; dosage is expressed as the salt.

	Results

Abstract Introduction Methods Results Discussion References

Cocaine caused leftward and upward shifts of the functions relating the logarithm of stimulation frequency (stimulation "dose": Yeomans, 1975) to response rate in each of the six groups of tested animals. The leftward and the upward shifts were each dose-orderly (fig. 1) and were observed in each group of tested animals. There were no significant differences (F_7,49 = 0.84, P > .55) in the slopes of the rising portions of the curves (estimated for each animal before averaging by computer algorithm from the threshold estimating procedure). A tendency to perseverate when receiving low stimulation frequencies, well characterized in the case of amphetamine (Olds and Travis, 1970), was seen with the higher cocaine dosages. The leftward shifts in the curves were reflected in a significant decrease in threshold as a function of cocaine dosage (fig. 2; F_7,49 = 12.92, P < .01); the upward shifts were reflected in a significant increase in maximum rate as a function of dosage (F_7,49 = 6.79, P < .01). At the 16 mg/kg dose (the dose producing the largest average decrease in threshold) cocaine reduced the threshold frequency by 0.47 log units, which represents a 3.1-fold increase in the rewarding potency of the stimulation.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Mean lever-pressing rate during the first 15 min after cocaine injection as a function of stimulation frequency and cocaine dosage. Frequency data were transformed to log difference-from-baseline values for each animal before averaging so that the slope of the mean rate-frequency functions would not be contaminated by between-animal differences in threshold.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Mean (± S.E.) self-stimulation frequency threshold (expressed as percent of baseline) as a function of cocaine dosage. Values are the means from the first threshold determination in the first hour after injection. For this and subsequent figures, each reference (baseline) value is the mean from the two threshold determinations taken just before the respective drug test.

No significant day-to-day changes in predrug (baseline) responding occurred in any of the six groups of animals.

Time courses of the effects of cocaine at various dosages are shown in figure 3. Thresholds were decreased (and maxima increased: not shown) for 1 to 2 h with the lower doses and for approximately 3 h with the higher doses. Mean thresholds for the first two determinations (first hour) were used to compare cocaine effects across doses.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Mean (± S.E.) self-stimulation frequency threshold as a function of time after injection with cocaine at different doses. The animals were tested after injections of the cocaine vehicle (saline) on the days preceding each of the seven drug tests; saline values are means across these seven vehicle tests.

The mean thresholds for the first two determinations (first hour after injection) were also used to compare cocaine effects (4 or 16 mg/kg) across days. There were no significant differences in the effects of cocaine on self-stimulation thresholds across days (fig. 4, top panel; Treatment × Days interaction: F_4,76 = 0.44, P > .05) or maximum response rates (fig. 4, bottom panel, F_4,76 = 0.59, P > .05). Moreover, thresholds were lowered, and maximum rates elevated, significantly in the two groups of animals injected with 16 mg/kg compared with the group injected with 4 mg/kg cocaine (main effect of Treatment: F_2,19 = 42.21, P < .05; F_2,19 = 6.28, P < .05, respectively).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Mean (± S.E.) self-stimulation frequency threshold (top) and maximum response rate (bottom) on successive days of testing for the first hour after repeated cocaine at 4 or 16 mg/kg. Circles (16 mg/kg group) and triangles (4 mg/kg group) indicate data for animals tested with saline (filled symbols) on days intervening drug tests; squares indicate data for animals kept in home cages on days intervening drug tests. Cocaine caused significant decreases (P < .01) in threshold and significant increases in maximum rates; there were no statistically significant Treatment × Days interactions.

In the comparison between thrice-daily and once-daily injection regimens, no significant differences occurred in the threshold-lowering effects of 10 mg/kg cocaine [when days 2 and 3 (20 mg/kg dose) are excluded from the analysis; fig. 5, top panel; F_10,60 = 0.79, P > .05]. Although maximum rates were significantly elevated in the cocaine group throughout the thrice-daily injection regimen this effect decreased markedly with once-daily injections (fig. 5, bottom panel; F_10,60 = 2.98, P < .01). The means of three threshold determinations were used for the comparisons across days. Moreover, there were no significant differences between groups during the once-daily injection regimen in the threshold-lowering effect of 10 mg/kg cocaine (F_1,6 = 0.00, P > .05) nor in the ability of cocaine to increase maximum rates (F_1,6 = 0.02, P > .05).

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Mean (± S.E.) self-stimulation frequency threshold (top) and maximum response rate (bottom) on successive days of testing for the first 45 min during a thrice-daily (filled symbols) and once-daily (open symbols) injection regimen. Cocaine [circles, 10 mg/kg (and 20 mg/kg, dashed box)] caused significant decreases in threshold and increases in maximum rates (P < .01). The effect of 10 mg/kg cocaine on threshold did not vary significantly across the treatment days.

All electrode tips were within the boundaries of the medial forebrain bundle at the level of the lateral hypothalamus (fig. 6).

View larger version (78K): [in this window] [in a new window]   Fig. 6.   Histological localization of stimulating electrode tips. The number beside each brain slice represents the distance from the bregma. Reconstructions based on the stereotaxic atlas of Swanson (1992). All electrode tips are shown in the left hemisphere to facilitate comparisons.

	Discussion

Abstract Introduction Methods Results Discussion References

As shown with several other habit-forming drugs (Wise, 1996), cocaine decreased brain stimulation reward thresholds by producing parallel leftward shifts in the functions relating response rate to stimulation frequency. These findings are consistent with findings from rate-duration (Frank et al., 1988, 1992) and rate-intensity (Kokkinidis and McCarter, 1990) paradigms and extend earlier studies to characterize the full range of reward-potentiating cocaine dosages. Inasmuch as the stimulation frequency determines the "dose" of rewarding stimulation per lever-press (Yeomans, 1975), the leftward shifts in the rate-frequency functions are equivalent to leftward shifts in the brain stimulation reward "dose-effect" function (Liebman, 1983; Wise, 1996). Thus cocaine increases the rewarding potency of the stimulation rather than the response capacity (Edmonds and Gallistel, 1974; Stellar and Neeley, 1982) of the animals. We interpret this finding as reflecting a summation between the rewarding effects of the drug and the rewarding action of lateral hypothalamic stimulation (Wise et al., 1992; Wise, 1996).

The present findings suggest that the threshold dose of cocaine for potentiating brain stimulation reward is approximately 2 mg/kg. This dose produced noticeable but marginally significant results in the first half-hour after injection. Maximal effectiveness was seen with the 16 mg/kg dose; this dose produced an average 65% reduction in the stimulation frequency required to sustain minimal responding. The 32 mg/kg dose produced longer lasting potentiation but did not appear to produce a greater magnitude of potentiation.

In the present experiment an 800-msec "time-out" was given after each 200-msec train of rewarding stimulation; this corresponds to a fixed interval (1-sec) schedule of reinforcement and was used to minimize the rapidly decaying (Gallistel, 1969; Gallistel et al., 1974; Reid et al., 1973) proactive "priming" effects of each earned stimulation train. There are at least two sets of consequences of stimulation that contribute to the control of rate of responding: the priming effect (Deutsch and Howarth, 1963; Gallistel, 1969; Gallistel et al., 1974; Reid et al., 1973) and the reinforcing effects (Deutsch and Howarth, 1963; Gallistel et al., 1974); it is for this reason that brain stimulation reward specialists tend to use the term "reward," implying the sum of the two factors, rather than the term "reinforcement," which does not reflect the contribution of the priming effect (Wise, 1989).

It is not clear whether cocaine augmented the rewarding effects of the stimulation by summating with the priming effect or one of the reinforcing effects of the stimulation. Cocaine clearly can augment the priming effect of stimulation (Esposito et al., 1978). The priming (Esposito et al., 1979; see also Wasserman et al., 1982)³ and the reinforcing (operant: Fouriezos and Wise, 1976; Franklin, 1978; Pavlovian: Ettenberg and Duvauchelle, 1988) effects of stimulation are all attenuated by dopamine antagonists, as are both the incentive-motivational (Spyraki et al., 1987) and reinforcing (operant: de Wit and Wise, 1977; Pavlovian: Spyraki et al., 1987) effects of cocaine. Inasmuch as cocaine was given independent of the behavior of the animal, it did not, by definition, qualify as an operant reinforcer (Skinner, 1937). Inasmuch as its effects were proactive (they were reflected in behavior that occurred after drug administration and absorption) it is reasonable to assume that the incentive-motivational (priming) effect of cocaine summated with the rewarding effect of the stimulation. That the priming effects of cocaine should summate with the rewarding effects of stimulation is consistent with the hypothesis that the neural substrates of incentive motivation and operant reinforcement are homologous (Bindra, 1972; Wise, 1989; Wise and Bozarth, 1987). However, we cannot rule out the possibility that cocaine could amplify the sensitivity of the reward system and thus have augmented brain stimulation reward by amplifying the response-contingent effects of the stimulation.

The hypothesized common denominator of incentive-motivation and reinforcement involves the brain mechanisms of forward locomotion, the common response to all positive reinforcers (Glickman and Schiff, 1967; Perkins, 1968; Schneirla, 1959; Wise and Bozarth, 1987). However, the effects of cocaine and other psychomotor stimulants show sensitization or "reverse-tolerance" with repeated administration (Babbini and Davis, 1972; Post and Rose, 1976; Stripling and Ellinwood, 1976), whereas the reward-potentiating effects of cocaine in the present experiment clearly did not. This finding is consistent with most previous reports of the effects of repeated treatment with cocaine (Frank, et al., 1988,1992), amphetamine (Wise and Munn, 1993; but see Predy and Kokkinidis, 1984), morphine (Bauco et al., 1993; Schenk et al., 1981), nicotine (Bauco and Wise, 1994) and phencyclidine (Carlezon and Wise, 1993a) on brain stimulation reward. While it might be argued that since brain stimulation reward experience is capable of sensitizing animals to amphetamine (Ben-Shahar and Ettenberg, 1994) we may have maximally sensitized our animals during brain stimulation reward training. However, the reward enhancing effects of amphetamine fail to show sensitization even in animals with extensive brain stimulation reward experience; indeed, sensitization to the locomotor-stimulating effects of amphetamine have been demonstrated in the same animals and in response to the same amphetamine injections that potentiated brain stimulation reward but failed to show sensitization of this reward-potentiating action (Wise and Munn, 1993). These findings would appear to falsify the narrow hypothesis (Wise and Bozarth, 1987) that the brain mechanisms of forward locomotion are homologous with the brain mechanisms of the drug potentiation of brain stimulation reward. Against this view, and also inconsistent with the hypothesis that the reward-potentiating and direct rewarding effects of cocaine are homologous, is the report that the direct rewarding effects of cocaine do undergo sensitization with repeated intoxication (Horger et al., 1990). Insofar as the mechanisms of forward locomotion and reward appear to have dopaminergic modulation and the region of the nucleus accumbens in common, and insofar as dopamine modulates multiple sets of parallel circuits in this part of the brain (Alexander and Crutcher, 1990), the present data suggest that it may be different subsets of cortico-striatal-thalamo-cortical subcircuitry that play roles in the two closely associated phenomena.

We cannot explain why maximum rates of responding during intoxication with 16 mg/kg of cocaine differed depending on whether the animals were given saline tests on the days intervening cocaine or were left in their home cages without testing on the intervening days. These findings suggest that the maximum response rate may be a function of more than simple motoric capacity, but do not suggest what other factors might be involved. Evidence that left-right and up-down shifts in rate-frequency functions are independent of one another (Edmonds and Gallistel, 1974; Rompré and Wise, 1989; Stellar and Neeley, 1982) is supported by the fact that cocaine-induced threshold shifts were not affected by whatever differentially altered cocaine-induced response maxima.

The present data are inconsistent with the commonly held notion that there is necessarily profound and long-lasting tolerance to the habit-forming actions of drugs of abuse. With the exception of acute or "within-session" tolerance (LeBlanc et al., 1975), where tolerance to the rewarding effects of cocaine have been clearly demonstrated (Emmett-Oglesby and Lane, 1992; Emmett-Oglesby et al., 1993; Fischman et al., 1985), recent evidence generally disputes the assumption of tolerance to the specific rewarding actions of both psychomotor stimulants and opiates. Prior experience with amphetamine (Piazza et al., 1990) or cocaine (Horger et al., 1990) is reported to increase rather than decrease the rewarding effectiveness of subsequent amphetamine or cocaine, reducing the threshold dose to establish an operant response habit or reducing the amount of training necessary to establish stable operant performance. Moreover, amphetamine and nicotine are reported to cross-sensitize rats to the reinforcing effects of cocaine (Horger et al., 1992). Prior amphetamine or morphine experience has also been reported to sensitize animals to the ability of amphetamine, morphine or cocaine to establish conditioned place preferences (Lett, 1989). Although the present data do not suggest sensitization to the reward-enhancing effects of cocaine, they also do not support the suggestion that there is between-session tolerance to these effects.

It is somewhat surprising that the present data also fail to offer any evidence for acute tolerance to the reward-enhancing effects of cocaine; inasmuch as acute tolerance has been shown for the direct rewarding effects of cocaine, the direct rewarding and reward-enhancing effects of cocaine have been argued to involve a common reward mechanism in the brain (Wise, 1996; Wise and Bozarth, 1987). An obvious factor is dosage. The treatment regimen in which acute tolerance has been shown in rats (Emmett-Oglesby and Lane, 1992; Emmett-Oglesby et al., 1993) is 20 mg/kg, thrice-daily for 7 days. This dosage regimen proved fatal for some of our animals, perhaps because the first dose each day was given during sessions of intracranial self-stimulation. Our failure to observe acute tolerance involved half the dosing regimen of Emmett-Oglesby et al. (1993). Still, when their animals were tested at several time points after the final injection, Frank et al. (1988, 1992) found no evidence of acute tolerance in animals treated with 25 or 30 mg/kg thrice- daily for 3 days. Tolerance is generally assumed to reflect drug-opposite neuroadaptations and dopamine depletion (Dackis and Gold, 1985), and elevated brain stimulation reward thresholds (Leith and Barrett, 1976) have been suggested as reward-relevant consequences of such neuroadaptations. Evidence of elevated self-stimulation thresholds has been found when animals are allowed to self-administer cocaine (Markou and Koob, 1991), and in this case the animals received approximately the same daily dose of cocaine (on average, 27 mg/kg/day) as given in the present study. In Markou and Koob's study, however, the drug was given intravenously every few minutes rather than intraperitoneally every 8 h as in the present study. Thus the Markou and Koob regimen involved more continuous intoxication. Markou and Koob saw immediately elevated brain stimulation reward thresholds after as little as 6 h of cocaine self-administration, when their animals were still intoxicated with satiating levels of cocaine. This is surprising because the acute effect of intoxicating doses of cocaine is a decrease in brain stimulation reward threshold and because the increased reward thresholds are assumed to be a rebound consequence of drug intoxication (Solomon and Corbit, 1973), which is expected only after the drug is metabolized and the drug effect wears off. The other known opponent-process neuroadaptation to cocaine, extracellular dopamine depletion (Parsons et al., 1991; Robertson et al., 1991), has been reported 10 days after termination of a cocaine treatment regimen of 20 mg/kg once-daily for 10 days (Parsons et al., 1991), and 7 days after termination of a regimen of 30 mg/kg once-daily for 18 days (Robertson et al., 1991). Compared with the treatment regimens in this study, these treatment regimens involve less daily cocaine in the first case and less chronic intoxication in both cases. Thus it would appear that while acute tolerance to cocaine can occur with high-dose and chronic treatment regimens, it seems not to be a simple consequence of the dopamine depletion reported with dosing regimens more modest than those used in the present study.