Psychoneuroimmunology Conditioning and Stress Ader20Conditioning20 20immunology20AnnRevPsy201993

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/14764205
Psychoneuroimmunology: Conditioning and Stress
Article in Annual Review of Psychology · February 1993

DOI: 10.1146/annurev.ps.44.020193.000413 · Source: PubMed
CITATIONS READS
284 3,955
2 authors, including:
Nicholas Cohen
University of Rochester
244 PUBLICATIONS 7,431 CITATIONS
SEE PROFILE
All content following this page was uploaded by Nicholas Cohen on 29 May 2014.
The user has requested enhancement of the downloaded file.

Annual Reviews
www.annualreviews.org/aronline
Annu.Rev. Psychol.1993. 44:53~5

Copyright
©1993by AnnualReviewsInc. All rights reserved
PSYCHONEUROIMMUNOLOGY:
CONDITIONINGAND STRESS
Annu. Rev. Psychol. 1993.44:53-85. Downloaded from arjournals.annualreviews.org
by UNIVERSITY OF ILLINOIS - CHICAGO on 11/01/05. For personal use only.
Robert Ader and Nicholas Cohen
Department
of Psychiatry and Microbiologyand Immunology,
Universityof Rochester
Schoolof Medicineand Dentistry, Rochester, NewYork14642
KEYWORDS:
psychoneuroimmunology, conditioning, stress
CONTENTS
INTRODUCTION
..................................................................................................................... 53
CONDITIONED
MODULATION OFIMMUNITY ................................................................ 54
Effectsof Conditioning onHumoral andCell-Mediated Immunity ..................... 55
Effectsof Conditioning onNonimmunologically SpecificReactions .................. 59
Antigen
asUnconditioned Stimulus ..................................................................... 60
BiologicImpact of Conditioned Changes in Immunity ........................................ 61
Conditioning
inHuman Subjects ......................................................................... 61
STRESS
ANDIMMUNITY
......................................................................................................63
Effects
ofStress onDisease .................................................................................. 64
Effectsof StressonHumoral andCell-Mediated Immunity ................................ 66
Effectsof StressonNonimmunologically SpecificReactions .............................. 69
MEDIATION OF BEHAVIORALLY INDUCED IMMUNE CHANGES .............................. 71
SUMMARY
.............................................................................................................................. 76
INTRODUCTION
During the past 10-15 years, psychoneuroimmunology--the study of the in-
teractions amongbehavior, neural and endocrine function, and immunepro-
cesses--has developed into a bona fide field of interdisciplinary research
(Ader 1981a, 1991a). Previously unknownand unsuspected connections be-
tween the brain and the immunesystem provide a foundation for the now
numerousobservations both (a) that the manipulationof neural and endocrine
functions alters immuneresponses, and the antigenic stimulation that induces
an immuneresponse results in changes in neural and endocrine function and
(b) that behavioral processes are capable of influencing immunologicreactiv-
53
0066-4308/93/0201-0053502.00
Annual Reviews
54 ADER & COHEN
ity and, conversely, the immunestatus of an organism has consequencesfor

behavior. This new research indicates that the nervous and immunesystems,
the two most complex systems involved in the maintenance of homeostasis,
represent an integrated mechanism contributing to the adaptation of the indi-
vidual and the species. Psychoneuroimmunology emphasizes the functional
significance of the relationship betweenthese systems--notin place of, but in
addition to the moretraditional disciplinary analysis of the mechanismsgov-
erning functions within a single system.
The range of phenomenathat bears on the relationship between behavior
and immunityis quite broad, and no attempt will be madeto provide even a
cursory summary of all this literature. Wefocus here on animal studies of the
effects of conditioning and stress in the modulationof immunefunction. There

are several more or less programmaticlines of research in humansthat the
reader maywish to explore. These deal with the immunologiccorrelates of
emotional states (primarily depression), personality traits as modulators
immunefunction, and the effects of stress on immunefunction. Fewgeneral-
izations are possible based on currently available data. Althoughthere is no
definitive evidence for the implied chain of events, the hypothesis that im-
munefunction maymediate the effects of psychosocial factors on the suscepti-
bility to or progression of somedisease processes remains tenable. Weconfine
this review, however, to the experimental literature on the modulation of
immunityby stress and conditioning. Other recent reviews (e.g.S. Cohen
Williamson 1991; Geiser 1989; Kemenyet al 1992; O’Leary1990) have dealt
with personality and emotional factors and immunityand/or disease. Someof
these have included an introductory outline of the immunesystem; an exten-
sive treatment of immunefunction can be found in any of several recent texts
(e.g. Stites &Ten"1991).
CONDITIONED MODULATIONOF IMMUNITY

Immuneresponses, like other physiological processes, can be modified by
classical conditioning. Conditioned modulation of host defense mechanisms
and antigen-specific immuneresponses were first explored by Russian investi-
gators and followed the Pavlovian conditioning principles and procedures of
the day (Metal’nikov&Chorine 1926, 1928). Typically, multiple pairings of
conditionedstimulus (CS; e.g. heat, tactile stimulation) were paired with injec-
tions of a foreign protein, the unconditioned stimulus (UCS); subsequent
presentation of the CSalone was reported to elicit conditioned increases both
in a variety of nonspecific defense responses and in antibody production.
There were English language reviews of this literature (Hull 1934; Kopeloff
1941), but they apparently attracted little attention, probably owingto the
nascent state of immunologyat the time and to the fact that by today’s
standards most of these early animal experiments were inadequately de-
scribed, poorly designed, lacked appropriate control groups, and constituted
Annual Reviews
PSYCHONEUROIMMUNOLOGY
55
little morethan preliminary observations. Althoughthe data on specific anti-

bodyresponses were not convincing,the studies of nonspecific cellular events
(e.g. changes in leukocyte number, phagocytosis, inflammatory responses)
were consistent and provided provocative evidence that conditioning could
modulatehost defenses. A detailed review (and somereanalyses) of these data
was provided by Ader (1981 b).
Effects of Conditioning on Humoral and Cell-Mediated

Immunity
Oneof the hallmarks of the immunesystem’s defense of the organismagainst

foreign, "nonself"material (antigens) is its specificity--its ability to recognize

precisely and then eliminate only the antigens it has confronted. Theseactivi-
ties are carried out by a variety of white bloodcells (leukocytes). Prominent
amongthese are T and B lymphocytesthat are capable of clonal proliferation
in response to antigens and retaining the "memory"of that encounter. Hu-
moral or antibody-mediated immunityinvolves the exposure of antigens to
bone marrow-derivedB cells. These cells effect the ultimate production of
antibodies that protect the organismagainst extracellular microorganismsand
reinfection. Cell-mediated immunityis provided by thymus-derived T cells
that protect against intracellular parasitic and viral infections. Anintegrated
immuneresponse to antigens, however, involves complexinteractions among
specialized subpopulationsof T cells (i.e. helper, suppressor, cytotoxic),
cells, other white blood cells such as macrophages,and substances (cytokines)
that are secreted by activated leukocytes. A variety of techniques have been
developedto measurethese cellular interactions in vitro. The essence of psy-
choneuroimmunology, however,is the recognition that in vivo these reactions
occur within a neuroendocrinemilieu that is demonstrably sensitive to the
organism’s perception of and adaptation to events occurring in its environ-
ment.
HUMORAL IMMUNITY Current interest in conditioned changes in immunologic

reactivity beganwith a study by Ader &Cohen(1975). Using a taste aversion
conditioning paradigm, a saccharin-flavored drinking solution, the CS, was
paired with an injection of cyclophosphamide(CY), an immunosuppressive
UCS.All rats were subsequently immunizedwith sheep red blood cells (SRBC).
Antibodytiters were measuredin conditioned animals that were injected with
CYon the day of immunization(to define the unconditioned immunosuppress-
ive effects of CY),in conditionedanimalsthat werenot reexposedto the CS(to
assess the influence of prior conditioning and the residual effects of CY),and
in conditioned animals that were reexposedto the CSon the day of immuniza-
tion and/or three days later (the critical experimental group). Subgroups
(nonconditioned)animalswere injected with CYfollowing the drinking of plain
water before immunizationand were subsequently provided with the presum-
ably neutral saccharin solution whenevera comparablesubgroupof conditioned
Annual Reviews
56 ADER & COHEN
animals received saccharin. A placebo-treated group was injected with vehicle

after consumingsaccharin or water before immunization and was exposed to
saccharin after immunization. As expected, conditioned animals showed an
aversion to the saccharin solution that had been paired with CY.Conditioned
animals that werereexposedto the CSat the time of, and/or three days after,
immunizationalso showedan attenuated anti-SRBCantibody response relative
either to conditioned animals that were not reexposed to saccharin or to
nonconditionedanimals that were similarly exposedto saccharin. Theseresults
were interpreted as reflecting a conditioned immunosuppressive response. The
acquisition and the experimentalextinction of a conditionedsuppression and/or
enhancementof antibody- and cell-mediated immuneresponses as well as
nonspecific host defenses have nowbeen observed under a variety of experi-

mentalconditions. Onlya brief overviewof this research is providedhere; more
detailedreviews are available elsewhere (Ader & Cohen1985, 1991).
In the conditioned taste aversion paradigm,animals learn to avoid flavored
solutions previously paired with the noxious or illness-inducing effects of a
variety of (pharmacologic)agents; that is, they reduce their consumption
the CSsolution. Therefore, to obviate the conflict ("stress") inducedby having
either to drink a solution paired with illness or to remain thirsty--and to
equate total fluid consumptionamongdifferentially treated animals~ondi-
tioning can be assessed with a two-bottle preference procedurethat permits the
animal to choose between plain water and the flavored CS solution. Under
these conditions, conditioned alterations of humoral(e.g. Ader et al 1982;
Bovbjerget al 1987b; N. Cohenet al 1979) and cell-mediated (Bovbjerg et
1982, 1984) immuneresponses are still obtained. The available literature
reveals no consistent relationship betweenconditioned behavioral (aversive)
responses and conditioned immunechanges in the taste aversion learning
paradigm; taste aversions can be expressed without concomitant changes in
humoral immunity, and conditioned changes in immunefunction can be ob-
tained without observable conditioned avoidance responses (Ader &Cohen
1975; Ader et al 1987; Bovbjerget al 1987b; Gorczynski1987; Gorczynskiet
al 1984; Rogerset al 1976; Schulze et al 1988; Wayneret al 1978). Moreover,
reexposure to the CS before rather than after immunizationwith SRBCalso
depresses in vivo antibody production (Ader et al 1982; Schulze et al 1988;
Kusnecovet al 1988). The latter results suggest that an antigen-activated
immunesystem is not necessary for the conditioning of an immunosuppress-
ive response.
Conditioned immunomodulatory effects are not confined to the use of CY
or the taste aversion conditioning situation. Conditionedchanges in various
parameters of immunologicreactivity have been observed using other im-
munomodulating substances (e.g. Ader & Cohen 1981; King et al 1987;
Kusnecovet al 1983; see also Hiramotoet al 1987; Husbandet al 1987). It is
not even necessary to use immunopharmacologicagents as UCSs; evidence
for a conditioned suppression of antibody-mediated responses has been ob-
Annual Reviews
57
tained using electric shock as the UCS(Sato et al 1984; Zalcmanet al 1989,

1991b).
There have been a few studies in which there were no observable condi-
tioned effects (Krank & MacQueen1988; MacQueen& Siegel 1989). For
theoretical reasons (see Eikelboom& Stewart 1982), these experiments were
conducted with the expectation of observing "paradoxical" or compensatory
conditioned responses (responses opposite in direction to the unconditioned
response). In conditioned animals reexposed to a CS previously paired with
CY,the anti-SRBCantibody response was higher than the response of condi-
tioned animals that were not reexposed to the CS and the response of animals
that experienced unpaired CS-UCS presentations, but the responses of condi-
tioned animalsreexposedto the CSdid not differ fromthose of a saline-treated

control group. Suchresults maypermit one to infer the existence of a condi-
tioned enhancementof antibody production based on the failure to observe
immunosuppression, but, as the investigators acknowledge,no direct evidence
of compensatoryconditioning was obtained. There is no obvious explanation
for the difference betweenthese results and those in the rest of the literature on
conditioned immunologicchanges.
Evidencefor the existence of the compensatoryconditioning of host de-
fense reactions, however,can be derived fromstudies on the role of condition-
ing in the developmentof pharmacologicaltolerance to repeated injections of
polyinosinic-polycytidylic acid (poly I:C; Dycket al 1986, 1987). In keeping
with a conditioning analysis of the developmentof tolerance to someother
pharmacologicagents (Siegel 1983), tolerance to the enhancingeffects of poly
I:C on natural killer cell activity is abrogatedby unreinforcedexposuresto the
CS; preexposure to the CSinterferes with the developmentof tolerance; and
tolerance is attenuated whenpoly I:C is injected in the absence of environmen-
tal cues previouslypaired with injections of the drug.
CELL-MEDIATED IMMUNITY The immunosuppressiveeffects of CYcan also be

used to condition changesin cell-mediatedresponses. In the studies by Bovbjerg
et al (1982, 1984), (Lewis x BrownNorwegian)Flrats were injected with Lewis
strain spleen cells to induce a local graft vs host (GvH)response. The basic
experimental protocol used by Ader& Cohen(1975) was modified so that: (a)
there was a 7-weekinterval betweenconditioning of the FI hybrid hosts and
injecting themwith donor cells (at whichtime there were no detectable residual
effects of CY),and (b) experimental animals were reexposed to the CS in
context of reexposure to a minimallyeffective injection of CY.(The Lewis
BrownNorwegian)F1 rats were first conditioned by pairing saccharin consump-
tion and CYand weresubsequentlyinjected ("grafted") with splenic leukocytes
1
Whensplenic leukocytes from an inbred strain of Lewis rats are injected into the footpad of
hybrid (Lewis x BrownNorwegian)F~ rats, the grafted cells recognize the host as "foreign," and
local inflammatory reaction (the GvH response) ensues and can be measured by weighing the
popliteal lymphnode that drains the injection site.
Annual Reviews
58 ADER&COHEN
from Lewis strain donors. On the day of grafting and on the following two days
they were reexposed to the CS; on the day after grafting they were also given a
low dose injection of CY. While low-dose injections of CY on Days 0, 1, and
2 dramatically suppressed the local GvHresponse, a single injection on Day l
caused only a modest decrease in the response. However, a single low-dose
injection of CYplus reexposure to the CS previously paired with CYsignifi-
cantly suppressed the GvHresponse relative to control groups that received only
the single low-dose injection of CY. As expected, unreinforced exposures to the
CS during the 7-week interval between conditioning and induction of the GvH
response resulted in extinction of the conditioned immunosuppressiveresponse.
Experimental extinction, one hallmark of a conditioned response, has also been
reported for other conditioned immunologic effects (Dyck et al 1986;

Gorczynski et al 1982; see also Lysle et al 1988).
Cyclophosphamide can enhance as well as suppress immunologic reactiv-
ity. In the case of a delayed-type hypersensitivity (DTH)2 reaction to SRBC,
low-dose treatment with CY at the time of sensitization can enhance DTHin
response to a subsequent challenge with the same antigen (Turk & Parker
1982); CYtreatment of sensitized animals just before antigenic challenge
decreases the DTHresponse (Gill & Liew 1978; Rodinone et al 1983). Al-
though CY decreases the DTbI reaction to an initial antigenic challenge, the
response to subsequent challenges is enhanced (Bovbjerg et al 1986).
Bovbjerg et al (1987a) did not observe any conditioning effects when sensi-
tized animals, previously conditioned with CY, were reexposed to the CS
before their initial challenge. However, reexposing conditioned animals to the
CS before two subsequent antigenic challenges resulted in enhanced DTH
responses. The conditioned enhancement could be a consequence of a selec-
tive conditioned immunosuppressive effect on suppressor cells (Gill & Liew
1978; Mitsuoka et al 1979). Another drug, levamisole, has been purported to
selectively depress cytotoxic/suppressor T cells. Thus, the conditioned eleva-
tion of the T-helper:T-suppressor-subset ratio in animals reexposed to a CS
previously paired with levamisole (Husband et al 1987) could be the phenom-
enological expression of a conditioned immunosuppressive response. Such an
interpretation would be consonant with data suggesting that T cell-dependent
reactions are especially sensitive to conditioning.
As in the case of antibody-mediated immune responses, conditioned
changes involving cell-mediated immunity are not confined to the use of the
taste aversion conditioning model. Mice exposed to a novel environment plus
a distinctive taste in conjunction with rotation on a turntable (the UCS)show
decreased ability to reject allogeneic skin grafts when reexposed to the com-
pound CS at the time of and following the antigenic challenge (Gorczynski
1992). In addition, when conditioned females (mated with nonconditioned
2
DTHis an in vivo inflammatory’reaction mediatedby sensitized T cells that is evokedby
contactwith the antigen withwhichthe animalhadbeenimmunized.
Annual Reviews
59
males) were reexposed to the CS on Days 13, 16, and 19 of gestation, there
was a depression of humoral and cell-mediated immunity in their un-
manipulatedoffspring.
Effects of Conditioning on Nonimmunologically Specific

Reactions
In addition to studying antibody- and cell-mediated immunity,immunologists

also study the nonimmunologically specific in vitro actions of natural killer
cells and mitogens. These are amongthe most frequently used measures in
behavioral experiments. Natural killer (NK)cells are large granular lympho-
cytes that nonspecifically attack and destroy certain virus-infected cells and
tumorcells and maybe involved in preventing tumor metastasis.
T and/or B lymphocytescan be induced to proliferate in vitro by stimula-
tion with various lectins (chemical substances obtained from plants), and the
lymphoproliferative response to such mitogens has been used to indicate an
alteration of the physiological state of T or B cells in a particular lymphoid
compartment(e.g. spleen, lymphnodes, peripheral blood). Changesin mitt-
genie responsiveness, however,do not necessarily reflect an organism’sabil-
ity to respond to antigens in vivo or in an immunologically specific manner.
In a comprehensiveseries of experiments, Lysle and his colleagues (1988,
1990a,b, 1991, 1992b)characterized the conditioned suppression of a variety
of nonspecific responses in rats rcexposcd to cues previously paired with
stressful stimulation. Comparedto nonconditioned animals, to conditioned
animals exposedto novel environmentalcues, and to animals exposedto cues
explicitly unpaired with the UCS,animals reexposed to auditory (or visual)
cues paired with electric shock stimulation showeda reliable suppression of
lymphoproliferative responses to concanavalin A (Con A) and phytohemag-
glutinin (PHA)(two T cell mitogens), lipopolysaccharide (LPS) (a
mitogen), interleukin-2 (IL-2)3 production, and NKcell activity. As expected,
exposure to the CSbefore conditioning retarded developmentof the condi-
tioned response, and unreinforced exposures to the CSresulted in experimen-
tal extinction (Lysle et al 1988). In addition, the conditionedresponsesvaried
as a function of both the timing of reexposureto the CSin relation to the time
of conditioning and the immunecompartment from which the cells were
obtained (Lysle et al 1990a,b; Lysle &Maslonek1991). In splenic lympho-
cytes there was a depressed response to T and B cell mitogens; in wholeblood,
there was a suppression of the response to Con A and PHA,but not to LPS;
3
Macrophages and activated lymphocytesproducesoluble products (cytokines), such as interleu-
kin-1 (IL-1), IL-2, interferons, and tumornecrosis factor that are involved in the proliferation,
differentiation, and effector functions of lymphocytes. In addition, they serve as "im-
munotransmitters"in the sense that some,if not all of themhave effects within the central nervous
systemand stimulate the release of hormones.
Annual Reviews
60 ADER& COHEN
cells obtained from mesenteric lymphnodes showedno effects of CSreexpos-

ure.
Usingeither a taste or odor aversion conditioning paradigmor a stressor as
the UCS,other studies have also observed the conditioned suppression of
lymphoproliferative responses both to mitogenic stimulation. (Drugan et al
1986; Kusnecov et al 1988; Neveu et al 1986, 1987), NKcell activity
(Gorczynski et al 1984; Hiramoto et al 1987; Lysle & Maslonek 1991;
O’Reilly &Exon 1986), and total white blood cell count (Klosterhalfen
Klosterhalfen 1987). The conditioned enhancementof NKcell activity has
also been reported (Hiramotoet al 1987; Solvasonet al 1988, 1991). However,
someof these latter studies suffer from major design flaws, and others have
been unable to repeat these observations (Ader & Cohen1991):

In a recent study, Coussonset al (1992) paired exposure to a distinctive
environmental setting with injections of morphine (which unconditionally
suppresses several nonspecific immuneresponses). Reexposureto the envi-
ronmentalcues resulted in a conditioned suppression of splenic and peripheral
blood lymphocyteresponse to T and B cell mitogens, splenic NKcell activity,
and IL-2 production. Theseresults are of additional interest in relation to the
issue of compensatory conditioning discussed above. The conditioned re-
sponse mimickedthe unconditioned response rather than inducing an opposite
or compensatoryresponse; induction of such a compensatoryresponse, how-
ever, characterizes someof the other behavioral and physiological effects of
morphine(e.g. Siegel 1976, 1983).
As is the case with antibody-mediatedresponses, the abovestudies uncov-
ered no consistent relationship betweenconditioned behavioral responses and
conditioned changes in nonspecific defense reactions (Klosterhalfen
Klosterhalfen 1987; Kusnecovet al 1988; Neveuet al 1986, 1987; Solvason et
al 1988).
Antigen as Unconditioned Stimulus

In behavioral terms, an antigen is an unconditionedstimulus for activation of
the immunesystem and has been used as the UCSin a few studies. A condi-
tioned release of histamine, a nonspecific mediator of an allergic reaction,
occurred in response to a CSassociated with the injection of the antigen,
bovine serum albumin(Dark et al 1987; Peekeet al 1987; Russell et al 1984),
and an increase in mast cell protease lI was observed in sensitized rats
reexposed to environmentalcues previously paired with exposure to egg albu-
min (MacQueenet al 1989). With respect to an immuneresponse, per se,
Gorczynski and his colleagues (Gorczynski et al 1982) repeatedly grafted
mice with allogeneic skin under a constant set of environmentalconditions.
Whenthese mice were given shamtransplantations, there was an increase in
the numberof precursor cytotoxic T lymphocytes. Subsequent unreinforced
exposures to the graft procedures resulted in extinction of the conditioned
response. Theseexperiments are amongthe relatively few that have addressed
Annual Reviews
61
conditioned immune responses as distinct from conditioned im-

munopharmacologic
effects.
Biologic Impact of Conditioned Changes in Immunity

Althoughhighly reproducible, the effects of conditioning in modulatingim-
muneresponses have been "relatively" small, and a recurring question has
been whether behaviorally induced alterations in immunocompetence have
any biological or clinical significance. So far, only a few studies have ad-
dressed this issue. In a study using micethat developa systemiclupus-erythe-
matosus-like autoimmunedisease (Ader &Cohen1982), conditioned stimuli
were substituted for half of the weekly treatments with active im-
munosuppressivedrug (CY). The onset of autoimmunedisease in the geneti-
cally susceptible (NZB× NZW)F1 mice was thereby delayed using a cumula-
tive dose of CYthat wasnot, by itself, sufficient to alter the progressionof the
lupus-like disease. Also, in lupus-prone mice that had previously been given
weeklytreatments with CY(paired with the taste of saccharin), reexposure
the CSafter discontinuation of active drug treatment prolongedsurvival rela-
tive to conditionedmice that received neither active drug nor reexposureto the
CS (Ader 1985). Analogousresults were obtained in studies of adjuvant-in-
duced arthritis in rats using either CY or cyclosporin as the UCS
(Klosterhalfen &Klosterhalfen 1983, 1990) and electric shock as the UCS
(Lysle et al 1992a), and in a study in which reexposure to a CSpreviously
paired with CYaccelerated mortality amongconditioned animals inoculated
with a syngeneic plasmacytoma(Gorczynskiet al 1985).
Recently the therapeutic potential of conditioning was examinedin differ-
ent transplantation modelsin an effort to prolonggraft survival. A/J recipient
mice typically reject allogeneic skin grafts from BALI3/cor C57BL/6donors
within two weeks. A low-dose injection of CYon the day of grafting, how-
ever, prolongs graft survival. In mice conditioned by the pairing of saccharin
and CY, reexposure to the CS alone on the day of grafting and at 5-day
intervals thereafter also prolongedsurvival of the skin allograft (Gorczynski
1990). In another recent study (Grochowiczet al 1991), reexposureof trans-
plant recipient rats to a CSpreviously paired with cyclosporin A extendedthe
survival of a heterotopic heart transplant. Experimentssuch as these hint at the
potential clinical significance of conditionedalterations in immune function.
Conditioning in Human Subjects

"As early as 1557, AmatusLusitanus related the case of a Dominicanmonk,
who,wheneverhe perceived the odor of roses or saw them at a distance, was
immediately seized with syncope and fell unconscious to the ground" (Mac-
kenzie 1896:51, italics added). Mackenziedescribed his ability to provoke
asthmatic symptomsby presenting an artificial rose to an allergic patient,
which"forcibly illustrates the role of purely psychical impressionsin awaken-
ing the paroxysmsof the disease familiarly knownas ’rose cold’" (Mackenzie
Annual Reviews
62 ADER & COHEN
1896:45). The literature contains descriptions of several similar cases of what

mayrepresent conditioning phenomena(Hill 1930; Smith & Salinger 1933;
Dekker et al 1957). Laboratory studies in humans(Dekker et al 1957; Khan
1977) and animals (e.g. Ottenberget al 1958) confirmthe clinical suggestions
that exposure to symbolic, nonallergenic stimuli (CSs) previously associated
with allergens are capable of inducing asthmatic symptomsin somesubjects.
Morerecent data provide preliminary evidence that conditioning may be
able to modify immuneresponses in humansubjects. In a study by Ikemi &
Nakagawa (1962), four subjects received cutaneous stimulation with a methy-
lene blue solution (the CS) containing the extract of a Japanese lacquer tree
that unconditionally induced eczemawithin 24 hr. After an unspecified num-

ber of CS-UCS pairings, the CS alone elicited a skin reaction in all four
subjects. Smith& McDaniels(1983) also tried to condition a DTHresponse
humansubjects. Healthy volunteers underwent tuberculin skin testing six
times at monthly intervals. In a counterbalanced manner, the "blinded" re-
search nurse administered tuberculin obtained from a green vial to one arm
and saline drawnfroma red vial to the other arm. Onthe test trial, the contents
of the colored vials were switched; neither the experimenternor the subject
was aware that tuberculin had nowbeen put into the red vial, saline into the
green. Saline administered to the ann that had prcviously been treated with
tuberculin did not evokea skin reaction, but there was a significant diminution
of the erythema and induration elicited by tuberculin in the arm previously
injected with saline. This finding is remarkablysimilar to that reported by
Moynihanet al (1989) in mice.
The failure of saline to evokea skin reaction could be instructive, particu-
larly for the strategy underlying such research. First of all, the immunologic
mechanismsthat could dampenan immunologicallyspecific response are not
necessarily the sameas those that might elicit or enhancethat sameresponse.
In the case of immunoenhancement,an immunogenicstimulus may be re-
quired, evenif it alone is not sufficient to elicit a discernible reaction. Amore
reasonable paradigmfor the behavioral modification of an immuneresponse,
then, mayrequire the application of a minimallyeffective immunogenic stim-
ulus-one barely able to elicit a measurableresponse but sufficient to initiate
someearly event in the immunologiccascade that leads to a typical immune
reaction. If "subthreshold" stimulation is capable of being potentiated by an
alteration of the neural and/or endocrine environment in which immunere-
sponses occur, one might be more likely to observe a behaviorally induced
alteration in immunologicreactivity under immunogenicstimulus conditions
that actually approximatenatural conditions. Thus, in the present example,a
sensory stimulus (e.g. saline) that does not affect the immunesystemdirectly
maynot suffice to elicit a tuberculin-inducedskin reaction. But the superimpo-
sition of an immunologicallyneutral stimulus previously paired with activa-
tion of the immunesystem or the response induced by what wouldordinarily
be a subthreshold immunogenic stimulus maybe sufficient to elicit a discern-
Annual Reviews
63
ible immuneresponse. This basic strategy was used by Bovbjerget al (1982,

1984) whodemonstrated that the combination of a CSfor immunosuppression
and a low dose of CYdepressed a GvHresponse to a significantly greater
degree than the low dose of CYwas able to accomplishalone.
The anticipatory nausea that occurs in 25-75%of patients undergoing
repeated chemotherapyfor cancer appears to reflect a classically conditioned
response (Andrykowskiet al 1985, 1988; Andrykowski& Redd 1987; Carey
& Burish 1988; Morrow & Dobkin 1988; Redd & Andrykowski 1982).
Bovbjerg et al (1990) studied womenwho had experienced at least three
sessions of chemotherapyto determineif cancer patients whodisplayed antic-
ipatory nausea and vomiting during the course of chemotherapywouldalso

showanticipatory immunosuppressivechanges. Peripheral blood was obtained
at the patients’ homes3-8 days before a scheduled chemotherapysession and
again in the hospital just before the intravenous drug infnsion. Nodifferences
in NKcell activity or in cell counts or cell subset numberswere observed, but
in vitro proliferation in response to the T-cell mitogensPHAand ConA were
significantly lower prior to chemotherapythan the responses measuredin the
homeenvironment. There was no evidence that the decreased mitogen re-
sponses were related to anticipatory nausea or to the increased anxiety that
occurred in the hospital. This observation of anticipatory immunosuppression
is entirely consistent with the proposition that chemotherapypatients who
receive an immunosuppressive drug under a relatively constant set of environ-
mental conditions would show conditioned changes in immunefunction as
well as in behavior. The extent to which these behavioral and immunere-
sponses reflect the same underlying mechanisms--or,as the animal literature
suggests, represent independently conditioned responses--remains to be de-
termined and might prove important in the clinical management of chemother-
apy patients.
STRESS AND IMMUNITY
In the present context, "stress" refers to any natural or experimentallycon-

trived circumstancesthat (intuitively, at least) pose an actual or perceived
threat to the psychobiological integrity of the individual. In nonhuman ani-
mals, environmentalconditions that are apparently perceivedas a threat to the
organism, and to which the organism cannot adapt, are accompaniedby both
transient and relatively long-lasting psychophysiologicchanges. Wepresume
that these changescontribute to the developmentof disease, especially if the
organismis at the same time exposed to potentially pathogenic stimuli. As
described below, however,the stress response is not uniformly detrimental to
the organism.
Annual Reviews
64 ADER& COHEN
Effects of Stress on Disease

Studies in humanshaveimplicated psychosocialfactors in the susceptibility to
and/or the progression of a variety of pathophysiologic processes including
bacterial, allergic, and autoimmunediseases that involve alterations in im-
munologicdefense mechanisms.Specific examplesinclude Epstein-Barr virus
infections, respiratory infections, streptococcal infections, asthma,and arthri-
tis. Comprehensivetreatments of this subject have been provided recently by
S. Cohen& Williamson (1991) and Weiner (1977, 1991). Most of the experi-
mental work has been conducted with animals and indicates that a variety of
behavioral manipulationsor stressors can influence susceptibility to a variety

of disease states in a variety of species. However,the stressors used do not all

producethe sameeffects. The impact and direction of the effects of stressful
stimulation dependuponthe disease process to which the organismis concur-
rently subjected. For example,physical restraint in rodents increases suscepti-
bility to infection with herpes simplex virus (Bonneauet al 1991a,b; Rasmus-
sen et al 1957 ) and the Maloneysarcomavirus (Seifter et al 1973), has
effect on the response to an experimentally induced lymphorna(Greenberg et
al 1984), and decreases susceptibility to allergic encephalomyelitis(Levine
al 1962). Likewise, electric shock stimulation increases susceptibility to
Coxsackie B virus but decreases susceptibility to malaria (Friedman et al
1965) and to the spontaneous developmentof leukemia in AKRmice (Plaut
al 1981). Using the same disease outcome(e.g. encephalomyocarditisvirus),
electric shockstimulation decreases susceptibility but the stimulation of han-
dling has no effect (Friedmanet al 1969); and, while both electric shock and
handling decrease susceptibility to collagen-induced arthritis, a different
stressor, auditory stimulation, increases susceptibility (Rogerset al 1980a,b).
Analogousresults are observed in studies on the developmentand progres-
sion of tumors in animals (Justice 1985; Sklar &Anisman1981). Psychosocial
factors can influence the developmentand/or growth of tumors, but the direc-
tion of the effects dependson the tumormodelchosen for study. The results
also dependon the quality, quantity, and duration of the stressor; the temporal
relationship betweenexposure to the stressor and the introduction of patho-
genic stimulation; socioenvironmental conditions (e.g. whether animals are
housedindividually or in groups); and a variety of host factors (e.g. species,
strain, and sex). Further, the observedeffects dependon the outcomemeasures
and samplingparametersselected for study. As described below, the effects of
stressful stimulation on immunefunction parallel those on stressful stimula-
tion and disease susceptibility. It seemsevident--or, at least likely--that the
ability to predict the outcomeof studies of the pathophysiologic effects of
stressful stimulation dependsupona moredetailed understandingof the inter-
action betweenresponses unconditionally elicited by specific kinds of poten-
tially pathogenicstimulation and the pattern of psychophysiologicalresponses
Annual Reviews
65
(including immunologicresponses) induced by specific forms of stressful

stimulation.
Whileillustrations like those abovedocument the effects of stressors on the
response to pathogenicagents, they do not provide evidence that these effects
are mediated by behaviorally induced alterations in immunefunction. Too
frequently, however,studies of the effects of stressors on tumordevelopment
or metastasis are cited as evidence of the influence of stress or behavioral
interventions on immunefunction--even in the absence of any measure of
immunefunction. Andeven when simultaneous measures of immunefunction
are being measured,it is not alwaysclear that the indexes chosenfor study are
related to the. pathophysiologicprocess under study. It is not yet clear that

behaviorally induced perturbations in immunefunction can influence or medi-

ate the effects of psychosocial factors on the developmentor progression of
disease in humans,but provocative data are nowbeing obtained (e.g. Glaser et
al 1987; Kemeny et al 1992), especially in relation to the experimentalinocu-
lation of respiratory viruses .(Broadbent et al 1984; S. Cohenet al 1991;
Totmanet al 1980).
Morerecent animal studies, however,are including biologically relevant
measures of immunefunction. Feng et al (1991), for example, infected mice
with influenza virus and observeda delay in the production of virus-specific
antibody levels in animals subjected to restraint. There were no detectable
differences in the magnitudeof the humoral response ultimately attained,
however--afinding that could be related to an interaction betweenthe tempo-
ral relationship of restraint and infection and the virulence of the pathogen
(Chaoet al 1990). At a clinical level, psychosocialfactors havebeen related
the manifestation of latent herpes virus infection (e.g. Glaser et al 1987).
Bonneauet al (1991a), studying the effects of a stressor on the murine re-
sponse to herpes simplex virus (HSV),found that repeated, prolonged periods
of physical restraint (16 hr/day for a varying numberof days around the time
of inoculation) suppressed NKcell activity and the primary developmentof
HSV-specificcytotoxic T lymphocytes. Also, higher titers of infectious HSV
at the site of infection were recovered from restrained than from unrestrained
mice. Similar results were obtained by Kusnecovet al (1992) using overnight
exposureto intermittent electric shock stimulation (1 shock every 30 rain, on
average) on the day preceding and for eight days following virus infection. In
another study (Bonneauet al 1991b),restraint inhibited the in vitro activation
and/or migration of HSV-specific cytotoxic T lymphocytes from previously
primed mice--i.e, activation of HSV-specificmemorycells was inhibited by
the physiological changes induced by the stressor. Other examplesrelate to
neoplastic disease (Ben-Eliyahuet al 1991; Brenneret al 1990). Usinga tumor
model in which lung metastasis is thought to be controlled by NKcells,
Ben-Eliyahu and his colleagues (1991) found that, depending upon whenthe
stressor was imposedin relation to the effect of NKcells on the metastatic
Annual Reviews
66 ADER & COHEN
process, forced swimmingdecreased NKcell activity and resulted in a two-

fold increase in lung metastases.
Effects of Stress on Humoral and Cell-Mediated Immunity

Unlike the measurementof hormonelevels or nonspecific defense reactions
that can be gaugedby samplingfactors circulating in blood, urine, saliva, etc,
the assessment of immuneresponses, such as most of those described abovein
the context of disease susceptibility, requires that experimental subjects be
immunized.This additional variable adds layers of complexityto the charac-
terization of the effects of behavioral and other interventions on immune
function and the mediation of these effects. In addition to the quality and
quantity of environmentalstimulation and the temporal relationship betweena
stressor and immunization,one must, for example, consider the nature as well
as the concentration of antigen and the "cascade" of immuneevents that
eventuate in the production of antibody or a cell-mediated inflammatoryre-
sponse. This complexityis reflected in muchof the data describing the effects
of stressful stimulation on different aspects of immunefunction.
The relevance of the concentration of antigen used for immunization in
studies of stress has beennoted in the past (Solomonet al 1974). In our view,
studies that address immunefunction per se use what might be described as
suprathreshold levels of antigenic stimulation---concentrations of antigen cal-
culated to induce a "robust" primary and secondary response. To the extent
that behaviorally induced neuroendocrine changes are capable of modulating
immuneresponses, the use of high levels of antigenic stimulation are not
comparable to the levels of antigen exposure experienced by the organism
under natural conditions and could be counterproductiveas a research strategy
designed to elaborate behaviorally induced alterations in immunocompetence.
For these reasons, Moynihanet al (1990a) studied the effects of a stressor
(electric shock) on the response to low-dose antigenic stimulation. An im-
munizing dose of as little as 1 btg of the protein antigen keyhole limpet
hemacyanin(KLH)was insufficient to elicit a detectable primary response but
was sufficient to prime mice. Whenshock was administered for seven days
before and after immunizationor only on one day, 24 hr after priming, mice
exposed to the stressor showeda depressed antibody response to challenge
with a second injection of 1 ~tg of KLHrelative to control mice that were
simply placed into the shock apparatus or that remained unmanipulated.When
mice were challenged with a higher dose of KLH(5 lag) there were no group
differences. It is thus possible that concentrationsof antigen that uncondition-
ally elicit robust immuneresponses could maskthe effects of behavioral
interventions.
For the most part, antibody production is suppressedby stressful stimula-
tion. Althoughdifferential housing (frequently mislabeled as "crowding"or
"isolation") maynot qualify as a stressor, the mannerin which animals are
housed or a change in housing conditions is sufficient to influence primary
Annual Reviews
67
and/or secondary antibody responses to several different antigens (e.g.

Cunnick et al 1991; Edwards & Dean 1977; Edwards et al 1980; Glenn &
Becker 1969; Rabin et al 1987a,b; Rabin & Salvin 1987; Solomon 1969;
Vessey1964). Antibodyresponses are also suppressed whensubmissive, in-
truder, and attacked and defeated rats are subsequently immunized(13eden
Brain 1982, 1984; Fauman1987; Fleshner et al 1989; Ito et al 1983). A recent
review of these data has been provided by Bohus& Koolhaas(1991). Eventhe
odors emittedby micesubjectedto stressful stimulation are sufficient to influ-
ence antibody production in consPeeific recipients. In one study (Zalcmanet
al 1991 a,b), the antibody response to SRBCwas suppressed in mice placed
into an apparatus vacated by conspecifics that had been exposedto a presumed
stressor; in another study (Cockeet al 1992), immunizationwith KLHfollow-

ing prolonged exposure to alarm signals emitted by conspecifics undergoing
concurrent stimulation enhancedantibody responses in the group-housedre-
cipients.
Antibodyresponses are influenced by early separation experiences. In rats,
brief handling and separation from the mother increased both primary and
secondaryantibody responses to flagellin, a bacterial antigen (Solomonet al
1968), In mice, early separations from the mother resulted in a decreased
anti-SRBCantibody response (Michaut et al 1981), and separation from the
mother and rearing environment influenced the antibody response of monkeys
to an antigenic challenge (e.g. Coeet al 1985, 1988).
Okimura and his colleagues (Okimura & Nigo 1986; Okimura et al
1986b,c) found that serum antibody titers and the numberof antibody-forming
splenic cells of mice were depressed if physical restraint (12 hr/day on two
consecutive days) was imposed after, but not before, immunization with
SRBC,a T-cell dependent antigen. 13oth adrenalectomyand chemical sympa-
thectomyblocked the stressor-induced immunosuppression.The response to a
hapten (TNP)conjugated to a type 2 T cell-independent carrier (Ficoll)
not influenced by restraint, but there was enhancedantibody production in
responseto a type 1 T cell-independent antigen (LPS). In this case, adrenalec-
tomy, but not sympathectomy,blocked the enhancedresponse. The differen-
tial sensitivity of type 1 and type 2 T cell-independentantigens, whichactivate
different populations of 13 cells, is of additional interest in relation to thc
differential effects of conditioning on the responseto these antigenic stimuli
(see N. Cohenet al 1979; Schulzeet al 1988; Wayneret al 1978).
Both primary and secondary antibody responses are suppressed when rats
are subjected to inescapable electric shockstimulation daily beginningbefore
and continuing after immunizationwith KLH(Laudenslager et al 1988). Ro-
tating mice after but not before immunizationwith SRBC also depressed the
immuneresponse to SRBC(Esterling & Rabin 1987). In contrast, avoidance
conditioning depressed the antibody response to SRBCrelative to yoked ani-
mals that experienced unavoidable shock and an apparatus control group
(Mormede et al 1988). The degree to which rats could avoid or escape shock
Annual Reviews
68 ADER & COHEN
influenced the immuneresponse, but not in the hypothesizeddirection. Zalc-

manand his colleagues (1988) subjected mice to a single session of footshock
0, 24, 48, 72, or 95 hr after immunization with SRBC.Both the numberof
antibody-forming cells and serum anti-SRBCantibody titers were depressed,
but only in micesubjected to the stressor 72 hr after immunization,suggesting
that there is a critical period for this acute stressor, at least, on antibody
production. Based on changes in norepinephrine activity within the central
nervous systemthat result from this sameschedule of electric shock treatment
(Anismanet al 1987; Zacharko& Anisman1988), Zakmanet al expected that
"controllability" would have influenced the immuneresponse. There were,
however, no differences between mice that were and were not able to escape
the electric shock.

A seemingly innocuous form of stimulation with the immediate and long-
teim ability to influence antibody production is the handling of animals. Moy-
nihan et al (1990b) found that antibody production in response to KLHwas
attenuated in adult mice that were immunizedafter a 2-weekperiod of daily
handling. Raymond et al (1986) found that, dependingon strain, rats that were
simply handled during the neonatal period had suppressed antibody titers in
response to subsequent immunization with SRBC.In response to flagellin
administration, adult rats handled during infancy showedan enhancedanti-
body response. In related research, Taylor &Ross (1989) found that neonatal
restraint (1 hr/day during the first 10 days of life) suppressed antibody re-
sponses to immunizationwith pneumococcalpolysaccharide in adult rats. In
adult animals, the same restraint imposedfor 3 weeksfollowed by a 3-week
interval before immunizationhad no discernible effects. 4 Moynihanand her
associates (Moynihanet al 1989) subjected mice to a different number
intraperitoneal (ip) injections of saline or just handling during the 2-week
period before the animals were injected with antigen. Daily handling as well
as the repeated ip injections of saline attenuated the productionof antibody in
response to KLHrelative to a group of unmanipulated mice. Whether there
was an attenuated response in the manipulated animals or an exaggerated
response in thc unmanipulated animals cannot be determined. Whether the
difference is a reflection of conditioning processes also remains to be deter-
mined. In either case, methodological questions arise about the control or
"resting" state of immunizedanimals.
With respect to cell-mediated immuneresponses, animals subjected to
stressors tend to showdelayed skin graft rejection (Wistar &Hildemann1960)
and suppressed DTHresponses (Okimuraet al 1986a). In a series of experi-
mentsby Blechaand his colleagues (1982a,b), restraint, heat, and cold, three
commonly used stressors, had different effects on two different cell-mediated
4
Thelimiteddata availablesuggestthat an exploration
of the long-term
effectsof earlylife
experiences
on immune flmctionwouldbe a fruitful avenueof research(Ader1983;O’Grady
Hall1991).
Annual Reviews
69
immuneresponses. Restraint imposedbefore sensitization or challenge sup-

pressed the delayed cutaneous hypersensitivity response to SRBC.Exposure
to cold for two days after sensitization also depressed the DTHresponse,
whereas exposure to cold for eight days increased the DTHresponse. Heat
exposure invariably increased the hypersensitivity response. Contact hyper-
sensitivity to dinitrofluorobenzenewas increased by each of these stressors.
Further, adrenalectomyor treatment with metapyroneabrogated the stressor-
induced suppression of DTHbut had no effect on the strcssor-induced en-
hancementof the contact sensitivity response. Again, the literature provides
little justification for generalizing fromone stressor to another---or fromone
aspect of immunefunction to another.
Anotherimmunologicallyrelevant reaction to stressors is a suppression of

macrophagefunction (Pavlidis & Chirigos 1980; Jiang et al 1990; Okimuraet

al 1986a; Qunidoset al 1986; Zwilling et al 1990), which maybe mediated by
a stressor-induced depression of IFN production (Glascr et al 1986;
Sonnenfeld et al 1992). The expression of class II major histocompatibility
complex antigens by macrophages(glycoproteins that, together with pro-
cessed antigen, are presented to and recognized by lymphocytes) was de-
creased by physical restraint (Zwilling et al 1990) and cold water exposure
(Jiang et al 1990); in the study by Sonnenfeld et al (1992), electric shock
suppressed IFN-gammaproduction and the expression of class II antigens
which could, in turn, influence antibody- and cell-mediated immunocompet-
ence.
Effects of Stress on Nonimmunologically Specific Reactions

As in the case of antibody- and cell-mediated immunity, the effects of
stressors on nonspecific defense reactions are generally suppressive, but not
uniformly so (Croiset et al 1987; Lysle et al 1990a,b; Monjan&Collector
1977; Rinneret al 1992;Weisset al 1989b).
The differential housing of rodents results in differences in
lymphoproliferativeresponses to mitogens, NKcell activity, and IL-2 produc-
tion (Ghoneumet al 1987; Jessop et al 1988; Rabin et al 1987a,b; Rabin
Salvin 1987), the direction and magnitudeof the effects being determinedby
species, strain, sex, and duration of housing. Otheragonistic social interac-
tions can also reliably influence the response to T and B cell mitogens, NK
cell activity, and IL-2 productionin fish as well as rodents (Bohus&Koolhaas
1991; Faisal et al 1989; Ghoneum et al 1988; Hardy et al 1990; Raab et al
1986). In addition to changesin antibody production, Cockeet al (1992) also
found reduced splenic NKcell cytotoxicity in mice exposed to the odors
emitted by footshockedconspecifics.
In general, the in vitro responseof laboratory rodents to T cell mitogensin
splenic and/or peripheral blood lymphocytesdecreases following relatively
acute exposureto a variety of stressors such as restraint, noise, and swimming
(Jiang et al 1990; Monjan& Collector 1977; Rinner et al 1992; Zha et al
Annual Reviews
70 ADER & COHEN
1992), and as a result of separation experiences (Ackermanet al 1988).

effects of separation experiences from mothers or peers in several nonhuman
primate species can have long-lasting effects (Coe et al 1989; Friedmanet al
1991; Gust et al 1992; Laudenslageret al 1985, 1990). Increases and decreases
in lymphocyteproliferation have also been observed in handled animals, de-
pending on the amount of handling and the age at which it is experienced
(Lown& Dutka 1987; Moynihanet al 1990b; Rinner et al 1992). Similarly,
NKcell activity is frequently, but not invariably, depressed following acute
exposure to a variety of stressors (Aarstad et al 1983; Jiang et al 1990;
Moynihan et al 1990a; Okimura et al 1986a; Steplewski & Vogel 1986;
Steplewskiet al 1985).
Electric shock stimulation has been the most frequently used stressor in
studies of nonspecifically stimulated defense reactions and thus provides the
more systematic data. Keller and his associates (1981) found that, with in-
creasing intensities of electric shock over a period of 18 hr, there was a
correspondingdecrease in the proliferation of peripheral blood lymphocytesin
response to PHA, even when the assay was based on a constant number of
cells to correct for the stressor-induced decrease in the total numberof circu-
lating lymphocytes.Less reliable effects were found in the responseof splenic
lymphocytes, although Weiss et al (1989b), using the same paradigm, found
significant suppression of lymphoproliferation of both splcnic and blood lym-
phocytes. Underthese conditions, they also found a decline in IL-2 and inter-
feron (IFN) production. Batumanet al (1990) obtained essentially the
results varying the numberof days of electric shock stimulation and measuring
the response to both Con A and PHA.The proliferative response to B cell
mitogens (pokeweedmitogen and LPS) was not influenced by the stressful
stimulation. These investigators also noted a decrease in the total numberof
splenic and peripheral blood lymphocytes,particularly of CD8+cells. In addi-
tion, IL-2 production was diminishedafter as little as one day of exposureto
the 3-hr period of stimulation.
Lysle et al (1987) also found that a single session of approximately1 hr
stressful stimulation during whichrats received 8 or 16 (but not 4) signaled
footshocks was sufficient to depress the proliferative response of splenic and
peripheral blood lymphocytesto Con A stimulation. After five days of stimu-
lation, the responseof spleen but not bloodcells had recovered.Similar effects
were obtained by Cunnicket al (1988). Thus, whether one varies the intensity
of electric shock, the numberof sessions of stimulation, or the numberof
shocks per session, the magnitudeof this form of stressful stimulation is
related to the magnitudeof the changes in at least somemeasuresof cellular
host defenses in some immunecompartments. However,results obtained by
Livnat et al (1985) and Lysle et al (1990b)with electric shockand by Weiss
al (1989b)with qualitative and quantitative differences in stressful stimulation
suggestthat this relationship maynot be linear.
Annual Reviews
PSYCHONEUROIMMUNOLOGY 71
The ability to predict or control the effects of stressful stimulation has

becomea major focus of stress-related research and, as a result, the effects
obtained by Laudenslager et al (1983) indicating that inescapable but not
escapable footshock depressed mitogen responsiveness in rats has been cited
frequently. Less frequently cited, however, is the report by these investigators
(Maier & Laudenslager 1988) of their inability to reproduce these effects.
contrast to the effects on antibody production, however, Mormedeand his
colleagues (1988) found that "control" did influence the response to mitogens.
Rats subjected to inescapable or unsignaled electric shock stimulation showed
a suppression of the lymphoproliferative response to T-cell mitogens. Also,
preliminary data obtained by Shavit et al (1983) indicated that inescapable but
not escapable shock suppressed splenic NKcell activity in rats. In another

preliminary study (Irwin & Custeau 1989), however, signaled electric shock
resulted in a more pronounced suppression of NKcell activity than unsignaled
5shock.
Electric shock stimulation also suppresses NKcell activity (Cunnick et al
1988; Keller et al 1988; Lysle et al 1990a; Lysle & Maslonek 1991; Shavit et
al 1984; Weiss et al 1989b), the kinetics of which are strain dependent in mice
(Zalcman et al 1991a). Shavit and his colleagues (1984, 1986) found
intermittent electric shock stimulation, which induces an opioid-mediated an-
algesia, also depresses NKcytotoxicity, whereas continuous shock, which
results in a nonopioid analgesia, has no such effect.
MEDIATION OF BEHAVIORALLY INDUCED IMMUNE

CHANGES
There are no definitive data on the mediation of behaviorally conditioned or
stressor-induced changes in either specific or nonspecific defense system re-
sponses. It is clear that different immunologic and neuroendocrine mecha-
nisms are involved in mediating the effects of different behavioral interven-
tions on different immune responses measured in different compartments of
the immune system.
The available data suggest that the effects of conditioning could be medi-
ated by a preferential effect on T cells. Conditioned suppression of
lymphoproliferative responses in rats and mice, for example, has been ob-
served in response to T-cell mitogens but not (or less reliably) in response
B-cell mitogens (Kusnecov et al 1988; Lysle et al 1990b, 1991; Neveu et al
1986). Adoptive transfer experiments also suggest that conditioning may be
mediated by T cell changes (Gorczynski 1987). Splenocytes from conditioned
or experimentally naive animals were transferred into (irradiated) naive
conditioned animals that were or were not subsequently reexposed to the CS.
5
Studies of stressor control in human subjects (Sieber et al 1991; Weisse et al 1990) yield
imilar contrasting effects.
Annual Reviews
72 ADER & COHEN
The observedincreases or decreases in the antibody-formingcell response to

SRBCdepended on the donor cells and the conditioning treatment experi-
enced by the recipient. The separate transfer of enriched T and B cells into
naive or conditioned mice suggested that conditioning effects were attribut-
able to the adoptively transferred T cells (Gorczynski1991). The question
whetherconditioning can modulatethe antibody response to different types of
T-cell independent antigens has not been resolved (N. Cohenet al 1979;
Schulze et al 1988; Wayneret al 1978).
Stressful stimulation also seemsto have greater effects on T cell than on B
cell function (e.g. Batumanet al 1990; Lysle et al 1990b;Mormede et al 1988;
Moynihanet al 1990b). Several investigators have demonstrated conditioned
and stressor-induced decreases in IL-2 production that might account for the
suppression of T cell proliferation (e.g. Batumanet al 1990; Ghoneum et al
1988; Hardy et al 1990; Kandil & Borysenko 1988; Weiss et al 1989b).
However, adding IL-2 to incubated lymphocytes did not normalize the
stressor-induced suppression of lymphoproliferation (Weiss et al 1989a).
These investigators did find that electric shock decreased the expression of
IL-2 receptors on lymphocytes, suggesting that a compromisedability to re-
spond to IL-2 may contribute to stressor-induced immunosuppression(and,
via neural and/or endocrine pathways, conditioned antibody- and/or cell-medi-
ated immuneresponses, as well).
The in vitro T cell effects of behavioralinterventions appear to be differen-
tially expressed by splenic and blood lymphocytesin the case of both condi-
tioning (Lysle et al 1990a; Lysle &Maslonek1991) and stressful stimulation
(Cunnicket al 1988; Keller et al 1983, 1988; Lysle et al 1987; Rinner et al
1992). This could reflect the trafficking of lymphocytesand a resulting differ-
ence in the kinetics of the response in these two compartments(Cunnicket al
1988; Lysle et al 1987); it could also relate to differences in the mannerin
which splenic and peripheral blood lymphocyteswere cultured. However,the
fact that there is noradrenergic innervation of the spleen (Felten et al 1987)
and that the stressor-induced suppression of splenic but not blood lymphocyte
proliferation is blocked by [~-adrenergic antagonists (Cunnick et al 1990)
suggests the operation of different mediating mechanisms.
Withrespect to neuroendocrineinfluences, it is reasonable to hypothesize
that conditioned alterations of immunologicreactivity could be mediated by
conditioned ncuroendocrinechanges, but data collected thus far are inconsis-
tent with the proposition that such effects are mediatedsimplyby nonspecific,
stressor-induced changes in hormonelevels. As reviewed elsewhere (Ader
Cohen1991), an extensive literature and other recent studies (e.g. Roudebush
&Bryant 1991) directly contradict or are inconsistent with predictions that
wouldfollow from a "stress mediation" hypothesisof the effects of condition-
ing or that conditioning effects are adrenal dependent. An elevation in
glucocorticoids is a common(and sometimes defining) characteristic of
stress response, and exogenouslyadministered glucocorticoids are generally
Annual Reviews
73
(but not uniformly) immunosuppressive(Munck& Guyre 1991). Nonetheless,

an endogenouselevation in adrenocortical steroids in response to different
stressors cannot be invoked to account for most of the stressor-induced
changes in immunocompetence.
In terms of antigen-specific immuneresponses, Okimuraet al (1986b)
found that adrenalectomy blocked the suppression of antibody responses to
SRBC in rats that had been subjected to restraint, but adrenalectomydid not
influence the immunosuppressive effects of rotation (Esterling &Rabin1987).
In studies using electric shock (Laudenslageret al 1988; Mormede et al 1988),
no relationship betweencorticosterone levels and suppressedantibody produc-
tion in rats was seen. Blechaet al (1982b)found that adrenalectomyattenuated
the suppression of a DTHresponse in restrained mice but did not affect the
restraint-induced enhancementof contact hypersensitivity. With respect to
nonspecific reactions, electric shock in rats induced a lymphopeniathat was
adrenal dependent (Keller et al 1981); the stressor-induced suppression
mitogen responsiveness, however,was unaffected by adrenalectomy(Keller et
al 1983). The suppressed lymphoproliferative response of splenic T lympho-
cytes, in particular, has beenshownconsistently to be independentof stressor-
induced adrenal activity (e.g. Lysle et al 1990b; Monjan&Collector 1977;
Mormede et al 1988; Rabinet al 1987a, b; Weisset al 1989b). Thestressor-in-
duced suppression of NKcell activity and IL-2 production are not blocked by
adrenalectomy(e.g. Weisset al 1989b).
Both conditioned and stressor-induced alterations of immuneand nonspe-
cific defense responseshave been attributed to the actions of catecholamines.
Gorczynski& Holmes(1989), for example, reported that both chlorpromazine
and amitriptyline abolished conditioned immunosuppressiveresponses based
on the pairing of a gustatory stimulus with CY.Lysle et al (1992a) reported
that the [3-adrenergicantagonist propranololblockedthe effects of conditioned
stressor effects on the developmentof adjuvant-inducedarthritis; but propran-
olol itself maybe capable of attenuating inflammatoryresponses (Kaplanet al
1980). Cunnickand her colleagues (1990) found that propranolo! and nadolol
(a ~-adrenergicantagonist that does not cross the blood-brain barrier) blocked
the electric shock-induced suppression of splenic but not peripheral blood
lymphocytesto ConA stimulation. Propranolol also blocked attenuation of the
lymphoproliferative response of splenic lymphocytes to a CS previously
paired with electric shock (Luecken&Lysle 1992). Neither propranolol nor
nadolol influenced the conditioned suppression of B cell mitogenesis of pe-
ripheral blood lymphocyteresponses or IL-2 production, but both antagonists
attenuated the conditioned suppression of IFN-gamma. The peripherally act-
ing [~-adrenergic receptor antagonist also blocked the suppression of IFN-
gammaproduction induced by electric shock stimulation (Sonnenfeld et al
1992).
Using a different stressor (repeated exposure to cold), Carr et al (1992)
described an enhancementof splenic antigen-specific antibody production
Annual Reviews
74 ADER & COHEN
and, at the same time, suppressed or unaltered serum levels of im-

munoglobulins. In this instance, phentolamine blocked the effects of cold
exposureand propranololpotentiated the effects of the stressor. Whilelittle is
knownof the immunologiceffects of catecholaminesin vivo, it is relevant to
note that Zalcmanet al (1991 a) wereunable to detect any meaningfulrelation-
ship betweenthe depression of brain levels of norepinephrineand the suppres-
sion of NKcell activity induced by one hour of uncontrollable electric shock
stimulation. It is evident that any involvementof catecholaminesin the media-
tion of stressor-induced alterations of immunityis determinedby interactions
involving the nature of the immuneresponse and the compartmentin which it
is measured and the neurochemical concomitants of the adaptive responses
inducedby the nature of the stressor.

Opioids acting within the central nervous system have also been implicated
in the mediation of conditioned and stressor (electric shock)-induced alter-
ations of nonspecific defense responses. Cunnickand her colleagues (1988)
found that naltrexone, an opioid receptor antagonist, blocked a shock-induced
suppression of splenic NKcell activity but did not obviate suppression of the
splenic lymphocyteresponse to T cell mitogens. Lysle et al (1992b) found that
naloxone (in doses higher than those used by Cunnick et al) blocked the
conditioned suppression of NKcell activity and the depressed response to both
T and B cell mitogens. Althoughthe association betweenan olfactory CS and
poly I:C was unaffected, subsequent enhancementof NKcell activity in re-
sponse to reexposure to the CSalso could be blocked by naltrexone (Solvason
et al 1989). N-methylnaltrexone,a quaternary form of naltrexone that does not
cross the blood-brain barrier, did not interfere with these conditioned re-
sponses.
That the effects of stressful stimulation on NKcell activity maybe medi-
ated by centrally acting opioids is suggested by the observations that the
immunosuppressioninduced by electric shock was prevented by pretreatment
with naltrexone but not by the peripheral actions of N-methylnaltrexone,and
that a dose-dependent suppression of NKcell activity could be induced by
morphine(Shavit et al 1984). However,Shavit et al (1986) found that NK
activity was suppressed to the same extent following 4, 14, or 30 daily ses-
sions of electric shock stimulation, whereastolerance to the immunosuppress-
ive effects of morphinedevelopedafter 14 sessions. Also, a stressor-induced
suppression of NKcell activity was observed in both morphine-naive and
morphine-tolerant rats. Thus, the suppression of NKcell activity in response
to morphineand to stressful stimulation is not mediatedby precisely the same
mechanisms.Of related interest is the report that serumobtained from mice or
rats subjected to physical restraint (Zha et al 1992) and electric shock (Weiss
et al 1989a) contained one or more factors that could suppress the
lymphoproliferative response of peripheral blood lymphocytes from un-
manipulated animals. Neither adrenalectomynor naltrexone treatment altered
this effect (Zhaet al 1992).
Annual Reviews
75
It mayalso be relevant that somestudies have uncovereddifferent changes

in mitogenesis and/or NKcell activity in animals subjected to avoidable or
unavoidable, escapable or inescapable, or signaled or unsignaled electric
shock (Irwin &Custeau 1989; Mormede et al 1988; Shavit et al 1983), despite
the fact that each of these stressful situations wouldfit the criterion for an
opioid-mediatedanalgesia that mightbe expected to suppress NKcell activity.
Otherstudies (e.g. Cunnicket al 1988; Keller et al 1988;Odioet al 1987)have
noted a dissociation betweenthe effects of the samestressor, electric shock
(albeit of different magnitudes,intensity, and duration), on NKcell and mito-
gen responses, both of which have been observed to be blocked by naltrexone
treatment in one study (Lysle et al 1992b). There is abundant evidence that
opioids, for which there are receptors on lymphocytes(Blalock 1988), can

influence immunefunctions. It is not likely, however,that centrally acting
opioids are the sole mediator of either conditioned or stressor-induced alter-
ations of immunocompetence.
The fact that someagonistic or antagonistic neurochemicalor pharmaco-
logic intervention can block or mimicthe immunologiceffects of reexposure
to a CSor stressful stimulationdoes not, in itself, constitute an explanationof
the behaviorally induced effects. It suggests only that such a pathwayis one
(of perhaps several) meansby whichsuch an effect could occur. Indeed, the
complexityof the multiple communicationchannels betweenthe brain and the
immunesystem provides any numberof other pathways through which behav-
iorally induced alterations in immunitycould be mediated. It is beyondthe
scope of this review, however,to detail these interactions (see Ader et al
199la); nor, because of the paucity of behavioral data, should we speculate on
their role in the psychophysiologicregulation of immunity.Theseinteractions
can, however, be summarizedas follows.
Primary (thymus, bone marrow)and secondary (spleen, lymphnodes, gut-
associated lymphoid tissues) lymphoidorgans are innervated by the sympa-
thetic nervous system (Felten & Felten 1991), and lymphoid cells bear
receptors for manyneuroendocrine and neurotransmitter signals that have
demonstrable immunomodulatingeffects (Blalock 1992). Moreover, microg-
lia cells of the central nervous systemcan producecytokines (e.g. Carr 1992;
Guillian et al 1986). Conversely, lymphocytes can produce neuroendocrine
factors such as proopiomelanocortin-derived peptides (Blalock 1988). Also,
cytokines produced by macrophagesand activated lymphocytes are signal
molecules ("immunotransmitters") that can energize the hypothalamo-pitu-
itary-adrenal axis (e.g. Besedovsky& del Rey1991). This networkprovides
the structural foundation for and (in combination, perhaps) the mediation
various kinds of relevant observation: that lesioning or electric stimulation of
areas within the hypothalamus results in alterations in immunefunction
(Felten et al 1991), and peak antibodyproductionis associated with changes
the electrical activity within the hypothalamus(Besedovskyet al 1977); that
hypophysectomyand changes in circulating levels of hormonesand neuro-
Annual Reviews
76 ADER & COHEN
transmitters can influence immuneresponses (e.g. Berczi &Nagy1991; Ader

et al 1990), and that immunologicreactions alter circulating levels of hor-
monesand neurotransmitters (e.g. Besedovsky & dcl Rey 1991); and that
behavioral manipulations(e.g. Pavlovian conditioning, stressful stimulation)
can, as described above, ~nodulate antigen-specific and nonspecific immune
responses, that the physiological effects of products of an activated immune
system can be conditioned (Dyck et al 1990), and that immunologic dys-
regulations have behavioral consequences(Ader et al 199 lb).
Wecannot yet specify the mechanismsunderlying the functional relation-
ship between the nervous system and the immunesystem--mechanisms
illustrated by conditioned and stressor-induced modulationsof different com-
ponents of immunologicaldefenses. In fact, for the most part we cannot yet

specify the functional significance of the neuroanatomical, neurochemical,
and neuroendocrine connections between the brain and the immunesystem.
Only in recent years, however, has the autonomyof the immunesystem been
seriously questioned. Behavioral research has played a central and enabling
role in provokingstudies of interactions betweenthe central nervous system
and the immunesystem, and now represents only one of several lines of
research that provide converging support for an integrated approach to an
understanding of adaptive processes.
SUMMARY
The acquisition and extinction of the conditioned suppression or enhancement

of one or another parameter of antigen-specific and nonspecific defense sys-
tem responses have been documentedin different species under a variety of
experimental conditions. Similarly, stressful stimulation influences antigen-
specific as well as nonspecific reactions. Moreover,both conditioning and
stressful stimulation exert biologically meaningfuleffects in the sense that
they can alter the developmentand/or progression of what are presumedto be
immunologically mediated pathophysiologic processes. These are highly re-
producible phenomenathat illustrate a functional relationship between the
brain and the immunesystem. However,the extent to which one can general-
ize from one stressor to another or from one parameter of immunologicreac-
tivity to anotheris limited. Fewgeneralizations are possible becausethe direc-
tion and/or magnitudeof the effects of conditioning and "stress ’~ in modulating
immuneresponses clearly dependon the quality and quantity of the behavioral
interventions, the quality and quantity of antigenic stimulation, the temporal
relationship betweenbehavioral and antigenic stimulation, the nature of the
immuneresponse and the immunecompartmentin which it is measured, the
time of sampling,a variety of host factors (e.g. species, strain, age, sex), and
interactions amongthese several variables. It seemsreasonable to assumethat
the immunologiceffects of behaviorally induced neural and endocrine re-
sponses depend on (interact with) the concurrent immunologicevents upon
Annual Reviews
which they are superimposed. Conversely, the efficacy of immunologic de-

fense mechanisms seems to depend on the neuroendocfine environment on
which they are superimposed. We seek to determine when and what im-
munologic (or neuroendocrine) responses could be affected by what neuroen-
docrine (or immunologic) circumstances. Wetherefore need studies that pro-
vide a parametric analysis of the stimulus conditions, the neuroendocrine
and/or immunologic state upon which they are superimposed, and the re-
sponses that are being sampled.
The neural or neuroendocrine pathways involved in the behavioral alter-
ation of immuneresponses are not yet known. Both conditioning and stressor-
induced effects have been hypothesized to result from the action of
adrenocortical steroids, opioids, and catecholamines, amongothers. Indeed, all

of these have been implicated in the mediation of some immunologic effects
observed under some experimental conditions. Weassume that different con-
ditioning and stressful environmental circumstances induce different constel-
lations of neuroendocrine responses that constitute the milieu within which
ongoing immunologic reactions and the response to immunologic signals
occur. Based on the complexity of the network of connections and regulatory
feedback loops between the brain and the immune system, these processes
appear to involve both neural and endocrine signals to the immunesystem and
signals from the immunesystem that are received by the nervous system and
provoke further neural and endocrine adjustments. Considering the variety of
immunologic processes involved, it does not seem likely--or biologically
adaptive--that there would be any single mediator of the diverse effects of
either conditioning or stressful stimulation. It does not seem likely, however,
that the ultimate elaboration of the mechanisms underlying behaviorally in-
duced alterations of immunefunction will have important clinical and thera-
peutic implications.
ACKNOWLEDGMENTS
Preparation of this review was supported by a USPHSResearch Scientist
Award (K05 MH-06318) to R.A.
Literature Cited
Aarstad, H. J., Gaudemack, G., Seljelid, R. Ader, R. 1983. Developmental psycho-

1983.Stress causesreducednatural killer neuroimmunology. Dev. Psychobiol.
cell activity in mice.Scand.J. Immunol. 16:251-67
18:461~54 Ader, R. 1985. Conditioned immuno-
Ackerman,S. H,, Keller,S. E., Schleifer,S. J., pharmacologiceffects in animals: im-
Shindledecker, R. D., Camerino, M., et al. plications for a conditioning modelof
1988. Prematurematernalseparation and pharmacotherapy. In Placebo:Theory,Re-
lymphocytefunction. BrainBehav.lmmun. search, andMechanisms, ed. L. White,B.
2:161-65 Tursky, G. Schwartz, pp. 306-23. New
Ader, R., ed. 1981a. Psychoneuroimmunology. York:Guilford
NewYork: Academic Ader,R., Cohen,N. 1975.Behaviorallycondi-
Ader,R. 1981b. Ahistorical accountof conditioned immunosuppression.Psychosom.
tioned immunobiologicresponses. See Med.37:333-40
Ader1981a,pp. 321-54 Ader, R., Cohen,N. 1981. Conditionedim-
Annual Reviews
78 ADER & COHEN
munopharmacologic responses. See Ader tumorin rats: evidencefor mediationby the

1981a, pp. 185-228 immune system. Brain Behav. lmmun.
Ader, R., Cohen,N. 1982. Behaviorally condi- 5:193-205
tioned immunosuppressionand murine sys- Berczi, I., Nagy, E. 1991. Effects of
temic lupus crythematosus..Science hypophysectomy on immunefunction. See
215:1534-36 Ader et al 1991a, pp. 339-75
Ader, R,, Cohen, N. 1985. CNS-immune sys- Besedovsky, H. O., del Rey, A. 1991. Physio-
tem interactions: conditioning phenomena. logical implications of the immune-neuro-
Behav. Brain Sci. 8:379-426 endocrine network. See Ader et al 1991a,
Ader, R., Cohen, N. 1991. The influence of pp. 589-608
conditioning on immuneresponses. See Besedovsky,H. O., Sorkin, E., Felix, D., Haas,
Aderet al 1991a, pp. 611-46 H. 1977. Hypothalamicchanges during the
Ader, R., Cohen, N., Bovbjerg, D. 1982. Con- immuneresponse. Eur. J. Immunol. 7:323-
ditioned suppression of humoral immunity 25
in the rat. J. Comp.Physiol. Psychol. Blalock, J. E. 1988. Production of neuroendo-
96:517-21 crine peptide hormonesby the immunesys-
Ader, R., Cohen,N., Grota, L. J. 1987. Adreno- tem. Prog. Allergy 43:1-13
cortical steroids in the conditioned sup- Blalock, J. E.,Chem.

ed. lmmunol.
1992. Neuroimmuno-
endocrinol. 52
pression and enhancement of immune
responses. In Hormones and Immunity, Blecha, F., Barry, R. A., Kelley, K. W. 1982a.
ed. I. Berczi, pp. 231-46. London:MTP Stress-induced alterations in delayed-type
Ader, R., Felten, D. L., Cohen,N. 1990. Inter- hypersensitivity to SRBCand contact sen-
actions between the brain and the immune sitivity to DNFB in mice. Proc. Soc. Exp.
system. Annu. Rev. Pharmacol. Toxicol. Biol. Med.169:239--46
30:561~502 Blecha, F., Kelley, K. W., Satterlee, D. G.
Ader, R., Felten, D. L., Cohen,N. 1991a. Psy- 1982b. Adrenal involvementin the expres-
choneuroimmunology. New York: Aca- sion of delayed-type hypersensitivity to
demic. 2nd ed. SRBCand contact sensitivity to DNFB in
Ader, R., Grota, L. J., Moynihan,J. A., Cohen, stressed mice. Proc. Soc. Exp. Biol. Med.
N. 1991b. l~ehavioral adaptations in auto- 169:247-52
immune disease-susceptible mice. See Bohus, B., Koolhaas, J. M. 1991. Psy-
Ader et al 1991a, pp. 685-708 choimmunology of social factors in rodents
Andrykowski, M. A., Jacobsen, P. B., Marks, and other subprimate vertebrates. See Ader
E., Gorfinkle, K., Hakes,T. B., et al. 1988. et al 1991a, pp. 807-30
Prevalence, predictors and course of antici- Bonneau,R. H., Sheridan, J. F., Feng, N., Gin-
patory nausea in womenreceiving adjuvant set, R. 1991a. Stress-induced suppression
chemotherapy for breast cancer. Cancer of herpes simplex virus (HSV)-specific
62:2607-13 cytotoxic T lymphocyteand natural killer
Andrykowski, M. A., Redd, W. H. 1987. Lon- cell activity and enhancement of acute
gitudinal analysis of the development of pathogenesis following local HSVinfec-
anticipatory nausea. J. Consult. Clin. Psy- tion, Brain Behav. Immun.5:170-92
chol. 55:36--41 Bonneau,R. H., Sheridan, J. F., Feng, N., Gla-
Andrykowski,M. A., Redd, W. H., Hatfield, A. ser, R. 1991b.Stress-inducedeffects on cell
1985. The development of anticipatory mediated innate and adaptive memory
nausea: a prospective analysis. J. Consult. components of the murine immune re-
Clin. Psychol. 53:447-54 sponse to herpes simplex virus infection.
Anisman,H., Irwin, J., Bowers, W., Ahluwalia, Brain Behav. Immun. 5:274-95
P., Zacharko, R. M. 1987. Variations of Bovbjerg, D., Ader, R., Cohen, N. 1982. Be-
norepinephrine concentrations following haviorally conditioned suppression of a
chronic stressor application. Physiol. graft-vs-host response. Proc. Natl. Acad.
Biochera. Behav. 26:653-59 Sci_ USA79:583-85
Batuman,O. A., Sajewski, D., Ottenweller, J. Bovbjerg, D., Ader, R., Cohen, N. 1984. Ac-
E., Pitman, D. L., Natelson, B. H. 1990. quisition and extinction of conditionedsup-
Effects of repeated stress on T cell numbers pression of a graft-vs-host response in the
and function in rats. Brain Behav. lmmun. rat. J. ImmunoL132:111-13
4:105-17 Bovbjerg, D., Ader, R., Cohen,N. 1986. Long-
Beden,S. N., Brain, P. F. 1982. Studies on the lasting enhancement of the delayed-type
effect of social stress on measuresof dis- hypersensitivity response to heterologous
ease resistance in laboratory mice. Aggres- erythrocytes in miceafter a single injection
sive Behav. 8:126-29 of cyclophosphamide.Clin. Exp. lmmunol.
Beden, S. N., Brain, P. F. 1984. Effects of at- 66:539-50
tack-related stress on the primary immune Bovbjerg, D., Cohen, N., Ader, R. 1987a. Be-
response to sheep red blood celts in cas- haviorally conditioned enhancementof de-
trated mice. IRCSMed. Sci. 12:675 layed-type hypersensitivity in the mouse.
Ben-Ellyahu, S., Yirmiya, R., Liebeskind, J. Brain Behav. lmmun. 1:64-71
C., Taylor, A. N., Gale, R. P. 1991. Stress Bovbjerg, 13., Kim, Y. T., Siskind, G. W.,
increases metastatic spread of a mammary Weksler, M. E. 1987b. Conditioned sup-
Annual Reviews
pression of plaque-forming cell response Psychological stress and susceptibility to

with cyclophosphamide. Ann. NY Acad. the common cold. N. Engl. J. Med.
Sci. 496:588-94 325:606--12
Bovbjerg, D. H., Redd,W. H., Maier, L. A., et Cohen,S., Williamson, G. M. 1991. Stress and
al. 1990. Anticipatory immunesuppression infectious disease in humans.PsychoLBull,
and nausea in womenreceiving cyclic che- 109:5-24
motherapyfor ovarian cancer. J. Consult. Coussons, M. E., Dykstra, L. A., Lysle, D. T.
Clin. Psychol. 5:153-57 1992. Pavlovian conditioning of morphine-
Brenner, G. J., Cohen, N., Ader, R., Moyni- induced alterations of immunefunction. J.
han, J. A. 1990. Increased pulmonaryme- Neuroimmunol.In press
tastases and natural killer cell activity in Croiset, G., Heijnen, C. J., Veldhuis,H. D., de
mice following handling. Life Sci. Wied, D., Ballieux, R. E. 1987. Modulation
47:1813-19 of the immune response by emotional
Broadbent, D. E., Broadbent, M. H. P., stress. Life Sci. 40:775-82
Phillpotts, R., Wallace, J. 1984. Some Cunnick,J, E., Cohen,S., Rabin, B. S., Carpen-
further studies on the prediction of experi- ter, A. B., Manuck,S. B., Kaplan, J. R.
mental colds in volunteers by psycho- 1991. Alterations in specific antibody pro-

logical factors. J. Psychosom. Res. duction due to rank and social instability.
28:511-23 Brain Behav. lmmun. 5:357-69
Carey, M. P., Burish, T. G. 1988. Etiology and Cunnick, J. E., Lysle, D. T., Annfield, A.,
treatment of the psychological side effects Rabin, B. S. 1988. Shock-induced modula-
associated with cancer chemotherapy: a tion of lymphocyteresponsiveness and nat-
critical review and discussion. Psychol. ural killer activity: differential mechanisms
Bull 104:307-25 of induction. Brain Behav. lmmun. 2:102-
Can’, D. J. J. 1992. Neuroendocrinepeptide re- 13
ceptors on cells of the immunesystem. Cunnlck, J. E., Lysle, D. T., Kucinski, B. J.,
Neuroimmunoendocrinol. Chem. lmmunol. Rabin, B. 1990. Evidence that shock-in-
52:84-105 duced suppression is mediatcd by adrcnal
Cart, D. J. J., Wooley,T. W., Blalock, J. E. hormonesand peripheral 13-adrenergic re-
1992. Phentolamine but not propranolol ceptors. Pharmacol. Biochern. Behav.
blocks the immunopotentiating effect of 36:645-51
cold stress on antigen specific IgMproduc- Dark, K., Peeke, H. V. S., Ellman, G., Salfi,
tion in mice orally immunizedwith sheep M. 1987. Behaviorally conditioned hista-
red blood cells. Brain Behav. lmmun.6:50- mine release. Ann. NYAcad. Sci. 496:578-
63 82
Chao, C. C., Peterson, P. K., Filice, G. A., Dekker, E., Pelser, H. E., Groen,J. 1957. Con-
Pomeroy,C., Sharp, B. M. 1990. Effects of ditioning as a cause of asthmatic attacks. Z
immobilization stress on the pathogenesis Psychosom. Res. 2:97-108
of acute murine toxoplasmosis. Brain Drugan, R. C., Mandler, R., Crawley, J. N.,
Be hav. lmmun.4:162-69 Skolnick,P., Barker,J. L., et al. 1986.Con-
Cocke, R., Moynihan,J. A., Cohen, N, Grota, ditioned fear induced rapid im-
L. J., Ader, R. 1992. Exposureto conspe- munosuppression in the rat. Neurosci.
ciflc alarm chemosignalsalters immunere- Abstr. 12:337
sponses in BALB/cmice. Brain Behav. Dyck, D. G., Driedger, S. M., Nemeth,R., Os-
lmmun.6: In press achuk, T. A. G., Greenberg, A. H. 1987.
Coe, C. L., Lubach, G. R., Ershlcr, W. B., Conditioned tolerance to drug-induccd
Klopp,R. G. 1989. Influence of early rear- (PolyI:C) natural killer cell activation: ef-
ing on lymphocyteproliferation response in fects of drug-dosageand context-specific-
juvenile rhesus monkeys. Brain Behav. ity parameters. Brain Behav. lmmun.
Immun. 3:47-60 1:251~66
Coe, C. L., Rosenberg,L. T., Levine, S. 1988. Dyck, D. G., Greenberg, A. H., Osachuk, T.
Effect of maternal separation on the com- 1986. Toleranceto drug-induced(Poly I:C)
plement system and antibody responses in natural killer (NK)cell activation: congru-
infant primates. Int. J. Neurosci. 40:289- ence with a Pavlovian conditioning model.
302 J. Exp. Psychol. Anim. Behav. Proc. 12:25-
Coe, C. L., Weiner, S. G., Rosenberg, L. T., 31
Levine, S. 1985. Endocrine and immune Dyck,D. G., Janz, L., Osachuk,T. A. G., Falk,
¯ responsesto separation and maternalloss in J., Labinsky, J., Greenberg, A. H. 1990.
nonhumanprimates. In The Psychobiology The Pavlovian conditioning of IL-l-in-
qf Attachment and Separation, ed. M. duced glucocorticoid secretion. Brain
Reite, T. Field, pp. 163-200. NewYork: Behav. lmmun. 4:93-104
Academic Edwards,E. A., Dean, L. M. 1977. Effects of
Cohen,N., Ader, R., Green, N., Bovbjerg, D. crowdingof mice on humoral antibody for-
1979. Conditioned suppression of a thy- mation and protection to lethal antigenic
mus-independent antibody response. Psy- challenge. Psychosom.Med. 39:19-24
chosom. Med. 41:487-91 Edwards,E. A., Rahe, R. H., Stephens, P. M.,
Cohen,S., Tyrell, D. A. J., Smith, A. P. 1991. Henry, J. P. 1980. Antibody response to
Annual Reviews
80 ADER & COHEN
bovine serum albuminin mice: the effects gressiveness in Tilapia. Dev. Comp.Im-
of psychosocial environmental change. munoL 12:595-602
Proc. Soc. Exp. Biol. Med. 164:478-81 Ghoneum,M., Gill, G., Assanah, P., Stevens,
Eikelboom,R., Stewart, J. 1982. Conditioning W. 1987. Susceptibility of natural killer
of drug-induced physiological responses. cell activity of old rats to stress. Immuno-
Psychol. Rev. 89:507-28 logy 60:461-65
Esterliug, B., Robin, B. S. 1987. Stress-in- Gill, H. K., Liew, F. Y. 1978. Regulation of
duced alteration of T-lymphocytesubsets delayed-type hypersensitivity. III. Effects
and humoral immunity in mice. Behav. of cyclophosphamide on the suppressor
Neurosci. 101:115-19 cells for delayed-type hypersensitivity to
Faisal, M., Chiapelli, F., Ahmed,I. I., Cooper, sheep erythrocytes in mice. Eur. J. Im-
E. L., Weiner, H. 1989. Social confronta- munol. 8:172-76
tion "stress" in aggressivefish is associated Glaser, R., Rice, J., Sheridan, J., Fertel, R.,
with an endogenous opioid-mediated sup- Stout, J., et al. 1987.Stress-related immune
pression of proliferative responses to mito- suppression: health implications. Brain
gens and nonspecific cytotoxicity. Brain Behav. Immun. 1:7~0

Behav. lmmun. 3:223-33 Glascr, R., Rice, J., Spcicher, C. E., Stout, J.
Fauman,M. A. 1987. The relation of dominant C., Kiecolt-Glaser, J. K. 1986. Stress
and submissive behavior to the humoral depresses interferon production by leuko-
immuneresponse in BALB/cmice. Biol. cytes concomitantwith a decrease in natu-
Psychiatry 22:771-76 ral killer cell activity. Behav. Neurosci.
Felten, D. L., Ackerman,K. D., Wiegand,S. J., 100:675-78
Felten, S. Y. 1987. Noradrenergic sympa- Glenn, W. G., Becker, R. E. 1969. Individual
thetic innervation of the spleen: I. Nerve versus group housing in mice: im-
fibers associate with lymphocytes and munological response to time-phased in-
macrophagesin specific compartments of jections. Physiol. Zool. 42:411-16
Gorczynski, R. M. 1987. Analysis of lympho-
the splenic white pulp. J. Neurosci. Res.
18:28-36 cytes in, and host environment of, mice
Felten, D. L., Cohen,N., Ader, R., Felten, S. showing conditioned immunosuppression
Y., Carlson, S. L., Roszman,T. 1991. Cen- to cyclophosphamide. Brain Behav.
tral neural circuits involved in neural-im- Immun, 1:21-35
Gorczynski, R. M. 1990. Conditioned enhance-
muneinteractions. See Ader et al 1991a, ment of skin allografts in mice. Brain
pp. 3-25 Behav. lmmun. 4:85-92
Felten, S. Y., Felten, D. L. 1991. Innervation Gorczynski, R. M. 1991, Conditioned im-
of lymphoidtissue. See Ader et al 1991a, munosuppression:analysis of lymphocytes
pp. 27-69 and host environment of young and aged
Feng, N., Pagniano,R., Tovar, A., Bonneau,R. mice. See Ader et al 1991a, pp. 647-62
H., Glaser, R., Sheridan, J. F. 1991. The Gorczynski, R. M. 1992. Conditionedstress re-
effect of restraint stress on the kinetics, sponses by pregnant and/or lactating mice
magnitude, and isotype of the humoral im- reduce immuneresponses of their offspring
muneresponse to influenza virus infection. after weaning. Brain Behav. Immun.6:87-
Brain Behav. lmmun.5:3713-82 95
Fleshner, M., Laudenslager,M. L., Simons,L., Gorczynski, R. M., Holmes, W. 1989. Neuro-
Maier, S. F. 1989. Reducedserum antibod- leptic and anti-depressant drug treatment
ies associated with social defeat in rats, abolishes conditioned immunosuppression
Physiol. Behav. 45:1183-87 in mice. Brain Behav. Immun.3:312-19
Friedman, E. M., Coe, C. L., Ershler, W. B, Gorczynski, R. M., Kennedy, M., Ciampi, A.
1991. Time-dependenteffects of peer sepa- 1985. Cimetidine reverses tumor growth
ration on lymphocyte proliferation re- enhancement of plasmacytoma tumors in
sponse in juvenile squirrel monkeys.Dev. mice demonstrating conditioned im-
Psychobiol. 24:159-73 munosuppression. J. lmmunol. 134:4261-
Friedman, S. B., Ader, R., Glasgow, L. A, 66
1965. Effects of psychological stress in Gorczynski, R. M., Macrae, S., Kennedy, M.
adult mice inoculated with CoxsackieB vi- 1982. Conditioned immuneresponse asso-
ruses. Psychosom. Med. 27:361-68 ciated with allogeneic
J. lmmunol. 129:704~9skin grafts in mice.
Friedman, S. B., Glasgow, L. A., Ader, R.
1969. Psychosocial factors modifyinghost Gorczynski, R. M., Macrae, S., Kennedy, M.
resistance to experimentalinfections. Ann, 1984.Factors involvedin the classical con-
NYAcad. Sci. 164:381-92 ditioning of antibody responses in mice. In
Geiser, D. S. 1989. Psychosocialinfluences on Breakdown in Human Adaptation to
human immunity. Clin. Psychol. Rev, "’Stress:" Towardsa Multidisciplinary Ap-
9:689-715 proach, ed. R. E, Ballieux, J, F. Fielding,
Ghoneum,M., Faisal, M., Peters, G., Ahmed, A. L’Abbate, pp. 704-12. Hingham, MA:
I. I., Cooper, E. L. 1988. Suppression of Martinus Nijhof
natural cytotoxic cell activity by social ag- Greenberg,A. H., Dyck, D. G., Sandier, L. S.,
Annual Reviews
Pohajdak, B., Dresel, K. M., Grant, D. once of time of stress application and type
1984. Neurohormonalmodulation of natu- of tumor. Psychol. Bull. 98:108-38
ral resistance to a murine lymphoma.J. Kandil, O., Borysenko, M. 1988. Stress-in-
Natl. CancerInst. 72:653-59 duced decline in immuneresponsiveness in
Grochowicz, P., Schedlowski, M., Husband, C3H/HeJmice: relation to endocrine alter-
A. J., King, M. G., Hibberd, A. D., Bowen, ations and tumor growth. Brain Behav.
K. M. 1991. Behavioral conditioning pro- lmmun. 2:32-49
longs heart allograft survival in rats. Brain Kaplan, R., Robinson, C. A., Scavulli, J. F.,
Behav. lmmun. 5:349-56 Vaughan,J. H. 1980. Propranolol and the
Guilian, D., Baker, T. J., Shih, L.-C. N., Lach- treatment of rheumatoidarthritis. Arthritis
man,L. B. 1986. Interleukin-1 of the cen- Rheum. 23:253-55
tral nervous system is produced by Keller, S. E., Schleifer, S. J., Liotta, A. S.,
ameboidmicroglia. J. Exp. Med. 164:594-- Bond,R. N., Farhoody, N., Stein, M. 1988.
6O4 Stress-induced alterations of immunityin
Gust, D. A., Gordon, T. P., Wilson, M. E., hypophysectomized rats. l)roc. Natl. Acad.
Brodie, A. R., Ahmed-Ansari, A., Mc- Sci. USA85:9297-9301

Clure, H. M. 1992. Removal from natal Keller, S. E., Weiss, J. M., Schleifer, S. J.,
social group to peer housing affects corti- Miller, N. E,, Stein, M. 1981. Suppression
sol levels and absolute numbersof T cell of immunityby stress: effects of a graded
subsets in juvenile rhesus monkeys.Brain series of stressors on lymphocytestimula-
tion in the rat. Science 213:1397-400
Behav. lmmun. 6:189-99
Hardy,C., Quay,J., Livnat, S., Ader, R. 1990. Keller, S. E., Weiss, J. M., Schleifer, S. J.,
Altered T lymphocyte response following Miller, N. E., Stein, M. 1983. Stress-in-
aggressive encounters in mice. Physiol. duced suppression of immunityin adrenal-
ectomized rats. Science 221:1301-4
Behav, 47:1245~ 1 Kemeny~ M. E., Solomon,G. F., Morley, J. E.,
Hill, L. E. 1930. Philosophy of a Biologist. Bennett, T. L. 1992. Psyclaoneuroimmuno-
London: Arnold logy. In A Comprehensive Textbook of
Hiramoto,R. N., Hiramoto,N. S., Solvason, H. Neuroendocrinology, ed. C. B. Nemeroff.
B., Ghanta, V, K. 1987. Regulation of nat- In press
ural immunity(NKactivity) by condition- Khan, A. U. 1977. Effectiveness of biofeed-
ing. Ann. NYAcad. Sci. 496:545-52 back and counterconditioniug in the treat-
Hull, C. L. 1934. The factor of conditioned ment of bronchial asthma. J. Psychosom.
reflex. In A Handbookof General Experi- Res. 21:97-104
mental Psychology, ed. C, Murchison. King, M. G., tlusband, A. J., Kusnecov,A. W.
Worcester: Clark Univ. 1987. Behaviourally conditioned im-
Husband,A. J., King, M. G., Brown,R. 1987. munosuppression using anti-lymphocyte
Behaviorally conditioned modification of serum:duration of effect and role of corti-
T cell subset ratios in rats. lmmunol.Lett. costeroids. Med.Sci. Res. 15:407-8
14:91-94 Klosterhalfen, S., Klosterhalfen, W. 1987.
Ikemi, Y., Nakagawa,S. 1962. A psychoso- Classically conditioned cyclophosphamide
matic study of contagions dermatitis. effects on white blood cell counts in rats.
KyushuJ. Med. Sci. 13:335-50 Ann. NYAcad.Sci. 496:569-77
Ito, Y., Mine, K., Ago, Y., Nakagawa, T., Klosterhalfen, S., Klosterhalfen, W. 1990.
Fujiwara, M., Ueki, S. 1983. Attack stress Conditioned cyclosporine effects but not
and IgE antibody response in rats. Phar- conditioned taste aversion in immunized
macol. Biochem. Behav. 19:883-86 rats. Behav. Neurosci. 104:716-24
Irwin, J., Custean, N. 1989. Stressor predict- Klosterhalfen, W., Klosterhalfen, S. 1983.
ability and immunefunction. Soc. Neu- Pavlovian conditioning of immunosuppres-
rosci. Abstr. 15:298 sion modifies adjuvant arthritis in rats.
Jessop, J, J., Gale, K., Bayer, B. M.1987. En- Behav. Neurosci. 97:663-66
hancementof rat lymphocyteproliferation Kopeloff, N. 1941. Bacteriology in Neuropsy-
after prolongedexposureto stress. J. Neu chiat~.. Springfield: Charles C Thomas
roimmunol. 16:261-71 Krank, M. D., MacQueen,G. M. 1988. Condi-
Jessop, J. J., Gale, K., Bayer, B. M. 1988. tioned compensatoryresponses elicited by
Time-dependent enhancement of lympho- environmental signals for cyclophospham-
cyte activation by mitogensafter exposure ide-induced suppression of antibody pro-
to isolation or water scheduling. Life Sci. duction in mice. Psychobiology16:229-35
43: l 133-40 Kusnecov,A. V., Grota, L. J., Schmidt, S. G.,
Jiang, C. G., Morrow-Tesch,J. L., Beller, D. Bonneau,R. H., Sheridan,J. F., et al. 1992.
S., Levy, E. M., Black, P. H. 1990. Im- Decreased herpes simplex viral immunity
munosuppressionin mice induced by cold and enhanced pathogenesis following
water stress. Brain Behav. Immun.4:278- stressor administration in mice. J. Neu-
91 roimmunol. 38:129-38
Justice, A. 1985. Reviewof the effects of stress Kusnecov,A. V., Husband,A. J., King, M. G.
on cancer in laboratory animals: import- 1988. Behaviorally conditioned suppres-
Annual Reviews
82 ADER&COHEN
sion of mitogen-inducedproliferation and Lysle, D. T., Luecken, L. J., Maslonek,K. A.

immunoglobulinproduction: effect of time 1992b. Modulation of immunefunction by
span between conditioning and reexposure a conditioned aversive stimulus: evidence
to the conditioned stimulus. Brain Behav. for the involvementof endogenousopioids.
lmmun. 2:198-211 Brain Behav. lmmun. 6:179-88
Kusnecov, A. V., Sivyer, M., King, M. G., Lysle, D. T., Lyre, M., Fowler, H., Rabin, B. S.
Husband. A. J., Cripps, A. W., Clancy, R. 1987. Shock-induced modulation of lym-
L. 1983. Behaviorally conditioned suppres- phocyte reactivity: suppression, habitua-
sion of the immuneresponse by antilym- tion, and recovery. Life Sci. 41:1805-14
phocyte serum. J. lmmunol. 130:2117-20 Lysle, D. T., Maslonek, K. A. 1991. Immune
Laudenslager,M. L., Capitanio, J. P., Reite, M. alterations induced by a conditioned aver-
1985. Possible effects of early separation sive stimulus: evidence for a time-depen-
experiences on subsequent immunefunc- dent effect. Psychobiology19:339-44
tion in adult macaque monkeys. Am. J. Mackenzie,J. N. 1896. The production of the
Psychiatry 142:862-64 -so-called "rose cold" by meansof an at~tifi
Laudenslager,M. L., Fleshner, M., Hofstadter, cial rose. Am. J. Med.Sci. 91:45-57
P., Held, P. E., Simons,L., Maier, S. F. MacQueen,G. M., Marshall, J., Perdue, M.,
1988. Suppression of specific antibody Siegel, S., Bienenstock, J. 1989. Pavlovian
production by inescapable shock: stability conditioning of rat mucosal mast cells Io
under varying conditions. Brain Behav. secrete rat mast cell protease II. Science
hnmun. 2:92-101 243:83-85
Laudenslager, M. L., Held, P. E., Buccia, M. MacQueen, G. M., Siegel, S. 1989. Condi-
L., Reite, M.L., Cohen,J. J. 1990. Behav- tional immunomodulation following train-
ioral and immunological consequences of ing with cyclophosphamide. Behav.
brief mother-infant separation: a species Neurosei. 103:638-47
comparison. Dev. Psychobiol. 23:247-64 Maier, S. F, Laudenslager, M. L. 1988. Ines-
Laudenslager, M. L., Ryan, S. M., Drugan, R. capable shock, shock controllability, and
C., Hyson,R. L., Maier, S. F. 1983. Coping rnitogen-stimulated lymphocyteprolifera-
and immunosuppression: Inescapable but tion. Brain Behav. lmmun.2:87-91
not escapable shock suppresses lympho- Metal’nikov, S., Chorine, V. 1926. R01edes
cyte proliferation. Science 221:568-70 rtflexes conditionne[s darts l’immunitt.
Levine, S., Strebel, R., Wenk,E. J., Harman,P. Ann. L’Inst. Pasteur40:893-900
J. 1962. Suppressionof experimentalaller- Metal’nikov, S., Chorine, V. 1928. ROle des
gic encephalomyelitisby stress. Proc. Soc. rdflexes conditionnels dans la formation
Exp. Biol. Med. 109:294-98 des anticorps. C. R. Soc. BioL 102:133-34
Livnat, S., Irwin, J, Ader, R., Gallo, K. Michaut, R.-J., Dechambre,R.-P., Doumerc,
Anisman,H. 1985. Immunologicreactivity S., Lesourd, B., Devillechabrolle, A.,
following acute and chronic stress. Sue. Moulias, R. 1981. Influence of early mater-
Neurosci. Abstr. 11:861 nal deprivation on adult humoral immune
Lown, B. A., Dutka, M. E. 1987. Early han- response in mice. Physiol. Behav. 26:189-
dling enhances mitogen responses of sple- 91
nic cells in adult C3Hmice. Brain Behav. Mitsuoka, A., Morikawa, S., Baba, M.,
lmmun. 1:356-60 Harada, T. 1979. Cyclophosphamideelimi-
Luecken,L. J., Lysle, D. T. 1992. Evidencefor nates suppressor T cells in age-associated
the involvement of beta-adrenergic recep- central regulation of delayedhypersensitiv-
tors in conditioned immunomodulation.J. ity in mice. J. Exp. Med. 149:1018-28
Neuroimmunol.In press Monjan,A. A., Collector, M. I. 1977. Stress-
Lysle, D. T., Cunnick,J. E., Fowler, H., Rabin, induced modulation of the immune re-
B. S. 1988. Pavlovian conditioning of sponse. Science 196:307-8
shock-induced suppression of lymphocyte Mormede, P., Dantzer, R., Michaud, B.,
reactivity: acquisition, extinction, and pre- Kelley, K., Le Moal,M. 1988. Influence of
exposureeffects. Life Sci. 42:2185-94 stressor predictability and behavioral con
Lysle, D. T., Cunnick, J. E., Kucinski, B. J., trol on lymphocytereactivity, antibody re-
Fowler, H., Rabin, B. S. 1990a. Character- sponses, and neuroendocrineactivation in
ization of immunealterations induced by a rats. Physiol. Behav.43:577-83
conditioned aversive stimulus. Psychobiol- Morrow,G. R., Dobkin, P. L. 1988. Anticipa-
ogy 18:220-26 tory nauseaand vomitingin cancer patients
Lysle, D. T., Cunnick, J. E., Rabin, B. S. undergoing chemotherapytreatment: prev-
1990b. Stressor-induced alteration of lym- alence, etiology and behavioral interven-
phocyte proliferation in mice: evidence fur tions. Clin. PsychoLRev. 8:517-56
enhancement of mitogen responsiveness. Moynihan, J. A., Ader, R., Grota, L. J.,
Brain Behav. lmmun. 4:269-77 Schachtman, T. R., Cohen, N. 1990a. The
Lysle, D. T., Luecken, L. J., Maslonek,K. A. effects of stress on the developmentof im-
1992a. Suppression of the development of munological memoryfollowing low-dose
adjuvant arthritis by a conditionedaversive antigen priming in mice. Brain Behav.
stimulus. Brain Behav. Immun.6:64-73 lmmun. 4:1-12
Annual Reviews
Moynihan,J., Brenner, G., Koota, D., Brene- Peeke, H. V. S., Ellman, G., Dark, K., Salfi,
man, S., Cohen, N., Ader, R. 1990b. The M., Reus, V. I. 1987. Cortisol and behav-
effects of handlingon antibodyproduction, iorally conditionedhistamine release. Ann.
mitogen responses, spleen cell number,and NYAcad. Sci. 496:583-87
lymphocyte subpopulations. Life Sci. Plaut, S. M., Esterhay, R. J., Sutbedand,J. C.,
46:1937-44 Wareheim,L. E., Friedman, S. B., et al.
Moynihan,J., Koota, D., Brenner, G., Cohen, 1981. Psychological effects on resistance
N., Ader, R. 1989. Repeatedintraperitoneal to spontaneous AKRleukemia in mice.
injections of saline attenuate the antibody Psychosom. Med. 42:72
responseto a subsequentintraperitoneal in- Qunidos, G., Ruis de Gordioa, J. C., Burgos,
jection of antigen. Brain Behav. lmmun. A., Ponton, J., Cistema, R. 1986. Fasting
3:90-96 and immobilizationstress cause a decrease
Munck, A., Guyre, P. M. 1991. in phagocytosis in male rats. Arch. lm-
Glucocorticoids and immunefunction. See munol. Ther. Exp. 34:573-76
Aderet al 1991a, pp. 447-513 Raab, A., Dantzer, R., Michaud,B., Mormede,
Neveu,P. J., Crestani, F., Le Moal,M. 1987.
P., Taghzouti, K., et al. 1986. Behavioral,

Conditioned immunosuppression: a new physiological, and immunological conse-

methodological approach. Ann. NY Acad. quencesof social status and aggression in
Sci. 496:595--601 chronically coexisting resident-intruder
Neveu, P. J., Dantzer, R., Le Moal, M. 1986. dyads of male rats. Physiol. Behav.
Behaviorally conditioned suppression of 36:223-28
mitogen-induced lymphoproliferation and Rabin, B. S., Lyte, M., Epstein, L. H.,
antibody production in mice. Neurosci. Caggiula, A. R. 1987a. Alteration of im-
Lett. 65:293-98
mune competency by number of mice
Odio, M., Brodish, A., Ricardo, M. J. 1987. housed per cage. Ann. NY Acad. Sci.
Effects on immuneresponses by chronic 469:492 500
stress are modulated by aging. Brain Rabin, B. S., Lyte, M., Hamill, E. 1987b. The
Behav. Immun. 1:204-15 influence of mousestrain and housing on
O’Grady, M. P., Hall, N. R. S. 1991. Long- the immuneresponse. J. NeuroimmunoL
term effects of neuroendocrine-immune in- 17:11-16
teractions during early development. See Rabin,B. S., Salvin, S. B. 1987. Effect of dif-
Ader et al 1991a, pp. 561-72 ferential housing and time on immunere-
Okimura, T., Nigo, Y. 1986. Stress and im- activity to sheep erythrocytes and Candida.
muneresponses I. Suppression of T cell Brain Behav. lmmun. 1:267-75
functions in restraint-stressed mice. Jpn. J. Rasmussen,A. F. Jr., Marsh,J. T., Brill, N. Q.
Pharmacol. 40:505-11
Okimura,T., Ogawa,M., Yarnauchi, T. 1986a. 1957. Increased susceptibility to herpes
Stress and immuneresponses III. Effect of simplex in mice subjected to avoidance-
learning stress or restraint. Proc. Soc. Exp.
restraint stress on delayedtype hypersensi- BioL Med. 96:183-89
tivity (DTH)response, natural killer (NK) Raymond, L. N., Reyes, E., Tokuda,S., Jones,
activity, and phagocytosisin mice. Jpn. J. B. C. 1986. Differential immuneresponse
Pharmacol. 41:229-35
Okimura, T., Ogawa, M., Yamauchi, T., in two handled inbred strains of mice.
Satomi-Sasaki, Y. 1986b. Stress and im- Physiol. Behav. 37:295-97
muneresponses IV. Adrenal involvement Redd, W. H., Andrykowski, M. A. 1982. Be-
in the alteration of antibody responses in havioral interventions in cancer treatment:
restraint-stressed mice. Jpn. J. Pharmacol. controlling aversive reactions to chemo-
41:237-45 therapy. J. Consult. Clin. Psychol.
Okimura, T., Satomi-Sasaki, Y., Ohkuma,S. 50:1018-29
1986c. Stress and immuneresponses II. Rinner, I., Schauenstein, K., Mangge, H.,
Identification of stress-sensitive cells in Porta, S. 1992. Opposite effects of mild
~nurine spleen cells. Jpn. J. Pharmacol. and severe stress on in vitro activation of
40:513-25 rat peripheral blood lymphocytes. Brain
Behav. lmmun. 6:130~.0
O’Lcary, A. 1990. Stress, emotion, and human
immunefunction. Psychol. Bull 108:363- Rodinone,S. N., Giovanniello, O. A., Barrios,
82 H. A., Nota, N. R. 1983. Effect of frac-
O’Reilly, C. A., Exon, J. H. 1986. tional cyclophosphamidedosage on sheep
Cyc lophosphamide-conditioned suppres- red blood cell-delayed-type hypersensitiv-
sion of the natural killer cell response in ity response in mice. J. lmmunoL
rats. Physiol. Behav. 37:759-64 1130:1600-3
Ottenberg, P., Stein, M., Lewis, J.. Hamilton, Rogers, M. P., Reich, P., Strom, T. B., Carpen-
C. 1958. Learnedasthma in the guinea pig. ter, C. B. 1976. Behaviorally conditioned
Psychosom. Med. 20:395-400 immunosuppression: replication of a recent
Pavlidis, N., Chirigos, M.1980. Stress-induced study. Psychosom. Med. 38:447-52
impairment of macrophage tumoricidal Rogers, M. P., Trentham, D. E., Dynesius, R.
function. Psychosom.Med. 42:47-54 A., Reich, P., David, J. R. 1980a. Exacer-
Annual Reviews
84 ADER & COHEN
bation of type II collagen-inducedarthritis tiveness and the conditionedreflex. Yale J.

by auditory stress. Clin. Res. 28:508 Biol. Med. 5:387402
Rogers, M. P., Trentham, D. E., McCune,W. Smith, G. R., McDaniels, S. M. 1983. Psycho-
J., Ginsberg, B. I., Rennke,H. G., et al. logically mediated effect on the delayed
1980b. Effect of psychological stress on hypersensitivity reaction to tuberculin in
the induction of arthritis in rats. Arthritis humans. Psychosom. Med. 45:65-70
Rheum. 23:1337-42 Solomon,G. F. 1969. Stress and antibody re-
Roudebnsh,R. E., Bryant, H. U. 1991. Condi- sponsein rats. Int. Arch. Allergy 35:97-104
tioned immunosuppressionof a murine de- Solomon, G. F., Amkraut, A. A., Kasper, P.
layed type hypersensitivity response: 1974. Immunity,emotions and stress. Ann.
dissociation from corticosterone elevation. Clin. Res. 6:313-22
Brain Behav. Immun. 5:308-17 Solomon,G. F., Levine, S., Kraft, J. K. 1968.
Russell, M., Dark, K. A., Cummins,R. W., Early experience and immunity. Nature
Ellman, G., Callaway, E., Peeke, H. V. S. 220:821-22
1984. Learned histamine release. Science Solvason, H. B., Ghanta, V. K., Hiramoto, R.
225:733-34 N. 1988. Conditioned augmentationof nat-

Sato, K., Flood, J. F., Makinodan,T. 1984. ural killer cell activity. Independencefrom
Influence of conditioned psychological nociceptive effects and dependenceon in-
stress on immunological recovery in mice terferon-~. J. lmmunol.140:661~65
exposed to low-dosex-irradiation. Radiat. Solvason, H. B., Ghanta, V., Soong, S.-J.,
Res. 98:381-88 Hiramoto, R. N. 1991. Interferon interac-
Schulze, G. E., Benson, R. W., Paule, M. G., tion with the CNSis required for the condi-
Roberts, D. W. 1988. Behaviorally conditioning of the NKcell response. Prog.
tioned suppression of murine T-cell depen- Neuroendocrinimmunol. 4:258~54
dent but not T-cell independent antibody Solvason, H. B., Hiramoto, R. N., Ghanta, V.
responses. Pharmacol. Biochem. Behav. K. 1989. Naltrexone blocks the expression
30:859-65 of the conditioned elevation of natural
killer
Behav. cell
Seifler, E., Rettura, G., Zisblatt, M., Levinson, activity in BALB/c mice. Brain
Irnmun. 3:247~52
S., Levine, N., et al. 1973. Enhancement of
tumor developmentin physically stressed Sonnenfeld, G., Cunnick, J. E., Armfield, A.
mice incubated with an oncogenic virus. V., Wood,P. G., Rabin, B. S. 1992. Stress-
Experientia 29:1379-82 inducedalterations in interferon production
Shavit, Y., Lewis, J. W., Terman,G. W., Gale, and class II histocompatibility antigen ex-
R. P., Liebeskind, J. C. 1984. Opioid pep- pression. Brain Behav. Immun.6:170-78
tides mediate the suppressive effect of Steplewskl, Z., Vogel, W. H. 1986. Total leu-
stress on natural killer cell cytotoxicity. kocytes, T cell subpopulation, and natural
Science 223:188-90 killer (NK)cell activity in rats exposed
Shavit, Y., Ryan, S. M., Lewis, J. W., restraint stress. Life Sci. 38:2419-27
Laudenslager, M. L., Terman,G. W., et al. Steplewski, Z., Vogel, W. H., Ehya, H.,
1983. Inescapable but not escapable stress Poropatich, C., Smith, J. M. 1985. Effects
alters immune function. Physiologist of restraint stress on inoculated tumor
26:A64 growth and immuneresponse in rats. Can-
Shavit, Y., Terman,G. W., Lewis, J. W., Zane, cer Res. 45:5128-33
C. J., Gale, R. P., Liebeskind, J. C. 1986. Stites, D. P., Terr, A. I. 1991. BasicandClini-
Effects of footshock stress and morphine cal Immunology. Norwalk, CT: Appleton
on natural killer lymphocytesin rats: stud- & Lange
ies of tolerance and cross-tolerance. Brain Taylor, C. E., Ross, L. L. 1989. Alteration of
Res. 372:382-85 antibody response to pneumococcalpoly-
Sieber, W.J., Rodin,J., Larson,L., Ortega, S., saccharide type III in rats by neonatal im-
Cummings,N. et al. 1992. Modulation of mobilization stress. Brain Behav. Immun.
humannatural killer cell activity by expo- 3:160-70
sure to uncontrollable stress. Brain Behav. Totman,R., Kiff, J, Reed, S. E., Craig, J. W.
lmmun. 6:141-56 1980. Predicting experimentalcolds in vol-
Siegel, S. 1976. Morphineanalgesia tolerance: unteers from different measuresof recent
Its situation specificity supports a Pavlov- life stress. J. Psychosom.Res. 24:155-63
ian model. Science 193:323 25 Turk, J. L., Parker, D. 1982. The effect of
Siegel, S. 1983. Classical conditioning, drug cyclophosphamide on immunological con-
tolerance, and drug dependence. In Re- trol mechanisms, lmmunol.Rev. 65:99-113
search Advances in Alcokol and Drug Vessey, S. H. 1964. Effects of grouping on
Problems, ed. Y. Israel, F. B. Gaser, H. levels of circulating antibodies in mice.
Kalant, R. E. Popham,W. Schmidt, R. G. Proc. Soc. Exp. BioL Med. 115:252-55
Smart, pp. 20746. NewYork: Plenum Wayner, E. A., Flannery, G. R., Singer, G.
Sklar, L. S., Anisman,H. 1981. Stress and can- 1978. The effects of taste aversion condi-
cer. Psychol. Bull. 89:369~,06 tioning on the primary antibody response
Smith, 13. H., Salinger, R. 1933. Hypersensi- to sheep red blood ceils and Brucella abor-
Annual Reviews
tus in the albino rat. Physiol. Behav. AnimalModels of Affective Disorders, ed.
21:995-1000 G. Koob,C. L. Ehlers, D. J. Kupfer, pp.
Weiner, H. 1977. Psychobiology and Human 204-38. Chicago: Univ. Chicago Press
Disease. NewYork: Elsevier Zalcman, S., Irwin, J., Anisman, H. 1991a.
Weiner, H. 1991. Social and psychobiological Stressor-induced alterations of natural
factors in autoimmunedisease. See Ader et killer cell activity and central catechola-
al 1991a, pp. 955-1011 mines in mice. Pharmacol. Biochem.
Weiss, J. M., Sundar, S. K., Becker, K. J. Behav. 39:361-66
1989a. Stress-induced immunosuppression Zalcman,S., Kerr, L., Anisman,H. 1991b. Im-
and immunoenhancement: cellular changes munosuppressionelicited by stressors and
and mechanisms. In Neuroimmune Net- stressor-related odors. Brain Behav.
works: Physiology and Disease, ed. E. J. hnmun. 5:262-73
Goetzl, N. H. Spector, pp. 193206. New Zalcman, S., Minkiewicz-Janda, A., Richter,
York: Liss M., Anisman,H. 1988. Critical periods as-
Weiss, J. M., Sundar, S. K., Becker, K. J., sociated with stressor effects on antibody
Cierpial, M. A. 1989b. Behavioral and neu- titers and on the plaque-formingcell re-
ral influences on cellular immuneresponse to sheep red blood cells. Brain
sponses: effects of stress and interleukin-1. Behav. Immun. 2:254-66
Z Clin. Psychiat. 50:43-53 Zalcman, S., Richter, M., Anisman,H. 1989.
Weisse,C. S., Pato, C. N., McAllister, C. G., Alterations of immunefunctioning follow-
Littman,R., Breier, A., et al. 1990. Differ- ing exposureto stressor-related cues. Brain
ential effects of controllable and uncontrol- Behav. lmmun. 3:99-109
lable acute stress on lymphocyte Zha, H., Ding, G., Fan, S. 1992. Serumfactors
proliferation and leukocyte percentages in inducedby restraint stress in miceand rats
humans. Brain Behav. Immun. 4:339-51 suppresses lymphocyteproliferation. Brain
Wistar, R. T., Hildemann,W. H. 1960. Effect Behav. lmmun. 6:18-31
of stress on skin transplantation immunity Zwilling, B. S., Brown,D., Christner, R., Faris,
in mice. Science 131:159-60 M., Hilberger, M., et al. 1990. Differential
Zacharko, R. M., Anisman, H. 1988. Pharma- effect of restraint stress on MHC class II
cological, biochemical and behavioral expression by murine peritoneal macro
analyses of depression: animal models. In phages. Brain Behav. Immun.4:330-38
View publication stats

Psychoneuroimmunology Conditioning and Stress Ader20Conditioning20 20immunology20AnnRevPsy201993

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Psychoneuroimmunology Conditioning and Stress Ader20Conditioning20 20immunology20AnnRevPsy201993

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Psychoneuroimmunology: Conditioning and Stress

Article in Annual Review of Psychology · February 1993

The user has requested enhancement of the downloaded file.

Annu.Rev. Psychol.1993. 44:53~5

Robert Ader and Nicholas Cohen

ity and, conversely, the immunestatus of an organism has consequencesfor

effects of conditioning and stress in the modulationof immunefunction. There

CONDITIONED MODULATIONOF IMMUNITY

little morethan preliminary observations. Althoughthe data on specific anti-

Effects of Conditioning on Humoral and Cell-Mediated

Oneof the hallmarks of the immunesystem’s defense of the organismagainst

foreign, "nonself"material (antigens) is its specificity--its ability to recognize

HUMORAL IMMUNITY Current interest in conditioned changes in immunologic

animals received saccharin. A placebo-treated group was injected with vehicle

nonspecific host defenses have nowbeen observed under a variety of experi-

tained using electric shock as the UCS(Sato et al 1984; Zalcmanet al 1989,

tioned animalsreexposedto the CSdid not differ fromthose of a saline-treated

CELL-MEDIATED IMMUNITY The immunosuppressiveeffects of CYcan also be

reported for other conditioned immunologic effects (Dyck et al 1986;

Effects of Conditioning on Nonimmunologically Specific

In addition to studying antibody- and cell-mediated immunity,immunologists

cells obtained from mesenteric lymphnodes showedno effects of CSreexpos-

been unable to repeat these observations (Ader & Cohen1991):

Antigen as Unconditioned Stimulus

conditioned immune responses as distinct from conditioned im-

Biologic Impact of Conditioned Changes in Immunity

Conditioning in Human Subjects

1896:45). The literature contains descriptions of several similar cases of what

that unconditionally induced eczemawithin 24 hr. After an unspecified num-

ible immuneresponse. This basic strategy was used by Bovbjerget al (1982,

ipatory nausea and vomiting during the course of chemotherapywouldalso

STRESS AND IMMUNITY

In the present context, "stress" refers to any natural or experimentallycon-

Effects of Stress on Disease

behavioral manipulationsor stressors can influence susceptibility to a variety

of disease states in a variety of species. However,the stressors used do not all

(including immunologicresponses) induced by specific forms of stressful

related to the. pathophysiologicprocess under study. It is not yet clear that

behaviorally induced perturbations in immunefunction can influence or medi-

process, forced swimmingdecreased NKcell activity and resulted in a two-

Effects of Stress on Humoral and Cell-Mediated Immunity

and/or secondary antibody responses to several different antigens (e.g.

stressor; in another study (Cockeet al 1992), immunizationwith KLHfollow-

influenced the immuneresponse, but not in the hypothesizeddirection. Zalc-

the electric shock.

immuneresponses. Restraint imposedbefore sensitization or challenge sup-

Anotherimmunologicallyrelevant reaction to stressors is a suppression of

macrophagefunction (Pavlidis & Chirigos 1980; Jiang et al 1990; Okimuraet

Effects of Stress on Nonimmunologically Specific Reactions

1992), and as a result of separation experiences (Ackermanet al 1988).

The ability to predict or control the effects of stressful stimulation has

not escapable shock suppressed splenic NKcell activity in rats. In another

MEDIATION OF BEHAVIORALLY INDUCED IMMUNE

The observedincreases or decreases in the antibody-formingcell response to

(but not uniformly) immunosuppressive(Munck& Guyre 1991). Nonetheless,

and, at the same time, suppressed or unaltered serum levels of im-

inducedby the nature of the stressor.

It mayalso be relevant that somestudies have uncovereddifferent changes

opioids, for which there are receptors on lymphocytes(Blalock 1988), can

transmitters can influence immuneresponses (e.g. Berczi &Nagy1991; Ader

ponents of immunologicaldefenses. In fact, for the most part we cannot yet

The acquisition and extinction of the conditioned suppression or enhancement

which they are superimposed. Conversely, the efficacy of immunologic de-