Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Cocaine administration activates the hypothalamic-pituitary-adrenal (HPA) axis in rats (Rivier and Vale, 1987; Saphier et al., 1993), rhesus monkeys (Broadbear et al., 1999; Sarnyai et al., 1996) and humans (Vescovi et al., 1992; Heesch et al., 1995; Ward et al., 1998), increasing the release of plasma adrenocorticotropic hormone (ACTH) and the glucocorticoids cortisol or corticosterone. In most of the studies that have examined the relationship between cocaine and HPA axis activation, the dose, route, and frequency of cocaine's administration were controlled by the investigator (e.g., Sarnyai et al., 1996; Spangler et al., 1997) using procedures that may have incorporated restraint, hypodermic injection, and anesthesia, each of which may itself have elevated HPA axis activity (Setchell et al. 1975; Elvidge et al., 1976; Puri et al., 1981; Reinhardt et al., 1990; Piazza et al., 1991). This can be problematic, because an increase in stress hormones before drug administration may blunt the effect produced by the drug (Dallman and Jones, 1973; Sarnyai et al., 1996). Therefore, it may be difficult to separate the contribution of cocaine to HPA activation from that of the experimental procedure. Procedural confounds may be problematic even for subjects apparently familiarized with the procedure before testing, because the elevation in HPA response may remain despite repeated presentation of a stressful handling procedure (Coe et al., 1983; Hattingh et al., 1988; Higley et al., 1992; Kirschbaum et al., 1995).

Does the method by which cocaine is administered affect the degree to which the HPA axis is activated? The issue of "control" over stimulus delivery, and how this may affect physiological and behavioral responses to the stimulus, has been addressed in a number of studies. Studies of this nature may employ a yoked-control design, where subjects in the "control" group respond to present, postpone, or terminate stimulus delivery, whereas those in the "yoked" group receive identical stimulus deliveries as their matched controls but are unable to alter their delivery. The HPA axis response to aversive stimuli has been found to vary between subjects in yoked-control designs. For example, rhesus monkeys that learned to press a lever to terminate high-intensity noise had lower cortisol levels than their yoked counterparts (Hanson et al., 1976), even though they were exposed to the same sound intensity and duration. Monkeys trained in the "control" group and then transferred to the "yoked" group had the highest cortisol levels. Similar findings were reported in rats in a yoked-control shock postponement paradigm (Herrmann et al., 1984). In the case of rewarding stimuli, two studies have examined the conditions under which animals will work to terminate the delivery of a formerly reinforcing stimulus. In the first, rats rapidly learned to terminate the delivery of i.c. self-stimulation when it was delivered in a pattern identical with what had been self-administered on an earlier occasion (Steiner et al., 1969). Similarly, squirrel monkeys worked simultaneously to obtain and terminate cocaine administration when drug delivery was controlled by two different, concurrent contingencies (Spealman, 1979). Both studies demonstrate that there are circumstances under which the same stimulus can maintain behavior that leads to both its delivery and its avoidance, and highlight how a reinforcer may be aversive under conditions where the subject does not instigate its delivery.

Previously, we demonstrated that self-administered cocaine produces dose-dependent increases in plasma ACTH and cortisol in male rhesus monkeys (Broadbear et al., 1999) and, in a pilot to the present study, that automatic infusions of cocaine using the same doses and pattern of delivery generated by each monkey the previous day led to identical cortisol levels (Broadbear et al., 1997). It was unclear whether the extensive cocaine self-administration history of these subjects may have influenced the results. The present study was designed in part to address this issue.

There were three purposes to this study. The first was to determine whether the effects of cocaine on the HPA axis were different when cocaine delivery was either response dependent or response independent. The second was to determine whether the effects of cocaine on the HPA axis were modified by a history of cocaine self-administration. The third was to determine whether parallels existed between HPA responsiveness to an infusion of corticotropin-releasing hormone (CRF) and to infusions to cocaine under the different contingencies.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Subjects

Five adult male rhesus monkeys (Macaca mulatta), four intact, and one castrated (monkey 1583), weighing between 9.0 and 14 kg, and five intact adult females, weighing between 4.0 and 7.4 kg, were used in this study. One of the males (2900) had a cocaine self-administration history. This monkey only took part in the CRF infusion experiment. Of the remaining nine monkeys, three males and four females had no prior experience with cocaine (ketamine, used for sedation, was the only psychoactive drug noted in the histories of these monkeys) and eight had no previous history of drug self-administration. One female monkey (2490) had a history of drug discrimination testing in another laboratory (involving mainly opioids), and the other, a male (1583), had scarring consistent with jugular vein catheterization before purchase, but there were no records available to indicate his drug history. The presence of jugular veins in this monkey has not yet been determined.

The monkeys were individually housed in stainless steel cages measuring 83.3 × 76.2 × 91.4-cm deep (Bryan Research Equipment Corporation, Bryan, TX) located in a laboratory that contained a total of 24 similarly housed monkeys. The monkeys were fed 8 to 12 Purina Monkey Chow biscuits twice daily to maintain normal adult weight and water was available ad libitum. Each monkey had an indwelling venous catheter in a femoral, internal, or external jugular vein. Catheters were inserted during aseptic surgery under ketamine (10 mg/kg) and xylazine (2 mg/kg) anesthesia. Following placement in the vein, the catheter was guided s.c. to the midscapular region where it exited the monkey. The external portion of the catheter was protected inside the cage by a flexible stainless steel arm, with one end attached to the polyester jacket (Lomir, New York) worn by the monkey and the other bolted to the rear of the cage.

Animals used in these studies were maintained in accordance with the University of Michigan Committee on Animal Care and Guidelines of the Committee on the Care and Use of Laboratory Animal Resources, National Health Council (Department of Health, Education and Welfare, ISBN 0-309-05377-3, revised 1996).

Apparatus

Each cage had a 15 × 20-cm panel fixed to its right wall. Each panel had three stimulus lights, two red and one central green light, placed above two response levers. The red stimulus light over the right lever signaled drug availability. Drug delivery was contingent on the monkey emitting the required response (30 lever presses). The green center light was illuminated for the duration of the drug infusion, 1 ml over 5 s. During each 10-min time out (TO), all stimulus lights were extinguished and responding had no programmed consequences.

The experiment was controlled by IBM/PS2 computers located in an adjacent room. The computers were programmed using Med Associates software (Georgia, VT).

Procedure

The temporal sequence of the experiments was as follows.

1. An i.v. injection of 1 mg/kg of cocaine was administered to the nine monkeys with no cocaine history, and blood was sampled from the i.v. catheter at various times before and following the injection.

2. The nine monkeys with no history of response-contingent cocaine administration received either no injections or nonsignaled, noncontingent injections of saline or cocaine every 10 min for a total of 13 injections of the same dose of cocaine or saline. Each dose was presented on successive days in the following sequence: No injection; saline; 0.01, 0.03, 0.1, 0.3, 0.3, 0.1, 0.03, and 0.01 mg/kg/injection of cocaine; saline; and no injection.

3. The nine monkeys with no history of cocaine self- administration were trained to self-administer 0.03 or 0.1 mg/kg/injection of cocaine. Initially, a fixed ratio (FR) 1, TO 10-s schedule was used, with 0.01 mg/kg/injection of cocaine as the reinforcer. Once a monkey had learned to lever press on this schedule, which took one session (n = 7), 1 week (monkey 2087), or 5 weeks (monkey 2490), the FR and TO were gradually increased to 30 and 10 min, respectively, over 1 to 3 weeks. During this time, the dose of cocaine was increased to 0.03 or 0.1 mg/kg/injection. Seven monkeys were maintained on 0.03 mg/kg/injection of cocaine, and 2 monkeys (2087 and 2490) were maintained on 0.1 mg/kg/injection of cocaine.

4. Following the acquisition of stable cocaine-maintained behavior (criteria listed below), the experiments described in section 2 were repeated, only this time cocaine delivery was contingent upon the monkey fulfilling the FR 30 (FR 20 for monkey 2490, because this subject was older and less mobile that the other subjects) response requirement.

5. An i.v. injection of 1 mg/kg of cocaine was administered to the previously cocaine-naive monkeys (n = 9) exactly as described in section 1 above, and blood was sampled in an identical fashion.

6. Intravenous injections of 1 and 10 µg/kg of CRF were administered to eight monkeys (7/9 of the monkeys that took part in sections 1-5 of this protocol, and 1 monkey that had an extensive drug self-administration history that predated this study), all of whom had experience with cocaine self-administration at the time of CRF administration. Blood was sampled before and after each CRF infusion. CRF infusion experiments took place at least 1 week apart.

Effect of Cocaine History on HPA Response to a Single Cocaine Infusion. Nine cocaine-naive monkeys (5 female and 4 male) took part in this study. Monkeys were given a single infusion of 1 mg/kg of cocaine on two occasions. These experiments commenced between 9 and 10 AM, and although the self-administration session normally scheduled in the room took place as usual, the monkeys in this study did not participate during the morning that this test was done. Blood was sampled at 15, 10, and 5 min before the infusion of cocaine, and then at 5-min intervals for the first hour and 10-min intervals for the second hour postinfusion. Details of the blood collection are described below.

Effect of Response-Contingent versus Noncontingent Cocaine Delivery on HPA Activation. The subjects for this study were four male and five female monkeys, none of whom had a prior history of drug self-administration. In experiments where noncontingent cocaine or saline was administered, neither the levers nor the stimulus lights were present, and cocaine (0.01, 0.03, 0.1, or 0.3 mg/kg/injection) or saline was injected at 10-min intervals, one dose/session, beginning at approximately 10 AM, for a total of 13 infusions in the sequence described above.

Comparison of Effect of CRF and Cocaine on HPA Axis Activity. Five male and four female monkeys were the subjects for this study. At the time the CRF study was conducted, each subject had at least several months' experience with cocaine self-administration.

Blood Collection and Handling

Each blood sample (1.1-1.4 ml) was placed in a 2-ml Vacutainer (Becton Dickinson and Company, Franklin Lakes, NJ) containing 0.04 ml of 7.5% EDTA and immediately placed on ice. After drawing each blood sample, 1.5 to 3 ml of 30 U/ml heparin saline solution was infused into the catheter and, when sampling was done during sessions in which cocaine was available, a volume of the cocaine solution equal to the catheter volume (0.6-1.5 ml) was injected after the heparin saline solution.

Blood samples were centrifuged at 5000 rpm for 5 min and then the plasma (0.7 ml) was pipetted into 2-ml Cryovials (Corning) and stored at 80°C until assay. Samples were sent on dry ice to Washington University (St. Louis, MO) where ACTH and cortisol levels were determined using radioimmunoassay kits (cortisol: Diagnostic Products Corporation, Los Angeles, CA; ACTH: Nichols Institute Diagnostics, San Juan Capistrano, CA).

Statistical Analyses

Data are presented as mean ± S.E.M. and also as AUC ± S.E.M. Both forms of data presentation were used because each highlights different aspects of the study. Presentation of data averaged for each sampling time provides the actual plasma measurements of cortisol and ACTH and shows the onset and duration of HPA changes following cocaine or CRF infusion. AUC data summarizes the overall changes in HPA activity, and normalizes the HPA response both between and within subjects for easier comparison. One- or two-way ANOVA and multiple ANOVA were conducted on normalized data, and where appropriate, post hoc pairwise comparisons using the Tukey honest significant difference (HSD) test of significance (p < .05) were carried out (Statistica v.5.0, Statsoft, Tulsa, OK). Where experiments were replicated within subjects, the mean response for each subject was used to calculate treatment effects across subjects. One subject (monkey 2087) showed no discernible HPA response to the 1-mg/kg cocaine infusions relative to his HPA response following a saline infusion (data not shown). His results were not included in the analysis of those data.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Effect of Response-Contingent versus Noncontingent Cocaine Delivery Studies on HPA Activation

One male monkey (2087) had an ACTH response to nonresponse contingent infusions of cocaine or saline that was the reverse of that observed for the other three male monkeys. The ACTH response for monkey 2087 was greatest at the lowest cocaine dose (0.01 mg/kg/injection) and saline-like at the highest cocaine dose (0.3 mg/kg/injection). His ACTH data were excluded from subsequent analyses.

Data Analysis Across Sampling Times. Rates of responding for saline and cocaine as well as the number of infusions earned during the response-contingent part of this study are shown in Table 1. Increases in cocaine dose are positively correlated with increases in rates of responding. Infusion number and rates of responding both peaked at a cocaine dose of 0.1 mg/kg/injection. Response rates and infusion numbers increased relative to saline responding for 0.01 mg/kg/injection of cocaine in most monkeys, and increased for all monkeys for 0.03 mg/kg/injection of cocaine. These increases in the levels of cocaine-maintained behavior occurred with or without any concomitant increase in HPA axis activity. There were no gender differences in either the rates of responding or the number of infusions of saline or cocaine that were earned (Table 1).

View larger version (42K): [in this window] [in a new window]   Fig. 1.   Mean (±S.E.M.) plasma cortisol (µg/dl) and ACTH levels (pg/ml) determined from blood sampled before, during, and after noncontingent and response-contingent administration of saline or of cocaine (mg/kg/injection). Upper panels, cortisol data following noncontingent (upper left) and response-contingent administration (upper right) for male (n = 4) and female (n = 5) monkeys. Lower panels, ACTH data following noncontingent (lower left) and response-contingent administration (lower right) for male (n = 3) and female (n = 5) monkeys. In the nonresponse contingent condition, an unsignaled infusion of saline or cocaine was injected every 10 min for a total of 13 infusions. In the response-contingent condition, cocaine (or saline) availability was signaled by a red stimulus light, which remained illuminated until the requisite number of lever presses had been emitted (FR 30, TO 600 s); drug or saline delivery was signaled by a green stimulus light. Each session lasted 130 min and a maximum of 13 infusions was possible. a, versus no infusion, saline, and 0.01 and 0.03 mg/kg/injection of cocaine (p < .05); b versus no infusion, saline, and 0.01 mg/kg/injection of cocaine (p < .05); c versus no infusion (p < .05); d versus 0.03 mg/kg/injection of cocaine (p < .05); e versus saline (p < .05); f versus no infusion and saline (p < .05).

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Mean (±S.E.M.) plasma cortisol (µg/dl) and ACTH levels (pg/ml) determined from blood sampled before, during, and after noncontingent and response-contingent administration of saline or cocaine (mg/kg/injection). Upper and middle panels, cortisol data following noncontingent and response-contingent administration was combined for male (n = 4, upper panel) and female (n = 5, center panel) monkeys as contingency of administration was not a significant determinant of cortisol levels. Lower panel, data for male (n = 3) and female (n = 5) monkeys were combined as there were no statistically significant differences in ACTH levels for either male and female monkeys or for noncontingent and response-contingent administration of cocaine or saline. a versus no infusion (p < .05); b versus no infusion and saline (p < .05); c versus no infusion, saline, and 0.03 mg/kg/injection of cocaine (p < .05); d versus no infusion and 0.03 mg/kg/injection of cocaine (p < .05); e versus no infusion, saline, and 0.01 and 0.03 mg/kg/injection of cocaine (p < .05). Details as for Fig. 1.

AUC Data Analysis. Cocaine administration had a significant effect on cortisol AUC (df = 5, F = 15.19, p < .001), with 0.1 and 0.3 mg/kg/injection of cocaine (n = 9) each producing a larger cortisol AUC than the no injection, saline, and 0.01 and 0.03 mg/kg/injection of cocaine conditions (p < .05; Fig. 3, top and center panels). There was a near-significant effect of contingency (p = .08), because response-contingent cocaine was associated with a larger cortisol response relative to noncontingent cocaine administration, particularly in male monkeys. There was a significant gender difference in the cortisol AUC data (df = 1, F = 5.30, p < .05), because male monkeys had a larger cortisol response to response-contingent cocaine administration than did female monkeys (Fig. 3, top and center panels).

View larger version (48K): [in this window] [in a new window]   Fig. 3.   Mean (±S.E.M.) cumulative release (AUC) of plasma cortisol (µg · min/dl) and ACTH (pg · min/ml) above basal (presession) levels during saline or cocaine (mg/kg/injection) administration. There was a gender difference in cortisol levels, and a near-significant difference due to contingency (p = .08). Upper panel, cortisol AUC for female monkeys (n = 5) for nonresponse contingent (hatched bars) and response contingent (solid bars) cocaine administration. Center panel, cortisol AUC for male monkeys (n = 4) for nonresponse contingent (hatched bars) and response contingent (solid bars) cocaine administration. *versus no infusion (p = .05), **versus saline (p < .05), ***versus no infusion and saline (p = .01), ****versus no infusion, saline, and 0.01 and 0.03 mg/kg/injection of cocaine (p < .001). Lower panel, ACTH AUC for male and female monkeys (n = 8) following nonresponse contingent (hatched bars) and response-contingent (solid bars) administration of cocaine. There was no gender difference in the ACTH response to cocaine. Response-contingent administration of 0.1 and 0.3 mg/kg/injection cocaine resulted in larger increases in ACTH levels than did noncontingent administration, attaining significance for 0.1 mg/kg/injection of cocaine (p < .05). In addition, the ACTH response to cocaine was more dose dependent under the response-contingent administration condition. a, ACTH for 0.3 coc (response contingent) greater than no infusion, saline, and 0.01 and 0.1 mg/kg/injection of cocaine (noncont.; p < .05). b, ACTH for 0.3 coc (noncont.) greater than no infusion (noncont.; p = .06). c, ACTH for 0.1 coc (resp. cont.) greater than for no infusion and 0.1 mg/kg/injection of cocaine (noncontingent; p < .05). d, ACTH for 0.3 and 0.1 coc (resp. cont.) greater than for no infusion and saline (resp. cont.; p < .05). Other details as for Fig. 1.

View larger version (43K): [in this window] [in a new window]   Fig. 4.   Cortisol (µg · min/dl) and ACTH AUC (pg · min/ml) for three subjects illustrating the individual variability of the HPA response to noncontingent and response-contingent cocaine administration (mg/kg/injection) or saline. The three upper panels represent the HPA response under the noncontingent infusion condition, and the lower three panels show the corresponding data for the response-contingent administration condition. Other details as for Fig. 1.

Effect of Cocaine History on HPA Response to a Single Cocaine Infusion

One male monkey (2087) had no increase in HPA activity following infusions of 1 mg/kg of cocaine relative to saline (data not shown) unlike the other monkeys in this study. His data were excluded from subsequent analyses.

There was a modest increase in plasma cortisol levels following i.v. cocaine administration (1 mg/kg) in male (n = 3) and female (n = 5) monkeys (Fig. 5, top panels). Plasma cortisol levels obtained from samples taken from 20 until 60 min after the cocaine infusion were significantly higher than the cortisol levels from earlier sampling times (n = 8; p < .05). There was no difference in the cortisol response to an infusion of 1 mg/kg of cocaine depending on whether the monkeys were cocaine-naive or -experienced. Males tended to have a larger cortisol response to the cocaine infusion, and this difference was significant in the cocaine-experienced males at a few sampling times (Fig. 5, top right panel).

View larger version (40K): [in this window] [in a new window]   Fig. 5.   Mean (±S.E.M.) plasma cortisol (µg/dl) and ACTH (pg/ml) subsequent to i.v. administration of 1 mg/kg of cocaine in male (n = 3) and female (n = 5) monkeys. The monkeys received 1 mg/kg of cocaine on two occasions: first as cocaine-naive subjects (left panels) and then as monkeys with a history of cocaine self-administration (right panels). Male () and female () cortisol and ACTH responses to a single injection of 1 mg/kg of cocaine are compared in each panel. There were no differences in the cortisol or ACTH response to cocaine between cocaine-naive and -experienced monkeys. *p < .05, **p < .01.

There was also an increase in plasma ACTH levels following i.v. cocaine administration (1 mg/kg) in male (n = 3) and female (n = 5) monkeys (Fig. 5, bottom panel). Plasma ACTH levels obtained from samples taken from 30 until 60 min after the cocaine infusion were significantly higher than the ACTH levels from earlier sampling times (n = 8; p < .05). There was no difference in the ACTH response to an infusion of 1 mg/kg cocaine depending on whether the monkeys were cocaine-naive or -experienced. Males tended to have a larger ACTH response than did females to the cocaine infusion, and this difference was significant in the cocaine-experienced males at several sampling times (Fig. 5, bottom right panel).

Comparison of Effects of CRF and Cocaine on HPA Axis Activity

Intravenous administration of CRF (1 and 10 µg/kg) resulted in significantly increased release of cortisol relative to preinfusion levels, from 30 to 80 min (1 µg/kg of CRF) and at all postinfusion sampling times (10 µg/kg of CRF; Fig. 6, top panel). The 10 µg/kg dose of CRF produced a 1.97 ± 0.29-fold greater release of cortisol than did 1 µg/kg of CRF (df = 1, F = 8.90, p < .01). There was a near-significant gender difference (p = .07), with male monkeys tending toward a larger cortisol response to CRF than female monkeys.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Mean (±S.E.M.) plasma cortisol (µg/dl) and ACTH (pg/ml) subsequent to i.v. infusion of human/rat CRF (1 and 10 µg/kg) in male (n = 5) and female (n = 4) rhesus monkeys. Upper panel, cortisol levels before and after 1 and 10 µg/kg of CRF for all monkeys (n = 9); **p < .01, ***p < .001. Middle panel, ACTH levels before and after 1 and 10 µg/kg of CRF in male monkeys (n = 5); ***p < .001. Lower panel, ACTH levels before and after 1 and 10 µg/kg of CRF in female monkeys (n = 4); there was an overall increase in cortisol levels subsequent to 10 µg/kg of CRF relative to 1 µg/kg (p = .05).

Intravenous administration of CRF (1 and 10 µg/kg) also resulted in significantly increased release of ACTH relative to preinfusion levels, from 10 to 50 min (1 µg/kg of CRF) and at all postinfusion sampling times up to 120 min (10 µg/kg of CRF; Fig. 6, center and bottom panels). The 10 µg/kg dose of CRF produced a 5.45 ± 1.01-fold greater release of ACTH than did 1 µg/kg of CRF (df = 1, F = 29.26, p < .001). There was a significant gender difference (df = 1, F = 6.36, p < .05), with male monkeys having a larger ACTH response to 10 µg/kg of CRF than did female monkeys. The difference in the ACTH response of female monkeys to 1 and 10 µg/kg of CRF was not significant at any time point.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The primary goal of this study was to investigate whether the ACTH and cortisol response to cocaine would differ when the cocaine delivery was controlled either by the investigator or by the monkey. It should be noted that response-contingent cocaine can function as a reinforcer without elevating plasma ACTH or cortisol, in confirmation of earlier findings (Broadbear et al., 1999). Nevertheless, we found that cortisol and ACTH levels were higher and more dose dependent following response-contingent cocaine administration than they were after noncontingent cocaine at higher cocaine doses. The secondary goal was to determine whether cocaine history had a significant impact on cocaine's effects on HPA axis activity. A single infusion of 1 mg/kg cocaine, administered when the monkeys were cocaine-naive and then again subsequent to the noncontingent and response-contingent experiments, did not differ in its effects on ACTH and cortisol. In addition, distinct gender differences were found during the course of this study, with male monkeys showing higher cortisol levels following noncontingent and response-contingent cocaine administration, and higher ACTH levels following administration of 1 mg/kg cocaine and 10 µg/kg of CRF infusions.

Although there were no overall differences in the HPA responses to the first and subsequent exposures to a single 1 mg/kg cocaine infusion for male or female monkeys, there was a distinct lack of correlation between both the ACTH and cortisol responses to cocaine on the two occasions that this dose was given (data not shown). This lack of correlation suggests that the HPA response of a cocaine-naive subject to a large noncontingent cocaine infusion is not at all predictive of that subject's HPA response to the same dose of cocaine once he has acquired extensive experience with cocaine. It was correctly anticipated that the first exposure to cocaine would increase HPA activity as it did in a study done in humans, where first-time administration of cocaine to male subjects produced an increase in plasma cortisol levels (Heesch et al., 1995). Similar results to ours were also found in a study by Sarnyai et al. (1996), in which male rhesus monkeys were infused with saline or 0.4 or 0.8 mg/kg of cocaine. Significant correlations were found between the behavioral response to cocaine and the cortisol and ACTH changes in these monkeys. One notable difference with this earlier study was that the baseline cortisol and ACTH values obtained in the hour before the cocaine infusion were considerably higher in some monkeys than those that were measured in the present study, particularly in monkeys that showed little or no HPA response to the cocaine infusion (ACTH > 40 pg/ml, cortisol  18 µg/dl). These high levels may have been a consequence of ketamine anesthesia (Elvidge et al., 1976; Puri et al., 1981), and the fact that testing took place in restraint chairs. Indeed, a negative correlation was found between baseline cortisol and subsequent changes in ACTH following infusion of 0.8 mg/kg of cocaine, possibly because of down-regulation of HPA axis activity via a negative feedback mechanism (Dallman and Jones, 1973). An earlier study using ovariectomized female rhesus monkeys (Sarnyai et al., 1995), found that acute infusions of 0.4 and 0.8 mg/kg of cocaine changed neither pulsatile ACTH nor cortisol release. CRF infusions given to these monkeys showed that their HPA axes were intact and responsive. This lack of effect of cocaine was attributed to the absence of normal levels of gonadal steroids. This may indeed be the case, because the five intact female monkeys in the present study each showed increased HPA activity in response to cocaine. The present study constitutes the first demonstration of cocaine's effects on HPA activity in intact, female rhesus monkeys.

Studies done in humans, comparing the pharmacokinetics and subjective effects of cocaine in males and females, have demonstrated that overall, phase of the menstrual cycle does not have any significant effect on these measures (Mendelson et al., 1998). The effect of 0.4 mg/kg of cocaine on ACTH was the same in male subjects as it was in 6/13 female subjects, and there was a sex difference in the cortisol response, with women having higher peak plasma levels (Sholar et al., 1998). The remaining seven women had elevated cortisol levels before the cocaine infusion, which once again may have contributed to the lack of effect of cocaine on HPA activity.

In the experiments that evaluated the effects of repeated noncontingent or response-contingent infusions of cocaine or saline, it was found that ACTH responses were similar for male and female monkeys. When combined, their data showed a distinctly dose-related response to cocaine over the course of the session. This is in agreement with our earlier findings in male monkeys with years of experience with cocaine self-administration (Broadbear et al., 1999). There was also an overall difference in the cortisol and ACTH responses to noncontingent versus response-contingent cocaine, which was particularly apparent with the higher doses of cocaine (AUC data). Overall, it appears that under the response-contingent condition, cocaine produced larger, more dose-dependent changes in cortisol and ACTH release. These differences may indicate that response-contingency and/or cocaine history "fine tunes" the HPA response to cocaine, from a more generalized "cocaine response" (as was seen in the non-dose-dependent nature of the ACTH AUC data following noncontingent administration) to a response that better reflects a high-dose cocaine response-contingent effect.

Unlike the ACTH data, there was a gender difference in the cortisol response to noncontingent and response-contingent cocaine administration, with male monkeys having a larger response to saline and cocaine over the entire dose range. When comparing these data with our earlier findings (Broadbear et al., 1999), the cortisol AUC for the male monkeys in the present study are only one-half the size of those originally reported, although the ACTH data are comparable. This difference could be due to the inclusion of monkeys in the present study that did not show a dose-dependent response to cocaine, or they could reflect the comparative lack of cocaine history of subjects in the present study. This raises the possibility that some of the differences in HPA response observed between noncontingent and response-contingent cocaine administration in the present study were due more to the extent of the monkeys' experience with cocaine than to the contingency of its administration, and that repeating these observations after a period of prolonged cocaine self-administration would result in an enhanced, more dose-dependent HPA response. In an earlier study, when cocaine was administered noncontingently to monkeys with an extensive self-administration history, their cortisol responses were no different from earlier occasions when cocaine was delivered response contingently (Broadbear et al., 1997).

It was surprising that some of the monkeys (2 males and 1 female) lacked a dose-dependent HPA response to cocaine under both noncontingent and response-contingent conditions. These monkeys were no different from the others with respect to response rate and cocaine intake. The CRF infusion also highlighted individual variability in ACTH and cortisol responses, revealing some interesting similarities in the monkeys' HPA responses to CRF and cocaine. Although regression analysis showed a significant positive correlation for both the cortisol and ACTH AUCs following nonresponse- contingent administration of 0.3-mg/kg/injection of cocaine and 1 or 10 µg/kg of CRF, there were no correlations between the HPA responses to CRF and/or cocaine when cocaine was administered as either a single injection or as repeated, non-response-contingent infusions (data not shown). This is perhaps an indication that differences exist in the sensitivity of each monkey to stimulation of the HPA axis directly by CRF, or indirectly, by cocaine. There was no gender difference in the cortisol response to a single infusion of cocaine or CRF, but male monkeys had a substantially larger ACTH response than females to 10 µg/kg of CRF, as was the case for a noncontingent infusion of a large (1 mg/kg) dose of cocaine. Past studies have demonstrated that plasma levels of ACTH are not always correlated with corticosteroid levels (Krieger and Allen, 1975). The fact that the larger dose of CRF stimulated a proportionally larger increase in ACTH relative to cortisol (as did cocaine) indicates that there may be a "ceiling effect" for the sensitivity of the adrenal cortex to ACTH (see also Cador et al., 1992).

In summary, both male and female rhesus monkeys showed an increased secretion of ACTH and cortisol following cocaine administration. This HPA response was larger and more dose dependent when cocaine was administered response-contingently then when it was delivered nonresponse-contingently to monkeys before acquisition of self-administration behavior. Male monkeys tended to have a larger HPA response to both cocaine and CRF infusions than did female monkeys. It is possible that the differences in the characteristics of the ACTH and cortisol release to cocaine under the two contingencies are due to the relative amount of experience with cocaine as well as to whether the delivery of cocaine is under the control of the monkey.