Content uploaded by Pu Feng
Author content
All content in this area was uploaded by Pu Feng
Content may be subject to copyright.
INVITED REVIEW
Effects of Opioid Tolerance and Withdrawal
on the Immune System
Toby K. Eisenstein & Rahil T. Rahim & Pu Feng &
Nita K. Thingalaya & Joseph J. Meissler
Published online: 23 May 2006
#
Springer Science+Business Media, Inc. 2006
Abstract Review of the robust literature using acute
drug injection paradigms points clearly to the conclu-
sion that morphine is immunosupressive. In contrast,
studies of the effect of subacute or chronic adminis-
tration of morphine on immune function is limited,
with variable results. In some cases tolerance to the
immunosupressive effects of the drug is clearly dem-
onstrated, but in other cases, selected immune param-
eters do not demonstrate tolerance. Discrepancies in
findings may result from differences in species or route
and manner of drug administration. Even fewer studies
(total of 10) have been published on the effects of
withdrawal on immune function. Most immune param-
eters tested are supressed following drug withdrawal.
Recovery time to baseline response levels varies in the
studies. In the single report of withdrawal in humans,
immune function was suppressed for up to 3 years. It is
clearly established that withdrawal suppresses capacity
of murine spleen cells to make an ex vivo antibody
response, which contrasts with evidence of polarization
of the lymphocytes towards a Th2 phenotype. Several
laboratories have shown that subacute and chronic
exposure to morphine, as well as drug withdrawal,
sensitize to the lethal effects of bacterial lipopolysac-
charide. Underlying sepsis, combined with morphine-
induced hypofunction of the hypothalamic-pituitary-
adrenal (HPA) axis, may be occult variables modulat-
ing immune responses during opioid administration
and withdrawal. As episodes of withdrawal are com-
mon among drug abusers, more intensive investigation
is warranted on the effects of withdrawal on immune
function, on mechanisms of immune modulation, and
on sensitization to infection.
Key words opioids
.
tolerance
.
withdrawal
.
immune responses
.
review
Abbreviations
CD cluster of differentiation
CD3 antigenic marker found on all T cells
Con A concanavailin A, a protein from the jack
bean (Canavalia ensiformis) that stimulates
T cells to divide
DTH delayed-type hypersensitivity
FIV feline immunodeficiency virus
fMLP a tripeptide composed of formyl-methionine,
leucine, and phenylalanine, which is chemo-
tactic and promotes inflammation
HPA hypothalamic–pituitary–adrenal axis
IFN-a interferon alpha
IFN-g interferon gamma
IgM immunoglobulin of the IgM type
IL interleukin
IL-1 interleukin 1
IL-2 interleukin 2
IL-12 interleukin 12
i.c.v. intracerebral vascular
i.p. intraperitoneal
J Neuroimmune Pharmacol (2006) 1: 237–249
DOI 10.1007/s11481-006-9019-1
T. K. Eisenstein (*)
:
P. Feng
:
N. K. Thingalaya
:
J. J. Meissler
Center for Substance Abuse Research
and Department of Microbiology and Immunology,
Temple University School of Medicine,
3400 North Broad Street,
Philadelphia, PA 19140, USA
e-mail: tke@temple.edu
R. T. Rahim
Center for Substance Abuse Research
and Fels Institute of Cancer Research and Molecular Biology,
Temple University School of Medicine,
3307 North Broad Street,
Philadelphia, PA 19140, USA
LPS lipopolysaccharide, a component of the out-
er membrane of gram-negative bacteria that
has endotoxic activity and can lead to a sep-
sis syndrome
Mitogen a substance that causes lymphocytes to di-
vide without ligating their antigen receptors
NK natural killer (cells or activity)
s.c. subcutaneous
Th
1
T helper type 1 cells
Th
2
T helper type 2 cells
TNF-a tumor necrosis factor alpha, a cytokine that
induces widespread inflammatory changes
and symptoms of sepsis
Overview: opioids and immune function
The pioneering work of Joseph Wybran laid the firm
experimental basis for the concept that opioids, both
exogenous and endogenous, can modulate immune
function, and that their effects were occurring via en-
gagement of opioid receptors on cells of the immune
system (Wybran et al. 1979). He showed that in vitro
addition of morphine or met-enkephalin up- or down-
regulated the binding of human T cells in vitro to sheep
red blood cells, and that opioid receptor antagonists
could block the effects. We now know that binding of
sheep red blood cells to T cells is mediated by CD2, a
T-cell surface antigen, and in retrospect we can hypo-
thesize that when opioid receptors are engaged, they
affect the level of this CD2 antigen. The important
conclusion from Wybran’s work is that it proved that
opioids can have direct effects on cells of the immune
system. Work over the next 25 years has provided
abundant evidence that cells of the immune system
express opioid receptors, mRNA for opioid receptors,
and that opioids can modulate a whole spectrum of
different immune functions (Eisenstein et al. 1996;
Eisenstein and Hilburger 1998;Carretal.1996;
McCarthy et al. 2001; Alonzo and Bayer 2002;Go´ mez-
Flores and Weber 1999; Roy and Loh 1996; Risdahl et
al. 1996; Mellon and Bayer 1998; Hall et al. 1998; Sharp
et al. 1998; Peterson et al. 1998). The literature support-
ing the effects of opioids on immune responses is quite
robust (at least 150 papers), coming from various
laboratories and using a spectrum of assays of immune
status including responses to mitogens, capacity to make
antibodies, natural killer (NK) cell activity, cytokine
levels, expression of surface antigens on cells of the
immune system, and susceptibility to infection. Not all of
these effects have been shown to occur when the opioid
is applied to cells of the immune system in vitro.Ifan
immunomodulatory effect of opioids has been demon-
strated in vivo, then the mechanism(s) mediating the
effect may or may not be a direct one via opioid recep-
tors on cells of the immune system. Other pathways
controlled by opioid administration may alter immune
function, such as opioid receptor-mediated release of
glucocorticoids by activation of the hypothalamic–pitui-
tary–adrenal (HPA) axis, or opioid activation of the
sympathetic nervous system. Dissection of these path-
ways is an important research direction in the field of
neuroimmunopharmacology.
Effects of subacute and chronic opioid administration
on immune function
Subacute effects
It is noteworthy that the overwhelming majority of
studies investigating the in vivo effects of opioids on
the immune system have used paradigms of acute
opioid administration, frequently a single injection of
drug followed some hours later (usually 2–24 h) by an
assay of immune responsiveness. Many further studies
have addressed longer lasting effects of opioid expo-
sure on immune function. The reason kinetic assess-
ments of effects of opioids on immune parameters are
important is because many pharmacologic responses to
opioids wane with continued drug administration, a
phenomenon called Btolerance.^ Thus, the addict may
require increasingly higher doses of heroin to maintain
the same level of Bhigh,^
and increasing doses of
morphine may be needed to block pain. The term
Bsubacute^ morphine administration can be applied to
the paradigm in which there is exposure to the drug
over the course of several days, but the dose or the
time of drug administration is sometimes not sufficient
for marked tolerance or dependence to have occurred.
In rodents, a limited number of studies have examined
the effect of subacute opioid administration on im-
mune function. This type of drug delivery has fre-
quently been accomplished via subcutaneous (s.c.)
implantation of a slow-release morphine pellet or of
an osmotic minipump. The pellet or pump supplies dr-
ug continuously and prevents episodes of withdrawal.
Effects on the immune system can be measured at var-
ious times after implantation of the continuous delivery
device. Placebo pellets, which contain the other ingre-
dients, but no drug, or saline pumps are used as controls.
By using a single 75-mg slow-release pellet in mice,
maximal immunosuppression was found to occur 48 h
after pellet implantation (Bussiere et al. 1993; Bryant et
al. 1987; Kelschenbach et al. 2005). The parameters
measured included responses to mitogens (Bryant et
238 J Neuroimmune Pharmacol (2006) 1: 237–249
al. 1987), the ability of ex vivo spleen cells taken from
the morphine-treated mice to mount an antibody res-
ponse to an antigen presented in vitro (Bussiere et al.
1993), or levels of the cytokines IFN-g and interleukin
(IL)-4 (Kelschenbach et al. 2005). Note that 75-mg
pellets release the drug slowly, and blood levels in m-
ice are estimated to be 2 mg/mL at 6–24 h and 0.6 mg/
mL at equilibrium 48–96 h (Bryant et al. 1987). Pumps
give strong, dose-dependent immunosuppression of in
vitro antibody formation by spleen cells harvested 48
h after pump implantation (Rahim et al. 2001), but a
kinetic analysis was not examined. The dose response
curve was U-shaped, with maximal immunosuppres-
sion found at 1 mg kg
j1
day
j1
(Rahim et al. 2001).
Chronic opioid administration and tolerance
A limited number of studies have addressed the ques-
tion of whether tolerance to the effects of opioids can
develop in the immune system (Table 1). In at least
Table 1 Effect of chronic morphine on immune responses and infection
Tolerance
Assay Yes No Reference Species Time of exposure
NK Shavit et al. 1986 Rats Daily injections for 14 days,
50 mg/kg, s.c.
Bhargava et al. 1994 Mice Two injections day 1, followed by
75 mg morphine pellet for 4 days
Carr and France 1993 Monkeys Daily injections for 2 years
West et al. 1997 Rats 0.6 mg/mL in drinking water for
20 days. Injection (s.c.) 1 h before
sacrifice
Mitogen responses
To Con A
To LPS Bryant et al. 1987 Mice 96 h post 75 mg slow-release pellet
West et al. 1997 Rats 0.6 mg/mL in drinking water for
20 days
To Con A Bayer et al. 1994 Rats Escalating s.c. doses 10–40 mg/kg,
twice daily for 4 days
To Con A Chuang et al. 1993 Monkeys Morphine, escalating up to
5 mg/kg, s.c., thrice daily for 3 months
Morphine-dependent switched to 2 mg/kg
LAMM twice daily for 3 more months
B cell Proliferation to anti-
IgM/IL-4
Bhargava et al. 1994 Mice Two injections day 1, 75 mg morphine
pellet for 4 days
Antibody formation (PFC
responses)
Bussiere et al. 1993 Mice 120 h post 75 mg morphine pellet
Rahim et al. 2002 Mice 96 h post 75 mg morphine pellet
Delayed-type hypersensitivity
responses to picryl
chloride
Bryant and Roudebush 1990 Mice 72 h post 75 mg slow-release pellet
Delayed-type hypersensitivity
responses to mycobacteria
Pellis et al. 1986 Rats Escalating i.p. injection 10–50 mg/kg,
twice daily for 20 days
Delayed-type hypersensitivity
to 2.4-dinitrofluorobenzene
Molitor et al. 1992 Pigs Subcutaneous 150 mg/kg depot morphine
in oil, 4-day interval up to 24 days
Cytotoxic T cell activity Bhargava et al. 1994 Mice Two injections day 1, 75 mg morphine
pellet for 4 days
Cytokine production
Anti-CD3 induced IL-2
or IL-4 production
Bhargava et al. 1994
Mice Two injections day 1, 75 mg morphine
pellet for 4 days
Proinflammatory cytokines Pacifici et al. 1993 Mice Escalating 20–80 mg/kg morphine
s.c. injection for 8 days
Macrophage antitumor
activity
Pacifici et al. 1993 Mice Escalating 20–80 mg/kg morphine
s.c. injection for 8 days
Simian immunodeficiency
virus (SIV) load
Donahoe et al. 1993 Macaque Two daily injections of 3 mg/kg
morphine for 2 years
J Neuroimmune Pharmacol (2006) 1: 237–249 239
two drug administration paradigms, tolerance to the
immunosuppressive effects of morphine has been
reported to occur. Figure 1A shows the kinetic curve
of immunosuppression in the antibody response of
mouse spleen cells to sheep red blood cells when the
cells were harvested and tested ex vivo at various times
after implantation of a 75-mg slow-release morphine
pellet. As described above, at 48 h there was maximal
immunosuppression, but as demonstrated in this fig-
ure, by 96 h, spleen cell plaque-forming cell responses
had returned to normal baseline values. At this time,
pellets are still releasing morphine (Bryant et al.
1987) and the animals have become Bdependent^ on
the drug—when pellets are removed and animals are
given an opioid antagonist, naloxone, they exhibit
typical signs of withdrawal, also called an Babstinence
syndrome^ (Rahim et al. 2002). Rodents exhibit
various symptoms of abstinence—jumping, writhing,
abdominal stretching, wet-dog shakes, diarrhea, teeth
chattering, and hyperactivity—which can be quantitat-
ed to provide evidence that they are drug-dependent.
Bryant et al. (1987) noted that, 96 h after pellet
implantation, tolerance developed to suppression of
responses in mouse spleen cells to the mitogens,
concanavalin A (Con A) and bacteria lipopolysaccha-
ride (LPS). Bayer et al. (1994) reported that adminis-
tration of morphine to rats twice daily for 4 days led to
tolerance to the suppressive effects of the drug on
proliferation of peripheral blood lymphocytes to Con
A. Shavit et al. (1986) gave rats injections of morphine
for 14 days and observed tolerance to suppression of
NK cell activity. In contrast, Bryant and Roudebush
(1990) showed that at 72 h after implantation of a
morphine slow-release pellet in mice, delayed-type
hypersensitivity (DTH) responses to injection of picryl
chloride were still depressed. Pellis et al. (1986) also
found that DTH responses to mycobacteria were
depressed after 7 days of intraperitoneal (i.p.) mor-
phine injections. Molitor et al. (1992) injected pigs
(s.c.) with a depot of morphine in oil and observed that
DTH responses to another chemical, 2,4-dinitrofluoro-
benzene, were depressed. In addition, Chuang et al.
(1993) found that monkeys given daily injections of
morphine for 3 months still had depressed proliferative
responses to Con A. Bhargava et al. (1994) induced
tolerance in mice by a combination of morphine
injections on day 1 followed by morphine pellet (75
mg) implantation for 4 days. Interestingly, some
parameters of immune function demonstrated toler-
ance whereas others did not. Thus, B cell proliferation
to anti-IgM/IL-4 stimulation was normal in tolerant
animals, as was cytotoxic T cell activity, but the ability
of spleen cells to produce IL-2 or IL-4 when stimulated
with anti-CD3 remained suppressed, as did NK cell
activity. The latter result agrees with the observation
of Carr and France (1993) that both NK activity and
CD4
+
/CD8
+
T cell ratios were suppressed in monkeys
given daily morphine injections for 2 years, leading to
the conclusion that tolerance did not develop to these
immune parameters. Pacifici et al. (1993) injected mice
with morphine for 8 days and found that peritoneal
macrophage antitumor activity as well as elaboration
of proinflammatory cytokines were suppressed, indi-
cating lack of tolerance to these parameters as well.
West et al. (1997) induced tolerance by feeding
morphine to rats in the drinking water for 20 days.
These investigators gave a morphine injection at the
end of the 20-day period. If rats are completely tolerant,
they should not respond to the morphine injection with
altered immune parameters in comparison to controls
that were drinking water. Other appropriate control
groups were included. These investigators also found
that morphine tolerance to some immune parameters
developed, but not to others. Thus, there was tolerance
to suppression of NK activity (i.e., NK activity was
normal in tolerant mice receiving an injection of
morphine on the 20th day), but no tolerance developed
to responses to B and T cell mitogens or to suppression
of production of IFN-g by the mitogen-stimulated cells.
Thus, when different parameters are examined in
different animal species, results of tolerance experi-
ments vary. The reasons for the diverse results are not
known. Perhaps, in some of the experiments where
tolerance did not develop, the drug administration
Fig. 1. Effects of morphine on immune responses in mice.
Plaque-forming cell (PFC) responses of mice implanted with a
morphine pellet for 96 h (A). (B) PFC responses of mice whose
morphine pellet was removed after 96 h (abrupt withdrawal) or
who had a pellet removed and were given naloxone (precipitated
withdrawal). (Reprinted from Journal of Neuroimmunology,
Vol. 147, Rahim et al., Paradoxes of immunosuppression in
mouse models of withdrawal, pp 114–129, copyright 2003, with
permission from Elsevier.).
240 J Neuroimmune Pharmacol (2006) 1: 237–249
regimen was not proven to be one that resulted in
tolerance as tested by using pharmacologic endpoints.
However, this is clearly not the case in every instance, as
Bryant et al. (1987) found tolerance to response to
mitogens, but not to development of DTH (Bryant and
Roudebush 1990). Using the same mouse strain and
similar slow-release pellets, Rahim et al. (2002)
observed tolerance to capacity to induce an in vitro
antibody response (Fig. 1A). Alternatively, divergent
results on different immune parameters could be due
to differential tolerance.
Pertinent to studies on tolerance are an interesting
set of experiments carried out by House et al. (2001)in
which rats with intracerebroventricular (i.c.v.) cannu-
lae were implanted with multiple 75-mg morphine
pellets for 5 days. On day 5, animals received IL-1b,
either i.c.v. or i.p. It was observed that chronic
morphine blunted the HPA axis response to IL-1b
injection. Evidence for this conclusion was based on
differences in the following parameters in rats treated
with chronic morphine and injected with IL-1b com-
pared to controls: (1) reduced induction of mRNA for
corticotrophin releasing factor in the hypothalamus
after i.c.v. IL-1b; and (2) reduced plasma corticoste-
rone levels after i.c.v. IL-1b. Mice on chronic morphine
also had increased leukocyte adhesion in mesenteric
venules in response to the chemotactic factor, fMLP,
or IL-1b, alone, or in combination. The conclusion
reached from these experiments was that chronic
morphine created a heightened inflammatory state in
the periphery due to failure to blunt responses of
proinflammatory cytokines or chemotactic stimuli by
the HPA axis. These results were supported by a
second paper from this group (Ocasio et al. 2004)
showing that chronic morphine treatment led to in-
creased levels of proinflammatory cytokines in serum
of rats given bacterial LPS.
Effect of withdrawal (abstinence) from opioids
on immune responses
Chronic drug administration can lead to drug depen-
dence. One measure of dependence is demonstration
of an abstinence syndrome when the drug is discon-
tinued. Drug cessation can be by Babrupt withdrawal^
(AW), in which drug administration is simply stopped,
or by Bprecipitated withdrawal^ (PW), in which an
antagonist to the opioid is administered, with or with-
out drug cessation. There are surprisingly few articles
in the literature on the effects of withdrawal on immune
function (Table 2). The paucity of experiments probing
Table 2 Effect of withdrawal on various immune responses and infections
Withdrawal: effect on
immune response
Assay Yes No Reference Species
Length of morphine administration
prior to withdrawal
NK Shavit et al. 1986 Rats Daily injection for 14 days, 50 mg/kg, s.c.
Bhargava et al. 1994 Mice Two injections day 1, followed by 75
mg morphine pellet for 4 days
Carr and France 1993 Monkeys Daily injections for 2 years
West et al. 1997 Rats 0.6 mg/mL in drinking water for 20 days.
Injection (s.c.) 1 h before sacrifice
Mitogen responses
To Con A
To LPS Bryant et al. 1987 C3HeB/FeJ
mice
96 h post 75 mg slow-release pellet
West et al. 1997 Rats 0.6 mg/mL in drinking water for 20 days
To Con A Bayer et al. 1994 Rats Escalating s.c. doses, 10–40 mg/kg,
twice daily for 4 days
To Con A Rahim et al. 2005 Mice 96 h post 75 mg slow-release pellet
B cell proliferation to
anti-IgM/IL-4
Bhargava et al. 1994 Mice Two injections day 1, 75 mg
morphine pellet for 4 days
Antibody formation
(PFC responses)
Rahim et al. 2002 Mice 96 h post 75 mg morphine pellet
Cytokine production
IFN-g
Kelschenbach et al. 2005 Mice 72 h post 75 mg morphine pellet
SIV load Donahoe et al. 1993 Macaque Twice daily injections of 3 mg/kg
morphine for 2 years
J Neuroimmune Pharmacol (2006) 1: 237–249 241
the effects of withdrawal on immune responses is an
important deficit in our understanding of the effects of
opioids on immunity, because addicts usually experi-
ence frequent episodes of withdrawal as they come
down from their Bhighs^ and are between Bhits.^ No one
has examined the potential effects of withdrawal on
HIV burden or on latent infections, for example, on
mycobacterial infection in addicts with tuberculosis.
These questions are of importance, as in the United
States 27% of HIV-infected people are intravenous
drug users (Centers for Disease Control and Prevention
2005).
There is only a single report on the effects of with-
drawal from heroin in human addicts, carried out by
Govitrapong et al. (1998) in Thailand. This group re-
ported that AW, unassisted by drugs to ameliorate the
symptoms of abstinence, resulted in depressed re-
sponses to phytohemagglutinin, a T cell mitogen, and
depression in NK cell numbers in the peripheral blood
mononuclear cell population. T cell numbers increased,
but the CD4/CD8 ratio decreased. Some of these pa-
rameters were not normalized until 3 years after the
initiation of abstinence. Donahoe et al. (1993) carried
out studies in macaque (Macaca mulatta) infected with
simian immunodeficiency virus (SIV) and maintained
on morphine. This group noted that as soon as animals
became tolerant to morphine, the progression of SIV
slowed. However, when animals were put into with-
drawal, their viral load increased (Donahoe et al. 1993).
Barr et al. (2003) examined the effect of chronic expo-
sure to morphine and withdrawal on feline immunode-
ficiency virus (FIV) infection in cats. Cats were given
morphine for 9 days and then infected with FIV. Mor-
phine treatment and withdrawal did not affect the sero-
conversion rates. However, morphine-treated animals
showed reduced alterations in brainstem auditory
evoked potentials, which are negatively affected by
FIV load, perhaps suggesting a protective effect of the
drug. Clear-cut evidence for alteration in FIV load was
not obtained because of the small number of animals
and the large variation in virmeia at various time points
as detected by real-time reverse transcription-polymer-
ase chain reaction (PCR).
There are only a limited number of studies inves-
tigating the effects of withdrawal from opioids on
immune responses in rodents. Bhargava et al. (1994)
examined the effect of AW from slow-release morphine
pellets on a number of immune parameters in mice.
Eight hours after the initiation of withdrawal, NK cell
activity was variable in regard to suppression. The lack
of clear suppression of NK activity stands in contrast to
the results obtained in morphine-tolerant animals, in
which NK activity was clearly suppressed. The func-
tional status of NK cells is also inconsistent with effects
on other immune parameters. Cytotoxic T cell activity,
B cell proliferative responses to anti-IgM plus IL-4, and
IL-2 production in supernatants of spleen cells stimu-
lated with Con A were depressed in morphine-with-
drawn animals. West et al. (1999) used a paradigm of
chronic opioid administration in drinking water to rats
for 20 days, at which time AW was initiated by dis-
continuing the drug. The NK cell activity measured at
12, 24, or 48 h after withdrawal was suppressed, but
was not suppressed at the 6-h time point. Responses to
mitogens were suppressed at 12 h but returned to nor-
mal levels by 24 h. Similarly, production of IFN-g
from Con A-stimulated spleen cell supernatants was
suppressed in cells taken at 12 h, but not at 24 h, after
withdrawal. Whether the difference in the effect on
NK activity between the studies of Bhargava et al. and
West et al. is attributable to differences in modes of
opioid administration, to the animal (mouse vs. rat), or
to the time points examined, is not known with cer-
tainty, but it is likely that the 8-h time point examined
by Bhargava et al. was not long enough to yield consis-
tent, significant differences.
Eisenstein’s laboratory undertook a detailed kinetic
study comparing the effects of AW and PW from slow-
release morphine pellets in mice on the capacity of
spleen cells to mount an ex vivo primary antibody re-
sponse to sheep red blood cells as the antigen (Rahim
et al. 2002). Figure 1B shows the kinetic curves that
were generated for the two withdrawal paradigms. In
both cases, profound immunosuppression was observed
by 24–48 h after the initiation of abstinence. However,
the shape of the curves for immunosuppression induced
by AW vs. PW differed. When the drug was discon-
tinued, via removal of the slow-release pellet 96 h after
it was implanted, mice showed a slow, but steady, de-
crease in splenic capacity to generate an
ex vivo anti-
body response. At 144 h after the start of withdrawal,
immune responsiveness was still over 60% below nor-
mal. The results were different if withdrawal was pre-
cipitated by injection of naloxone or administration of
minipumps dispensing naltrexone after pellet removal.
In this paradigm of PW, there was an initial period of
immunoenhancement that peaked at 3 h, followed by a
phase of immunosuppression, which reached the same
nadir as that observed with AW; however, animals’ im-
mune responses recovered by 72 h. These results were
surprising for several reasons. First, biological re-
sponses in withdrawal are usually opposite to that which
is observed during acute or subacute opioid adminis-
tration, which means that overall immunoenhancement
was what was expected. Second, in PW, mice show
many signs of physical dependence, including jumping
242 J Neuroimmune Pharmacol (2006) 1: 237–249
and wet-dog shakes, whereas following AW these overt
signs are not observed, indicating that PW is a more
dynamic physical event than AW. Yet, mice experienc-
ing PW showed a more transient level of immunosup-
pression and recovered much faster than those
experiencing AW. Body weight was depressed in both
AW and PW at 24–48 h (Rahim et al. 2002). Spleen
weight was more severely depressed in mice undergo-
ing PW, even though they had less immunosuppression
than AW mice. Therefore, the degree of suppression
of immune responses appears to be uncorrelated with
the severity of other physical changes accompanying
withdrawal. The cellular composition of the spleen was
also analyzed in mice that were undergoing withdraw-
al. No significant difference in percentages of B cells, T
cells, or macrophages was observed between spleens of
normal mice, mice that were tolerant to morphine, or
mice 24 h after withdrawal (Rahim et al. 2003). This is
in contrast to the relative cell numbers in the spleens
of mice 48 h after morphine pellet implantation when
the number of B cells and macrophages are reduced
and the number of T cells is increased (Hilburger et al.
1997a). Further studies were carried out in Eisenstein’s
laboratory to identify mechanism(s) of inhibition of
immune function, either the cell(s) whose activity(ies)
are compromised during withdrawal or alterations in
cytokines or costimulatory molecules (Rahim et al.
2003, 2005). To approach this problem, splenic cocul-
tures containing mixtures of spleen cells from placebo
withdrawn or morphine withdrawn mice with normal
cells were used. In the first set of experiments, normal
cells were added to cells from mice in AW or PW for
24 h. Controls were included for cell crowding, by com-
paring cocultures containing morphine withdrawn cells
to similar cultures that had equal numbers of normal or
placebo-withdrawn cells added. It was observed that
addition of unfractionated normal cells to morphine
withdrawn cells in a 1:3 ratio resulted in restoration of
immune responses. A series of fractionation steps were
then carried out to identify the restorative cell popula-
tion. Simple separation of splenic macrophages from
lymphocytes by plastic adherence showed that the
adherent, macrophage-rich population, but not the
nonadherent lymphocyte-rich population, was active.
Macrophages purified by magnetic columns or by cell
sorting in a flow cytometer for expression of CD11b
+
restored immune responses to withdrawn spleen cells,
but CD11b
j
cells (T- and B-cell-enriched) did not
(Rahim et al. 2003). Furthermore, addition of the
macrophage cytokine, IL-1b, or the macrophage acti-
vating cytokine, IFN-g, restored immune responsive-
ness to pure cultures of cells of AW mice (Rahim et al.
2003). These results suggested that spleens of with-
drawn mice had a deficit of macrophage function, which
could be a possible cause of their incapacity to make in
vitro antibody responses. This conclusion was strength-
ened by RNase protection assays showing that mRNA
levels for the proinflammatory cytokines IL-1b and
tumor necrosis factor alpha (TNF-a) were depressed
in spleens of AW and PW mice compared to placebo-
withdrawn animals. In addition, macrophages obtained
from spleens of AW mice had depressed expression of
the costimulatory molecule B7.2, which is needed to
initiate immune responses by stimulation of naı¨ve T
cells. B7.2 signals tend to bias toward Th
2
cytokine
production, so that a depression in expression of this
antigen would be consonant with reduced antibody
production (Allison 1994; King et al. 1995; Rulifson et
al. 1997). Another approach used to assess the immune
status of spleen cells of withdrawn mice was to measure
cytokine production after stimulation of the cells with
either Con A or LPS. It was observed that IL-12p40 and
IFN-g were depressed when either mitogen was used as
the stimulus in cells taken from morphine-withdrawn,
as compared to placebo-withdrawn animals. IL-1b and
TNF-a were also significantly depressed between pla-
cebo- and morphine-withdrawn groups stimulated with
LPS (Rahim et al. 2003). These results support the
conclusion that withdrawal from morphine results in
macrophages that are deficient in capacity to support
the process of antibody formation by B and T cells.
Another set of experiments was carried out by
Rahim et al. (2005), who used the inverse paradigm
in which cells taken from spleens of animals 24 h after
they entered withdrawal from morphine were added to
cultures of normal cells in a 1:3 ratio. Unexpectedly, it
was found that cells from the morphine-withdrawn
spleens could suppress in vitro antibody responses of
normal spleen cells in culture. This observation would
imply that there are active suppressor cells in the
spleens of withdrawn animals. Extensive fractionation
studies carried out using magnetic cell separation
showed that addition of purified macrophage or B cell
fractions from withdrawn spleens suppressed normal
cells, but T cell fractions did not. These results were
confirmed by adding in macrophage or B cell fractions
from withdrawn spleens that were sorted by flow cyto-
metry to yield 99% purity. It was also shown that B cells
from withdrawn mice could suppress responses to Con
A, but not to LPS, showing that the suppressor activity
could be demonstrated in two systems, in vitro antibody
formation and responses to mitogens.
Thus, the observations obtained when cultures of
withdrawn cells and normal cells are in a ratio of 3:1,
compared to a ratio of 1:3, lead to different conclu-
sions—a finding that is paradoxical (Rahim et al. 2004).
J Neuroimmune Pharmacol (2006) 1: 237–249 243
The first paradigm suggests that immunosuppression is
attributable to a deficit in macrophage function. The
second paradigm suggests that there are active sup-
pressor macrophages and B cells. The complexity is
mostly in regard to the macrophages, as suppressor
macrophages have typically been defined as ones that
produce nitric oxide, a condition that is usually associ-
ated with maximal macrophage activation (Eisenstein
2001). Preliminary studies have shown that addition of
an inhibitor of nitric oxide, N-monomethyl-
L-arginine,
to cocultures of withdrawn and normal cells (3:1 ratio)
blocks the immunosuppression induced by the mor-
phine-withdrawn cells; however, no nitric oxide is
detectable, and inducible nitric oxide synthase is not
elevated (unpublished observations). Suppressor B
cells have been described in the literature (Cuff et al.
1989), but have never been fully characterized, and
remain a footnote to immunoregulatory processes.
Chronic exposure to opioids and withdrawal in vitro
Morphine has been tested for its capacity to inhibit
various immune responses when cells are exposed to
the drug in vitro (for a review, see Eisenstein et al.
1996). However, there are few systematic studies
examining whether time of in vitro exposure to an
opioid alters any observed effects. Renaud’s laborato-
ry has developed a paradigm in which murine perito-
neal macrophages phagocytize sheep red blood cells
opsonized with IgG antibody via Fc receptors. Acute
exposure of macrophages to morphine (50–100 nM) in
vitro for 30 min inhibits phagocytosis (Casellas et al.
1991). However, chronic exposure for 6–16 h had no
inhibitory effect, and at some time points, a slight
stimulatory effect (Tomei and Renaud 1997), a state
called Bputative tolerance.^ Removal of morphine by
washing, followed by a 5-h period before retesting,
resulted in reinstatement of inhibition of phagocytosis
(Tomei and Renaud 1997;La´zaro et al. 2000), a phe-
nomenon interpreted to be an in vitro analog of
Bdependence and withdrawal.^ These phenomena
were demonstrable in peritoneal cells obtained from
C3HeB/FeJ and C57BL/6 mice, but not from mu re-
ceptor knockout mice on the C57/BL6 background,
demonstrating the pharmacological specificity of the
biological effect (Tomassini et al. 2003).
Recently, Wang et al. (2005) undertook similar
studies in which human hepatic cell lines infected with
hepatitis C virus (HCV) were exposed to morphine in
vitro for 4 days. On day 5, the opioid was washed out
of the culture, with or without addition of naloxone,
thereby simulating AW or PW. Hepatitis C virus RNA
and HCV protein levels were assessed in two different
cell lines, and in both, removal of morphine enhanced
HCV replication from 50% to 100%. Addition of nal-
oxone to simulate PW further increased RNA levels to
2.5-fold above that of untreated controls. The adverse
effect of morphine withdrawal in promoting HCV rep-
lication was attributed to down-regulation of IFN-a in
the hepatic cells.
Effects of opioid withdrawal on response to bacterial
LPS and infection
Additional paradoxical observations regarding the im-
mune status of animals in withdrawal have been uncov-
ered by carrying out in vivo experiments. Eisenstein’s
laboratory asked whether mice in withdrawal are more
susceptible to infection or to toxicity of microbial
products. In particular, sensitivity to the toxic effects of
bacterial LPS and to infection with the gram-negative
bacterium, Salmonella typhimurium, were tested. Un-
der the experimental design, bacterial LPS was injected
to mice undergoing AW from morphine. Lipopolysac-
charide causes a septic syndrome and mortality when
given in high doses. To assess whether morphine en-
hanced sensitivity to this microbial extract, groups of
mice were injected i.p. with what would normally be a
sublethal dose of LPS (80 mg/mouse) 24 h after the re-
moval of either placebo or morphine pellets. As ex-
pected, the placebo-withdrawn animals all survived (17
survived/17 total). In sharp contrast, all morphine-
withdrawn animals died 50 h after LPS injection (0 sur-
vived/17 total) (Feng et al.
2005a). Mice that received a
morphine pellet and were not withdrawn, but kept their
pellet in place for the extra 24 h, and were then given
80 mg of LPS, had minimal mortality (7 survived/10
total), indicating that it is not just morphine adminis-
tration that sensitized the animals to LPS, but also the
withdrawal process. Sepsis and septic shock are medi-
ated by TNF-a, and mice in withdrawal from morphine
had elevated levels of this cytokine in the blood at 3 and
6 h after the challenge with LPS. In contrast, there was
no detectable TNF-a in the blood of mice that were
withdrawn from morphine and given saline instead of
LPS. Placebo-withdrawn mice given LPS had slightly
elevated TNF-a. Levels of other proinflammatory
cytokines and chemokines were also measured by using
the cytometric bead assay (BD Bioscience), which
allows the simultaneous assessment of a panel of these
mediators in small amounts of mouse serum. IL-6,
another proinflammatory cytokine, had significantly
sustained elevation in levels in serum of morphine-
withdrawn mice receiving LPS compared to similarly
treated placebo-withdrawn mice, at 6 and 12 h after
244 J Neuroimmune Pharmacol (2006) 1: 237–249
challenge. A significant and reproducible result was that
IL-12p70 was depressed over the 12-h period examined
in blood of morphine-withdrawn, LPS challenged mice
compared to the placebo-withdrawn, LPS-challenged,
group. These differences in cytokine expression were
supported by data examining splenic mRNA levels by
using real-time PCR at 3 h post-LPS administration. IL-
10 and iNOS mRNA were also significantly elevated in
the morphine-withdrawn group receiving LPS, as was
serum nitric oxide (Feng et al. 2005a). All of these data
indicate that mice in withdrawal from morphine, but
not from a placebo pellet, were in a sensitized state
that made them hyperreactive to LPS. Heightened pro-
duction of proinflammatory cytokines, such as TNF-a
and IL-6, as well as elevated levels of NO and iNOS,
are indicative of a state of macrophage activation. IL-
10 is frequently elevated under these conditions to
limit the potentially toxic proinflammatory response of
the macrophages. The role of TNF-a in the mortality
rate of morphine-withdrawn, LPS-challenged mice was
demonstrated by using passive administration of a
goat antibody to TNF-a. When this antibody was given
just before the LPS injection, with a second injection 3
h later, the mortality rate was reduced from 100% to
50%, which was highly significant (Feng et al. 2005a).
The conclusion that withdrawal results in macrophage
activation seems to contradict some of the findings
described in the first part of this review, where
evidence from several approaches supported the
conclusion that there is a deficit in macrophage
function in spleen cells tested 24 h postwithdrawal
without exposure to LPS in vivo. Among these lines of
evidence was the observation that when LPS was
added in vitro to these morphine-withdrawn cells,
proinflammatory cytokine production was lower than
that of cells taken from placebo-withdrawn animals
(Rahim et al. 2003).
Other data also bear on the interpretation of the
complex events that must be occurring after morphine
withdrawal to lead to in vivo sensitization to LPS. Mice
were challenged 24 h after morphine withdrawal with
live S. typhimurium, either i.p. (Feng et al. 2005b)or
orally (unpublished data). In both cases, it was found
that morphine-withdrawn animals were sensitized to
this pathogen compared to placebo-withdrawn mice. A
major finding was that IL-12p70 and IL-12p40 were
suppressed in sera of infected, morphine-withdrawn
animals compared to the infected, placebo-withdrawn
group. Thus, mice in morphine withdrawal failed to
respond to either LPS or Salmonella challenge with an
elevation in IL-12, which would normally be part of
the cytokine response to these types of insults. Placebo-
withdrawn animals manifested the expected elevation
in IL-12. IL-4 was measured in sera of morphine-
withdrawn mice 48 h after Salmonella infection. When
the infection was i.p., an increase in IL-4 was observed,
but overall levels of the cytokine were low. In animals
that were not infected, IL-4 levels of morphine-with-
drawn mice were depressed compared to those of
placebo animals (unpublished observations). Similar
experiments have been carried out in Dr. Roy’s lab-
oratory in which mice were implanted with slow-release
morphine pellets that were removed after 72 h, rather
than 96 h (Kelschenbach et al. 2005). Spleen cells were
harvested after 24 h and stimulated in vitro with Con
A. Supernatants were analyzed for cytokines that po-
larize the immune response toward Th
1
or Th
2
re-
sponses. It was found that morphine-withdrawn, Con
A-stimulated supernatants had increased levels of IL-4
and increased levels of mRNA for IL-4 compared to
placebo-withdrawn supernatants. In contrast to the
findings of Rahim et al. (2003), IFN-g levels were not
depressed. The molecular basis for increased IL-4 pro-
duction was pursued, and it was observed that mor-
phine withdrawal increased GATA-3 protein levels in
Western blots. GATA-3 is a transcription factor that
regulates IL-4 mRNA levels, so an increase would be
compatible with the hypothesis of polarization toward
aTh
2
response. Electrophoretic mobility shift assays
showed that GATA-3 binding, as well as that of another
IL-4-associated transcription factor, Stat-5/6, were ele-
vated in spleen cells of morphine-withdrawn mice. The
authors conclude that morphine withdrawal polarizes
the immune response to a Th
2
phenotype. Whether or
not Th1 is down-regulated remains to be settled. IFN-g
levels were not decreased, but T-bet, the transcription
factor that up-regulates the Th
1
pathway, was depress-
ed in cells of morphine-withdrawn mice. Interestingly,
this group also reported that IL-12p70 and IFN-g levels
in serum were decreased in mice challenged with LPS.
Feng et al. (2005a), in a kinetic analysis, found that at
6 h after LPS in morphine-withdrawn mice, IFN-g
was depressed, but by 12 h the levels of this cytokine
were the same as those of the placebo-withdrawn plus
LPS group. There were differences in the protocols
between the two groups. Eisenstein’s laboratory ad-
ministered 80 mg LPS 24 h postwithdrawal (96 h after
initial pellet implantation) and monitored the blood at
3, 6, and 12 h post-LPS (Feng et al. 2005a). Meanwhile,
Kelschenbach et al. (2005) group gave 25 mg LPS at the
time of withdrawal (72 h after initial pellet implanta-
tion) and assessed serum cytokine levels at 24 h after
both LPS and the start of withdrawal.
An intriguing question is what blocks the IL-12
response during morphine withdrawal? It cannot be a
simple explanation such as inhibition of mobilization
J Neuroimmune Pharmacol (2006) 1: 237–249 245
of NF-kB to the nucleus, as other proinflammatory
cytokines, such as TNF-a (which also utilize NF-kBasa
transcription factor), are up-regulated. An intriguing
piece of data in Kelschenbach et al.’s work is that ad-
ministration of the pan-opioid antagonist, naltrexone,
at the time of morphine withdrawal, blocked the in-
crease in GATA-3 levels. The interpretation of this
finding by the investigators is that endogenous opioids
may be providing a bias toward Th
2
immune responses
during withdrawal. It is difficult to know how to re-
concile these findings with those from Eisenstein’s lab-
oratory showing definitively that spleen cells of mice in
withdrawal are highly immunosuppressed in their ca-
pacity to form an antibody response to sheep red blood
cells when cultured ex vivo. If the immune system is
being biased toward a Th
2
response, one would expect
that antibody production would be enhanced, not de-
pressed, because Th
2
cytokines promote antibody for-
mation. Both groups agree that IL-12 is inhibited when
immune cells are stimulated in vivo or in vitro after
morphine withdrawal. Inhibition of IL-12 would be
expected to interfere with Th
1
responses, which may or
may not have a direct bearing on the extent of Th
2
responses. Eisenstein’s group did not examine IL-4
levels as intensively as Roy’s group, but their labora-
tory did not find elevated IL-4. Two other groups have
examined immune responses after withdrawal in mice.
Bhargava et al. (1994) measured IL-4 levels in super-
natants of spleen cells taken from mice 8 h after ini-
tiation of withdrawal and stimulated in vitro for 48
with anti-CD3 antibody. They reported no difference
in IL-4 levels in morphine-withdrawn supernatants as
compared to placebo-withdrawn, spleen supernatants.
However, spleen cells of morphine-tolerant mice were
suppressed in IL-4 production with respect to placebo-
tolerant spleens. Therefore, the observation of no
difference in IL-4 levels after initiation of withdrawal
actually represents an increase in IL-4 over the pre-
withdrawal levels, even though the levels in abstinence
are not different from placebo. Perhaps, if the kinetics
had been followed further, IL-4 levels might have con-
tinued to rise as the time after withdrawal increased.
Bhargava et al. also measured IL-2 levels produced by
T cells stimulated with anti-CD3 and placed in a bio-
assay for cytotoxic T cells. Both tolerant and abstinent
mice had markedly suppressed IL-2 levels. IL-2 is a
Th
1
polarizing cytokine, which would fit with the ob-
servations made by Eisenstein et al. and Roy et al. that
IL-12 is depressed. Bhargava et al. also reported that
cytotoxic T cell activity was suppressed in mice in with-
drawal, but not in tolerant mice, whereas NK cell ac-
tivity was suppressed in tolerant mice, but variably
suppressed in abstinent animals. In contrast, B cell pro-
liferation in response to anti-IgM/IL-4 showed no sup-
pression in tolerant animals and marked suppression in
morphine-withdrawn animals. Rahim et al. (2005) found
that responses of T cells to Con A were suppressed
24 h after morphine withdrawal, but there were no
differences in proliferation to LPS.
Is there a unifying hypothesis on effects of tolerance
and withdrawal from opioids on the immune system?
As discussed in the Overview, there is general agree-
ment that acute exposure to morphine leads to sup-
pression of various immune responses. The existing
literature on effects of chronic exposure to morphine
supports the conclusion that some immune parameters,
which are suppressed during acute exposure to mor-
phine, return to normal levels as the time of exposure to
the drug is increased, i.e., tolerance to the effects of
opioids can develop in the immune system. However,
depending on the parameter measured and the species
being tested, results vary.
The scanty literature on effects of withdrawal from
opioids on immune responses is in agreement that im-
munosuppression follows morphine withdrawal, and
that the cytokine, IL-12, is depressed. In addition, a
major point of agreement between several laboratories
seems to be that acute or chronic exposure to morphine,
or withdrawal from morphine, sensitizes subjects to the
toxic effects of bacterial LPS. Hilburger et al. (1997b)
was the first to observe that at 48 h after implantation
of a 75-mg morphine pellet, mice became septic, as
evidenced by the presence of normal flora of the
gastrointestinal tract in the peritoneal cavity, spleen,
and liver. Moreover, morphine-treated animals were
found to be sensitized to a sublethal dose of bacterial
LPS, resulting in 100% mortality. These findings were
confirmed by Roy et al. (1999). As described above, a
similar phenomenon of sensitization to sublethal doses
of LPS was observed by Feng et al. (2005a) in mice
undergoing abrupt withdrawal. Furthermore, antibod-
ies to TNF-a protected them from mortality after LPS
challenge, suggesting that morphine withdrawal sensi-
tizes to LPS lethality by increasing the production of
TNF-a. Ocasio et al. (2004) reported that rats exposed
to morphine for 5 days have measurable endotoxin in
their serum. When they were challenged with LPS, they
demonstrated coagulopathy in the microvasculature,
and a diminished mean arterial blood pressure, both
signs of endotoxemia and sepsis. The proinflammatory
cytokines, IL-1, IL-6, and TNF-a, were elevated follow-
ing LPS administration. Recent work in Eisenstein’s
laboratory shows that mice in withdrawal are also
246 J Neuroimmune Pharmacol (2006) 1: 237–249
septic, with enteric bacteria being culturable from the
blood and liver (manuscript in preparation). Collective-
ly, these results suggest that morphine given acutely or
chronically, or when it is withdrawn, can result in a state
of sepsis. These observations may help to explain the
seemingly conflicting observations that antibody
responses are suppressed in mice in withdrawal (Rahim
et al. 2002), but Th
2
cytokines and Th
2
cytokine trans-
cription factors are increased (Kelschenbach et al.
2005). The elevation in Th
2
cytokines observed in mice
in withdrawal by Kelschenbach et al. may be attribut-
able to an underlying level of LPS present in the circu-
lation that results from occult sepsis. LPS is known to
elicit Th
2
rather than Th
1
responses (Eisenbarth et al.
2002). The immunosuppression of antibody responses
observed by Eisenstein may be mediated by other
mechanisms that override the Th
2
polarization.
Little work has been done to elucidate the mech-
anisms that control levels of immune responses fol-
lowing withdrawal. Studies on neural pathways, and
neuropeptide, neurotransmitter, and endocrine media-
tors activated during withdrawal provide many candi-
date mechanisms and molecules that could modulate
immune responses. These include serotonin, norepine-
phrine, and adenosine. Chang has presented evidence
supporting a hypofunctional HPA axis in rats exposed
to chronic morphine. Eisenstein’s laboratory is address-
ing the question as to whether elevated glucocorticoids
mediate the immunosuppression of antibody formation
in spleen cells extracted from morphine withdrawn
animals. Unpublished data show that RU486 does not
block the immunosuppression that is observed, a result
that argues against a role for glucocorticoids as the sup-
pressor mechanism. Kelschenbach et al. assayed serum
of morphine-withdrawn mice 24 h after withdrawal
from a morphine pellet and did not find elevated cor-
ticosterone levels, leading Kelschenbach et al. (2005)
to conclude that cortisone is not the inducer of the
immune suppression. In cats maintained on morphine
and subjected to periods of withdrawal, with or with-
out FIV infection, plasma cortisol levels rose during
periods of withdrawal. Within the limits of the assays
used, this increase did not seem to have an adverse
effect on FIV load or disease progression (Barr et al.
2003). West et al. (1999) reported that clonidine, an
a
2
-adrenergic receptor agonist, blocked the suppression
of NK activity, mitogen responses, and depressed IFN-g
in rats in withdrawal from orally administered mor-
phine, leading to the hypothesis that the sympathetic
nervous system may be indirectly causing the immuno-
suppression. Clearly, much more work needs to be
directed toward elucidating how opioid withdrawal
alters immune responses. Finally, other provocative
pieces of evidence are worth considering when thinking
about neural-immune connections during withdrawal.
Dafny et al. reported as early as 1986 that mice
subjected to whole-body irradiation, and thereby de-
pleted of immune cells, failed to express signs and
symptoms of withdrawal (Dafny and Pellis 1986; Pellis
et al. 1987; Dougherty et al. 1986). Moreover, IFN-a
and cyclosporine have been reported to ameliorate the
behavioral effects of withdrawal (McVaugh et al. 1989).
These studies raise the possibility that during with-
drawal, not only may neural pathways modify immune
responses, but cells of the immune system may modify
neuronal activity.
Future avenues of inquiry regarding effects
of withdrawal on immune responses
There is a need for more descriptive studies in animal
models, and if possible, in humans, documenting what
occurs during withdrawal with respect to immune re-
sponses and infection. Certainly, further studies on
mechanisms by which immune perturbations occur
would be beneficial in understanding how opioids affect
cells of the immune system. Mechanistic studies could
occur at the level of documenting the changes that occur
in the immune system. Examples of this approach are
those used by Kelschenbach et al. (2005) to examine T
cell polarization, by Feng et al. (2005a) demonstrating
IL-12 depression, and by Wang et al. (2005) showing
that in vitro withdrawal alters expression of IFN re-
gulatory factor 7, which is a strong transactivator of the
IFN-a promoter. The other type of mechanistic studies
comprises those that try to dissect the pathways leading
to immunomodulation following withdrawal. West et al.
(1999) have shown that clonidine blocks the suppres-
sion of several immune parameters following withdraw-
al, suggesting a sympathetic nervous system mediation
of immunosuppression. Other tools should be used to
investigate in greater depth the pathways that might
be involved. As described above, the little evidence that
is available does not support the hypothesis that the
HPA axis is the major mediator of withdrawal-induced
immunosuppression.
Acknowledgements ThisstudywassupportedbyNational
Institute of Drug Abuse grants DA14223, DA11134, DA13429
and DA06650. The authors thank Dr. Martin W. Adler for a
critical reading of the manuscript.
References
Allison JP (1994) CD28–B7 interactions in T-cell activation.
Curr Opin Immunol 6:414–419
J Neuroimmune Pharmacol (2006) 1: 237–249 247
Alonzo NC, Bayer BM (2002) Opioids immunology and host
defenses of intravenous drug abusers. Infect Dis Clinic N
Am 16:553–569
Barr MC, Huitron-Resendiz S, Sanchez-Alavez M, Hendriksen
SJ, Phillips TR (2003) Escalating morphine exposures
followed by withdrawal in feline immunodeficiency virus-
infected cats: a model for HIV infection in chronic opiate
abusers. Drug Alcohol Depend 72:141–149
Bayer BM, Brehio RM, Ding XZ, Hernandez MC (1994)
Enhanced susceptibility of the immune system to stress in
morphine-tolerant rats. Brain Behav Immun 8:173–184
Bhargava HN, Thomas PT, Thorat S, House RV (1994) Effects
of morphine tolerance and abstinence on cellular immune
function. Brain Res 642:1–10
Bryant HU, Roudebush RE (1990) Suppressive effects of
morphine pellet imlants on in vivo parameters of immune
function. J Pharmacol Exp Ther 255:410–414
Bryant HU, Bernton EW, Holaday JW (1987) Immunosuppres-
sive effects of chronic morphine treatment in mice. Life Sci
41:1731–1738
Bussiere JL, Adler MW, Rogers TJ, Eisenstein TK (1993)
Cytokine reversal of morphine-induced suppression of the
antibody response. J Pharmacol Exp Ther 264:591–597
Carr DJJ, France CP (1993) Immune alterations in morphine-
treated Rhesus monkeys. J Pharmacol Exp Ther 267:9–15
Carr DJ, Rogers TJ, Weber RJ (1996) The relevance of opioids
and opioid receptors on immunocompetence and immune
homeostasis. Proc Soc Exp Biol Med 213:248–257
Casellas AM, Guardiola H, Renaud FL (1991) Inhibition by
opioids of phagocytosis in peritoneal macrophages. Neuro-
peptides 18:35–40
Centers for Disease Control and Prevention (2005) In: HIV/
AIDS surveillance report. US Dept. of Health and Human
Services, Atlanta
Chuang LF, Killam KF Jr, Chuang RY (1993) Opioid dependency
and T-helper functions in Rhesus monkey. In Vivo 7:159–166
Cuff CM, Packer BJ, Rogers TJ (1989) A further characteriza-
tion of Candida albicans-induced suppressor B-cell activity.
Immunology 68:80–86
Dafny N, Pellis NR (1986) Evidence that opiate addiction is in
part an immune response. Destruction of the immune
system by irradiation-altered opiate withdrawal. Neuro-
pharmacology 25:815–818
Donahoe RM, Byrd LD, McClure HM, Fultz P, Brantley M,
Marsteller F, Ansari AA, Wenzel D, Aceto M (1993)
Consequences of opiate-dependency in a monkey model of
AIDS. Adv Exp Med Biol 335:21–28
Dougherty PM, Harper C, Dafny N (1986) The effect of alpha-
interferon, cyclosporine A, and radiation-induced immune
suppression on morphine-induced hypothermia and toler-
ance. Life Sci 39:2191–2197
Eisenbarth SC, Piggott DA, Huleatt JW, Visintin I, Herrick CA,
Bottomly K (2002) Lipopolysaccharide-enhanced, Toll-like
receptor 4-dependent T helper cell type 2 responses to
inhaled antigen. J Exp Med 196:1645–1651
Eisenstein TK (2001) Implications of Salmonella-induced nitric
oxide (NO) for host defense and vaccines: NO, an
antimicrobial, antitumor, immunosuppressive and immuno-
regulatory molecule. Microbes Infect 3:1223–1231
Eisenstein TK, Hilburger ME (1998) Opioid modulation of
immune responses: effects on phagocyte and lymphoid cell
populations. J Neuroimmunol 83:36–44
Eisenstein TK, Hilburger ME, Lawrence DMP (1996) Immuno-
modulation by morphine and other opioids. In: Friedman H,
Klein TW, Specter S (eds) Drugs of abuse, immunity and
infections. CRC Press, Boca Raton, pp 103–120
Feng P, Meissler JJ Jr, Adler MW, Eisenstein TK (2005a)
Morphine withdrawal sensitizes mice to lipopolysaccharide:
elevated TNF-alpha and nitric oxide with decreased IL-12.
J Neuroimmunol 164:57–65
Feng P, Wilson QM, Meissler JJ Jr, Adler MW, Eisenstein TK
(2005b) Increased sensitivity to Salmonella enterica serovar
Typhimurium infection in mice undergoing withdrawal from
morphine is associated with suppression of interleukin-12.
Infect Immun 73:7953–7959
Go´ mez-Flores R, Weber RJ (1999) Opioids, opioid receptors,
and the immune system. In: Plotnikoff NP, Faith RE,
Murgo AJ, Good RA (eds) Cytokines, stress and immunity.
CRC Press, Boca Raton, pp 281–314
Govitrapong P, Suttitum T, Kotchabhakdi N, Uneklabh T (1998)
Alterations of immune functions in heroin addicts and heroin
withdrawal subjects. J Pharmacol Exp Ther 286: 883–889
Hall DM, Suo J-L, Weber RJ (1998) Opioid mediated effects on
the immune system: sympathetic nervous system involve-
ment. J Neuroimmunol 83:29–35
Hilburger ME, Adler MW, Rogers TJ, Eisenstein TK (1997a)
Morphine alters macrophage and lymphocyte populations in
the spleen and peritoneal cavity. J Neuroimmunol 80:106–114
Hilburger ME, Adler MW, Truant AL, Meissler JJ Jr,
Satishchandran V, Rogers TJ, Eisenstein TK (1997b)
Morphine induces sepsis in mice. J Infect Dis 176:183–188
House SD, Mao X, Wu GD, Espinelli D, Li WX, Chang SL (2001)
Chronic morphine potentiates the inflammatory response by
disrupting interleukin-1b modulation of the hypothalamic–
pituitary–adrenal axis. J Neuroimmunol 118:227–285
Kelschenbach J, Barke RA, Roy S (2005) Morphine withdrawal
contributes to Th cell differentiation by biasing cells toward
the Th
2
lineage. J Immunol 175:2655–2665
King CL, Stupi RJ, Craighead N, June CH, Thyphronitis G
(1995) CD28 activation promotes Th
2
subset differentiation
by human CD4
+
cells. Eur J Immunol 25:587–595
La´zaro MI, Tomassini N, Gonza´lez I, Renaud FL (2000)
Reversibility of morphine effects on phagocytosis by murine
macrophages. Drug Alcohol Depend 58:159–164
McCarthy L, Wetzel M, Sliker J, Eisenstein TK, Rogers TJ
(2001) Opioids, opioid receptors, and the immune response.
Drug Alcohol Depend 62:111–123
McVaugh W, Lawrence B, Kulkarni A, Pizzini R, Van Buren C,
Rudolph F, Wolinsky I, Dafny N (1989) Suppression of
opiate withdrawal by cyclosporin A and dietary modifica-
tion. Life Sci 44:977–983
Mellon RD, Bayer BM (1998) Evidence for central opioid receptors
in the immunomodulatory effects of morphine: review of
potential mechanism(s) of action. J Neuroimmunol 83:19–28
Molitor TW, Morilla A, Risdahl JM, Murtaugh MP, Chao CC,
Peterson PK (1992) Chronic morphine administration im-
pairs cell-mediated immune responses in swine. J Pharmacol
Exp Ther 260:581–586
Ocasio FM, Jiang Y, House SD, Chang SL (2004) Chronic
morphine accelerates the progression of lipopolysaccharide-
induced sepsis to septic shock. J Neuroimmunol 149:90–100
Pacifici R, di Carlo S, Bacosi A, Zuccaro P (1993) Macrophage
functions in drugs of abuse-treated mice. Int J Immuno-
pharmacol 15:711–716
Pellis NR, Harper C, Dafny N (1986) Suppression of the
induction of delayed hypersensitivity in rats by repetitive
morphine treatments. Exp Neurol 93:92–97
Pellis NR, Kletzly NE, Dougherty PM, Aronowski J, Dafny N
(1987) Participation of lymphoid cells in the withdrawal
syndrome of opiate dependent rats. Life Sci 40:1589–1593
Peterson PK, Molitor TW, Chao CC (1998) The opioid-cytokine
connection. J Neuroimmunol 83:63–69
248 J Neuroimmune Pharmacol (2006) 1: 237–249
Rahim RT, Meissler JJ Jr, Cowan A, Rogers TJ, Geller EB,
Gaughan J, Adler MW, Eisenstein TK (2001) Administra-
tion of mu-, kappa- or delta
2
-receptor agonists via osmotic
minipumps suppresses murine splenic antibody responses.
Int Immunopharm 1:2001–2009
Rahim RT, Adler MW, Meissler JJ Jr, Cowan A, Rogers TJ,
Geller EB, Eisenstein TK (2002) Abrupt or precipitated
withdrawal from morphine induces immunosuppression.
J Neuroimmunol 127:88–95
Rahim RT, Meissler JJ Jr, Zhang L, Adler MW, Rogers TJ,
Eisenstein TK (2003) Withdrawal from morphine in mice
suppresses splenic macrophage function, cytokine produc-
tion, and costimulatory molecules. J Neuroimmunol
144:16–27
Rahim RT, Feng P, Meissler JJ Jr, Rogers TJ, Zhang L, Adler MW,
Eisenstein TK (2004) Paradoxes of immunosuppression in
mouse models of withdrawal. J Neuroimmunol 147:114–120
Rahim RT, Meissler JJ Jr, Adler MW, Eisenstein TK (2005)
Splenic macrophages and B-cells mediate immunosuppres-
sion following abrupt withdrawal from morphine. J Leukoc
Biol 78:1185–1191
Risdahl JM, Peterson PK, Molitor TW (1996) Opiates, infection
and immunity. In: Friedman H, Klein TW, Specter S (eds)
Drugs of abuse, immunity, and infections. CRC Press, Boca
Raton, pp 1–42
Roy S, Loh HH (1996) Effects of opioids on the immune system.
Neurochem Res 21:1373–1384
Roy S, Charboneau RG, Barke RA (1999) Morphine synergizes
with lipopolysaccharide in a chronic endotoxemia model.
J Neuroimmunol 95:107–114
Rulifson IC, Sperling AI, Fields PE, Fitch FW, Bluestone JA
(1997) CD28 costimulation promotes the production of Th
2
cytokines. J Immunol 158:658–665
Sharp BM, Roy S, Bidlack JM (1998) Evidence for opioid
receptors on cells involved in host defense and the immune
system. J Neuroimmunol 83:45–56
Shavit Y, Terman GW, Lewis JW, Zane CJ, Gale RP,
Liebeskind JC (1986) Effects of footshock stress and
morphine on natural killer lymphocytes in rats: studies of
tolerance and cross-tolerance. Brain Res 372:382–385
Tomassini N, Renaud FL, Roy S, Loh HH (2003) Mu and delta
receptors mediate morphine effects on phagocytosis by
murine peritoneal macrophages. J Neuroimmunol 136: 9–16
Tomei EZ, Renaud FL (1997) Effect of morphine on Fc-
mediated phagocytosis by murine macrophages in vitro.
J Neuroimmunol 74:111–116
Wang CQ, Li Y, Douglas SD, Wang X, Metzger DS, Zhang T,
Ho WZ (2005) Morphine withdrawal enhances hepatitis C
virus replicon expression. Am J Pathol 167:1189–1191
West JP, Lysle DT, Dykstra LA (1997) Tolerance development
to morphine-induced alterations of immune status. Drug
Alcohol Depend 46:147–157
West JP, Dykstra LA, Lysle DT (1999) Immunomodulatory
effects of morphine withdrawal in the rat are time depen-
dent and reversible by clonidine. Psychopharmacology
146:320–327
Wybran J, Appelboom T, Famaey J-P, Govaerts A (1979)
Suggestive evidence for receptors for morphine and methi-
onine–enkephalin on normal human blood T lymphocytes.
J Immunol 123:1068–1070
J Neuroimmune Pharmacol (2006) 1: 237–249 249
