Peptides. Vol. 4, pp. 19-24, 1983. ~ Ankho International. Printed in the U.S.A.

Shuttlebox Avoidance in Rhesus Monkeys: Effects on Plasma Cortisol

and/3-Endorphin

A N D R E W C. S C A L L E T , 2 D O U G L A S R. R U S H , N E D H. K A L I N , S A M U E L C. R I S C H , 3 S T E P H E N J. S U O M I A N D T E R R Y Z A A S

Depar tment o f Psychology Primate Laboratory, University o f Wisconsin, Madison, I4/1 53706

Rece ived 7 D e c e m b e r 1982

SCALLET, A. C., D. R. RUSH, N. H. KALIN, S. C. RISCH, S. J. SUOMI AND T. ZAAS. Shuttlebox avoidance in rhesus monkeys: Effects on plasma cortisol and ~-endorphin. PEPTIDES 4(1) 19-24, 1983.--Groups of monkeys either extensively pretrained to avoid shocks in a shuttlebox or with minimal prior experience were compared for plasma cortisol and/~-endorphin levels immediately following: (I) an exposure to the box with no shock, (2) the box providing repeated inescapable shocks or (3) a re-exposure to the box, again with no shock presentation. Mere exposure to the unfamiliar box elevated plasma cortisol just as much as exposure + shock did when inexperienced monkeys were tested. However, animals with a history of previously successful shock avoidance showed smaller elevations when exposed to the box alone, than they did when inescapable shock was received. Plasma/3-endorphin levels following shuttlebox exposure showed only a sporadic pattern of elevations in either inexperienced or pretrained monkeys. However, levels of/3-endorphin as deter- mined under control conditions in the home cage were lower in pretrained animals, as were plasma levels of cortisol. The results indicate that behavioral factors may effect plasma cortisol and /3-endorphin following both acute and chronic shuttlebox experience.

Plasma cortisol Plasma/3-endorphin Stress Conditioning Controllability Avoidance Primates

IN rats, conditioned fear or sometimes mere exposure to an unfamiliar environment can activate the hypothalamo- pituitary-adrenal (HPA) axis, as indicated by elevated plasma levels of cortisol or/3-endorphin-like immunoreac- tivity [9, 17, 22]. These behavioral techniques can experi- mentally separate the neural activation accompanying actual pain from stressful experiences involving simply an anticipation of a negative consequence [22]. The present ex- periment employed similar behavioral methodology applied to rhesus monkey subjects in order to compare a range of stimuli as to their effectiveness as activators of the HPA axis, as measured by plasma cortisol response. Although blood levels of both/3-endorphin and ACTH respond to pain- ful stress in a parallel fashion in rodents [8], and/3-endorphin responds to even nonpainful conditioned fear in rodents [17,22], no such studies have been reported in primate spe- cies. Accordingly, we also wished to examine blood levels of /3-endorphin-like immunoreactivity accompanying such be- havioral protocols in rhesus monkeys.

Since we have previously reported on sex differences in cortisol response to agonistic encounters [21], both males and females were studied for possibly similar differences in their response to novelty and conditioned fear. Also, an ad-

ditionai group of males that had been extensively trained on shuttle avoidance as part of a separate study (D. Rush et al., unpublished) was available. Since the psychological dimen- sion of controllability of shock has been considered an im- portant determinant of aversiveness in behavioral studies in general [24] as well as in the phenomenon of naioxone re- versible stress-induced analgesia [7], these males were in- cluded in the present study, even though an equivalent group of females (in order to allow a complete factorial design) was unavailable. This group allowed us to study the hypothalamo-pituitary effects of extensive experience of control over shock as well as the subsequent experimental loss of that control.

A shuttlebox designed for primates [20] that delivered 3 mA constant current shocks was employed. The experi- menters determined that this amount of current was aversive but not tissue damaging by placing their own hands across the grids of the shuttlebox. The three groups of monkeys (n=4 per group) received an identical series of exposures to the shuttlebox environment. One group of males (loss of control or LC males) had been previously trained over a period of months immediately prior to the present study on a Sidman avoidance schedule to avoid or escape shocks by

~This study was supported by USPHS Grants MHi 1894 and MH28485. 2Current address for reprint requests: Department of Psychiatry, Washington University School of Medicine, 4940 Audubon, St. Louis, MO

63110. 3Department of Psychiatry, University of California, San Diego, CA 92037.

19

20 SCALLET ET AL.

shuttling between halves of the box. The two other groups (termed naive males and naive females) had never previously been exposed to the shuttlebox environment. All monkeys received an initial half-hour exposure to the shuttlebox with the shocker turned off, immediately followed by a blood draw to assess plasma cortisol (and in some cases plasma /3-endorphin) response to the unfamiliar (naive males, naive females) or highly familiar (LC males) shuttlebox environ- ment. Two days later, they were again exposed to the shuttlebox which now provided unavoidable shocks at vari- able intervals averaging 30 sec apart (VI-30 schedule) repre- senting a loss of control experience for LC males who could now no longer avoid shocks by shuttling, but merely an ini- tial experience with noxious shock for the naive monkeys. Two to three days after this, all monkeys were exposed for a third time to the shuttlebox, which (as in trial l) again had the shocker turned off, as a test of whether the environment had acquired any property of conditioned fear.

The study was directed towards several questions. First, would naive monkeys ' plasma cortisol levels be greater when the animals were exposed to the shuttlebox following an experience of aversive shock than they had been when the box represented merely a neutral unfamiliar environment? Would there be any significant differences in hormonal re- sponses between the sexes? Would the LC monkeys who had a great deal of experience of control over shock and familiarity with the shuttlebox show an exaggerated cortisol response to a loss of that control? Would LC monkeys differ in hormonal response as a result of their extensive prior training? And finally, would plasma fl-endorphin in monkeys be elevated by conditions of painful or nonpainful stress or evidence any co-regulation with plasma ACTH levels, as inferred from the cortisol results?

M E T H O D

Behavioral

Subjects were 12 laboratory born rhesus macaques (8 males and 4 females) between 4 and 6 years of age---all had been reared with mothers and peers. One group of 4 males had prior experience in the shuttlebox and had achieved stable baseline rates of shuttling on a Sidman avoidance schedule, e.g., repeated shuttling would avoid all shocks and any shuttle would terminate a shock already in progress. They each had received one hour test sessions every after- noon, five days per week, for at least three months immediat- ly preceding Day 1 of the experiment. These LC males shuttled sufficiently so that in the 2 days of sessions im- mediately prior to Day I, only 3 shocks of 3.7 sec duration were received collectively. The remaining 4 males and 4 females formed two groups with no prior exposure to the shuttlebox or to shocks. Trials were conducted in a two-way shuttlebox constructed of stainless steel (sides) and plexi- glass (top) measuring 140x70x70 cm, with entry doors at each end. The two halves of the box were separated by a barrier 30 cm high, and switches in the floors of each half registered the location of the monkey and when a shuttle had occurred. A shock generator and scrambler (Lehigh Valley Models 1531 and 1311 SS) delivered (when activated) a con- stant current 3 mA shock to grid bars located throughout the box on floors and walls, and the entire apparatus was housed in a separate experimental room lined with sound-absorbing material. A noise generator provided 65 db of white noise as measured inside the shuttlebox both to mask extraneous

noises and to act as a salient environmental cue for condi- tioning responses to the box environment. Programing of shock schedules was automated by equipment in an adjacent room. Further details of the apparatus are described elsewhere [20].

Four monkeys per day were tested between the hours of 1600 and 1900. They each received blood draws immediately following removal from the shuttlebox or from their home cage. A record was kept of the time from initial handling to blood withdrawal for each sample, and these times were al- ways between 30 and 180 seconds. Such sampling technqiues adequately reflect plasma cortisol levels at the time point of initial handling, since there is a delay of 2-3 rain before exog- enous ACTH perfused through the catheterized renal portal artery elevates portal venous cortisol [27] and plasma cor- tisol has a relatively long (90-120 rain) half-life [16, 25, 26]. Each monkey received blood draws on four occasions: Days 1-3 were consecutive and Day 4 was a retention test 48-72 hr later. Day 1 constituted a "'Box Alone I" condition during which animals were placed in the shuttlebox for 30 min but received no shock. Day 2 was a " H o m e Cage Control" con- dition during which monkeys received no treatment except rapid withdrawal from their home cage and venipuncture at the same time of day they had been bled on Day I. Day 3 was a "Box + Shock" condition during which each monkey re- ceived a total of 50 unavoidable 5-sec 3 mA shocks on a VI-30 schedule for 25 min, following a 5-min adaptation to the box. Finally 48 hr (half of each group) or 72 hr (the other half) after the Box + Shock session, monkeys received a Day 4 session identical to Day l, again with the shocker off, e.g., a "'Box Alone I I" or conditioned retention trial. Lim- ited availability of the shuttlebox because of other concur- rent studies necessitated the two retention intervals. How- ever, there were no systematic differences in any of the data for these time points, and they were combined for statistical analysis.

Biochemical

All blood samples were collected into chilled EDTA treated vacutainers which were subsequently kept on ice until centrifuged at 3°C within 30 rain of sampling. Plasma was stored in coded plastic vials at -70°C until subsequent assay. Cortisol determinations were performed on coded samples using radioimmunoassay kits (New England Nu- clear, 549 Albany St, Boston, MA 02118) employing a rabbit antiserum to cortisol 21-succinyl bovine albumin. The antiserum cross-reacts 25.3% with corticosterone. Unex- tracted 10/~1 plasma samples are assayed with a lower limit of sensitivity of 2/~g%, and the basal and stress-responsive levels reported here and in unpublished pilot work appear quite comparable to results we have obtained with previ- ously described competit ive protein binding techniques [2,21]. In brief, the procedure consists of a 15-30 rain incu- bation of samples and standards with I ~2s labeled cortisol and antiserum, followed by centrifugation and gamma counting of the resulting pellet. Unlabeled cortisol standards are supplied pre-weighed in vials ready for reconstitution and use. The intraassay coefficient of variation is about 6%, and the interassay coefficient is 7%.

17 _ or assay offl-endorphin-like immunoreactivity, I to 2 ml

aliquots of monkey plasma were acidified with HCI to pH 4 and extracted with 60 mg/ml of silica gel. The silica was washed with cold water and cold acid, and then the B-endorphin-like immunoreactivity was eluted from the gel

AVOIDANCE, CORTISOL, AND fl-ENDORPHIN 21

500-

z 400- cn ¢n uJ u~

300-

..1

200-

<

~ ~00-

LOSS OF CONTROL MALES

NAIVE MALES

I I NAIVE FEMALES

Box Alone Sox +Shock I

(Novelty)

BOX AlOne 2

(RetenhOn)

FIG. 1. The total number of shuttles performed by each group during each thirty minute test session under the various conditions is plot- ted. The loss of control male monkeys previously well trained in many avoidance sessions shuttle more than naive males, and are the only group to increase their shuttle rate in response to uncontrolla- ble shocks. Naive female monkeys showed a tendency towards more shuttling activity than naive males, resulting from two individ- uals with high shuttle rates. Bars illustrate means plus SEMs, and *p<0.01 from LC males' rates during the Box Alone I and II condi- tions.

into cold 50% acetone. The acetone was dried under nitrogen at 45°C in siliconized glass tubes, and the immunoreactivity was reconstituted into 0.25 ml of buffer (50 nM Na monobasic phosphate, pH 7.4, 5 nM NaCI, 0.1% BSA, 0.02% azide and 0.6% Triton X-100) for subsequent assay of 0.10 mi aliquots in duplicate as described by Ghazarossian et al. [6] utilizing an antiserum directed against human /3-endorphin showing 100% molar cross-reactivity with/3-1ipotropin pro- vided by V. Hrllt. This antibody has shown cross- reactivities of less than 0.1% with met-enkephalin, ct-endorphin (/3-LPH~,_7~) and ACTH,_3, but cross reacts 10-20% with porcine fl-LPH78 ~,, indicating a C-terminal antigenic determinant and thus a probable crossreactivity with N-acetylated/3-endorphin [i I].

One-tenth ml volumes of standard or sample in buffer were incubated at 4°C with 0.10 ml of antiserum (diluted 1:4000) for 24 hr, followed by the addition of 0.10 ml of trace for 24 hr and separation of free from bound hormone with 1.0 ml of a dextran-charcoal slurry. The minimum detectable amount of/3-endorphin-like immunoreactivity was about 2 fmoles per tube, binding of iodinated synthetic human fl-endorphin was 50% inhibited (IC.~,) by 7 fmoles of standard human fl-endorphin, and the Iog-probit plot of the standard curve was linear from 3-30 fmoles. Several monkey plasma extracts assayed at different dilutions in buffer were found to parallel the standard curve, suggesting an absence of any nonspecific interference. Recovery was 72_+4% as deter- mined by spiking fl-endorphin-free (e.g., previously ex- tracted) plasma with synthetic fl-endorphin followed by the silica extraction and radio-immunoassay. All values reported are corrected for recovery and expressed as f-moles of fl-endorphin-like immunoreactivity per ml of the original monkey plasma sample. The assay for /3-endorphin de- scribed here is the same as previously utilized by me to demonstrate a range of elevations of rat plasma ~-endorphin following footshock, conditioned fear or exposure to novel environments [221.

50-

cD 40 -

3 0 -

20-

I0.

gO1 Alone I

(NOvelty)

LOSS OF CONTROL MALES NAIVE MALES

I I NAIVE FEMALES

,t.

i i i

Home Coqe Box + Shock Box Control Alone 2

(Retenhon)

FIG. 2. Plasma cortisol levels in/~g per 100 ml are plotted for all groups under the various conditions. The naive male and naive female groups show similar elevations over their own home cage control values following each exposure to the shuttlebox environ- ment, regardless of the presence or absence of uncontrollable shock. The loss-of-control males have consistently lower cortisol levels than the other groups, and show a significant plasma elevation in the shuttlebox only when uncontrollably shocked. Statistical compari- sons are by analysis of variance with Fisher's Least Significant Difference Test employed for individual comparisons, *p<0.01 from that group's home cage control mean, tp<0.05 from the mean of the other groups within a given condition, and ttp<0.05 from that group's home cage control mean.

RESULTS

Behavioral

The pattern of shuttling shown by the three groups of monkeys (naive females, naive males, and LC males) is illus- trated in Fig. 1. Although there was no systematic shuttling pattern shown by the naive monkeys, the rate of shuttling by the LC males increased from a baseline mean of 123 per session to 283 during the session of uncontrollable shock, and subsided to 153 during the retention test (groups by re- peated sessions interaction, F(4.18)=4.83, p<0.01, individ- ual comparisons of Box + Shock to Box Alone I and Box Alone II, t(3)'s=7.09 and 5.89, respectively, both p's<0.01). The LC males tended to shuttle more during the retention trial (Box Alone II) than during their Box Alone I sessions, t(3)=2.97, p<0.10. These data indicate that the shocks were indeed aversive, since the LC males were willing to shuttle more to attempt to avoid them. In fact, during the retention trial with the shock inactivated, the shuttle rates of each LC male declined gradually between the beginning and end of the trial (data not shown), suggesting that for well-trained monkeys the rate of shuttling may be a useful index of the aversiveness of the situation. There was a trend for LC males and naive females to shuttle more than naive males, F(2,9)= 1.99, N.S. Although for the females this was largely due to two individuals with high shuttle rates, we have pre- viously observed a similar tendency for greater activity and locomotion by female monkeys in novel environments ([21] and Scallet, unpublished observations).

Plasma Cortisol

Plasma cortisol results (see Fig. 2) showed main effects for groups, F(2,9)= 10.7, p<0.01, as well as repeated meas-

22 SCALLET ET AL.

TABLE I

PLASMA/3-ENDORPHIN-LIKE IMMUNOREACTIVITY IN MALE MONKEYS FOLLOWING 30 MIN EXPOSURE TO THE SHUTTLEBOX WITH OR WITHOUT THE

PRESENCE OF UNCONTROLLABLE SHOCK

Home Cage Box + Box Alone 2 Box Alone I Control Shock (Retention)

fmole/ml fmole/ml fmole/ml fmole/ml fiE /3E /3E /3E

LC Males AD48 <2 5 > 176 2 AD69 6 6 3 > 176 AC37 <2 9 - - - - AC42 5 6 25 2

Mean ± SEM 3.75 ± 1.0 6.5 ± 0.86* 68 ± 54 60 ± 58

Naive Males AA92 AA82 AA32 AC57

Mean ± SEM

11 46 13 24 29 6

8 8 > 176 18 14 10

15.25 ± 3.6 24.25 ~ 8.5 51.25 ± 41

*p<0.05 from the mean of naive males.

ures, F(3,27)=14.39, p<0.001, but no overall groups by re- peated measures interaction, F(6,27)=1.25, p>0.10. The group main effect indicated that the LC males had lower cortisol values than either the naive males or females when considered throughout all four repeated trials in the shuttlebox. The repeated measures effect demonstrated that cortisol levels varied with trials, while the absence of a group by sessions interaction suggests a similar response to trials by all three groups. Fisher's Least Significant Difference Tests (a technique that preserved a per comparison, [5]) were applied for individual comparisons between groups within a given trial (LSD.0~ (9 df)= 19.7/~g) and for comparisons of each group's mean during one session with its mean during each other session (LSD.0~ (27 dr)= 10 p.g%; LSD.01 (27 dJ)= 13.5/xg%). Both the naive females and naive males showed elevated cortisol levels (all p 's<0.01) following Box Alone I, Box + Shock, and Box Alone II sessions as compared to their Home Cage Control means. However, the novel experience of their first exposure to the shuttlebox was just as effective as repeated shocks or re-expsoure to the box at elevating plasma cortisol e.g., no means for naive males or females in the three box conditions were significantly different from any other. By contrast, LC males showed a significant elevation over con- trol levels only when they were exposed to unavoidable shock during the Box + Shock condition (,o<0.05). While they showed trends towards elevations on their other trials in the shuttlebox as well, these were not statistically reliable (p's>0.10). Such results suggest that extensive familiarity with the box can reduce the magnitude of the cortisol re- sponse to a 30 rain session, as long as shock is omitted.

Plasma fl-Endorphin

There was insufficient plasma remaining to assay /3-endorphin from all animals in all conditions by our proce- dures although the LC males' samples were adequate (e.g.

>1 ml). Therefore, six weeks after their initial trials, the naive males group was tested for a second time. Since we were interested in establishing a conditioned fear of the shuttlebox environment, a Box Alone I condition was con- sidered undesirable in this series, because it would act as an extinction trial. Therefore, Day 1 was a Home Cage Control blooddraw, Day 2 was a re-exposure to the Box + Shock condition and 48-72 hr later we tested the (originally) naive males' response to the shuttlebox alone. All samples from this repeat testing of the naive males group were analyzed for both cortisol and/3-endorphin, and the samples from the LC males were analyzed for/3-endorphin. Table I lists the indi- vidual/3-endorphin values from the original set of trials of the LC males, and from the second test series of the naive males group.

The naive males group again showed cortisol elevations when Box + Shock (mean=29.7 + _ 1.8 .~g%) and Box Alone (mean=27.5+_ 1.7 p.g%) conditions were compared to Home Cage Control values (mean=15_+3 tzg%; F(2,6)=12.66, p<0.01). As compared to levels for the naive males, second test series, the Home Cage Control cortisol values of the LC males were still depressed (mean=6.7__+ 1.3 p.g%, t(6)=2.59, p<0.05). Since co-regulation of ACTH and /3-eudorphin under certain circumstances in rats and humans has been described [8], one might predict lower/3-endorphin as well as cortisol levels among the LC males, and this was in fact observed, t(6)=2.41, p <0.05 one-tailed, and see Table 1. In- spection of Table 1 also shows that high individual values of /3-endorphin were seen only among the samples collected during the more stressful conditions of Box + Shock and Box Alone II, although high variability between subjects in these conditions precludes any statistical evaluation.

DISCUSSION

For naive subjects, the Box Alone I condition provided a first experience in the unfamiliar shuttlebox environment

AVOIDANCE, CORTISOL, AND f l -ENDORPHIN 23

which proved to be a potent activator of the HPA axis for each such monkey. In classical conditioning terminology, the shuttlebox itself was a sufficient unconditioned stimulus (UCS) to evoke a near maximal unconditioned response (UCR)---namely, activation of the HPA axis as measured by plasma cortisol levels. Since the shuttlebox itself was just as effective a UCS as the series of uncontrollable shocks, no conditioning interpretation is possible. Although the shock schedule probably was more aversive than the box a lone- -as judged by video remote observation of monkeys barking and threatening during shock-- the cortisol results failed to re- flect any difference. The utility of plasma cortisol as a con- sistently reliable index of the degree of behavioral arousal has been previously criticized [18,26], and our results simi- larly demonstrate the presence of a potent ceiling effect for naive monkeys ' responses under the conditions of this exper- iment.

There were no significant differences between naive males and females in control levels of cortisol or in response to any of the Box conditions, although there was a trend for females to show higher cortisoi levels, in agreement with our previous observations [2 I]. Preliminary results with ovariec- tomized and castrated adults (Scallet et al. , unpublished data) suggest that male-female adrenocortical differences are not as prominent and are not primarily estrogen mediated in rhesus monkeys as they are in rodents.

In contrast to the naive monkeys, the LC males showed low basal values and only small, nonsignificant cortisol eleva- tions following a routine Box Alone I exposure to an en- vironment in which shock had been avoidable for several months. When they experienced unexpected shocks and loss of control over shock, however, their elevations were larger, although still not as large as the naive monkeys ' elevations from any of the conditions. Thus while the LC males' cor- tisol levels may be varying in a different range than the naive monkeys, their acute responses do appear sensitive to the degree of familiarity or controllability of the box, as demon- strated previously for rats [9]. Perhaps plasma cortisol re- sponses are reliable indicators of behavioral arousal when examined in certain situations devoid of ceiling effects, while they appear insensitive under more extreme conditions, e.g., in studies which have varied the number or timing of multi- ple shock schedules and assayed cortisol response [1,4].

For the LC males, the retention (Box Alone II) trial can be viewed as a combined reflection of their extensive previ- ous experience of control over shock in the shuttlebox plus their more recent loss of control for a single session. The absence of significant cortisol elevations following either Box Alone condition as well as the declines of shuttle rate across the 5 min blocks of the LC monkeys ' Box Alone II trail suggest that extensive prior experience of control over shock can reduce at least the acute indications of the aver- siveness of the shuttlebox. Both the behavioral and cortisol data show no evidence of a prolonged conditioning effect at the time of the retention trial.

It is noteworthy that the LC males evidence reduced levels of plasma cortisol throughout all experimental condi- tions as compared to the groups of naive monkeys. Mason [15] has previously demonstrated reductions in plasma cor- tisol levels that were attributed to chronic stress following repeated sessions of Sidman avoidance utilizing a bar press- ing paradigm in rhesus monkeys. Our results with a similar behavioral protocol suggest that repeated sessions of shuttlebox avoidance may have the same effect, namely a chronic inhibition of the HPA axis [15]. The apparent

chronic inhibition of the HPA axis in the LC males does not compromise our observation of a differentially sensitive re- sponse to the familiar box vs. Box + Shock. However it does limit comparisons between the adrenocortical effects of un- controllable shocks in the naive monkeys versus effects of the same shock schedule on LC monkeys since the latter start at a lower cortisol level and may vary over a different range. Our finding of decreased plasma fl-endorphin-like immunoreactivity as well as cortisol in the LC monkeys suggests that the effect can be observed at the level of the pituitary as well as the adrenal. If we can infer lower plasma ACTH levels from the cortisol results for the LC monkeys then the data would also suggest a co-variation of plasma ACTH and/3-endorphin in response to repeated avoidance trials. However, direct assay of plasma ACTH levels would clearly be a useful addition to future investigations, since alternative explanations, for example, a chronic reduction of cortisol leading to a pituitary shift toward production of ACTH and away from/3-endorphin, cannot be ruled out by the present data.

Under conditions involving stressful activation of the HPA axis, plasma immunoassayable /3-endorphin values were inconsistent and variable, sometimes elevated above the range of sensitivity of the assay and sometimes as low or lower than the Home Cage Control values for that animal (see Table 1). Since the half-life for/3-endorphin in plasma has been estimated to be much shorter than for cortisol (e.g., 5 min, [19]), the value of/3-endorphin observed in a given plasma sample would be more reflective of the immediately preceding state of pituitary activation than is plasma cortisol, which provides a picture of pituitary activation more inte- grated over some considerable length of time, due to its slow response and long half-life. Since pituitary secretion of ACTH has been shown to be subject to rapid onset rate- sensitive negative feedback effects from rising plasma cor- tisol levels [13], a plausible explanation of the results might be that /3-endorphin is similarly subject to such negative feedback and that 30 min following the onset of the stressful conditions, some of the monkeys ' pituitaries had been inhib- ited by rising cortisol levels but some had not. Future studies may have to take account of the rapid alterations of /3-endorphin levels by providing for more frequent time points sampled under stable psychological conditions, e.g., from monkeys adapted to chronic indwelling catheters for unobtrusive blood draws.

The present study has compared several behavioral methodologies for activation of the HPA axis in monkeys, as indexed by plasma measures of cortisol and /3-endorphin. The conditioning approach taken here indicated that the po- tent properties of any novel environment as a UCS for plasma cortisol limit the possibilities of employing it as a CS, except perhaps by a process of differentially conditioning responses to " sa fe" or "nonsafe ," but equally familiar, rooms or complexes of stimuli within a single room. In fact, since simple exposure to a novel room is such a strong stimulus, the results also suggest it may be unnecessary to employ more noxious stimuli such as shock in primate studies of HPA activation. Prolonged experience of control over shock (and familiarity with an environment) may reduce the magnitude of the acute cortisol response to a routine avoidance session as compared to a period of loss of control. These results indicate that plasma cortisol response may be a sensitive index of behavioral arousal when sampled under conditions minimizing the likelihood of ceiling effects. In addition, prolonged experience of control over shock

24 SCALLET ET AL.

through repeated avoidance sessions may lower the starting levels of plasma cortisol as well as /3-endorphin, at least when measured at times during the afternoon when sessions were routinely conducted. Basal levels of plasma /3-endorphin-like immunoreactivity in naive rhesus monkeys during the afternoon (15.0_+3.6 fmol/ml) appear lower than levels I reported for rats (25. l _+2.3 fmol/ml, 22) shortly after lights-on in the morning (roughly equivalent time points, jud- ing by the circadian literature on cortisol and fl-endorphin in plasma, [3, 10, 23] but were generally higher than levels found in humans by the same assay ([6], J. Dahl, personal communication). This observation may correspond with the presence in rhesus monkeys of a substantial pituitary inter- mediate lobe [12] containing immunoreactive /3-endorphin ([14], and S. J. Watson, oral presentation, NIDA meeting 1982). Under several conditions involving clear prior activa-

tion of the HPA axis as indexed by consistent plasma cortisol elevations in all cases, plasma /3-endorphin was elevated sporadically only in some animals under some conditions. It remains to be determined whether the sporadic/3-endorphin responses seen represent individual differences in respon- sivity to the behavioral conditions or if they result instead from blood sampling at a time of individual differences in the degree of pituitary inhibition by negative feedback of cor- tisol.

ACKNOWLEDGMENTS

I would like to thank Drs. Robert Bowman and June Dahl for providing laboratory space and consultation necessary to the con- duct of this research. The assistance of Pat Tollund, Wendy Ehrman, and Bob Tyllo in data collection, and Sophia Goodman in preparation of the manuscript is also appreciated.

REFERENCES

1. Bassett, J. R., K. D. Calrncross and M. G. King. Parameters of novelty, shock predictability and response contingency in corti- costerone release in the rat. Physiol Behav Ilk. 901-907, 1973.

2. Bowman, R. E. and R. F. DeLuna. Assessment of a protein binding method for cortisol determination. Anal Biochem 26: 465-469, 1969.

3. Dent, K. R. M., C. Guilleminault, L. H. Albert, B. I. Posner, B. M. Cox and A. Goldstein. Diurnal rhythm of plasma im- munoreactive fl-endorphin and its relationship to sleep stages and plasma rhythms of cortisol and prolactin. J Clin Endocrinol Metab 52: 942-947, 1981.

4. Friedman, S. B., R. Ader, L. J. Grota and T. Larson. Plasma corticosterone response to parameters of electric shock stimu- lation in the rat. Psychosom Med 29: 323-328, 1967.

5. Fryer, H. C. Concepts and Methods of Experimental Statistics. Boston: Allyn and Bacon, 1966.

6. Ghazarossian, V. E., R. R. Dent, K. Otsu, M. Ross, B. Cox and A. Goldstein. Development and validation of a sensitive radioimmunoassay for naturally occurring /3-endorphin-like peptides in human plasma. Anal Biochem 102: 80-89, 1980.

7. Grau, J. W., R. L. Hyson, S. F. Maier, J. Madden IV and J. D. Barchas. Long-term stress-induced analgesia and activation of the opiate system. Science 213: 1409-1411, 1981.

8. Guillemin, R., T. Vargo, J. Rossier, S. Minick, N. Long, C. Rivier, W. Vale and F. Bloom. /3-Endorphin and adrenocor- ticotropin are secreted concomitantly by the pituitary gland. Science 197: 1367-1369, 1977.

9. Hennessy, M. B. and S. Levine. Sensitive pituitary-adrenal re- sponsiveness to varying intensities of psychological stimulation. Physiol Behav 21: 295-297, 1978.

10. Holaday, J. W., H. M. Martinez and B. H. Natelson. Syn- chronized ultradian cortisol rhythms in monkeys: persistence during corticotropin infusion. Science 198: 56--58, 1977.

11. H611t, V., O. A. M011er and R. Fahlbusch. /~-Endorphin in human plasma: basal and pathologically elevated levels. Life Sci 25: 37--44, 1979.

12. Holmes, R. L. and J. N. Ball. The Pituitary Gland:A Compara- tive Account. London: Cambridge University Press, 1974.

13. Jones, M. T., B. Gillham, B. D. Greenstein, U. Beckford and M. C. Holmes. Feedback actions of adrenal steroid hormones. Curr Top Endocrinol 2: 45-68, 1982.

14. Lowry, P. J. and A. P. Scott. The evolution of vertebrate cor- ticotropin and melanocyte stimulating hormone. Gen Comp Endocrinol 26: 16--23, 1975.

15. Mason, J. W. Organization of psychoendocrine mechanisms. Psychosom Med 30: 565-808, 1968.

16. Meyer. J. S. and R. E. Bowman. Rearing experience, stress, and adrenocorticosteroids in the rhesus monkey. Physiol Behav 8: 339-343, 1972.

17. Meyerhoff, J. L., B. N. Bunnell and E. H. Mougey. Plasma /3-endorphin in rats is increased by psychological stress. Soc Neurosci Abstr 7: 166, 1981.

18. Natelson, B. H., N. Krasnegor and J. W. Holaday. Relations between behavioral arousal and plasma cortisol levels in mon- keys performing repeated free-operant avoidance sessions. J Cornp Physiol Psychol 90: 958-969, 1978.

19. Pezalla, P. D., M. Lis, N. G. Seidah and M. Chretien. Lipotro- pin, melanotropin, and endorphin: in vivo catabolism and entry into cerebrospinal fluid. Can J Neurol Sci 5: 183-188, 1978.

20. Rush, D. R. A two-way shuttlebox for testing nonhuman pri- mates. Behav Res Methods lnstrum. 1982, in press.

21. Scallet, A. C., S. J. Suomi and R. E. Bowman. Sex differences in adrenocortical response to controlled agonistic encounters in rhesus monkeys. Physiol Behav 26: 385-390, 1981.

22. Scallet, A. C. Effects of conditioned fear and environmental novelty on plasma fl-endorphin in the rat. Peptides 3: 203-206, 1982.

23. Seggie, J., B. Shaw, I. Uhlir and G. M. Brown. Baseline 24-hour plasma corticosterone rhythm in normal, sham-operated, and septally-lesioned rats. Neuroendocrinology 15: 51-61, 1974.

24. Seligman, M. E. P. Helplessness. San Francisco: U. H. Freeman, 1975.

25. Setchell, K. D. R., L. H. Rees and R. L. Himsworth. Effect of acute inhibition of adrenocorticotrophin secretion on plasma corticosteroids in the rhesus monkey (Macaca mulatta). J Endocrinol 67:251-257, 1975.

26. Smith, G. P. Adrenal hormones and emotional behavior. Prog Physiol Psychol 5: 299-351, 1973.

27. Urquhart, J. and C. C. Li. Dynamic testing and modeling of adrenocortical secretory function. Ann NY Acad Sci 156: 756- 778, 1969.

28?Wolf, R. C. and R. E. Bowman. Adrenal function in the infant rhesus monkey. Endocrinology 72: 146-148. 1963.