0022-3565/94/2703-1127$03.

O0/O
Tus Jouw.L OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS vol 270, No. 3
Copyright C 1994 by The American Society for Pharmacology and Experimental Therapeutics Printed in U.S.A.
JPET 27.1127-1133, 1994

A Mechanism of Action for Morphine-Induced

Immunosuppression : Corticosterone Mediates Morphine-Induced
Suppression of Natural Killer Cell Activity1

DAVID 0. FREIER2 and BRUCE A. FUCHS

Department of Pharmacology and Toxicology, Medical College of Virginia, Virginia Commonwealth UnWerslty, Richmond, Wrginia
Accepted for publicadon May 5, 1994

ABSTRACT
Morphine is a drug of abuse with an ability to down-regulate naloxone and naltrexone completely blocked morphine-induced
immune responsiveness that could have potermally serious con- suppression of NK activity, whereas naloxone methiodide, a
sequences in both heroin addicts and in the clinical environment. congener that crosses the blood-brain barrier much more slowly
The exact mechanism of achon by which morphine induces than naloxone, produced very lfttle blockade. Implantation of the
immunosuppression has yet to be dearly determined. A direct 75-mg morphine pellets produced a significant slevation in serum
mechanism of action is suggested to operate through lympho- corticosterone levels. In vitro exposure to corticosterone is
cyte opiate receptors, but the nature of such receptors is still in known to suppress NK activity directly, whereasin vitro morphine
question. The alternative, an indirect mechanism of action is was unable to alter directly NK activity. The glucocorticoid recap-
proposed to be mediated by two possible pathways, hypothal- tor antagonist RousseRiclaf 38486 blOCked morphine-Induced
amlc-pituitary-adrenal axis activation with increased production suppression of NK activity in a dose-responsive fashion. Naltrex-
of adrenal corticosteroids, or activation of the sympathetic nary- one (1 0-mg pallet) antagonized the morphine-Induced elevation
ous system and concomitant catecholamine release. Natural in serum corticosterone. These results suggest that the prinCiple
killer (NK) cell activity was used to determine potential indirect mechanism of action for morphine-Induced suppression of NK
mechanisms of action for morphine. NK activity in the B6C3F1 activity is through a centrally located, opiate receptor-mediated,
mouse was suppressed between 1 2 and 48 hr after implantation elevation in drculating corticosterone which directly suppresses
of 75 mg timed-ralease morphine pellets. Morphine suppressed NK activity.
NK activity in a dose-responsive manner. The opiate antagonists

Opiates have been shown to exert an immunosuppressive in regulating morphine’s effects on NK activity (Weber and
effect in laboratory animals, and data suggest they may have Pert, 1989). The strong central nervous system-mediated effect
similar effects in man. The ability of NK cells to lyse tumor of morphine suggests that suppression of NK activity by mor-
targets is one of the immune parameters that is suppressed by phine is primarily mediated by an indirect pathway of action,
morphine administration. Shavit et at. (1984) showed a dose- beginning with central opiate receptors. To determine the
responsive suppression of NK activity in female Fischer rats mechanism of action of morphine-induced immunosuppression,
after daily injection of 10, 30 and 50 mg/kg/day for a 4-day studies must be undertaken to establish the exact pathway by
period. Direct central administration of morphine to rats has which morphine works.
Morphine-induced immunosuppression by a direct mecha-
shown that an opiate receptor within the central nervous sys-
nism of action through opiate receptors upon lymphocytes is
tem is involved in morphine-induced immunosuppression
not consistently supported by experimental evidence (Sibinga
(Shavit et at., 1986). Direct central administration of morphine
and Goldstein, 1988). It has been suggested that glucocorticoids
to the periaqueductal gray matter of the hypothalamus of male
and/or agents of the sympathetic nervous system (i.e., cate-
Fischer rats showed this as the principal site for this receptor
cholamines) may serve as the mediators of morphine’s immu-
Received for publication December 2, 1993. nosuppressive actions (Bryant et aL, 1990; Shavit et aL, 1986).
1 Work supported by NIDA Grants DA00490 and R01DA07292 to B. F. Intracerebroventricular administration of morphine induces
2Present address: Minneapolis MediCal Research Foundation, 914 S. 8th St. increased hypothalamic corticotropin-releasing factor levels
D-3, Minneapolis, MN 55408. and in turn produces an increase in adrenocortical activity

ABBREVIATiONS: NK, natural killer; HPA, hypothalamic-pituitary-adrenal; FBS, fetal bovine serum; NBCS, newborn calf serum; RPMI, Roswell Park
Memorial Institute; HBSS, Hank’s balanced salt solution; DMSO, dirnethyl sutfoxide; PBS, phosphate-buffered saline; P1, propldlum iodide; RU486,
Roussel-Uclaf 38486.

1127
1 128 FrIar and Fuchs Vol. 270

(Suemaru et at., 1986). The ability of the periaqueductal gray (Palo Alto, CA) were used for timed release of naloxone hydrochloride
matter to regulate morphine-induced suppression of NK activ- and naloxone methiodide (1 or 5 mg/kg/day). Both drugs were prepared
in Ringer’s solution (vehicle). The pumps were implanted using the
ity suggests that activation of the HPA axis is important in
same surgical procedures described for the pellets. All mice for pump
mediating morphine-induced immunosuppression.
studies weighed at least 20 g. Before implantation, the micro-osmotic
Implantation of 75-mg morphine pellets induces glucocorti-
pumps were incubated for 4 hr in physiological saline at 37C to
coid-mediated thymic apoptosis that is responsible for the eliminate the lag time which would have occurred had the pumps been
dramatic thymic atrophy seen in morphine-dependent mice implanted directly. This allowed for immediate release of antagonist
(Fuchs and Pruett, 1993). This effect was completely blocked into the animal.
by the opiate antagonist naloxone and the glucocorticoid recep- Splenocyte culture. Spleens were aseptically removed from naive
tor antagonist, RU 486. The thymic atrophy was shown to be mice and placed into HBSS. Spleens were then mashed between the
primarily due to the induction of apoptosis (Freier and Fuchs, frosted ends of microscope slides to produce a single cell suspension.
1993) and thymocyte apoptosis is a glucocorticoid-sensitive Splenocytes were washed once in HBSS and twice in RPMI 1640
process (McConkey et at., 1989). A comparison study of adre- supplemented with 10% PBS, penicillin/streptomycin, HEPES and
sodium bicarbonate. Splenocytes were suspended in 5 ml of RPMI in
nalectomy and RU 486 after implantation of 75-mg morphine
25-cm2 culture flasks at a concentration of 1 x iO cells/mi and placed
pellets in C3H/HeN mice indicated that adrenal activation was
into an incubator (5% C02; 37#{176}C) for 3 hr. Morphine sulfate prepared
a critical component in the mechanism of action of morphine- in Ringer’s solution was added to cultures just before incubation.
induced immunosuppression (Bryant et at., 1991). To determine Corticosterone was prepared in DMSO, and necessary dilution was
clearly the mechanism of action of morphine in mediating the done to bring the final DMSO level in culture to less than 0.1%.
suppression of NK activity, the direct actions of adrenal corti- Vehicle-cultured cells received both Ringer’s and DMSO. After 3 hr
costeroids and the action of morphine on central opiate recep- cells were washed in RPMI 1640 1- 10% NBCS, counted and diluted to
tor’s must be linked. appropriate concentration for use in the chromium-release procedure.
Our purpose was to determine the pharmacologic
parameters Chromium-release assay. The chromium-release assay was done
of morphine-induced suppression of NK activity
in mice and as described (Puccetti et aL, 1980) with some modifications. Yac-1
thymic lymphoma cells (American Type Culture Collection, Rockville,
to evaluate the ability of glucocorticoid hormones to serve as
MD) are grown in continuous culture (RPM! 1640 + 10% PBS) and
the direct mediators ofthis morphine-induced suppression. The
kept in log growth phase for use as target cells. On the day of assay
opiate antagonists naloxone and naltrexone were used to eval-
cell number is determined by hemacytometer count and 1 x iO Yac-1
uate the importance of an opiate receptor in this phenomenon cells are placed into a conical tube and pelleted at 175 x g. Supernatant
and naloxone methiodide was used to show the central location is removed and the cells are resuspended in 500 Ci of [5Na]CrO4 in a
of the opiate receptor involved. Furthermore, the glucocorti- volume of 0.5 ml. Yac-1 cells are then incubated with the chromium
coid-specific antagonist, RU 486, was used to evaluate the role for 90 mm in a 37#{176}C water bath. Cells were washed five times in RPM!
of adrenal corticosteroids in morphine-induced immuno- 1640 + 10% NBCS and then counted and diluted to working concen-
suppression. Serum levels of corticosterone and the chromium trations. Experimental release was determined by measuring radioac-
release assay for NK activity were used to assess the effective- tivity from a 100-il sample of supernatant using the average of four
replications for each treatment at each effector/target ratio. Maximum
ness of these treatments.
release was determined using 100 Ml of Yac cells (1 x 10) plus 100 d
of 2 N HC1. Spontaneous release was 10% or less for all experiments.
Materials and Methods Only Yac cell preparations >90% viable were used for the determina-
Animals. Female B6C3F1 mice (virus free, Taconic Farms, Ger- tion of chromium release. To calculate the percentage of release the
mantown, NY), 6 to 12 weeks old, were housed four per cage on a 12- following formula was useth
hr light/dark schedule (light on at 6 A.M.). Mice were quarantined for
% Specific Release
1 week and tested to ensure their specific pathogen-free status before
use. Food and water were provided ad libitum. All animals were cared - Experimental Release - Spontaneous Release < #{174}
for as described in Guide for Care and Use of Laboratory Animals as Maximum Release - Spontaneous Release
adopted by the National Institutes of Health. All mice were sacrificed
between 7 and 10 A.M. to maintain experimental consistency. Data were analyzed for effector/target ratios of 400:1, 200:1, 100:1,
Drugs/chemicals. Timed-release morphine pellets (75, 25 and 8 50:1, 25:1 and 12.5:1. Each sample is run in quadruplicate and serial
mg), placebo pellets, naltrexone pellets (10 mg) and morphine sulfate dilution is done to provide the desired effector/target ratios. Once serial
were obtained from the National Institute on Drug Abuse (NIDA) in dilutions are completed, chromium-labeled Yac-1 cells are added. Ef-
Rockville, MD. The drug corticosterone and all media and supplements fector/target ratio curves are analyzed using the Allfit program (Dc-
were purchased from Sigma Chemical Co. (St. Louis, MO). RU 38486 Lean et aL, 1978). This program fits an equation to the sigmoideffector/
(RU 486) was a gift of Rousael-Uclaf(Romaineville, France). Naloxone target response curves. This equation is then used to determine lytic
hydrochloride and naloxone methiodide were purchased from Research units. For our purposes a lytic unit is defined as the number of cells
Biochemicals Inc. (Natick, MA). PBS was purchased from Hyclone necessary to induce 10% specific release and is expressed as lytic units,
(Logan, UT). NBCS was purchased from ICN BiochemiCals Inc. (Cam- io splenocytes.
bridge, MA). Culture media was RPMI 1640 + 10% FBS supplemented linmunofluorescent staining for flow cytometric analysis. An
with 2 nM gluta.mine, 100 U/ml of penicillin, 100 ig/ml of streptomy- aliquot of 1.5 x i0 cells was removed from each splenocyte sample
cm, 2.25 mg/ml of sodium bicarbonate and 25 mM HEPES buffer. before preparation for the NK assay. Each aliquot was placed into a
Implantation ofpellets and osmotic pumps. Pellet implantations well of a 96-well plate in a volume of 200 zl of RPM!. Plates were then
were done as previously described (Freier and Fuchs, 1993). Briefly, centrifuged and cells washed twice in 150 il of buffer (PBS + 1%
mice were anesthetized, an incision was made in the lower back of the bovine serum albumin and 0.1% sodium aside). Cells were pelleted by
mouse and the pellet was inserted. The woundwas closed with a surgical centri.fugation for 2 mm at 300 x g. Supernatants were aspirated via a
staple and swabbed with a 2% polyvinyl pyrrolidone iodine solution 25-gauge needle, the plate briefly vortexed to resuspend the cell pellet
used in place of Betadine, as previously described. and then 5 l of mouse IgG (Sigma Chemical Co.) diluted 1:5 from
Micro-osmotic pumps (model 1007D; lot #046101; mean pumping stock, was added to each well. This was done to eliminate nonspecific
rate, 0.49 jsl/hr, mean fill volume, 98 l) obtained from Aiza corporation staining of NK 1.1 on splenocytes. The plate was allowed to sit on ice
1994 Activity 1129

for 5 mm, washed once in 150 tl ofPBS buffer and 150 l of biotinylated . Placebo
NK 1.1 antibody added (diluted 1:150 from stock). This was incubated
on ice, in the dark, for 20 mm, washed three times in PBS buffer and
150 Ll of phycoerythrin-conjugated strepavidin (diluted 1:150 from a
(1)

stock) was added and allowed to incubate on ice for 20 mm. Cells are 0
0
then washed
mg/ml; diluted
three times in 150
1:20 in PBS) was
of PBS buffer and 150 tl of P1 (0.1
added and the plate incubated on ice
aa.
C

U)
for 4 mm. After a final PBS wash, the pellet was resuspended in 200 N.
l of PBS and cells analyzed on a FACSCAN flow cytometer (Becton/
Dickinson, San Jose, CA). A gate was set up to exclude nonviable cells
In
from the analysis based upon P1 stain and 5000 viable cells were
C
analyzed. P1 staining was also used to determine viability for in vitro-
0
cultured splenocytes. Staining was done as described above, but only
P1 was used to stain the cells. .1
RU 486 preparation and administration. Mice were given RU
486 at 5, 25 or 100 mg/kg in a suspension of 0.2% Tween 80 and 0.25%
methylcellulose in Ringer’s solution. The mixture was vortexed over-
Time (hours)
night to produce a consistent suspension of RU 486 for administration.
RU 486 was administered by oral gavage at 1 hr before pellet implan- Fig. 1. Time course of effect of morphine on splenic NK cell cytotoxic
tation at a volume of 0.2 ml. Vehicle-treated animals received a 0.2-mi activity. Mice were implanted with morphine (75-mg) or pcebo pellets
volume of the 0.2% Tween 80, 0.25% methylcellulose in Ringer’s and sacrificed at time zero. X axis represents time in hours after implan-
solution without RU 486. This preparation has been previously de- tation, whereas y axis represents lytic units. Lytic units are defined as
scribed by others (Philibert, 1984). the number of cells needed to produce 1 0% lysis of Yac-1 tumor targets
in a 4-hr chromium release assay. Data are presented as lytic units per
Corticosterone radioimmunoassay. All assays were conducted
100 million splenocytes. Each bar represents pooled results of multiple
using standardized kits purchased from Diagnostic Products Corpora-
animals, where n is the number of animals, where n = 1 1 (naive), n = 5
tion (Los Angeles, CA). Serum samples were taken by cardiac puncture (1 68, 48 hr and for placebo-treated at 96 hr) and n = 4 (24, 12 and 3 hr
at the time of sacrifice. Sacrifice was conducted by CO2 inhalation. and for morphine-treated at 96 hr). Experiment represents one of four
Blood samples were placed into glass test tubes and allowed to sit at separate repetitions. Error bars represent S.E.M.; P < .01 difference
room temperature for 1 hr. This allowed clot formation to begin. Clots from naive treatment.
were freed from the tube wall using wooden applicator sticks. Tubes
14
were then refrigerated for
clots to contract. Samples
at least 2 hr, but no more than 24 hr to allow
were then centrifuged at 500 x g and serum a
Cl)
12
0
removed from red blood cells and placed into microcentrifuge tubes. 0
Thbes were then stored frozen at -70’C and defrosted on day of assay. aa.
C

All samples were assayed within 1 month of initial freezing. Unknown (1)8 C.
N.
serum concentrations were compared to standard curves determined
from kit provided known samples. Final corticosterone levels are as-
sessed as nanograms per milliliters of corticosterone. In 4 C,

Statistical analysis. A parametric one-way analysis of variance as C

2
C.

performed to determine whether there were significant differences 0

between treatment groups. When such differences were found post-hoc 1

8 25 75
analysis was done using either Dunnett’s t test or Duncan’s New
Multiple Range test to determine significance. Dunnett’s t test was
Placebo Morphine
used in studies to determine whether morphine was inducing a signifi- Fig. 2. Dose-response of effect of morphine on NK cell cytotoxic activfty.
cant suppression of control (placebo) responses. Duncan’s Multiple Morphine (8-, 25- 75-mg) or pellets were implanted in mice
Range test was used to determine the efficacy ofantagonists in blocking 48 hr before sacrifice. Results are expressed as lytic units per 1 0 mIllion
the effects of morphine. Results were considered significant at P s .05. splenocytes. Lytic units are defined as the number of cells needed to
produce 10% lysis of Yac-1 tumor targets; n = 3 for each treatment
group and experiment represents one of three repetitions. < .01
**)

Results difference from placebo-treated group.

Mice were implanted with either 75-mg morphine timed- caused a time-dependent decrease in NK 1.1k splenocytes, but
release pellets or placebo pellets at the indicated time before the decrease in total splenic NK cells was only significant at
sacrifice. Spleens were aseptically removed and a single cell 48 hr after morphine pellet implantation (fig. 3). This suggests
suspension prepared. NK cytotoxic activity was measured by that a loss in splenic NK cell number is only partially respon-
the 51Cr-release method. The earliest significant effect of mor- sible for the decrease in splenic NK activity.
phine on NK activity was at 12 hr and a significant suppression Splenocytes were exposed to increasing concentrations of
was maintained through 48 hr (fig. 1). NK activity began to morphine in in vitro culture to determine whether morphine
recover by 96 hr after morphine, and by 7 days (168 hr) NK was directly causing the suppression of NK activity. At concen-
activity had returned to placebo control values. Morphine’s trations of 0.1, 1.0, 10 and 100 zM of morphine there was no
effect on splenic cellularity followed a similar pattern, but suppression of NK activity (fig. 4). A slight enhancement
recovery was only about 75% of placebo controls by 7 days appeared to be in evidence, but this was not significant. Cell
(data not shown). Implantation of 8-, 25- and 75-mg morphine viabilities, measured after the 3-hr culture period, were no
pellets dose-responsively suppressed NK activity at 48 hr (fig. different from untreated or vehicle-treated controls (data not
2). To determine whether the loss in NK activity was the result shown). This suggests that morphine is acting through an
of a reduction in total splenic NK cells, the NK 1.1 antibody indirect mechanism to induce the observed suppression of NK
was used to assess splenic NK cell populations. Morphine activity.
I 130 Freler and Fuchs Vol. 270

12
-0-- PLACEBO

-0-- MORPHINE
10
In
. NAIVE 0
a 0
C
0
a
8 a.
Cl)
Ca N.

6
-.5

U)

4 C

2 -a Placebo Vehicle 1 mg/kg 5 mg/kg 1 mg/kg 5 mg/kg

#{149}
+ Naloxone
Vehicle Naloxone Methlodide
. 12 #{149} 24 48 96 168 MORPHINE
Time (hours)
Fig. 5. Involvement of opiate receptors in morphine-induced suppression
Fig. 3. Alterations of spienic NK cell population by morphine. The effect of NK cell cytotoxicity. Naloxone and naloxone methiodide were adrnin-
of the 75-mg morphine pellet on spienic NK cells was evaluated using a istered for 16 hr using the Aizet mrniosmotic pump, implanted simulta-
biotinylated NK 1 .1 antibody and strepavidin-phycoerythnn, followed by neously with placebo or morphke (75-mg) pelletS. Each treatment group
FACS analysis. Raw percentage data were corrected for total splenic has n = 5. Figure represents one of two repetitions. Lytic units are
cellularity. Bars represent groups of pooled data (n = 5 for all groups defined as the number of cells needed to produce 10% lysis of Yac-1
except 48- and 24-hr morphine in which n = 4 and 12-hr placebo in targets. < .01 ;
#{149}P < .05 difference from placebo + vehicle.
which n = 3). Statistical significance determined by two-way analySiS of
variance. Error bars represent S.E.M. < .01 difference from naive 2000
controls.
8
C,

0
6
a
C

C
I 000
0 0
C a
a U)
a. 8
U)
N.
I!
T 0
C.)

II)
C

-a 0.
Vehicle Vehicle 5 25 100
+ KU 48b (mglKg)
Placebo
Morphine Pellet Implanted
0
Fig. 6. Effect of morphine and RU 486 on serum corticosterone levels.
0.1 1.0 10 0.1 1.,
0

a
a o
a Groups of four mice
pellets 16 hr before
were implanted with morphine (75-mg) or placebo
sacrifice. RU 486 was given orally I hr before pellet
implantation. Blood was drawn by cardiac puncture at the time of
: Corticosterone Morphine (pM)
(pM) sacrifice. Error bars represent S.E.M. Significance determined by one-
way analysis of variance. < .01 difference from placebo + vehicle
determined by Dunnett’s test.
Fig. 4. In vitro NK cell cytotoxic response after morphine or corticoster-
one prainoubation. Naive splenocytes were cultured for 3 hr at 1 x 10 tation of 10-mg naltrexone pellets produced a comparable
cells/mI in a volume of 5 ml with the indicated concentration of drug.
Drug was added in a volume of 50 MI. Corticosterone was prepared in antagonism (data not shown). The administration of naloxone
1% DMSO, where the final concentration of DMSO in culture was 0.1% methiodide produced only a partial blockade of morphine-
or less. The same concentration of DMSO was added to all morphine- induced suppression, which was not dose responsive (fig. 5).
treated cultures. Three separate flasks were used for each treatment. This supports an indirect mechanism of action through a
Experiment represents one of three repetitions. Error bars represent
centrally located opiate receptor.
S.E.M. < .01 difterence from vehicle treatment.
Morphine is a known activator of the HPA axis, and induces
The involvement of opiate receptors in morphine’s immu- increased adrenocortical output. Corticosterone has been im-
nosuppressive effects was assessed using the opiate antagonists plicated as the mediator for morphine’s immunosuppressive
naloxone and naloxone methiodlide. Naloxone methiodide, a effects and was able to dose-dependently suppress NK activity
quaternary congener of naloxone which penetrates the blood- in vitro (fig. 4). Measurement of serum corticosterone levels in
brain barrier at a much slower rate, was used to determine the mice implanted with the 75-mg morphine pellets showed a
possible location of the opiate receptor, either central or pe- significant increase in circulating corticosterone (fig. 6).
ripheral. Continuous infusion of naloxone completely antago- To determine the importance of the increase in serum corti-
nized the suppressive effects of morphine (fig. 5) and implan- costerone in mediating morphine’s suppressive effects, the glu-
1994 INKACtIVItY 1131
200a
cocorticoid receptor antagonist RU 486 was used. RU 486 had
no effect on the morphine-induced elevation in serum corticos-
terone (fig. 6). This is not an unexpected outcome because RU
486 prevents feedback inhibition and therefore would prevent
a decrease in corticosterone output after morphine-induced
elevation. RU 486 dose-responsively antagonized the in vivo 1500
morphine-induced suppression of NK activity (fig. 7). This
supports adrenocortical activation as the mechanism of mor-
phine-induced suppression of NK activity. Naltrexone was able
E
to block the induced elevation in serum corticosterone by C

morphine, further supporting corticosterone as the mediating a

C

agent and supporting

the adrenocortical
the involvement
activation (fig. 8).
of an opiate receptor in a
0

U)
1000

0
0
C.
0
Discussion C.)

C,
Morphine has strong immunosuppressive potential and the 500
consequences of its effect’s in clinical use and in opiate abuse
C,
need to be clearly understood. Suppression of NK cell cytotoxic
activity by morphine was chosen for evaluating a mechanism 1
of action for morphine-induced immunosuppression. It has
been shown that morphine
et at., 1983; Weber
can suppress
et aL, 1987; Ho et aL, 1986),
NK activity (Shavit
but the
0 -I Naive Placebo
+
Placebo
+
Naltrexone
+
Naltrexone
+
mechanism of action is still unclear. Time-dependent effects of Placebo MorphIne Placebo Morphine
morphine on splenic parameters have been demonstrated as
Fig. 8. Naltrexone antagonizes mOrphine-induced increases of serum
early as 24 hr (Bryant et aL, 1988) and maximum effects were
corticosterone. Groups of four mice were pIanted with placebo, mor-
seen at 3 to 5 days after pellet implantation with full recovery phke (75-mg) or naltrexone (10-mg) pellets 16 hr before saciiflce. Biood
by day 28 (Arora et aL, 1990). Implantation of 75-mg morphine samples were taken by cardiac puncture at the time of sacrifice. Exper-
pellets at 168, 96, 48, 24, 12 and 3 hr before sacrifice was done iment represents one of three repetitions. Error bars represent S.E.M.
Significance determined by one-way analysis of variance. < .05 and
to examine the kinetics of morphine’s effect on NK activity.
**P < .01 difference from placebo + morphine determined by Duncan’s
The strongest suppressive effect was seen at 12, 24 and 48 hr Multiple Range test.
after morphine pellet implantation, with recovery by 7 days
(168 hr) after morphine pellet implantation. This time course vided the necessary parameters for further investigations on
was shorter than those for mitogenic proliferative effects the mechanism of action of morphine.
(Bryant et at., 1988) and alterations in thymic subpopulations Morphine causes a rapid loss in splenic cellularity, and the
(Freier and Fuchs, 1993). Morphine dose-dependently sup- loss in splenic NK cells could be the reason for the observed
pressed NK activity in mice implanted with 8-, 25- or 75-mg decrease in NK activity. The NK 1.1 monoclonal antibody was
morphine pellets. The time- and dose-dependent studies pro- chosen to measure NK cell number in spleens of morphine-
and placebo-treated mice.This antibody recognizes a cell sur-
6 face protein found on the majority of splenic NK cells (Koo
and Hatzfeld, 1980). Aliquots of splenic NK cells were assessed
using the biotin NK 1.1 antibody and strepavidin-phyco-
erythrin. A significant decrease in NK 1.1 splenocytes was in
evidence at 48 hr after morphine pellet implantation (fig. 3). It
is important to note that morphine suppressed NK activity at
12 hr (fig. 1), but NK 1.1 analysis indicates that NK cell number
was not different from placebo controls at this time. This

:. #{149} jTIi 25
suggests the observed suppression
tially due to a loss in NK cells from the spleen.
activity is strongly suppressed
Because
at 12 hr, when NK 1.1k cell
of NK
NK
activity is only par-

numbers are still comparable to placebo control levels, the

Vehicle Vehicle RU486(mgkg)
suppression of NK activity is most likely to be due to an effect
Plabo Morphine Pellet Implanted of morphine on the NK cell function, as opposed to a direct
loss in splenic NK cell number.
Fig. 7. RU 486 dose-responsive antagonism of morphine-induced Splenocytes were incubated for 3 hr in vitro at concentrations
suppression of NK caN cytotoxicity. Groups of four mice were implanted
with morphine (75-mg) or placebo pellets 16 hr before saclifice. RU 486 of 0.1, 1.0, 10 and 100 jM of morphine. This was done to
was given orally at the indicated dose, 1 hr before pellet implantation. determine whether morphine had a direct immunosuppressive
Lytic units are defined as the number of cells necessary to produce 10% effect on NK activity. There was no effect of morphine at any
tysis. Experiment represents one of three repetitions. Error bars repre- concentration in comparison to untreated or vehicle controls
sent S.E.M. Significance determined by one-way anaiysis of vanance.
(fig. 4). Similar results have been observed during in vitro
**P < .01 and P < .05 difference from morphine and vehicle and from
morphine + RU 486 5 mg/kg; P < .05 difference from morphine + RU exposure of purified human peripheral blood lymphocytes to
486 25 mg/kg, determined by Duncan’s Multiple Range test. morphine. A 4-hr culture had no significant effect on NK
I 132 Frslar and Fuchs Vol. 270

activity except at the highest concentration of 1 x iO M (100 ing a no effect level. RU 486 did not alter elevated serum
M) (Yeager et aL, 1992). Cell viability was not assessed by corticosterone levels. These results support the conclusion that
Yeager and co-workers, but we observed no effect on splenocyte morphine stimulates increased adrenal corticosteroid activa-
viability in our in vitro culture system. Therefore it seems tion to produce the suppression of NK activity.
unlikely that morphine has the capacity to directly suppress The induced elevation in serum corticosterone was also an-
NK activity. tagonized by implantation of 10-mg naltrexone pellets (fig. 8).
The opiate receptor antagonist naloxone was used to deter- This is a key result because it links the ability of morphine to
mine whether an opiate receptor-mediated mechanism was operate at a central opiate receptor with morphine-induced
involved. Its congener, naloxone methiodide, which has a lim- elevation in serum corticosterone. Because naltrexone corn-
ited ability to cross the blood-brain barrier (Iorio and Frigeni, pletely antagonizes the morphine-induced elevation of serum
1984, Russell et at.,, 1982) was used to determine the location corticosterone and the suppression of NK activity, this strongly
of opiate receptor’s involved. Continuous administration of 1 supports an indirect mechanism of action through central opi-
or 5 mg/kg/day of naloxone completely antagonized morphine- ate receptors to activate the HPA axis and increase adrenocor-
induced suppression of NK activity. Implantation of 10-mg tical production of corticosterone.
naltrexone pellets also antagonized morphine-induced suppres- Adrenalectomy is another alternative for assessing the role
sion of NK activity (data not shown). Naloxone methiodide of adrenocortical activation in morphine’s immunosuppressive
was only able to antagonize partially the effects of morphine mechanism, but removal of the adrenal gland also removes a
(fig. 5) and the doses of 1 and 5 mg/kg/day produced an source of catecholamines, as well as other hormones, and thus
equivalent level of antagonism. This supports the interpreta- limits the ability to rule out these other hormones as being
tion of a central opiate receptor by which morphine acts to involved in morphine’s suppressive actions. The thymic invo-
induce a suppression of NK activity. Intraperitoneal injection lution produced by morphine was abrogated by adrenalectomy
of N-methyl morphine, a quaternary form of the agonist, failed (Sei et aL, 1991), but Bryant and colleagues (1991) state that
to produce a significant suppression of NK activity (Shavit et the sympathetic nervous system cannot be ruled out on the
zL, 1986). This suggests morphine is acting on a central opiate basis of adrenalectomy studies alone.
receptor to suppress NK activity. Direct injection of morphine Two recent reports have suggested that the mechanisms
into the periaqueductal gray matter induced suppression of NK involved in mediating morphine’s effects on immune function
activity in mice (Weber and Pert, 1989) and morphine stereos- may be even more complicated than hypothesized. Using a dose
pacifically suppressed antibody production to the antigen TNP- of 15 mg/kg of morphine in male Lewis rats and examining a
ovalbumin (Weber et aL, 1987). Recently the mitogenic re- number of immune parameters 1 hr after morphine, a dose-
sponse of peripheral blood lymphocytes was suppressed by dependent suppression of Con A induced proliferation of
acute morphine administration (Hernandez et at. 1993). The splenic and blood lymphocytes, production of IL-2 and IFN by
suppressive effect of morphine was centrally mediated, but was spleen and lymph nodes and splenic NK cell activity (Lysle et
not linked to gluocorticoid elevation. These studies support a aL, 1993). Suppression ofthese parameters were all antagonized
role for a central opiate receptor in mediating morphine-in- by naltrexone administration. In the companion paper the beta
duced immunosuppression. receptor antagonists nadolol, atenolol and ICI-118,551 were
Activation of the HPA axis and production of adrenocortical used to assess the role of the sympathetic nervous system in
steroids or activation of the sympathetic nervous system have mediating morphine’s suppressive effects. All three were able
been suggested as a possible indirect pathway for morphine’s to antagonize morphine-induced suppression of Con A- and
immunosuppressive effects (Bryant et aL, 1987; Shavit et at., PHA-induced splenic lymphocyte proliferation but were unable
1986). HPA axis activation increases central corticotropin- to block the suppression of NK activity (Fecho et aL, 1993).
releasing factor and production of adrenocorticotropic hormone Lysle and co-workers (1993) suggest that morphine has corn-
and produces an increase in serum corticosterone levels. Mor- partment specific effects on immune function, depending upon
phine can activate this pathway through central opiate recep- the origin of the lymphocytes. The beta antagonists used by
tors (Suemaru et aL, 1986) and adrenocortical activation has Fecho et aL (1993) were unable to antagonize morphine-induced
been linked to morphine’s immunosuppressive effects (Bryant suppression of NK activity. Considering the results ofour work
et aL, 1991). Other studies have also shown a strong role for and that of Fecho and co-workers (1993), the suggested corn-
glucocorticoids in mediating morphine-induced effects on partment specificity may operate in such a way that different
suppression of antibody production and thymic apoptosis, re- immune functions are effected by different regulatory path-
spectively (Pruett et aL, 1992; Fuchs and Pruett, 1993). ways, e.g., NK cell activity seems to be controlled by the HPA
To assess the importance of adrenal corticosteroids in mor- axis and corticosteroid production, whereas Con A lymphocyte
phine-induced suppression of NK activity several methods were proliferation is under sympathetic nervous system regulation.
used. The 75-mg morphine pellets produced a significant in- These studies have important considerations for future work
crease in circulating corticosterone 16 hr after pellet implan- evaluating the effects of morphine upon immune parameters.
tation (fig. 6) and corticosterone directly suppressed NK activ- Morphine-induced suppression of NK activity is maximal
ity in a dose-responsive fashion (fig. 4). A number of studies in within 12 to 48 hr after exposure and is a dose-dependent effect.
mice and in humans (using peripheral blood lymphocytes), This suppression is mediated by a centrally located opiate
show reduction in NK activity directly induced by glucocorti- receptor, which induces activation of the HPA axis and in-
coids (Matera et aL, 1988; Gatti et at., 1987; Pedersen and creases the circulating levels of corticosterone in serum. These
Beyer, 1986; Cox et aL, 1982). Oral administration of 5, 25 and high levels of serum corticosterone parallel the suppression of
100 mg/kg of RU 486 dose-dependently antagonized morphine- NK activity by morphine, and the effects are dose dependently
induced suppression of NK activity (fig. 7). RU 486 (5 mg/kg) blocked by the steroid antagonist RU 486. From all of these
did not significantly differ from morphine and vehicle, indicat- studies, a picture of the mechanism of morphine’s immunosup-
1994 INkActivily 1133

pressive pathway can be formed. In suppressing NK activity, Lvsiz, D. T., CoussoNs, M. E., WATTS, V. J., Bgrnm’r, E. H., Dyxsm, L
A.: Morphine-induced alterations of immune status: Dose dependency, com-
morphine is acting through an indirect path, by activating the partment specificity and antagonism by naltrexone. J. PhSrmSCOL Exp. Ther.
HPA axis through a central opiate receptor and inducing high 365: 1071-1078, 1993.
MATERA, L., CaDoso, E., Vscu, F., CESANO, A., BELuoNE, G., VuoLo, A.
corticosterone levels that produce the observed suppression. AND MOUNA’I’I’I, G.: Effect of cortisol on the native and in vitro induced non-
MHC restricted CytOtoxiCty of large granular lymphocytes. J. Clin. Lab. Im-
Aeknowledgments
niunoL 27: 77-81, 1988.
The authors would like to thank Dr. Louis S. Harris for the gift of the timed- MCCONKEY, D. J., H&rrzgLL, P., JONDAL, M. AND Oiutziiius, S.: Inhibition of
release pellets ofmorphine sulfate and both Dr. Harris and Dr. Albert E. Munson DNA fragmentation in thymocytes and IsOlated thymocyte nuclei by agents
for advice and discussion during the course of these studies. that Stimulate protein kinase C. J. BioL Chem. 264: 13399-13402, 1989.
PEDERSEN, B. K. AND BEYEn, J. M.: Characterization of the in vitro effects of
References glucocorticosteroids on NK cell activity. Allergr 41: 220-224, 1986.
ARORA, P. K., FRIDE, E., Pwrrrro, J., WAGGlE, K., AND SKOLNICK, P.: Morphine- PmLmERT, D.: RU 38486: An original multifaceted antihormone in two. In
induced immune alterations in uiuo. CelL InununoL 126: 343-353, 1990. Adrenal Steroid Antagonism, ad. by M. K. Agarwal, pp. 77-101, WaIter de
BRYANT, H. U., BERteroN, E. W. AND HOLADAY, J. W.: Immunosuppressive Gruyter & Co. Berlin, 1984.
effects ofchronic morphine treatment in mice. Life Sci. 41: 1731-1738, 1987. Pnua’rr, S. B., HAN, Y.-C. D FucHs, B. A.: Morphine suppresses primary
BRYANT, H. U., BweroN, E. W. AND HOLADAY, J. W.: Morphine pellet-induced humoral immune responses by a predominantly indirect mechanism. J. Phar-
immunomodulation in mice Temporal relationships. J. PhSrIBSCOL Exp. Ther. macel Exp. Ther. 262: 923-928, 1992.
245: 913-920, 1988. PUccETTI, P., SANTON1, A., RiccAiwi, C. AND HERBERMAN, R. B.: Cytotoxic
BRYN’r, H. U., BnnToN, E. W. AND HOLADAY, J. W.: Immunomodulatory effector cells with the characteristics ofnatural killer cells in the lungs of mice.
effects of chronic morphine treatment: Pharmacologic and mechanistic studies. hit. J. Cancer 25: 153-158, 1980.
NatI. Inst. Drug Abuse Bee. Monogr. Ser. 96: 131-149, 1990. RussELL, J., Bass, P., GOLDBERG, L I., SCHUSTER, C. R. AN M. HERa:
BRY*NT, H. U., BanwroN, E. W., KENNER, J. R. AND HOLADAY, J. W.: Role of Antagonismofgut, but notcentraleffectsofmorphinewith quaternary narcotic
adrenal cortical activation in the immunosuppressive effects of chronic mor- antagonists. Eur. J. PhSrmSCOL 78: 2-261, 1982.
phine treatment. Endocrinology 128: 3253-3258, 1991. SE:, Y., YosHIMoro, K., McINTyRE, T., SKOLNICK, P. AND ARORA, P. K.:
Cox, W. I., HoLsaooK, N. J., Gn*.sso, R. J., SPEcTsR, S. AND FRIEDMAN, H.: Morphine-inducedtbymic hypoplasia is glucocorticoid-dependent J. ImmunoL
Suppression ofthe natural killer cell activity of murine spleen cell cultures by 146: 194-198,1991.
dexamethaaone (41489). Proc. Soc. Ezp. BioL Med. 171: 146-150, 1982. SHAVIT, Y., LEwis, J. W., TERwi, G. W., GLn, it P. AND LIEBESKIND, J. C.:
DnLwi, A. P., MuNsoN, P. J. AND RODBARD, D.: Simultaneous analysis of Endogenous opioids may mediate the effects of stress on tumor growth and
families of sigmoidal curves: Application to bioassay, radioligand assay, and immune function. Proc. West. PharmaCOL Soc 26: 53-56, 1983.
physiological does-response curves. Am J. Physiol 235: E9’7-E102, 1978. SHAV1T, Y., DspAuus, A., MARTIN, F. C., Tgnr, G. W., PECHNICK, It N.,
FECHO, K., D’ncsm, L A. AND LY8LE, D. T.: Evidence for beta adrenergic ZANE, C. J., Gii, R. P. AND LIEBE8KIND, J. C.: Involvement of brain opiate
receptor involvement in the immunomodulatory effects of morphine. J. Phar- receptors in the immune-suppressive effect ofmorphine. Proc. NatL Aced. Sci.
macoL Exp. Ther. 265:1079-1087,1993. U.S.A. 83: 7114-7117, 1986.
Faniza, D. 0. AND FUCHS, B. A.: Morphine-induced alterations in thymocyte SHAVIT, Y., LEwis, J. W., ThRMAN, G. W., GALa, It P. AND LIEBESKIND, J. C.:
subpopuiations ofB6C3F1 mice. J. PharmaCOL Exp. Ther. 265:81-88,1993. Opioid peptides mediate the suppressive effects of stress on natural killer cell
FUCHS, B. A. AND PRUETT, S. B.: Morphine induces apoptosis in murine thy- cytotoxicity. Science (Washington, DC) 223: 188-190, 1984.
mocytes in viuo but not in vitro: Involvement of both opiate receptors and SIBINGA, N. E. S. AND GOLDsTEIN, A.: Opioid peptides and opioid receptors in
glucocorticoid receptors. J. Pharniacol. Exp. Ther. 266: 417-423, 1993. cells of the immune system. Annu Rev. ImmunoL 6: 219-249, 1988
GA’rri, G., CAVALLO, R., S.RTom, M. L, DEL P0N’rE, D., MASERA, R., SALVA- SuEM*.Ru, S., HA5HIMOTO, K. AND 0Th, Z.: Effect of morphine on hypothalamic
DORI, A., CARIGNOLA, K AND ANGEL!, A.: Inhibition by cortisol of human corticotropin-releasing factor (CRF) and pituitary-adrenocortical activity. En-
natural killer (NK) cell activity. J. Steroid Biochem. 26: 49-58, 1987. docrinoL Jpn. 33: 441-448, 1986.
HERNANDEZ, M. C., Fioans, L. R. D BAYER, B. M.: Immunosuppression by Wasan, R J., IKajmi, B., RIcE, K. C., PERT, A. AND HAGAN, A. A.: Opiate
morphine is mediated by central pathways. J. PharmaCOL Exp. Ther. 267: receptor mediated regulation ofthe immune response in vivo. NatL Inst Drug
1336-1341, 1993. Abuse Res. Monogr. Ser. 76:341-348,1987.
Ho, W. K., CHEUNG, K. W., LEUNG, K. N. AND WaN, H. L: Suppression of Wsan, It J. AND PERT, A.: The periaqueductal gray matter mediates opiate-
immunological functions in morphine addicted mice. NatL Inst Drug Abuse induced immunosuppresaion. Science (Washington, DC) 245: 188-190, 1989.
Rca. Monogr. Ser. 75: 599-602, 1986. YEAGER, M. P., Yu, C. T., CAMPBELL, A. S., MoscHEu.A, M. AND Guyna, P.
Ioitio, M. A. Ni FRIGENI, V.: Narcotic agonist/antagoulat pmperties of quater- M.: Effect of Morphine and fl-endorphin on human Fe receptor-dependent and
nary diastereomers derived from oxymorphone and naloxone. Eur. J. Med. natural killer cell functions. Clin. ImmunoL ImmunopathoL 62: 336-343, 1992.
Chem. 19: 301-303, 1984.
Koo, G. C. AND HATZFELD, A.: Antigenic phenotype of mouse natural killer Send reprint requests to: David 0. Freier, Ph.D., Minneapolis Medical Re-
cells. In Natural and Cell-Mediated Immunity Against Thmors, ad. by R B. search Foundation, 914 S. 8th Street, D-3, Minneapolis, MN 55404.
Herberman, pp. 106-116, Academic Press, New York, 1990.