Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Lower levels of synaptic serotonin are associated with greater ethanol seeking and intake (Sellers et al., 1992). Synaptic serotonin levels are regulated by serotonin release and serotonin reuptake. Fluvoxamine is a highly selective serotonin reuptake inhibitor (Wong et al., 1983; Jordan et al., 1994). Thus, fluvoxamine might be expected to reduce ethanol-seeking by increasing synaptic serotonin levels as a result of blocking reuptake.

Evidence for the role of serotonin in regulating ethanol consumption includes data showing reduced serotonergic activity in alcoholics, comparisons of strains of rats differing in ethanol consumption, and studies on the effects of serotonergic agents.

There are at least four lines of evidence indicating altered serotonergic function in alcoholics. Alcoholics have lower levels of the serotonin deamination product 5-hydroxyindoleacetic acid in their cerebrospinal fluid (Ballenger et al., 1979). Second, there is a decreased availability of the serotonin precursor tryptophan in alcoholics (Buydens-Branchey et al., 1989). Third, alcoholics have reduced serotonin concentrations in their platelets (Baily et al., 1990), whereas serotonin platelet uptake is increased (e.g., Daoust et al., 1991). Finally, Heinz et al. (1998) found a lower density of serotonin transporters in the raphe nucleus of alcoholics. These studies suggest disruptions of the serotonin function in alcoholics.

There are a number of differences in the serotonergic systems of alcohol-preferring and -nonpreferring rats (see McBride et al., 1992 for a review). For instance, alcohol-preferring rat strains (P and HAD) have lower central nervous system serotonin and 5-hydroxyindoleacetic acid levels than their nonpreferring counterparts (NP and LAD). Additionally, alcohol-preferring P rats have higher densities of 5-HT1A, but not 5-HT2, receptor binding than nonpreferring NP rats.

Preclinical studies provide evidence that increasing synaptic serotonergic throughput decreases ethanol-seeking and consumption. For instance, the selective serotonin reuptake inhibitor (SSRI) fluoxetine selectively decreases ethanol intake (compared with water intake) in a limited access procedure (Gardell et al., 1997; see Murphy et al., 1988 for similar results using an operant procedure). Similarly, the 5-HT1A agonist/5-HT2A antagonist FG5938 selectively decreases ethanol intake (compared with water intake) in a limited access procedure (West et al., 1998) and the 5-HT1A agonist/5-HT2A antagonist FG5974, as well as the 5-HT1A agonist 8-OH-DPAT, decreases operant responding for ethanol at lower doses than saccharin-maintained responding (Roberts et al., 1998). Thus, a number of well controlled studies consistently show that increasing synaptic serotonin throughput decreases ethanol-seeking and consumption.

Studies in heavy drinkers show that under certain conditions citalopram, zimelidine, and viqualine reduced drinking, compared with placebo (Naranjo et al., 1992). Fluoxetine has produced more equivocal decreases in ethanol consumption (Naranjo et al., 1990; Gorelick and Paredes, 1992). Alcoholism treatment studies examining the effects of SSRIs (Kranzler et al., 1995; Naranjo et al., 1995; Kabel and Petty, 1996; Angelone et al., 1998; Pettinati et al., 2000) have had mixed results. These mixed results may reflect a variety of factors: 1) differing efficacy between SSRIs (see Naranjo and Bremner, 1992); 2) loss of efficacy of SSRI treatment over time (e.g., Gorelick and Paredes, 1992; Naranjo et al., 1995); 3) a dose dependence of SSRI effects; 4) floor effects (Kranzler et al., 1995); and 5) differential drug treatment responses between subgroups of alcoholics (e.g., Johnson et al., 2000; Pettinati et al., 2000).

Fluvoxamine is extremely potent at blocking serotonin reuptake, while being much less potent at blocking norepinephrine reuptake and relatively inactive at blocking dopamine reuptake (Wong et al., 1983). Fluvoxamine is one of the most selective drugs at blocking serotonin reuptake, having a ratio of 44 for the (IC₅₀ norepinephrine)/(IC₅₀ 5-HT), compared with ratios of 29, 12, and 10 for fluoxetine, clomipramine and zimelidine, respectively. Fluvoxamine also has a low affinity for -adrenergic, histamine, serotonin, and dopamine receptors (Wong et al., 1983). Fluvoxamine is essentially inactive at concentrations up to 10 µM at muscarinic, opiate, -adrenergic, -aminobutyric acid, and benzodiazepine receptors. Thus, fluvoxamine is an extremely selective blocker of serotonin reuptake. Fluvoxamine increases serotonin levels in the forebrain and raphe nuclei (Bel and Artigas, 1993; Jordan et al., 1994), while minimally affecting norepinephrine levels (Jordan et al., 1994).

Given the effects of fluvoxamine on serotonin reuptake and the central role of serotonin reuptake in regulating synaptic serotonin levels and the relationship between synaptic serotonin levels and alcohol consumption, we decided to examine the effects of fluvoxamine upon operant behavior maintained by ethanol. Because sufficient doses of any drug will reduce responding, we compared these effects of fluvoxamine with the effects of fluvoxamine upon food-maintained behavior.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects

Male Harlan Lewis rats were individually housed in a colony room with an average temperature of 20°C and a 12-h light/dark cycle (rats were trained and tested during the light phase). Upon arrival to our institution, the animals were between 270 and 300 g, and were handled daily and habituated to the vivarium routines for 2 weeks. During this period, food (Purina Rat Chow, Purina, St. Louis, MO) was gradually restricted to approximately 12 g/day for all rats; thus, maintaining final body weights at approximately 335 g. Restricted feeding has been shown to be beneficial to the rat's health (see Pugh et al., 1999) and is sufficient to maintain food-reinforced responding. Water was freely available.

Apparatus

Food maintained behavior and drug discrimination training and testing were conducted in eight operant chambers (Med Associates, St. Albans/Georgia, VT) capable of delivering food reinforcement, constructed of Plexiglas and aluminum, equipped with two response levers, house and lever lights, and a grid floor. Each chamber was enclosed within sound- and light-attenuating boxes, which were equipped with an exhaust fan. These chambers were connected to an IBM-compatible PC via an LVB interface (Med Associates, East Fairfield, VT).

Ethanol self-administration training and testing were conducted in two operant chambers (Camden Instruments, Ltd., London, UK), equipped with two retractable response levers, and capable of delivering liquid reinforcement, connected to an IBM-compatible PC via an LVB interface. Behavioral sessions and data collection were conducted using a software program (Med-PC, Med Associates, St. Albans, VT).

Food-Maintained Responding

Training. Rats were trained to respond for food (45-mg Noyes pellets) under a fixed-ratio (FR) 10 schedule of reinforcement. Under this schedule, when the house light is off, and the stimulus light above the response lever lit, completion of 10 presses on the response lever resulted in the delivery of two 45-mg food pellets, illumination of the house light, turning off the stimulus light over the lever, and the initiation of a 10-s time-out period (during which responding was recorded, but had no programmed consequences). At the end of the 10-s time-out period, the stimulus light above the lever was lit, the house light turned off, and the FR 10 schedule of reinforcement reinstated. Experimental sessions were 10 min long and conducted once each weekday.

Testing. When the rate of responding over the last 5 days was stable, animals were tested for drug-induced changes in responding. Drug tests were conducted on Tuesdays and Fridays; vehicle tests were conducted on Thursdays. Regular training sessions were conducted on Mondays and Wednesdays. Test sessions were 10 min in duration and conducted similarly to training sessions, except that doses of test drugs or vehicle were administered before session onset. Each test compound was examined in all eight rats comprising this group.

Ethanol-Saline Discrimination

Training. Rats were magazine trained and shaped to lever press for food reinforcement until they responded 10 times for each reinforcer (FR 10). Each reinforcement consisted of two 45-mg Noyes pellets. Then, they were trained on a two-choice task to respond on drug- or vehicle-appropriate levers once daily 5 days a week. The position of drug-appropriate levers was randomly assigned among subjects so that it was to the right of the food cup for half the subjects. Animals were administered ethanol (1.2 g/kg) or vehicle intraperitoneally 10 min (see Hiltunen et al., 1989) before session onset. Throughout the session, an FR 10 schedule was in effect so that 10 presses on the appropriate lever delivered two food pellets. Presses on the wrong lever were recorded, but had no programmed consequences. The schedule of drug or saline administration was random, with no more than two consecutive drug or saline trials. To avoid the influence of odor cues left in a chamber by a preceding subject (see Extance and Goudie, 1981), the order in which drug and saline training sessions were conducted for animals trained in the same chamber was randomized. Training sessions were conducted Monday through Friday, and lasted 10 min. Training was continued until animals had reached the acquisition criterion of selecting the state-appropriate lever on at least 8 of 10 consecutive training days. A correct selection was defined as total presses before the first reinforcement being equal to, or less than 14 (i.e., an animal did not press the wrong lever more than 4 times before pressing 10 times on the appropriate lever). Customarily, the median first reinforcement value was 10 once the discrimination was acquired (Järbe and Mathis, 1992).

Testing. After animals had reached acquisition criterion, test sessions were conducted on average 3 times each fortnight; on interim days, training sessions were conducted. Half the test sessions were preceded by a drug training session; the other half were preceded by a saline session. Tests were conducted only if responding during the preceding training sessions had been correct. During testing, animals were reinforced for 10 presses on either lever until 10 min had elapsed or 6 reinforcers had been delivered, whichever occurred first. There was one session per test day. For each dose tested, the percentage of responding on the drug-appropriate position/lever was calculated from the ratio of the number of presses on the ethanol-associated lever to the total number of presses in a test session. Additionally, response rate (responses per second) was calculated. The order of tests was mixed within subsets of test protocols. The first subset examined ethanol and fluvoxamine alone; and the second subset examined ethanol and fluvoxamine in combination. The third subset examined the specificity of the discrimination by examining the ability of ⁹-tetrahydrocannabinol (⁹-THC) and d-amphetamine to occasion ethanol-appropriate responding.

Ethanol Self-Administration

Seven rats were trained to self-administer ethanol using a sucrose-fading procedure (Samson, 1986). Briefly, rats were initially exposed to a concurrent variable-time (VT) 60-s, FR 1 schedule of dipper presentation. In other words, after a variable interval of time that averaged 60 s, the dipper was presented. The dipper was also presented following each response on the left lever. Responses on the right lever had no programmed consequences. The experimental sessions were initially 180 min long, dipper presentations were 30 s, and dippers contained 5% sucrose (w/v) and 1% ethanol. During the concurrent VT 60-s, FR 1 schedule, the stimulus lights above the response levers were lit and the white noise generator was turned on. During dipper presentations, the stimulus lights above the response levers and the white noise generator were turned off, the house light was lit, and the response levers were inaccessible. Following acquisition of drinking from the dipper, the VT schedule of dipper presentations was discontinued, and the parameters gradually changed. The final values of these parameters were: session length 10 min, FR 10, dipper presentations 10 s, sucrose concentration 0%, and ethanol concentration 8%. Rats were placed in the experimental chambers 10 min before the experimental session began. During this 10-min period, the chamber was dark, no white noise presented, and the response levers inaccessible.

Later, the ethanol concentration was varied. Each ethanol concentration was available for at least 2 weeks, with no upward or downward trends in individual response rates over the most recent week before testing with fluvoxamine began.

Data and Analysis

Food-Maintained Responding. Response rate was the primary measure and was calculated by dividing the number of responses by the time available to make these responses (i.e., responses/second). For some analyses, response rate was converted from responses/second to percentage of control responding by dividing the rate of responding by the mean vehicle control rate for that individual rat. ED₅₀ values were calculated using linear regression upon the mean data points that were monotonic and produced between 15 to 85% of the maximal response (Tallarida and Murray, 1987). The confidence limits (CLs) for these points were determined as suggested by Bliss (1967). ANOVA and regressions were calculated using SYSTAT (version 5.2.1; Systat, Inc., Evanston, IL) run on a Macintosh G4.

Drug Discrimination. Dependent measures include the percentage of presses on the drug-appropriate lever. ED₅₀ values were calculated using linear regression upon the data points that were monotonic and produced between 15 to 85% of the maximal response (Tallarida and Murray, 1987). The CLs for these points were determined as suggested by Bliss (1967).

Self-Administration of Ethanol. Self-administration results were analyzed by examining response rate. Response rate was expressed as percentage of control rate. Response rate was converted from responses/second to percentage of control responding by dividing the rate of responding by the mean vehicle control rate for that individual rat. ED₅₀ values were calculated using linear regression upon the data points that were monotonic and produced between 15 to 85% of the maximal response (Tallarida and Murray, 1987), i.e., those points for which the mean value was between 15 to 85%. This corresponds to doses of 3, 10, and 30 mg/kg fluvoxamine for ethanol concentrations of 4 and 32%, and to doses of 1, 3, and 10 mg/kg fluvoxamine for a concentration of 16% ethanol. For a concentration of 8% ethanol, fluvoxamine doses of 3, 5.6, and 10 mg/kg were used. Although the fluvoxamine dose of 30 mg/kg also met criterion for inclusion, this point was excluded to maintain equal dose spacing in the regression. The CLs for these points were determined as suggested by Bliss (1967).

Drugs

Ethanol was administered i.p. (food maintained responding and drug discrimination learning) or by drinking. Ethanol (95% v/v) was diluted with normal physiological saline (0.9%) to a concentration of 10% (v/v). Thus, injection volume ranged from 2.5 to 20.0 ml/kg; vehicle injections were 10 ml/kg. Ethanol doses were achieved by varying volumes administered, rather than by varying the solution's concentration (Linakis and Cunningham, 1979). For the drinking solutions, ethanol (95% v/v) was diluted with tap water to obtain 4 to 32% (w/v) solutions. All solutions were prepared at least 24 h before training or testing and stored in covered bottles at room temperature. Fluvoxamine maleate [5-methoxy-4'-(trifluoromethyl)valerophenone-(E)-O-(2-aminoethyl) oxime maleate (1:1)] was dissolved in physiological saline and administered i.p. in a volume of 1 ml/kg, 30 min pre-session. Fluvoxamine was a gift from Solvay Pharmaceuticals (Marietta, GA). d-Amphetamine sulfate was purchased from Sigma (St. Louis, MO), dissolved in saline, and administered i.p. in a volume of 1 ml/kg. ⁹-THC, dissolved in ethanol, was provided by the National Institute on Drug Abuse and stored at 20°C until used. Before use, the ethanol was evaporated under a stream of nitrogen; the residue then dissolved in a solution of propylene glycol and Tween 80. Shortly before being used, the solute was diluted with normal (0.9%) saline after the solute had been sonicated for 20 to 30 min. Ethanol was administered 10 min before experimental sessions, fluvoxamine and ⁹-THC were administered 30 min before, and d-amphetamine was administered 15 min before experimental sessions. All doses are expressed in the forms indicated above.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Table 1 shows the data from the sessions following saline injections in rats responding for 8% ethanol. Ethanol maintained responding (i.e., reinforced behavior) as evidenced by substantially more responding occurring on the drug lever than on the other lever. This was true for all seven rats. Little responding occurred during the time-out periods.

Figure 1 illustrates the effects of fluvoxamine on behavior maintained by delivery of either 0.1 ml of an 8% ethanol solution (panel A) or two 45-mg food pellets (panel B). As seen in Fig. 1, fluvoxamine dose dependently decreased both ethanol-maintained behavior and food-maintained behavior.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   A, effects of fluvoxamine behavior maintained by 8% ethanol. B, effects of fluvoxamine on behavior maintained by food. The vertical axis represents response rate as a percentage of control rate, and the horizontal axis represents fluvoxamine dose in milligrams per kilogram on a log scale. Points represent mean values. Vertical bars on the far left of each graph represent the range of the mean control values. Bars around points represent standard errors of the mean. Rats responded under an FR 10 schedule for either ethanol (0.1 ml, 8% w/v ethanol; N = 7) or food (2 × 45-mg pellet; N = 8). Fluvoxamine was given i.p. 30 min before the start of the 10-min session.

The bars on the graph represent the range of the control values. As seen in Fig. 1A, doses of 1 and 3 mg/kg decreased responding to levels just below this range in the ethanol-maintained rats. The rate of responding was at 71% of control level after both of these fluvoxamine doses, and the lowest control value was 72%. Higher doses of fluvoxamine decreased responding in a monotonic manner. All animals responded at doses of 5.6 mg/kg and lower. At a dose of 10 mg/kg, three of the seven subjects emitted no responses during the 10-min session. At a dose of 30 mg/kg, five of the seven subjects emitted no responses during the 10-min session, whereas one subject responded at a rate of 37% of control and the other subject responded at a rate of 89% of control. Clearly, this last outlier significantly influenced the rate reported following 30 mg/kg. The ED₅₀ for the rate-decreasing effects of fluvoxamine was 5.9 mg/kg (4.0-8.7 mg/kg, 95% CL).

Effects of fluvoxamine upon responding on both the drug and the other lever in the presence of the FR stimulus and during the time-out, number of dipper presentations, and absolute response rate are shown in Table 2 for the rats responding for 8% ethanol. As seen in Table 2, fluvoxamine dose dependently decreases the number of FR responses, the number of dipper presentations, and FR response rate. Responding during the time-out periods and on the nonfunctional lever only occurred at very low rates and was little affected by fluvoxamine.

As seen in Fig. 1B, the rate of responding following 1.8, 3.0, and 5.6 mg/kg fluvoxamine ranged between 77 and 84% of control rate, values just below the low end of the range of control values (85%). Doses of 10, 18, and 30 mg/kg fluvoxamine decreased responding to 41, 24, and 41%, respectively, of control. All subjects responded during the 10-min session following doses of 10 mg/kg or less. However, two of the subjects receiving 10 mg/kg responded at extremely low levels. Four of the eight subjects emitted no responses during the 10-min session following 18 mg/kg fluvoxamine. Responding for the other four subjects was at levels of 1, 48, 65, and 76% of control. Following 30 mg/kg fluvoxamine, three of the eight subjects emitted no response during the 10-min session. Responding for the other five subjects was at 17, 52, 69, 92, and 98% of control. The two animals responding at near control levels are clearly responsible for the up-turn in the dose-response curve. The ED₅₀ for the rate-decreasing effects of fluvoxamine was 10.0 mg/kg (8.1-12.2 mg/kg, 95% CL).

Comparison of the two ED₅₀ values for fluvoxamine on behavior maintained by 8% ethanol delivery and on behavior maintained by food pellet delivery indicates that the ED₅₀ for fluvoxamine for ethanol-maintained behavior was lower than the ED₅₀ for fluvoxamine for food-maintained behavior. The food ED₅₀ was not within the 95% confidence limits for the ethanol ED₅₀. Similarly, the ethanol ED₅₀ was not within the 95% confidence limits for the food ED₅₀.

The selective effects of fluvoxamine that we observed were modest. Thus, these selective effects might be limited to the particular ethanol concentration (8% w/v) studied, a concentration that resulted in modest levels of ethanol earned. Thus, we examined the effects of fluvoxamine in these same rats while we varied the concentration of ethanol available from 4 to 32% (w/v). In Fig. 2, increasing the concentration of ethanol available resulted in rats earning increasing amounts of ethanol. The median amount earned at 4% ethanol was 0.18 g/kg, at 8% the median amount earned was 0.31 g/kg, at 16% the median amount earned was 0.59 g/kg, and at 32% the median amount earned was 0.72 g/kg. Thus, at the two higher ethanol concentrations, 16 and 32% (w/v), relatively substantial amounts of ethanol were earned in the 10-min period of availability.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Median amount of ethanol earned by each rat on days before fluvoxamine administration as a function of the concentration of ethanol available for self-administration. The vertical axis represents the amount of ethanol earned in grams per kilogram, and the horizontal axis represents the percentage (w/v) concentration of ethanol available on a log scale. Individual points represent the median for individual rats, and the solid line represents the median of these values.

As can be seen in Fig. 3, on behavior maintained by each of these ethanol concentrations, fluvoxamine produced very similar decreases in responding. The effects of fluvoxamine on ethanol-maintained behavior did not seem to have a clear dependence upon the concentration of ethanol maintaining behavior or the amount of ethanol that was being earned in the experimental session. This can be seen even more clearly in Fig. 4 in which the ED₅₀ values for fluvoxamine are plotted against the ethanol concentration being used to maintain behavior. As seen in Fig. 4, all four of these ED₅₀ values are outside the 95% CL of the fluvoxamine ED₅₀ for food-maintained behavior [5.7 mg/kg (3.3-9.9), 5.9 mg/kg (4.0-8.7), 3.4 (1.6-7.1), and 6.3 (2.4-16.5); ED₅₀ (95% CL) for 4, 8, 16, and 32% ethanol, respectively].

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effects of fluvoxamine on the self-administration of 4% (), 8% (), 16% (), and 32% () ethanol (w/v). The vertical axis represents response rate as a percentage of control rate, and the horizontal axis represents fluvoxamine dose in milligrams per kilogram on a log scale. Points represents mean values. Rats (N = 6-7) responded under an FR 10 schedule for ethanol (0.1 ml per dipper presentation). Fluvoxamine was given i.p. 30 min before the start of the 10 min session.

View larger version (35K): [in this window] [in a new window]   Fig. 4.   ED50 values for fluvoxamine on food-maintained and ethanol-maintained behavior. The vertical axis represents the fluvoxamine ED50 in milligrams per kilogram, and the horizontal axis represents the maintaining event. The stippled area represents the 95% confidence limits of the fluvoxamine ED50 value for food-maintained behavior.

The differential effects of fluvoxamine on ethanol- and food-maintained behavior could potentially reflect the interactions of fluvoxamine and ethanol. In other words, because the rat is exposed to both ethanol and fluvoxamine when we studied the effects of fluvoxamine upon ethanol-maintained behavior, but only to fluvoxamine when we examined the effects of fluvoxamine upon food-maintained behavior, the effects of fluvoxamine and ethanol in combination might explain the seemingly selective effects of fluvoxamine. To examine this possibility, we examined the behavioral effects of combinations of ethanol and fluvoxamine on food-maintained behavior.

The effects of fluvoxamine and ethanol in combination can be seen in Fig. 5. In Fig. 5A, the effects of ethanol in combination with 1 mg/kg fluvoxamine are shown. In Fig. 5B, the effects of ethanol in combination with 3 mg/kg fluvoxamine are shown, and in Fig. 5C, the effects of ethanol in combination with 10 mg/kg fluvoxamine are shown. From observation of the results shown in panels A and B, one can see that pretreatment with either 1 or 3 mg/kg fluvoxamine had little effect on the rate-decreasing actions of ethanol. This is confirmed by statistical analysis. For instance, an ANOVA with rate as an independent variable and ethanol dose (0.4, 0.8, 1.2, and 1.6 g/kg) and fluvoxamine dose (0 or 1 mg/kg) as factors show that combining 1 mg/kg fluvoxamine with ethanol did not significantly alter the rate-decreasing actions of ethanol (ethanol dose, F = 15.68, df = 3,56, P < 0.01; fluvoxamine dose, F = 0.97, df = 1,56, P > 0.10; the interaction between ethanol and fluvoxamine dose, F = 1.28, df = 3,56, P > 0.10). Similar results were obtained with 3 mg/kg fluvoxamine in combination with ethanol (ethanol dose, F = 18.84, df = 3,56, P < 0.01; fluvoxamine dose, F = 2.05, df = 1,56, P > 0.10; the interaction between ethanol and fluvoxamine dose, F = 1.78, P > 0.10).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Effects of ethanol alone and in combination with fluvoxamine upon food-maintained behavior. The vertical axis represents response rate in responses/second, and the horizontal axis represents ethanol dose in grams per kilogram on a log scale. Bars around points represent the standard error of the mean. A, effects of ethanol in combination with 1 mg/kg fluvoxamine. B, effects of ethanol in combination with 3 mg/kg fluvoxamine. C, effects of ethanol in combination with 10 mg/kg fluvoxamine. Open circles are the effects of ethanol and an injection of saline. Filled circles are the effects of fluvoxamine combined with ethanol. The filled circle above "fluvoxamine + 10 ml/kg saline" represents the effects of the dose of fluvoxamine shown in that panel tested in combination with an injection of saline. Fluvoxamine was given 30 min before the start of the session, and ethanol was given 10 min before the start of the session. Both drugs were given i.p. Behavior was maintained under an FR 10 schedule of food presentation during the 10-min session.

As can be seen in Fig. 5C, by examining the filled circle above "10 mg/kg fluvoxamine + 10 ml/kg saline", 10 mg/kg fluvoxamine had substantial rate-decreasing effects when given alone. These effects of fluvoxamine alone need to be accounted for when examining the effects of fluvoxamine and ethanol in combination. Thus, whereas the effects of 0.4 and 0.8 g/kg ethanol in combination with 10 mg/kg fluvoxamine are greater than the effects of these doses of ethanol alone, these effects of ethanol and fluvoxamine in combination are about the same as, or slightly less than, those of 10 mg/kg fluvoxamine alone. The effects of 10 mg/kg fluvoxamine in combination with 1.2 or 1.6 g/kg ethanol are also similar to the effects of fluvoxamine alone. Therefore, ethanol and fluvoxamine in combination clearly do not synergistically decrease responding.

The effects of ethanol and fluvoxamine alone and in combination in rats discriminating 1.2 g/kg ethanol from vehicle are shown in Fig. 6. As seen in Fig. 6A, ethanol dose dependently occasioned responding on the ethanol-appropriate lever. As can be seen in Fig. 6B, fluvoxamine did not reliably occasion ethanol-appropriate responding. At the highest dose of fluvoxamine tested, 10 mg/kg, only 4 of the 8 rats tested completed enough responses to earn food pellets.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Percentage of ethanol-appropriate responding by rats trained to discriminate 1.2 g/kg (i.p., 10-min pre-session) ethanol from vehicle. A, effects of ethanol alone (). B, effects of fluvoxamine alone (). C, effects of ethanol + saline (), ethanol + 1 mg/kg fluvoxamine (), or ethanol + 3 mg/kg fluvoxamine (). D, effects of fluvoxamine combined with 1.2 g/kg ethanol (). Bars around points represent the standard error of the mean (when no bars can be seen standard error is within the point). Ethanol injections were given i.p. 10 min before, and fluvoxamine injections were given i.p. 30 min before the beginning of the 10-min experimental session.

In Fig.6, C and D, the effects of combinations of ethanol and fluvoxamine are shown. As can be seen in Fig. 6C, the ethanol dose-response curve was not substantially altered by coadministration of either 1 mg/kg fluvoxamine (filled squares) or 3 mg/kg fluvoxamine (filled triangles). This can also be seen by comparing the ED₅₀ values for ethanol under these three conditions. The ED₅₀ for ethanol alone was 0.48 g/kg (95% CL, 0.24-0.97), whereas the ED₅₀ for ethanol in combination with 1 mg/kg fluvoxamine was 0.55 g/kg (0.39-0.79), and in combination with 3 mg/kg the ED₅₀ was 0.46 g/kg (0.36-0.59). Thus, fluvoxamine did not appear to alter the potency of ethanol at occasioning ethanol-appropriate responding.

As can be seen in Fig. 6D, fluvoxamine also did not alter the ability of 1.2 g/kg ethanol to occasion ethanol-appropriate responding in these rats. This dose of ethanol alone occasioned 95% ethanol-appropriate responding; and in combination with 1 to 10 mg/kg fluvoxamine this dose occasioned 93 to 97% ethanol-appropriate responding. At the highest dose of fluvoxamine tested in combination with 1.2 g/kg ethanol, only 3 of the 8 rats tested responded a sufficient number of times to earn food pellets.

Table 3 shows that neither ⁹-THC nor d-amphetamine reliably occasioned ethanol-appropriate responding, including doses that markedly reduced responding. Thus, ethanol-appropriate responding was below 30%, except in rats tested with 3 mg/kg d-amphetamine, where 45% ethanol-appropriate responding occurred. However, at this dose, only two of the eight rats responded a sufficient number of times to earn food pellets.

View this table: [in this window] [in a new window]   TABLE 3 Effects of d-amphetamine and 9-tetrahydrocannabinol in an ethanol discrimination Substitution test results with d-amphetamine and 9-THC in the ethanol (1.2 g/kg i.p.) drug discrimination assay. %RDP refers to percentage of ethanol-appropriate responding and is based on the average of responding animals, i.e., to be included in this calculation an animal must have obtained at least one reinforcement out of a total of six possible reinforcements per test session. Rate refers to responses per second (resp/s) and reflects the mean of all rats tested. NT/NR reflects the number of animals tested and the number of animals earning at least one reinforcement.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Fluvoxamine decreased ethanol-maintained responding in this study. Although fluvoxamine also decreased food-maintained responding, this effect required higher doses of fluvoxamine. These observations appear reliable and replicable. Furthermore, these findings are not an artifact of the combined rate-decreasing effects of fluvoxamine and ethanol. Interestingly, whereas fluvoxamine selectively decreases ethanol-maintained responding, fluvoxamine did not appear to alter the discriminative effects of ethanol. Thus, these results are consistent with previous studies finding a relationship between synaptic levels of serotonin and alcohol consumption (see Sellers et al., 1992 for a review), and indicate that fluvoxamine is an interesting lead compound for alcoholism medication development.

The decrease in ethanol-maintained behavior following fluvoxamine that we observed is consistent with the results of other studies showing decreases in either ethanol-maintained behavior or ethanol drinking following the administration of either fluvoxamine or other serotonin reuptake inhibitors (Lyness and Smith 1992; Gulley et al., 1995; Maurel et al., 1999a,b). The specificity of these earlier observations is often difficult to assess. In many cases, either ethanol-maintained behavior or ethanol drinking alone was assessed (e.g., Lyness and Smith 1992; Gulley et al., 1995).

In this study, the effects of fluvoxamine on lever-pressing maintained by either ethanol or food were compared. Fluvoxamine decreased rates of lever-pressing for ethanol at lower doses than the doses needed to decrease food-maintained behavior. Other studies have compared the effects of drugs affecting serotonergic function upon behavior maintained by ethanol to that maintained by food or water. However, in many of these studies comparing ethanol- and water-maintained behavior, the animals were not water deprived, and water was not clearly reinforcing behavior (e.g., Hodge et al., 1993; Cohen et al., 1999; Maurel et al., 1999a). This makes interpreting these experiments difficult.

There are at least two studies, however, that examined the selectivity of the effects of serotonergic agents upon ethanol-maintained behavior. In the first, Murphy and coworkers (1988) compared the effects of fluoxetine on water- and ethanol-maintained responding. Fluoxetine was administered repeatedly, and the selective effects of fluoxetine appeared to increase with repeated administration. In the second, Roberts and coworkers (1998) studied the effects of FG 5974, 8-OH-DPAT, and amperozide on ethanol- or saccharin-maintained behavior. The mixed 5-HT1A agonist/5-HT2A antagonist FG 5974 selectively decreased ethanol-maintained behavior. The effects of the 5-HT1A agonist 8-OH-DPAT also appeared to be selective. In contrast, the 5-HT2A antagonist, amperozide, decreased ethanol- and saccharin-maintained behavior at similar doses. The results of these two studies are consistent with the results of the present study.

One possible interpretation of our results and the results of Roberts and coworkers (1998) is that these drugs act synergistically with ethanol to decrease responding. At least for fluvoxamine, this interpretation does not appear to be correct. When we examined the effects of ethanol and fluvoxamine administered jointly, the two compounds did not produce synergistic effects. The effects of 1 and 3 mg/kg fluvoxamine in combination with ethanol were almost identical to the effects of ethanol alone or ethanol administered in combination with vehicle. The effects of 10 mg/kg fluvoxamine administered in combination with ethanol were similar to the effects of 10 mg/kg fluvoxamine alone. Thus, synergistic rate-decreasing effects from combinations of ethanol and fluvoxamine cannot explain our observations.

Another potential confound of our findings is that the selective effects of fluvoxamine might be an artifact of unequally matched levels of food and ethanol used to maintain behavior. This explanation appears unlikely. The ED₅₀ values for fluvoxamine on ethanol-maintained behavior did not appear to vary across a wide range of ethanol concentrations (4-32% w/v), and all of these ED₅₀ values differed significantly from the ED₅₀ of fluvoxamine for food-maintained behavior.

Drugs such as ethanol can come to serve as discriminative stimuli, i.e., the presence of ethanol can come to set the occasion for particular responses. These discriminative stimulus effects are typically pharmacologically specific. For instance, in this study, ethanol dose dependently occasioned ethanol (1.2 g/kg)-appropriate responding, but d-amphetamine and ⁹-THC did not dose dependently occasion ethanol-appropriate responding. In drug self-administration procedures, the drug being self-administered can not only reinforce behavior, but also serve a discriminative function. For instance, a particular level of drug effect might inform the subject of impending satiation, thus leading to cessation of responding; alternatively, the absence of the expected drug effect might inform the subject of extinction conditions, thus leading to cessation of responding.

Fluvoxamine did not alter the discriminative effects of ethanol in this study. The potency of ethanol alone and in combination with either 1 or 3 mg/kg fluvoxamine was similar. Fluvoxamine 1 to 10 mg/kg also did not antagonize the ability of the training dose (1.2 g/kg) of ethanol to occasion ethanol-appropriate responding. Finally, fluvoxamine did not occasion ethanol-appropriate responding. These results would seem at variance with the results of two studies studying fluoxetine, one of which also studied the effects of paroxetine (Maurel et al., 1997; Risinger, 1997). Maurel and coworkers found that fluoxetine and paroxetine dose dependently occasioned ethanol-appropriate responding in rats trained to discriminate the effects of ethanol (1 g/kg i.p., 15-min pretreatment). Neither our study in rats trained with 1.2 g/kg ethanol, nor Risinger's (1997) study in mice trained with 1.0 g/kg ethanol, are consistent with the results from Maurel and coworkers. In the Risinger (1997) study, 10 mg/kg fluoxetine did not occasion ethanol-appropriate responding. In our study, fluvoxamine in doses from 1 to 10 mg/kg did not occasion ethanol-appropriate responding.

Risinger (1997) did find that 10 mg/kg fluoxetine increased the effectiveness of lower doses of ethanol to occasion ethanol-appropriate responding. This differs somewhat from our findings. We did not find the ethanol ED₅₀ to differ when it was combined with fluvoxamine from the ethanol ED₅₀ when it was combined with the fluvoxamine vehicle. This difference between Risinger's study and ours is, however, less than it would at first appear. In the Risinger study, fluoxetine's potentiation of ethanol would not have been significant had the comparison been made to the initial ethanol dose-response curve rather than the ethanol dose-response curve determined in combination with fluoxetine vehicle. In our study, fluvoxamine did not potentiate the effects of ethanol when compared with the ethanol dose-response curve determined in combination with fluvoxamine vehicle, but would have been marginally significant if compared with our initial ethanol dose-response curve.

The serotonergic component of the ethanol discriminative stimulus is crucially dose-dependent (Grant and Colombo, 1993). The ability of the serotonin agonist, trifluoromethylphenylpiperazine, to occasion ethanol-appropriate responding is greater when the ethanol training dose was 1.5 g/kg (intragastrically) than when the training dose was 1.0 or 2.0 g/kg (intragastrically). Thus, slight differences in training conditions between the three studies may account for the differences between these studies.

One conclusion of the present work might be that fluvoxamine has selective actions upon ethanol reinforcement. Certainly, fluvoxamine decreased ethanol-reinforced behavior at lower doses than food-reinforced behavior; and thus fluvoxamine has selective effects upon ethanol-reinforced behavior. However, these selective effects of fluvoxamine on ethanol-reinforced behavior are not necessarily due to a selective action of fluvoxamine upon ethanol reinforcement. We examined several possible alternative interpretations of this observation in this study. We found that this observation was not easily explained by inadequacies in matching the magnitude of reinforcement across conditions, because the effects of fluvoxamine did not appear to vary with the concentration of ethanol available. We also found that synergistic rate-decreasing effects of ethanol-fluvoxamine combinations did not easily explain this observation. Finally, we found that this observation was not easily explained by fluvoxamine-induced alterations in the discriminative stimulus effects of ethanol. However, many other possible explanations for our observation of selective effects of fluvoxamine upon ethanol-reinforced behaviors exist. Only by systematically exploring these explanations will the validity of the conclusion that fluvoxamine has selective actions upon ethanol reinforcement be proved or disproved.