Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The physiological responses that enable one to effectively cope with short-term stress can have dire consequences when continually activated (see Sapolsky, 1994). Prolonged psychosocial stress and trauma appear to be particularly harmful, precipitating psychological disturbances, increasing vulnerability to an array of disorders from the common cold to cancer and accelerating death (Ballieux, 1997; Bergsma, 1994; Dorian and Garfinkel, 1987; Friedman and Yehuda, 1995; Leeming, 1994; Sapolsky, 1994). Drug abuse is not typically listed among stress-related disorders. Nevertheless, a growing body of clinical and epidemiological evidence exists linking this behavioral pathology to stress. Again, as for other stress-related disorders, it is psychosocial stress and trauma that are associated with drug abuse (Barnet et al., 1995; Cottler et al., 1992; Helzer et al., 1987; Kosten et al., 1986; Lindenberg et al., 1993; O'Doherty, 1991; Stine and Kosten, 1995).

Observations from animal studies also have linked stress to drug use and, in addition, suggest that glucocorticoids released from the adrenal glands might play a significant role. For example, repeated exposure to stressful stimuli (Deroche et al., 1995; Leyton and Stewart, 1990; Molina et al., 1994; Shaham et al., 1995) or adverse living conditions (Deroche et al., 1993a, 1994) enhances the psychomotor effects of psychostimulants and morphine. This stress-induced enhancement of the pharmacological effects of morphine appears to be dependent on corticosterone secretion (Deroche et al., 1993a, 1994, 1995; Piazza et al., 1994) and can be mimicked by repeated injections of corticosterone (Deroche et al., 1992). In addition, stress or adverse living conditions potentiate opiate (e.g., Alexander et al., 1981; Carroll et al., 1979; Dib, 1985; Shaham et al., 1993) and psychostimulant (Bozarth et al., 1989; Goeders and Guerin, 1994; Piazza et al., 1990; Ramsey and Van Ree, 1993) self-administration and reinstatement of self-administration after a drug-free period (Shaham 1993; Shaham and Stewart, 1995). Furthermore, the propensity to self-administer appears to be positively related to reactivity to stress and corticosterone secretion. That is, rats with higher locomotor and corticosterone responses to novelty have a greater propensity to self-administer morphine and psychostimulants (e.g. Deroche et al., 1993b; Goeders and Guerin, 1994; Maccari et al., 1991; Piazza et al., 1991). Last, it has been reported that rats will orally and intravenously self-administer corticosterone itself, suggesting that it may have some role in reinforcing drug use (Deroche et al., 1993b; Piazza et al., 1993). Thus, it is becoming increasingly apparent that stress and glucocorticoids may play a significant role in the etiology of drug abuse.

Recently, we (Nock et al., 1997) reported that chronic exposure of male rats to morphine markedly increases the concentration of CBG (transcortin), the major glucocorticoid binding protein in blood. Elevated CBG levels occurred by 3 days and appeared to be maximal at 7 days after morphine pellet implantation. Thereafter, CBG levels returned toward normal as morphine cleared from the general circulation. However, when a supplemental pellet was implanted on day 7 to maintain morphine levels, the concentration of CBG remained elevated through 14 days, the longest time interval examined. Morphine effects on CBG, therefore, appear to persist for at least as long as the drug is detectable in blood.

The morphine-induced increase in CBG in turn appears to greatly reduce the amount of free, physiologically active corticosterone available to intracellular receptors. This deficit is most conspicuous after exposure to mild stress. Essentially, the morphine-induced increase in CBG deadens the hormonal response to stress, with levels of free hormone at the peak of the stress response suppressed by as much as 90%. However, basal levels of free corticosterone are even more dramatically reduced (>95%). Based on these findings, we suggested that chronic exposure to morphine impairs a principal hormonal mechanism for coping with stress and thereby might contribute to the perpetuation of drug use and to adverse effects of drug exposure (Nock et al., 1997).

Although there remains a paucity of information pertaining to drug abuse by women, the available evidence indicates that the sexes may differ substantially in terms of motives for drug use, severity of abuse, effective treatment strategies and treatment outcome (Bowker, 1977; Griffin et al., 1989; Lukas et al., 1996; Marsh and Miller, 1985; Marsh and Simpson, 1986; Robbins, 1989; Ryser, 1983). Psychosocial factors undoubtedly play a role in these gender differences. Whether and how biological factors contribute are largely unknown. However, we (Cicero et al., 1996, 1997) and others (Baamonde et al., 1989; Islam et al., 1993; Kepler et al., 1989) have reported a marked gender difference in the antinociceptive effects of morphine in rodents that does not appear to be due to a sex difference in morphine pharmacokinetics (Cicero et al., 1997). Sex differences in the discriminative stimulus properties of morphine, the severity of naloxone precipitated withdrawal and the effects of stress on tolerance in rats have also been reported recently in abstract form (Craft and Bartok, 1997; Craft and King, 1997; Ratka et al., 1997). Thus, inherent biological factors may play a significant role in determining differences in the drug abuse/liability profiles for men and women. Because a clear awareness of how morphine and glucocorticoids interact is likely to be important for a full understanding of the causes and consequences of drug use, in the study reported here, we compared the effects of chronic morphine exposure on CBG in female and male rats to determine whether drug exposure produces deficits in glucocorticoid action in females as it does in males.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals

Male and female rats purchased from Harlan Sprague-Dawley (Cumberland, IN) were used for all experiments (10-12 rats/group). The rats were housed in groups of two or three/cage with lights in the colony room on from 6 a.m. to 6 p.m. They were 60 to 90 days old at the beginning of drug treatment.

Bilateral ADX, CAST, OVX and sham operations were performed with the animals under anesthesia (40 mg/kg Brevital). ADX rats were maintained on a 0.9% saline/1% sucrose solution containing 20 µg of corticosterone/ml for 7 days after surgery. Thereafter, ADX rats were provided a salt lick. For one experiment, specified in the Results section, ADX rats were maintained on the corticosterone drinking solution for the duration of the experiment.

Trunk blood was collected after decapitation. All rats were killed between 10 a.m. and 11 a.m., when corticosterone secretion is normally low in the males of our colony. The rats were individually carried by hand to a procedural room that was isolated from the colony room. Rats were killed within 15 sec of being picked up, and no more than 60 sec elapsed between the first and last rats in a cage. A number of additional precautions were taken to minimize the disturbance to the animals before death. First, all preparations in the procedural room and in the colony room were carried out on the preceding day. Second, the door to the colony room was baffled to minimize any sound on opening or closing. Third, all manipulations of the cages were done as quietly as possible, and all other sounds were kept to a minimum. Fourth, routine daily maintenance in the colony room was conducted after the experiments.

Drug Treatments

Placebo or morphine (75 mg) pellets were implanted subcutaneously with the animals under anesthesia (40 mg/kg Brevital) and were allowed to remain in place for the duration of the experiment.

All pellets were implanted on day 0 with one exception. For the five-pellet paradigm, pellets were implanted sequentially, with one implanted on day 0 and two additional pellets implanted on days 3 and 5.

Serum Preparation

Serum was prepared by standard procedures from trunk blood. All steps were carried out at 0 to 4°C. Blood was allowed to clot for ~45 min, and serum was decanted after centrifugation for 15 min at 1910 rcf in an IEC PR-7000 M preparative centrifuge.

CBG Levels

CBG was measured using a protein binding assay modified from procedures described by McCormick et al. (1995). All steps were carried out at 0 to 4°C.

Sample preparation. Endogenous steroids were removed from serum by charcoal adsorption. A 5-µl aliquot of serum was diluted to 1 ml with buffer (30 mM Tris·HCl, 1 mM disodium ethylenediaminetetraacetic acid, 10 mM sodium molybdate, 1 mM dithiothreitol, 10% glycerol) containing dextran charcoal (Norit-A; 5.0 mg/ml final). The mixture was vortexed, allowed to sit at 4°C for 20 min and centrifuged for 15 min at 2500 rpm. Aliquots of the supernatant fluid were diluted 1:5 (males) or 1:10 (females) with buffer and assayed.

Binding procedures. A 100-µl aliquot of steroid-free sample was incubated overnight at 4°C with 7.0 nM [³H]corticosterone with or without 16 µM corticosterone to define specific binding (final volume = 150 µl). Ethanol (1% final) was included in the incubation mixture to minimize steroid binding to glass. Bound and free [³H]corticosterone were separated using Sephadex LH-20 columns. Columns were made from 1.0-ml plastic disposable micropipette tips stopped with a 4-mm glass bead. The column was filled with LH-20 to a height of 3.2 cm from the middle of the glass bead and equilibrated with 400 µl of buffer (30 mM Tris·HCl, 1 mM disodium ethylenediaminetetraacetic acid, 10 mM sodium molybdate, 1 mM dithiothreitol, 10% glycerol). A 100-µl aliquot of incubation mixture was placed onto the column and washed in with 100 µl of buffer. Thirty minutes later, the sample was eluted using 600 µl of buffer. Specific binding was expressed as picomoles of specific binding/milligrams of serum protein. Protein content was determined according to the protein-dye binding method of Bradford (1976).

Morphine Levels in Serum

Morphine was assayed in serum using an RIA kit purchased from Diagnostic Products (Los Angeles, CA). The antibody is highly specific for morphine with <0.03% cross-reactivity for morphine-3-glucuronide and <0.1% cross-reactivity for morphine-6-glucuronide. The lower limit of sensitivity of the assay was 5 ng/ml, and the standard curve was linear over the range of 5 ng to 1000 ng/ml. Intra-assay and interassay variabilities were <5%.

Serum Corticosterone Levels

Corticosterone levels refer to the total amount of corticosterone in serum. Corticosterone was assayed using an RIA kit purchased from Diagnostic Products. The antibody is highly specific for corticosterone with <3% cross-reactivity for 11-deoxycorticosterone and <1% cross-reactivity for 18-hydroxydeoxycorticosterone, cortisol, progesterone, 17-hydroxyprogesterone, dehydroepiandrosterone, aldosterone, testosterone and estradiol. The lower limit of sensitivity of the assay was 20 ng/ml, and the standard curve was linear over the range of 20 ng to 2000 ng/ml. Intra-assay and interassay variabilities were <5%.

Statistical Analysis

All data are expressed as mean ± S.E.M.; t tests were used to compare two groups, and otherwise, data were subjected to analysis of variance followed by post-hoc analysis. Differences between groups were considered statistically significant when the probability that they occurred by chance was <.05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

CBG binding was markedly higher in serum of untreated adult female than of male rats (fig. 1). Scatchard analysis of pooled serum indicated that this sex difference is attributable to a difference in B_max (males = 7.1 pmol/mg protein; females = 16.6 pmol/mg protein) rather than K_d (males = 0.5 nM; females = 0.5 nM). That is, CBG has a similar affinity for corticosterone in males and females, but the number of binding sites is much higher in females than in males. Untreated females also had higher corticosterone levels than males at the time of death (fig. 1).

View larger version (43K): [in this window] [in a new window]   Fig. 1.   CBG (top) and corticosterone (bottom) in serum of untreated adult male and female rats. The figure shows that females had significantly higher levels of both CBG [t(26) = 14.2, P = .0001] and corticosterone [t(24) = 4.9, P = .0001] than males. *, P < .05. n = 12-15 rats/group.

Figure 2 shows CBG levels in serum of adult male and female rats at 7 days after the implantation of one, two or five (implanted sequentially as described in Materials and Methods) placebo or morphine pellets. All of the morphine treatments significantly increased CBG binding in males but had no apparent effect on CBG in females (fig. 2). Conversely, all of the morphine dosages resulted in significantly lower corticosterone levels in females but not in males (fig. 3). At the time of death, males and females had similar serum morphine levels (one pellet: males = 343 ± 37, females = 349 ± 36; two pellets: males = 772 ± 37, females = 681 ± 39; five pellets: males = 1521 ± 87, females = 1804 ± 196 ng/ml).

View larger version (35K): [in this window] [in a new window]   Fig. 2.   CBG levels in serum of adult male and female rats implanted with one, two or five (implanted sequentially as described under Materials and Methods) placebo or morphine pellets 7 days earlier. The figure shows that all of the morphine treatments significantly increased CBG binding in males but did not significantly affect CBG in females [one pellet Ftreatment x sex(1,47) = 19.8, P = .0001; two pellets Ftreatment x sex(1,42) = 13.5, P = .0007; five pellets Ftreatment x sex(1,50) = 45, P = .0001]. *, P < .05 vs. the placebo group. n = 10-15 rats/group.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Corticosterone levels in serum of adult male and female rats implanted with one, two or five (implanted sequentially as described under Materials and Methods) placebo or morphine pellets 7 days earlier. The figure shows that all of the morphine treatments resulted in significantly lower corticosterone levels in females at the time of sacrifice but did not significantly affect corticosterone levels in males [one pellet Ftreatment x sex(1,48) = 5.3, P = .03; two pellets Ftreatment x sex(1,49) = 11.2, P = .002; five pellets Ftreatment x sex(1,37) = 19.2, P = .0001]. *, P < .05 vs. the placebo group. n = 10-15 rats/group.

In an independent study, Scatchard analysis of pooled serum from animals implanted with two placebo or morphine pellets for 7 days indicated that the morphine-induced increase in CBG binding in males was attributable to an increase in the maximal number of binding sites (B_max = 5.9 and 15.9 pmol/mg protein for placebo and morphine groups, respectively) rather than binding affinity (K_d = 0.64 and 0.67 nM for placebo and morphine groups, respectively). Morphine had little effect on either the maximal number of binding sites in females (B_max = 21 and 23 pmol/mg protein for placebo and morphine groups, respectively) or binding affinity (K_d = 0.69 and 0.63 nM for placebo and morphine groups, respectively).

Figure 4 shows the concentration of CBG, corticosterone and morphine in serum of females at selected time intervals after implantation of two placebo or morphine pellets. Morphine levels were highest on the day after pellet implantation and decreased to relatively low levels by day 14. Morphine had no significant effect on CBG in females at any time interval examined. It might be noteworthy, however, that CBG levels were markedly depressed in both the placebo and morphine groups on day 1, perhaps as a result of the stress associated with pellet implantation on the preceding day. Corticosterone levels in the morphine-treated females were significantly suppressed at the time intervals examined.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   CBG (top), total corticosterone (middle) and morphine (bottom) in serum of females at selected time intervals after implantation of two placebo or morphine pellets. The figure shows that exposure to morphine did not significantly affect CBG at any time interval but resulted in significantly lower corticosterone levels [Ftreatment(1,82) = 33.5, P = .0001]. *, P < .05 vs. the placebo group. n = 11-12 rats/group.

In an attempt to determine whether the sex difference in the effects of morphine on CBG might be attributable to the sex difference in corticosterone levels and/or the effects of morphine on those levels, male and female rats were ADX or sham-operated. Two weeks later, rats from each group were implanted with one morphine or placebo pellet, and CBG was measured 7 days later. ADX increased the concentration of CBG in both males [sham placebo = 7.0 ± 0.4 pmol/mg protein vs. ADX placebo = 19.5 ± 1.3 pmol/mg protein; t(33) = 7.9, P = .0001] and females [sham placebo = 16.0 ± 1.0 pmol/mg protein vs. ADX placebo = 26.4 ± 1.0 pmol/mg protein; t(28) = .5, P = .0001]. Morphine increased the concentration of CBG in sham-operated males [sham placebo = 7.0 ± .4 pmol/mg protein vs. sham morphine = 13.4 ± .9 pmol/mg protein; t(24) = 7.0, P = .0001] but not sham-operated females [sham placebo = 16.0 ± 1.0 pmol/mg protein vs. sham morphine = 14.5 ± .9 pmol/mg protein; t(27) = 1.1, P = .29]. We were unable to assess the effects of morphine on CBG in ADX animals because a large proportion of ADX males (15 of 22) and females (12 of 18) treated with morphine died. None of the ADX males (0 of 21) or ADX females (0 of 16) that were implanted with placebo pellets died. Thus, ADX appears to increase sensitivity to morphine to the extent that a single pellet is lethal to most rats.

In a follow-up study with females, we repeated the experiment described above but provided corticosterone in the drinking water of the ADX rats throughout the experiment. This produced relatively low corticosterone levels (ADX placebo = 27 ± 7.5 ng/ml; ADX morphine = 37.1 ± 4.8 ng/ml) similar to those seen in males (see fig. 3). Furthermore, the ADX morphine and placebo-treated females that were maintained on corticosterone had similar hormone levels (see above) in contrast to intact morphine and placebo-treated females (see fig. 3) and the sham-operated morphine and placebo-treated females in this experiment (sham placebo = 173 ± 28 ng/ml; sham morphine = 92 ± 20 ng/ml). When corticosterone was provided in the drinking water, ADX did not significantly affect CBG levels. Morphine did not significantly affect CBG levels in sham-operated or ADX females (fig. 5).

View larger version (53K): [in this window] [in a new window]   Fig. 5.   CBG levels in serum of adrenalectomized and sham-operated female rats 7 days after implantation of one morphine or placebo pellet. Adrenalectomy was performed two weeks before drug exposure, and the females were provided corticosterone in their drinking water throughout the experiment. The figure shows that morphine had no effect on CBG in adrenalectomized or sham-operated females [Ftreatment(1,43) = .584, P = .45]. n = 13-16 rats/group. Also, note that in contrast to the first experiment with adrenalectomized rats (see text), adrenalectomy did not cause a marked increase in CBG levels when corticosterone was provided in the drinking water for the duration of the experiment.

To determine whether the sex difference in the effects of morphine on CBG might be attributable to gonadal hormones and/or the effects of morphine on those hormones, male and female rats were gonadectomized or sham-operated. Two weeks later, rats from each group were implanted with two morphine or placebo pellets, and CBG was measured 7 days later. Gonadectomy did not significantly affect CBG levels, and morphine continued to exert gender-typical effects in gonadectomized animals; that is, morphine significantly increased CBG in sham-operated and castrated males but had no apparent effect in sham-operated or ovariectomized females (fig. 6). With regard to corticosterone levels, there was no significant effect of either CAST or morphine in males (sham placebo = 65 ± 13; sham morphine = 40 ± 12; CAST placebo = 59 ± 11; CAST morphine = 47 ± 10 ng/ml). OVX decreased corticosterone levels in females [F(1,46)_surgery = 6.1, P = .02]. Morphine decreased corticosterone levels in both sham-operated and OVX females [sham placebo = 238 ± 28; sham morphine = 56 ± 12 ng/ml; OVX placebo = 153 ± 30; OVX morphine = 35 ± 6 ng/ml; F(1,46)_{drug treatment} = 49, P = .0001]. Morphine levels were similar in all of the morphine-treated groups (male sham = 617 ± 34; male CAST = 609 ± 55; female sham = 578 ± 33; female OVX = 687 ± 52 ng/ml).

View larger version (61K): [in this window] [in a new window]   Fig. 6.   CBG in serum of gonadectomized or sham-operated adult male and female rats 7 days after implantation of two morphine or placebo pellets. Gonadectomy was performed 2 weeks before drug exposure. The figure shows that morphine significantly increased the concentration of CBG in castrated and sham-operated males [Fdrug treatment(1,42) = 103, P = .0001; Fsurgery(1,42) = 2.7, P = .11; Fdrug treatment x surgery(1,42) = 3.1, P = .09] but had little or no effect on CBG levels in ovariectomized or sham-operated females [Fdrug treatment(1,46) = .05, P = .47; Fsurgery(1,46) = .07, P = .80; Fdrug treatment x surgery(1,46) = .3, P = .87]. Gonadectomy had no apparent effect on CBG levels in either sex. *, P < .05 vs. the placebo group. n = 12-15 rats/group.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Chronic exposure to morphine increased the concentration of CBG in serum of adult male but not female rats. In males, exposure to a single pellet for 7 days was sufficient to cause a marked increase (120%) in CBG. That appeared to be a maximal increase for 7 days of exposure because two morphine pellets caused a similar increase (110%), as did sequential implantation of five pellets (115%). None of these morphine treatments affected the concentration of CBG or its affinity for corticosterone in females. Furthermore, in a time-response experiment, CBG levels did not differ from control values at 1, 3, 7 or 14 days after the implantation of two morphine pellets. Thus, there is a pronounced sex difference with regard to the effects of morphine on CBG.

We did not carry out extensive experiments comparing the pharmacokinetics of morphine in males and females in the present study. However, at the time of death, serum morphine levels were similar in the two sexes. In addition, we reported elsewhere that a single morphine pellet produces similar drug levels in serum and brain of male and female rats. Moreover, the disappearance rate from serum and brain after morphine pellet withdrawal is similar in males and females (Cicero et al., 1997). We have also reported that acute injections of morphine over the dose range of 2.5 to 15.0 mg/kg produce similar peak levels of drug in blood and brain and that there is no sex difference in the half-elimination rate from blood or in the disappearance of morphine from brain (Cicero et al., 1997). Thus, it appears unlikely that there is a sex difference in morphine pharmacokinetics or metabolic tolerance that is sufficiently large to explain the sex difference in the effects of morphine on CBG. This conclusion is further supported by the finding that morphine treatments that resulted in serum morphine levels five times higher than needed to induce a maximal increase in males did not affect CBG levels in females.

Another factor that could conceivably underlie the differential effect of morphine on CBG in males and females is the sex difference in CBG concentration. On the average, we found CBG levels to be more than twice as high in females than in males, which is in good agreement with the findings of others (Gala and Westphal, 1965; Mataradze et al., 1992). One possibility is that morphine does not increase CBG in females because levels are already maximal. Although we cannot completely dismiss that possibility, we did find that adrenalectomy increased CBG in females by 60% to 65%. Thus, morphine failed to increase CBG in females even though up-regulation appears to be possible.

Corticosterone levels were also substantially higher in females than in males at the time of sacrifice. In agreement with our previous report (Nock et al., 1997), chronic exposure to morphine had no apparent effect on corticosterone levels in males. However, corticosterone levels were significantly lower in morphine-treated females than in placebo-treated females. Although only crude dose-response studies can be conducted using pellets, this effect may be dose dependent. On the average, corticosterone levels at 1 week after implantation of a single morphine pellet were ~45% below control levels, whereas levels were ~85% lower after two pellets. Corticosterone levels were 82% lower than control values on day 7 after sequential implantation of five morphine pellets. Thus, two pellets appear to produce a maximal effect. In a time-response study, lower corticosterone levels were seen at 1, 3, 7 and 14 days after implantation of two morphine pellets. However, the magnitude of the difference decreased markedly from day 7 to day 14 as the amount of morphine in blood decreased to relatively low levels. Thus, corticosterone levels appear to return toward normal as morphine clears from the general circulation.

We are uncertain at present as to whether the sex difference in the effects of morphine on corticosterone are entirely attributable to effects on basal corticosterone in females or whether the known greater reactivity of females to stressors also contributed (Handa et al., 1994). Although every precaution was taken to disturb the rats as little as possible before sacrifice (see Materials and Methods), we cannot rule out the possibility that some amount of adrenal activation might have occurred in females. If so, that activation was suppressed by morphine. That, in itself, would be rather interesting because we previously reported that chronic exposure of males to morphine has little or no effect on the corticosterone response to mild stress (Nock et al., 1997).

Although CBG levels are known to be influenced by corticosterone (Gala and Westphal, 1966), it does not seem likely that the sex difference in corticosterone levels or the effects of morphine on those levels underlies the sex difference in the effects of morphine on CBG. First, in a study to directly assess that possibility, adrenalectomized females were maintained on corticosterone in the drinking water. This treatment is known to produce corticosterone levels in adrenalectomized rats that follow naturally occurring circadian rhythms (Jacobson et al., 1988). In addition, it produced corticosterone levels in females that were similar to those in males and, like those in males, were unaffected by exposure to morphine. Thus, the addition of corticosterone to the drinking water essentially eliminated the sex difference in corticosterone levels and in the effects of morphine on those levels. Nevertheless, morphine did not affect CBG levels under these conditions. Second, corticosterone exerts a negative influence on CBG (Gala and Westphal, 1966); that is, a decrease in corticosterone increases CBG. It is difficult to envision how a morphine-induced decrease in corticosterone would prevent an increase in CBG. Thus, the sex difference in the effects of morphine on CBG does not appear to be due to sex differences in corticosterone levels.

Perhaps the most obvious possibility is that gonadal steroids underlie the differential effect of morphine on CBG in males and females. As an initial test of that possibility, we examined the effects of morphine on CBG in males and females that had been gonadectomized for 2 weeks before drug exposure. Gonadectomy did not have a statistically significant effect on CBG levels. Furthermore, morphine continued to exert gender typical effects in gonadectomized animals; it significantly increased CBG in castrated males but had no apparent effect in ovariectomized females. Thus, the sex difference in the effects of morphine on CBG does not appear to be due to short-term actions of gonadal steroids in adulthood. Testosterone has been reported to masculinize the CBG phenotype during puberty, and thereby underlies the large sex difference in adult CBG concentrations (Mataradze et al., 1992). Whether testosterone action during puberty also masculinizes the responsiveness of CBG to morphine remains to be established.

In summary, our data clearly establish a sex-dependent effect of morphine on CBG. From the discussion, it is apparent that we can eliminate a number of potential factors that could contribute to this sex difference, but we are not yet in a position to speculate about the underlying mechanism. Nevertheless, there seems to be little question that morphine does not decrease the amount of free, physiologically active corticosterone through CBG in females as it appears to do in males. On the other hand, chronic exposure to morphine decreases corticosterone levels per se in females but not in males. Thus, chronic exposure to morphine lowers the amount of physiologically active corticosterone in rats of both sexes but by different mechanisms.

Morphine may have a similar effect on free glucocorticoid levels in humans. Although there has been no systematic study of that possibility as yet, Garrel (1996) recently referred to data that indicate an inverse correlation between free cortisol levels and amount of morphine given to patients. That is, the more morphine they had received, the lower free cortisol levels. Thus, chronic exposure to morphine may produce deficits in physiologically active glucocorticoid in humans, as well as rats. If so, it is likely to have adverse consequences for the mental and physical health of the individual over the long run. Glucocorticoids exert both permissive and suppressive actions to protect the body against stress (Munck and Náray-Fejes-Tóth, 1994). Permissive actions are exerted by basal glucocorticoid levels that "prime" defense mechanisms, enabling them to be fully expressed in response to stress (Ingle, 1954). Suppressive actions are exerted at stress levels and function to protect the body against damage from its own defense reactions, which can be harmful if they continue unconstrained. In the face of a morphine-induced decrease in physiologically active hormone, whether it results from an increase in CBG or a decrease in the hormone per se, basal and stress glucocorticoid levels may be inadequate to prime defense mechanisms or prevent them from overshooting in response to stress.

It is not difficult to envision ways by which such a situation might foster drug-seeking behavior. There is no question that a deficit in glucocorticoids impairs the ability to cope with stress and thereby increases its impact. The extreme example would be patients with Addison's disease who typically die within 2 years of diagnosis without hormonal therapy (Dunlop, 1963). A long-term deficit in physiologically active corticosterone brought on by chronic exposure to morphine may not have such dramatic results, but it is, nevertheless, likely to intensify the impact of a stressor. In this regard, it may be relevant that chronic opiate abuse is known to be a significant risk factor for the development of post-traumatic stress disorder (a long-term maladaptive behavioral syndrome precipitated by a severe stressor) and exacerbates post-traumatic stress disorder symptoms (Cottler et al., 1992; Stine and Kosten, 1995). Considering that stress has been shown to increase the propensity for opiate self-administration (Alexander et al., 1981; Carroll et al., 1979; Dib, 1985; Shaham et al., 1993), it seems reasonable to speculate that an opiate-induced increase in the intensity/impact of stress might be one factor contributing to the perpetuation/maintenance of drug use.