Science Watch

Psychoneuroimmunology
The Interface Between Behavior, Brain, and Immunity
Steven F. Maier, Linda R. Watkins, and Monika Fleshner

Psychoneuroimmunology is the study of interactions be- have been adaptive in evolution, rather than simply being
tween behavior, the brain, and the immune system. This a curiosity. These are the questions that psychologists
article is designed to provide an overview of this new field typically ask when they are presented with work in this
for the general psychologist. The existence of bidirectional area. Discussions such as these often will be, of necessity,
communication pathways between the brain and the im- quite speculative. However, it is our belief that the psy-
mune system and the implications of this network for be- chologist will be as captivated by this field as we are only
havior are emphasized. Implications are that behavioral- if some of these evolutionary-functional possibilities are
psychological processes ought to be capable of altering elaborated, so that the connections between behavior and
immune function and that events that occur as part of immunity come to make intuitive sense.
immune responses should modulate behavior. Evidence PNI is the study of interactions between behavior,
for influences in both of these directions is reviewed. The the nervous system, and the immune system. It grew from
discussion of psychological modulation of immunity fo- the realization that the immune system does not operate
cuses on classical conditioning and stress, whereas that autonomously, as had often been supposed. The typical
of immune modification of behavior highlights behavioral view, held as recently as 10 years ago, was that the immune
effects produced by substances released by the immune system was a closed system. It was thought to be driven
system. Finally, the adaptive role that such changes might by challenges from foreign substances (antigens) and reg-
play is considered. ulated by soluble products produced and released by im-
mune cells (lymphokines or cytokines, more generally).
These products serve both to communicate between im-

T he purpose of this article is to provide psychologists mune cells both locally and at distant sites and to control
with an overview of the new field of psychoneuro- the progress of the immune response. Although antigens
immunology (PNI), which has developed over the do initiate immune responses and cytokines do regulate
past 10 to 15 years. A detailed review is not possible here; immune processes, a wide array of recent research dem-
various aspects of PNI have recently been given extensive onstrates that there are bidirectional communication
review (Ader & Cohen 1993; Ader, Felten, & Cohen 1991; pathways between the immune system and central ner-
Cohen & Williamson, 1991; Kemeny, Solomon, Morley, vous system (CNS), with each providing important reg-
& Bennett, 1993; Plotnikoff, Murgo, Faith, & Wybran, ulatory control over the other.
1991). Neither is the purpose to review our own work in As will be noted, immune function can require
this area. Instead, our goals are to provide (a) a sketch of global alterations involving the entire organism (e.g., a
the basic core facts that led to the coalescence of a new shift in energy balance) as well as the more usually con-
discipline, (b) some indication of the possible functional sidered local processes (e.g., selective rapid multiplication
significance of the basic aspects of organization that have of T cells in a lymph node in response to a detected an-
been discovered, (c) a feel for some of the exciting pos- tigen). Only the CNS can orchestrate such widespread
sibilities provided by PNI, and (d) some cautions to note. outcomes in a coordinated fashion. Thus the CNS must
We will concentrate on the "whys" rather than provide
a list of studies. For example, in our discussion of stress Lyle E. Bourne served as action editor for this article.
Steven F. Maier, Department of Psychology, University of Colorado;
and immunity we will not provide extensive documen- Linda R. Watkins, Department of Psychology, University of Colorado;
tation that stress can alter immune function—that is well- Monika Fleshner, Department of Psychology, University of Colorado.
known and has often been reviewed. The "hows" (what Preparation of this article was supported by Grants MH-45045 to
type of stressor, which hormone is the critical mediator, Steven F. Maier and Linda R. Watkins and RSA MH-00314 to Steven
F. Maier.
etc.) have also been reviewed elsewhere. Instead, we will Correspondence concerning this article should be addressed to
attempt to rationalize why it is that stress alters immunity Steven F. Maier, Department of Psychology, Campus Box 345, University
and why this might be adaptive and functional, or might of Colorado, Boulder, CO 80309-0345.

1004 December 1994 • American Psychologist

Copyright 1994 by the American Psychological Association, Inc. O003-O66X/94/$2.0O
Vol. 49. No. 12, 1004-1017
be able to exert control over some aspects of the immune types of cells that have to interact with each other in very
response. Conversely, in order to accomplish this function, circumscribed ways. It is a dynamic process over time,
the CNS must receive information about events in the not a discrete or punctate response.
body (e.g., an infectious agent has penetrated the skin) The immune system is so-called from the Latin term
and the status of immune processes. Indeed, the immune immunis, meaning "exempt," and is the body's defense
system serves as a diffuse sensory organ to provide the against invading pathogenic microorganisms and tumors,
brain with a variety of input. Thus the immune system as well as being an important component of tissue repair
controls neural function, and the CNS controls the im- processes after injury. It is divided into innate (or non-
mune system. Of course, the existence of neural-immune specific) and specific acquired immunity. Innate immu-
interactions permits behavioral-psychological events to nity refers to one's resistance to pathogens, which is pres-
enter the matrix; if neural processes regulate immune ent from birth and which operates in a nonspecific way
processes, then there is a pathway by which psychological without regard to the exact nature of the pathogen (e.g.,
factors could impact immunity. Conversely, if immune whether it is a pneumococcus bacterium, a polio virus,
processes alter neural function, then they can also poten- or some other). There are a wide variety of innate immune
tially impact on behavior, emotion, and thought. PNI, defenses. Some are anatomical (e.g., the skin prevents the
then, is the study of these complex interactions between entry of many pathogens and its acidity limits bacterial
neural, immune, and behavioral processes. growth), some are physiological (e.g., mucus contains
The multidisciplinary nature of PNI makes this re- substances that can destroy bacterial cell walls), and some
view somewhat complex. We ask the reader to bear with are phagocytic (e.g., macrophages can engulf and destroy
us, for we believe that this field is of great potential sig- microorganisms that they contact). Perhaps the most im-
nificance for psychologists. A brief preface might help portant innate defenses are provided by the inflammatory
orient readers and keep them aware of why specific topics and acute phase responses, which will be described later.
are included and presented in the order that they are. We Specific immunity is acquired, rather than innate.
begin with a brief description of the immune system, as It involves two separate but related processes—recogni-
many readers may not be familiar with its organization, tion of foreign, "nonself" substances called antigens (de-
and without some understanding, it is not possible to see rived from "antibody generator") and destruction (re-
how interactions between behavioral and immunological moval of antigen). T and B lymphocytes are critical to
processes might be adaptive. (Readers familiar with basic these processes. T cells arise from progenitor cells in the
immunology can easily skip this section.) We next de- bone marrow and migrate to the thymus where they ma-
scribe why it is that the immune system is now thought ture. After maturation, the T cells circulate through the
to be under neural regulation. After establishing the basis blood and lymph and often reside in secondary immune
for thinking that there is a communication pathway be- organs, such as the spleen, and lymph nodes. Each T cell
tween the CNS and the immune system, we briefly review has an exquisitely selective receptor on its surface that
what is known about psychological modulation of im- can recognize and bind only a single antigen. A given T
mune function. We then consider the other direction in cell has many receptor sites, but they are all specific for
the bidirectional pathway between the immune system a same single antigen.
and the CNS and review connections from the immune T cells cannot recognize antigens by themselves. In-
system to the CNS and immune modulation of behavior. stead, the antigen must be presented to T cells in a pro-
The final section focuses on the functional significance cessed form. Antigen processing and presentation is most
or the reasons why—why would it be reasonable for stress, often accomplished by immune cells called macrophages.
for example, to impact on immunity and why the ob- Macrophages engulf and digest the antigens. They then
served pattern of behavioral changes that occur during excrete chunks of the digested antigen, and these antigen
immune challenge are both logical and adaptive. This fragments bind onto the exterior surface of the macro-
section delves into the evolution of immunity, the costs phages. It is this processed macrophage-bound form of
of immunity (any biological process has costs as well as antigen with which the T cells interact. Thus the mac-
benefits), and issues of energy balance to attempt to come rophages bearing the antigen must contact those few T
up with a rational set of reasons. The point of view that cells that happen to have the receptor for that antigen.
we develop is speculative, frankly, but it makes sense to Because there are on the order of 10'5 different T-cell
us. We end with some conclusions and the hope that we receptors in humans and each T cell has only one type
have convinced the readers that this field is on to some- of receptor, it follows that there cannot be very many T
thing. cells with a receptor for any particular antigen. Obviously,
to be able to defend oneself against foreign invasion, one
The Immune System needs to quickly create many T cells with a receptor for
The purpose of this section is to provide the necessary the antigen that has now invaded the body, so that the
basics to understanding interactions between behavior and antigen can be attacked effectively.
immune function. The major thought is that the specific T cells circulate in an inactive form, so the first task
immune response to an invading pathogen (virus, bac- is to activate the T cells with the appropriate receptor. To
teria, etc.) is a process that extends over many days and do this, the macrophages release a cytokine called inter-
requires complex coordination between many different leukin-1 (interleukin refers to chemicals that are used to

December 1994 • American Psychologist 1005

 communicate between leukocytes or white blood cells), cific to the antigen have to multiply greatly before a de-
which activate T cells. There are several different types fensive response can occur, and the biology of cell pro-
of T cells. One, the T helper, becomes activated and se- liferation is such that it requires several days. Fortunately,
cretes other cytokines that control the progress of the the memory processes built into the specific immune re-
immune response. For example, the activated T helper sponse allow a further encounter with the antigen to be
cell releases interleukin-2, which promotes multiplication dealt with much more rapidly. This is where the specific
and maturation of T cells that are specialized to fight the response is most effective.
antigen that began the process. The cytokines released
by the T helper cells also help cytotoxic T cells to multiply, Connections From the Central Nervous
if they are specialized for killing the invading pathogen System to the Immune System
(i.e., have the appropriate receptor), whether it is an an- The purpose of the following discussion is to provide the
tigen-bearing microbe or antigen-infected cell. Thus the basics of how it is that we know that the brain regulates
effector of this type of immune defense, called cellular the immune system. Two conditions would have to be
immunity, because the killing is being done by a cell, is demonstrated. First, the brain would have to make phys-
the activated cytotoxic T cell with a receptor that can ical contact with the immune system in some way. Second,
detect and bind the antigen. Note that this whole process alterations in the activity or integrity of these connections
takes several days, inasmuch as it extends from detecting would have to affect the course of immune responses to
an invader to creating an army of cells to fight it. Finally, antigens.
memory T cells develop that have a very long life span The first question one might ask concerns how the
and can rapidly recognize the antigen if it is encountered brain is able to connect with and control other peripheral
again. processes. The brain has two ways to control peripheral
B cells mature in the bone marrow. They also have organs and processes. One is through the peripheral ner-
specific receptors on their surface, but they have a different vous system. The autonomic nervous system, composed
structure than the T-cell receptor and are called antibody of sympathetic and parasympathetic branches, innervates
molecules. When a B cell encounters the antigen that can visceral organs such as the stomach and the heart. Re-
bind its receptor (membrane-bound antibody), it begins search conducted during the past dozen years (e.g., D. L.
to divide, and its progeny differentiate into memory B Felten, Ackerman, Wiegand, & Felten, 1987) demon-
cells and plasma cells. This process is aided by a large strated that the sympathetic nervous system innervates
number of different cytokines secreted by activated T immune organs such as the thymus, bone marrow, spleen,
helper cells. This is a complex multiday process that is and, even, lymph nodes. Sympathetic nerve terminals
orchestrated by interleukins secreted by T helper cells. release the catecholamine, norepinephrine, and immune
The T helper cells release different interleukins in a very organs and cells contain catecholamine receptors. Fur-
specific sequence across days, which control the matu- thermore, the terminals of sympathetic nerves in these
ration of memory and plasma B cells. The plasma cells immune organs make contacts with lymphocytes them-
develop a new form of its surface receptor that does not selves, and these contacts have the ultrastructural features
remain membrane bound but rather is secreted as a sol- of synaptic contacts (S. Y. Felten & Felten, 1991). Thus
uble receptor into the circulation. These antigen-specific the brain is physically connected to the immune system.
receptors released from plasma cells are called antibodies. The other way in which the brain can communicate
The antibody will bind the antigen wherever it comes to peripheral organs is by releasing factors that cause en-
into contact with it and functions as the effector of this docrine glands to secrete hormones into the circulation,
form of immunity, called humoral immunity. The process thereby enabling the hormones to reach the various organs
of binding the antigen by antibody is sometimes sufficient and bind to hormone receptors on the organs. An example
to eliminate it. In addition, the bound antigen-antibody that will be of particular relevance later in this article
complex can activate a blood system called complement, concerns hormones produced by stress. Many of the
which can destroy the antigen. The process of generating bodily effects of stress are produced by steroid hormones
antibody takes roughly five or more days. If the antigen called glucocorticoids, which are released from the outer
is encountered again in the future, the antigen-specific portion (cortex) of the adrenal glands. Indeed, the pres-
memory cells can produce a much more rapid and potent ence of stress is often defined by the existence of high
reaction. levels of these hormones in the blood. The sequence of
In sum, innate immune mechanisms operate as a events is that both physical and purely psychological
first line of defense against invading pathogens. However, stressors lead cells in the paraventricular nucleus of the
the nonspecific nature of the processes allows some hypothalamus to synthesize and release a substance called
pathogens to escape. Specific immunity enables the de- corticotropin releasing hormone into the portal blood
velopment of defense responses to a staggering number system at the base of the brain. This hormone then reaches
of potential antigens. Unfortunately, the specificity of the the anterior lobe of the pituitary gland where it leads to
mechanisms involved requires that the response will be the synthesis and release of adrenocorticotrophic hor-
slow and delayed, because receptor specificity means that mone into the blood. This pituitary hormone ultimately
there cannot be many lymphocytes with receptors for arrives at the adrenal gland where it causes release of the
any particular antigen. This means that lymphocytes spe- glucocorticoids. The concept is that the brain released

1006 December 1994 • American Psychologist

something that led to hormones being released into the an immunosuppressive drug in a Pavlovian manner ac-
general circulation. T and B cells have receptors for many quired the ability to suppress antibody responses to an
of these hormones, including the stress hormones just antigen. A large amount of subsequent research (see Ader
noted (Plaut, 1987). Activation of the sympathetic ner- & Cohen, 1993, for a review) has confirmed the generality
vous system by stressors also leads to the release of cat- of this finding across conditioned stimuli, immuno-
echolamines (i.e., norepinephrine and epinephrine) from modulatory unconditioned stimuli, and immune mea-
the inner portion (medulla) of the adrenal gland into the sures. The initial studies used animal subjects, but con-
blood; lymphocytes have catecholamine receptors as well. ditioned modulation of immunity has also been dem-
It is important to appreciate that immune cell function onstrated in humans (Smith & McDaniels, 1983).
is altered by the action of these hormones and transmitters Furthermore, both immune suppression and enhance-
at receptors on the lymphocytes. ment of the immune response can be conditioned as well
In sum, the anatomical arrangements are such that (Solvason, Ghanta, & Hiramoto, 1988).
the brain could control immune cells and organs in the Two questions are at the heart of current research
same ways it controls other peripheral structures. How- concerning conditioned immunomodulation. One con-
ever, the fact that the brain can does not mean that it cerns the mechanisms involved: Are the immune changes
does. Is there evidence that the brain does participate in "directly" conditioned or is something else conditioned
controlling normal immune responses? If the brain par- (e.g., fear, anxiety, aversion, glucocorticoid release) that
ticipates in the regulation of the immune system, then is then responsible for the immune alterations? The sec-
brain lesions and stimulation at some brain site(s) ought ond involves the potential practical implications of con-
to modulate some aspect(s) of immune responses. The ditioned immunomodulation. Could conditioned im-
hypothalamus plays a key role in integrating neural con- mune responses occur in real-life settings and influence
trol of visceral processes in general, and so it is not sur- disease processes, and could conditioning procedures be
prising that lesions of the hypothalamus alter the course used in clinical settings?
of a variety of immune processes. This is true for in vivo Indeed, there is promise of clinical application. In a
measures of immune function such as antibody produc- classic study, Ader and Cohen (1982) explored the de-
tion and rejection of tissue transplants (Macris, Schiavi, velopment of systemic lupus-erythematosus-like autoim-
Camerino, & Stein, 1970) and in vitro measures such as mune disease in genetically prone mice. The onset of lu-
stimulated lymphocyte proliferation (Roszman, Cross, pus symptoms can be retarded and survival prolonged by
Brooks, & Markesbery, 1985). Moreover, lesions in other immunosuppressive drugs such as cyclophosphamide. A
regions can also alter immune function (Nance, Rayson, neutral stimulus (saccharine flavored water) was paired
& Carr, 1987). Conversely, electrical stimulation of hy- with cyclophosphamide in a Pavlovian manner. That is,
pothalamic regions has been reported to augment several rats were allowed to drink the solution and then were
immune parameters (Korneva, 1967). With regard to the immediately injected with cyclophosphamide. Weekly
autonomic nervous system, chemical destruction with 6- treatments with cyclophosphamide are required to delay
hydroxydopamine can impair some aspects of immune the onset of lupus symptoms; administration of cyclo-
function (Livnat, Felten, Carlson, Bellinger, & Felten, phosphamide every other week has no measurable effect.
1985). The point is that destruction or stimulation of However, Ader and Cohen (1982) found that the sac-
neural pathways that are connected to the immune system charine solution could be substituted for cyclophospha-
do, in fact, alter the function of the immune system, and mide every other week and delay the onset of the autoim-
so the connection between the CNS is of real significance, mune disorder, but only if it had been paired with cyclo-
not merely an anatomic curiosity. Similarly, blocking the phosphamide (also see Klosterhalfen & Klosterhalfen,
hormone receptors on lymphocytes alters the course of 1983). Moreover, reexposure to the saccharin cue after
immunity (Blalock, Smith, & Meyer 1985). the cyclophosphamide treatment was discontinued pro-
longed survival (Ader, 1985). Because cyclophosphamide
Psychological Modulation of Immunity is quite toxic, as are many chemotherapeutic agents, this
The interactions between the CNS and immunity sum- sort of use of conditioned immune change may be of
marized above suggest that psychological events should considerable benefit. In more recent work (Grochowitz
be capable of altering immunity, because such events both et al., 1991) it has also been suggested that conditioned
alter and are expressed in neural activity and neural events immunosuppression can delay rejection of tissue trans-
make contact with the immune system. Research con- plants, suggesting a use in organ transplantation.
cerning psychological modulation of immunity has cen- In more general terms, conditioned immuno-
tered on two topics—classical conditioning of immunity suppression might be expected to occur whenever an or-
and the impact of stress. ganism repeatedly encounters an immunosuppressive
agent in a particular environment. Chemotherapy for
Classical Conditioning cancer is a particularly important example. Chemo-
Processes under the control of the CNS are generally therapeutic drugs are chosen for their ability to inhibit
modifiable by associative processes. Modern interest in the cell division of rapidly replicating cells, among which
the conditioning of immune responses stems from a study are cancer cells. However, these agents also inhibit the
by Ader and Cohen (1975) in which a taste paired with replication of other rapidly dividing cells such as immune

December 1994 • American Psychologist 1007

 cells. Thus chemotherapeutic agents are immunosup- also been examined after exposure to a stressor (Zwilling
pressive. Indeed, cyclophosphamide is an often used drug et al., 1990), as has the secretion of soluble mediators
for cancer chemotherapy. The repeated chemotherapy is such as the interleukins and the development of surface
typically done in the same room in the same hospital receptors for them (Weiss, Sundar, Becker, & Cierpial,
setting, and so conditioned immunosuppression might 1989). Stressors have also been shown to alter the migra-
be expected to develop. Consistent with this argument, tion pattern of immune cells between and into compart-
Bovjberg et al. (1990) found that women who had un- ments of the immune system such as spleen, thymus, and
dergone a number of chemotherapeutic treatments for lymph nodes (Fleshner, Watkins, Bellgrau, Laudenslager,
ovarian cancer displayed immunosuppression after simply & Maier, 1992). Indeed, it is difficult to think of an aspect
being brought to the hospital prior to chemotherapy. This of immunity that has not been found to be altered by
learned immune change could easily exacerbate the un- some stressor.
conditioned effects of the drugs on the immune system, One question that psychologists often ask when pre-
further compromising the ability of the patients to fight sented with work such as this concerns whether the in-
the cancer. The psychologist of learning could suggest fluence of stress on immunity merely reflects a simple
many procedures that should reduce the conditioning physical impact of painful or arousing events or whether
(e.g., giving the chemotherapy in different environmental the interaction between stress and immunity embodies
contexts, rather than always in the same context). Such some of the more subtle aspects of psychological modu-
work is currently under way. lation of stress effects that can be observed in other areas
Stress and Immunity of stress research. The answer is that the impact of stress
on immunity cannot be explained in simple physical
Much of the current interest in PNI stems from the pos- terms. As an example we chose some experiments from
sibility that exposure to environmental stressors might our own research. If several male rats are allowed to live
interfere with the immune response, thereby providing a together in a large enclosure, one will become dominant,
potential link between stress and physical disease. Because that is, the alpha male. If a stranger is introduced into
stressors activate both the sympathetic nervous system the environment, the alpha male will attack the intruder.
and the hypothalamo-pituitary-adrenal axis (see Stan- The intruder begins by engaging in defensive aggression
ford & Salmon, 1993, for a recent comprehensive review), but invariably gives up and adopts species-typical defeat
it is not surprising that stressors can impact on immunity. postures. Intruders never beat residents. Being placed into
This is because, as noted earlier, plasma catecholamines the established territory of the other rat, even for a brief
released by sympathetic terminals and the adrenal me- period of time, severely inhibits the production of anti-
dulla, as well as hormones released by the pituitary and body to antigen administered before the intruder is in-
the adrenal cortex, participate in the regulation of the troduced into the territory (Fleshner, Laudenslager, Si-
immune response. mons, & Maier, 1989). This effect is all the more im-
Indeed, numerous studies conducted over the past pressive when it is recognized that antibody is not
30 or so years have demonstrated that a wide variety of measured until one to three weeks later. However, the
stressors can alter many aspects of the immune response important point is that it can be determined whether the
(see Solomon, 1969, for an excellent early example). In inhibition of antibody is produced by the physical aspects
animals, acute exposure to electric shocks (Keller, Weiss, of attack, such as being bitten or pushed, or the psycho-
Schleifer, Miller, & Stein, 1983), social defeat (Fleshner, logical factor of being defeated. This is because adopting
Laudenslager, Simons, & Maier, 1989; review by Bohus and maintaining defeat postures tend to inhibit the attacks
& Koolhaas, 1991), maternal separation (Coe, Rosenberg, from the alpha male. That is, the two are negatively cor-
& Levine, 1988; Laudenslager, Held, Boccia, Reite, & related. Thus one can ask whether it is engaging in sub-
Cohen, 1990), rotation (Esterling & Rabin, 1987)), the missive behaviors or being bitten and assaulted that is
odor of a stressed conspecific (Zalcman, Richter, & An- correlated with the reduction in antibody formation. It
isman, 1989), immersion in cold water (Jiang, Morrow- was the adoption of submissive behaviors—the correlation
Tesch, Beller, Levy, & Black, 1990), restraint (Bonneau, between time spent in defeat posture and antibody was
Sheridan, Feng, & Glaser, 1991a, 1991b), handling -.80. Indeed, there were a small number of animals who
(Moynihan et al., 1990), intraperitoneal injection of saline did not submit at all and received numerous bites. An-
(Moynihan, Koota, Brenner, Cohen, & Ader, 1989), and tibody in these animals was unaffected.
loud noise (Monjan & Collector, 1977) have all been On the human level, acute stressors, such as final
shown to suppress some aspect of immunity. Chronic examinations (Dorian et al., 1982), battle task vigilance
stressors such as crowding (Rabin, Lyte, Epstein, & Cag- (Palmblad et al., 1976), and sleep deprivation (Palmblad,
giula, 1987) have also been examined. Immune measures Petrini, Wasserman, & Akerstedt, 1979), have been shown
have ranged from effectors of cellular immunity, such as to impact on immune parameters. More chronic condi-
the development of cytotoxic T cells to an antigen, to tions, such as divorce (Kiecolt-Glaser, Fisher, et al., 1987),
effectors of humoral immunity, such as the development bereavement (Schleifer, Keller, Camerino, Thornton, &
of antibody to an antigen, to nonspecifc measures, such Stein, 1983), and Alzheimer caregiving (Kiecolt-Glaser,
as mitogenic stimulation of lymphocyte proliferation. The Glaser, et al., 1987), also alter measures of immunity.
function of specific cell types, such as macrophages, has The literature demonstrating these links has been the

1008 December 1994 • American Psychologist

subject of numerous recent reviews (Ader & Cohen, 1993; iety, as measured in the hospital environment (Fredrik-
Cohen & Williamson, 1991; Weiner, 1991) and will not son, Furst, Lekander, Rotstein, & Blomgren, 1993). Thus
be reviewed here yet again. However, it should be em- different stressors and other psychological events that do
phasized that the study of the psychological modulation impact on immunity may do so in different ways, pro-
of immunity has only scratched the surface of the rela- ducing different outcomes.
tionships that probably exist. Obviously, psychological Yet another complexity stems from the fact that
factors cannot directly contact white blood cells and are the specific immune response involves a complex cascade
capable of altering immunity because they modulate au- of events that extends over many days. The peripheral
tonomic function and the release of peripheral hormones products of stress play numerous roles in regulating this
that modulate immunity. Thus any psychological event cascade, and so the effects of stress will of necessity be
that alters these neural and hormonal factors is capable variable. There will be conditions under which stressors
of modulating immunity. As an example, mood states interfere with immunity, have no effect, or even enhance
such as depression are associated with dysregulation of immune measures (e.g., Croiset, Heijnen, Veldhuis,
the pituitary-adrenal system (Holsboer, Von Bardeleben, deWied, & Ballieux, 1987; Lysle, Cunnick, & Rabin,
Gerken, Stalla, & Muller, 1984), and depressed individuals 1990; Lysle, Lyte, Fowler, & Rabin, 1987; Rinner,
often have chronically elevated levels of glucocorticoids Schauenstein, Mangge, & Porta, 1992). The effects ob-
in blood (Carrol, 1978). As would therefore be expected, served will depend on the precise blend, duration, and
immune system dysfunction has often been reported to timing of hormones and sympathetic activation pro-
exist in depressed populations (Irwin, Daniels, Bloom, duced by the stressor. For example, Fleshner et al., 1992,
Smith, & Weiner, 1987; Schleifer et al., 1984). Emotions and Zalcman, Minkiewicz-Janda, Richter, and Anisman,
such as anger and anxiety might also be expected ulti- 1988, have found that a stressor will interfere with the
mately to impact on immunity (Fleshner et al., 1993). production of antibody to an antigen (measured 1 -3
Indeed, thoughts ought to be capable of altering immu- weeks after antigen administration) only if stress occurs
nity. Thinking about or encountering a learned signal for within narrow time ranges relative to antigen exposure.
an aversive or unpleasant event can activate autonomic In addition, different aspects of the immune response
outflow and the release of hormones and so are capable are differently affected by autonomic function and hor-
of impacting on immunity. For example, presentation of mones, and so the effects of a particular stressor on im-
a previously neutral stimulus (a light, a tone, etc.) that munity might be quite selective, impacting on one kind
has come to signal an aversive event can suppress a num- of immunity but not on another. It is quite possible for
ber of aspects of immunity (Lysle, Cunnick, Kucinski, a stressor to alter antibody generation, for example, but
Fowler, & Rabin, 1990). not alter T-cell proliferation (Maier & Laudenslager,
This line of reasoning suggests more than the fact 1988). Research exploring these sorts of interactions is
that emotions and thoughts impact on immunity. It fur- in its infancy. However, efforts in these directions will
ther suggests that these effects will be subtle and selective. doubtlessly uncover a rich matrix of psychological in-
Stressors are not generic events that have identical pe- fluence.
ripheral outcomes. Different stressors produce different This discussion also leads to a caution. Not infre-
mixes of autonomic activation and hormones (Mason, quently, investigators have drawn sweeping conclusions
1971). For example, one stressor might lead to intense such as "stress suppresses immune function" from studies
autonomic activation and consequent plasma catechol- that have measured but one aspect of immunity at one
amine release but relatively little activation of the pi- point in time. This is akin to measuring a single aspect
tuitary and adrenal glands and their hormones. Another of neural function (e.g., the release of a single transmitter
might produce the opposite pattern. In addition, the time in a single nucleus) and making claims about what "stress
course of these changes will differ for different stressors, does to the brain." Furthermore, these measures have
emotions, and thoughts. Moreover, personality, coping often been nonspecific and assess some intermediate as-
processes, and the like modulate the autonomic and pect of the immune response (e.g., production of inter-
hormonal consequences of exposure to stressors (Ursin leukins, proliferation of cells in response to mitogens)
& Olff, 1993). They too will then modulate the immune rather than an effector endpoint that detects and clears
consequences of stressors (e.g., Mormede, Dantzer, Mi- antigen, recognizes and destroys tumors or virally infected
chaud, Kelly, & LeMoal, 1988). For example, the impact cells, and so forth. The immune system contains a high
of final examinations depends on the student's level of degree of redundancy, and so the fact that an event might
loneliness (Kiecolt-Glaser et al., 1984), and the effects alter an intermediate product or step does not provide
of divorce depend on the degree of prior attachment to convincing information about whether the event in ques-
the partner (Kiecolt-Glaser, Fisher, et al., 1987). This tion would impact on a normal endpoint of the immune
sort of modulation by psychological variables is not re- response (e.g., the production of antibody). Indeed, there
stricted to the effects of stress. For example, we noted are instances in which a condition influences one but not
earlier that repeated chemotherapeutic treatments result the other (Cunnick, Lysle, Armfield, & Rabin, 1991;
in conditioned immunosuppression to cues that regu- Sheridan et al., 1991). It will take a considerable amount
larly precede chemotherapy. However, this conditioning of research to distill the truly general principles from the
may be restricted to patients who are high in state anx- specifics.

December 1994 • American Psychologist 1009

 Implications for Disease major effects on disease. The next few years of research
The popular press is replete with conclusions concerning will determine the role of these links in human disease
stress-induced immunomodulation as the mechanism and will explore whether psychological interventions are
mediating between stress and disease. There is no question capable of modifying the course of disease. A particularly
that stress can alter immunity and that stress can alter promising study of this sort was reported by Fawzy et al.
disease, but there is actually very little work directed at (1990). They found that cancer patients who received
determining whether the effect on immunity is causal to psychiatric group intervention showed an increase in NK
the effect on disease. This is an issue because stressors cell activity, compared with untreated control partici-
can modulate many factors other than immunity that pants. Furthermore, this change was correlated with
can impact on disease directly without intervention by changes in anxiety. Much more work like this is needed
the immune system. Some of these factors are biological. and is under way.
For example, numerous experiments have shown that ex- Connections From the Immune System
posure to a stressor can accelerate the growth of implanted to the Central Nervous System
tumors (Sklar & Anisman, 1981). The natural assumption
has been that this is because the stressor impacted on Thus far we have been concerned with the influence of
immune processes involved in tumor control. However, the CNS and psychological processes on immune func-
stressors also alter blood flow, levels of hormones such as tion. We now turn to the other direction of influence in
prolactin, body temperature, and so forth, all of which the bidirectional interactions between behavior and im-
can directly affect the rate of growth of a tumor. In ad- munity—the impact of immune responses on brain and
dition, human studies allow mediation by behavioral behavior. We begin with the immune-to-CNS link. The
variables. Most chronic stressors (e.g., bereavement) immune response occurs outside the nervous system in
doubtlessly change behavior patterns (e.g., eating, sleep- the periphery in response to peripheral antigen. For the
ing, drug intake, interaction with others), which can brain to participate in the regulation of this response it
modulate the course of a tumor. must therefore receive information that an immune re-
sponse is in fact occurring. Moreover, the generation of
Clearly, careful analytic work is required to tie the a specific immune response is a complex affair extending
impact of a stressor on disease to mediation by immunity. over a number of days and involving the interaction of
Fortunately, there are a small number of studies that do many different cell types and mediators. Thus it might
just that. An elegant example is provided by a series of even be necessary for the CNS to receive detailed infor-
studies conducted by Ben-Eliyahu, Yirmiya, Liebeskind, mation about the course of the response. As would be
Taylor, and Gale, 1991. They worked with a tumor cell expected, both the electrical and chemical activity of the
that is known to be very sensitive to regulation by natural brain do change as immune responses occur. For example,
killer (NK) cells. (NK cells are a type of lymphocyte that Besedovsky, Sorkin, Felix, and Haas (1977) found hy-
does not have to be activated for it to be able to destroy pothalamic neural activity to increase at the time of peak
cells and responds in a relatively nonspecific way to a B-cell proliferation to an administered antigen. Similarly,
variety of tumor- and virally infected cells.) This gave neurotransmitters in the hypothalamus, such as norepi-
them the advantage of knowing which aspect of immunity nephrine, also show profound changes at this time (Carl-
to measure. They implanted these tumor cells in rats and son, Felten, Livnat, & Felten, 1987). The antigen used in
found that a stressor would exaggerate tumor growth and these experiments was a harmless protein, not an agent
metastasis if it was given within 24 hours of tumor im- that produces illness or disease. Thus it was not illness
plantation. The stressor also reduced the ability of NK or disease that altered neural activity; rather, the activity
cells to kill tumor cells as measured directly in an assay. of the brain changed with the progress and course of an
The question was then whether the effect on NK cells immune response per se.
mediated the effect on tumor growth. Ben-Eliyahu et al.
approached this question in two ways. First, they blocked How could this occur? After all, the cells of the im-
the effect of the stressor on NK cells using a pharmaco- mune system, such as T cells, B cells, and the like, have
logical agent that had no direct effect on the tumor. How- only limited access to the brain, because of the blood-
ever, the agent blocked the effect of the stressor on the brain barrier. This is a key question that has generated
tumor. They then used an antibody directed against NK considerable excitement recently. Much of the focus has
cells to produce the same change in NK activity that the been on the soluble proteins released by immune cells,
stressor had produced, but without administering the the cytokines, during the course of the immune response.
stressor. This enhanced tumor growth. Thus the change These have always been thought of as messengers between
in NK cells was both necessary and sufficient to produce cells of the immune system, but they may also converse
the facilitation of tumor development. The same sort of with the nervous system. Space does not permit a dis-
conclusion can be drawn from work by Bonneau et al. cussion of the many cytokines; thus only one, interleukin-
(1991a, 1991 b) using the herpes simplex virus. More work 1 (IL-1), will be used as an example.
of this sort will be needed before we can make confident IL-1 is synthesized and secreted by a number of dif-
assertions about links to disease. Nevertheless, it is clear ferent cells. However, activated macrophages are the major
that stress can alter immunity and that this can exert source of IL-1 during the specific immune response.
Macrophages are activated either by engulfing antigen or

1010 December 1994 • American Psychologist

by a number of chemical signals that bind to surface re- Tilders, & Besedovsky, 1987). It is important to recognize
ceptors on the macrophage. Activation of macrophages that in these studies an immune response was elicited by
with virus or bacterial endotoxin produces a release of administration of a harmless protein, not a pathogen.
IL-1 and a subsequent alteration in the electrical activity Thus it is the immune response itself that produced what
of the brain (Saphier, 1992), as well as metabolism of the appeared to be a stress response.
neurotransmitters norepinephrine, serotonin, and do- The foregoing description might suggest that behav-
pamine in a number of discrete brain regions (Dunn, ioral manifestations of stress will appear at some stages
Powell, & Small, 1989). These CNS changes to virus and of the immune response. Animal experiments support
endotoxin are known to be produced by IL-1, because this contention. Animals exposed to fear or anxiety-
an antagonist to the IL-1 receptor blocks them and the arousing stimuli engage in a well-characterized set of be-
peripheral or central administration of IL-1 produces haviors including reductions in the following: activity,
them (Dunn, 1993). Thus IL-l and other cytokines may tendencies to explore novel objects, social interaction,
well be the communicators between the immune system food and water intake, and willingness to engage in sexual
and the brain, with potent effects on neural activity. behavior. The administration of IL-1 or of substances that
However, an interesting question remains. IL-1 and stimulate immune cells, such as macrophages, to release
other cytokines are large lipophobic proteins and are IL-1 produce all of these behavioral changes (Dantzer,
therefore unlikely to readily cross the blood-brain barrier. Bluthe, Kent, & Goodall, 1993).
Several possibilities have been proposed with regard to The behavioral effects of immune products are not
how cytokines could then alter neural activity. One is limited to stress and stress-related behaviors. For example,
that there is an active transport mechanism to carry IL- we have recently begun to study potential relationships
1 across the barrier (Banks, Kastin, & Durham, 1989). between immune processes and pain. The CNS contains
Another is that IL-1 is able to cross the vascular endo- circuitry that, when activated, enhances the pain that re-
thelium in regions of the brain where the barrier is weak sults from a painful stimulus, above and beyond that
or absent, such as in the organum vasculosum lateralis which the stimulus normally produces (see Coderre, Katz,
terminalis (Katsuura, Arimura, Koves, & Gottchall, Vaccarino, & Melzack, 1993, for a review). These mech-
1990). A final possibility is that IL-1 can stimulate pe- anisms in the brain and spinal cord are especially im-
ripheral nerves, such as the vagus, that send afferent input portant because they are thought to be involved in the
to the brain (Watkins et al., 1994). This suggests that the production of chronic pain pathologies that create so
immune system may truly act as a sensory organ con- much human misery. All of the studies of these neural
veying information to the brain. mechanisms had activated them through direct electrical
or chemical stimulation of the neurons involved. We
Immune Modulation of Behavior wished to determine whether there were naturally occur-
The focus in PNI has been on psychological modulation ring environmental events that would activate these neural
of immunity. However, there are recent suggestions that hyperalgesia circuits. We reasoned backwards and asked
events in the immune system can also modify behavior. whether there were circumstances under which it would
Behavior, thoughts, and emotions vary across time in ways be adaptive to experience exaggerated pain from a painful
that often seem unpredictable. At the very least, psycho- stimulus. It seemed to us that illness or injury might be
logical processes sometimes change, even though events such a condition. During these times it might be useful
in the external environment seem to have been constant. to be especially attentive to sites of pain. Pain could serve
Internal events are doubtlessly responsible for some of to guide recuperative behaviors, such as licking the site
these dynamics, and events in the immune system may of injury (Bolles & Fanselow, 1980), and lead to the con-
well be a previously unsuspected part of this matrix. servation of energy during illness (see below). We con-
The first hint of an immune-to-behavior causal link ducted a series of experiments in which rats were made
was provided by studies indicating that there is an increase ill by administering agents that induce illness, and in all
in autonomic nervous system activity and the levels of cases a long-lasting hyperalgesic response to pain stimuli
pituitary-adrenal stress hormones in blood at various was induced (Wiertelak et al., 1994). This does not by
stages of an immune response to an antigen (Besedovsky, itself implicate the immune system. However, in further
Sorkin, Keller, & Muller, 1975). That is, the occurrence studies we demonstrated that the hyperalgesia occurred
of an immune response leads to the peripheral physio- because the illness-inducing agents stimulated macro-
logical equivalent of a stress response. In addition, the phages to release IL-1 (Maier, Wiertelak, Martin, & Wat-
pituitary-adrenal response is activated by the same kins, 1993) and that IL-1 did activate the hyperalgesia
mechanisms that activate their response to stressors. The circuitry in the brain and spinal cord (Watkins et al.,
paraventricular nucleus of the hypothalamus is induced 1994). Note that injury as well as pathogenic agents will
to release corticotrophin releasing factor into the portal produce the release of cytokines such as IL-1 (see below).
blood; that is, the immune response communicates with We are unaware of comparable human studies.
the brain in order to release the peripheral stress hor- However, it would be intriguing to determine whether
mones. The communication link is provided by IL-1 and mood, emotional reactivity, pain, attention, and other
perhaps other cytokines released by immune cells during processes might be affected by the status of the immune
the immune response (Berkenbosch, VanOers, Del Rey, system, even when the immune system is merely re-

December 1994 • American Psychologist 1011

sponding to harmless as well as pathogenic agents that Assume that a microbe enters the body or that tissue
we all encounter in our daily experience. This could ac- injury occurs. This activates a number of systems that
count for some of the seemingly unmotivated mood ultimately lead a variety of cells to migrate to and enter
swings that we all experience. the injured or affected area. The macrophage is perhaps
the most important cell. Chemical signals that macro-
Functional Significance—Inflammation phages resident in the area receive from the initiating
and the Acute Phase Response events in inflammation and signals that macrophages
newly arrived in the area receive from resident macro-
We next turn to a consideration of some of the "whys."
phages activate the macrophages to produce a variety of
Why would it be useful for stressors and other behavioral
products. Some are enzymes that help destroy pathogens
events to impact on immunity, and why should immune
and cellular debris produced by injury, and others regulate
responding alter behavior? In particular, how could it be
the activity of a number of other cells. Macrophages can
adaptive to interfere with the immune response? Is this
also engulf and destroy microbes. An important point is
simply pathophysiological or does stress-induced im-
that many macrophage products act back on the mac-
munosuppression play a physiological role? Another way
rophage itself in a positive feedback fashion. For example,
to inquire into this issue is to wonder why stress hormones
IL-1 released from macrophages stimulates IL-6 released
are immunosuppressive. Conversely, why should immune
from fibroblasts and macrophages themselves, which
responding produce what looks like a stress response?
stimulates further IL-1 release. IL-1 can even induce itself,
Until now we have focused on the specific immune upregulating IL-1 gene transcription (Schindler, Bhezzi,
response. However, specific immunity is a more recent & Dinarello, 1991). This is noted here because many
evolutionary development than innate immunity; it macrophage products are highly dangerous and can kill
evolved from processes present in innate immunity. In- healthy tissue as well as pathogens; proteases and lyso-
deed, specific immunity is present only in vertebrates somal enzymes are examples. The message is that the
(Reinisch & Litman, 1989). The stress response is far inflammatory response has to be limited in some way
older in evolutionary time. Innate immune defenses are because of the positive feedback properties involved.
also quite old. For example, phagocytic cells that engulf
and destroy particles are present in the sponges, the most The Acute Phase Response
primitive multicellular organism known. Thus an un-
derstanding of the physiological role of stress-induced The inflammatory process is localized at the site of injury
immune changes might be illuminated by considering or infection. It can trigger a general or systemic response,
innate immunity and the role of stress. the acute phase response, that both supports the local
reaction and fights infections that are no longer localized
Inflammation and the acute phase response (see and have become systemic. Some of the support involves
Baumann & Gauldie, 1994, for a review) are particularly delivering more needed building blocks, such as amino
important aspects of innate immunity. These are innate acids, to the site of inflammation. Other aspects of the
or nonspecific because the cells involved do not respond support involve the delivery of additional mediators, and
only to a specific antigen or molecule but act on a broad still other aspects entail more global metabolic changes
range of substances. There is no antigen-driven multipli- that facilitate the cellular processes of destruction of
cation of cells specific for an antigen. pathogen and repair of tissue, at the same time limiting
Inflammation pathogen growth. IL-1, IL-6, and tumor necrosis factor
(TNF) produced by macrophages at the site are critical
Inflammation is a local response to tissue injury, micro- triggers of this acute phase response (Baumann & Gaul-
bial invasion or infection, and irritants. The purpose of die, 1994). These products stimulate further cytokine re-
the inflammatory response is to limit damage caused by lease from local endothelial cells and fibroblasts, and when
injury to a local site. In the case of a pathogen, inflam- they accumulate in sufficient quantity they enter the cir-
mation limits its spread and kills and removes the patho- culation and orchestrate the elements of the acute phase
gen through phagocytosis, primarily by macrophages and response.
neutrophils called to the site. In addition, inflammation The acute phase response involves numerous pro-
involves the initiation of repair processes designed to fix cesses. Leukocytes are produced in bone marrow and
any tissue damage. This involves proliferation of con- then enter the circulation and can migrate to the local
nective tissue, production of collagen and elastins, and site. In addition, a number of enzymatic reactions cause
so forth. The time course of these responses is on the reductions in plasma iron and zinc, both of which are
order of hours—inflammatory responses can be observed required for the growth of certain pathogens. Acute phase
within 1 to 2 hours after infection, and the acute phase proteins are synthesized and released by the liver. These
response (see below) occurs 8 to 12 hours after local in- are a group of about 30 plasma proteins that play diverse
fection. Recall that the specific immune response requires roles, such as scavenging and removing cellular debris,
days to generate effectors that can kill antigen. Thus in- promoting destruction of bacteria by activation of com-
flammation and the acute phase response are the first line plement, for example. Importantly, fever is produced.
of defense against agents that have penetrated anatomic Fever is a phylogenetically old mechanism and is
and physiological barriers. highly adaptive. Numerous experiments have demon-

1012 December 1994 • American Psychologist

strated that reducing fever by various means decreases the organism's entire energy balance, and this can only
survival after infection. The increased body temperature be accomplished by involving the CNS, so that the array
produced by fever accelerates a number of enzymatic re- of changes from metabolism to behavior can be orches-
actions at the site of inflammation that operate sub- trated. This is another important reason why immune
optimally under normal circumstances because they can products must be able to communicate with the CNS.
be damaging to healthy tissue, slows replication of mi- Again, IL-1, TNF, and IL-6 are key elements of the com-
crobes, and enhances the rate of proliferation of immune munication (Kent et al., 1992). Clearly, the suggestion is
cells. It is important to understand that fever is not just that the behavioral consequences of IL-1 and immune
an increase in body temperature. Rather, the set point responding may function as behavioral energy conser-
for temperature is increased in hypothalamic temperature vation and may have evolved for that purpose.
control centers by IL-1 action at the brain (Rothwell, G/ucocorticoids
1992). Thus mechanisms to increase temperature are en-
gaged that decrease temperature loss (e.g., peripheral va- The level of adrenal glucocorticoids in blood rises sharply
soconstriction and huddling) and increase temperature during inflammation and is considered part of the acute
production (e.g., muscular activity involved in shivering). phase response. Again, cytokines released by macrophages
and other cells are the mediators of the increased glu-
Energy Demand and Balance cocorticoid synthesis and release from the adrenal cortex
All of this creates a tremendous energy demand. For ex- (Baumann & Gauldie, 1994). Similar to the sequence of
ample, it is estimated that each degree increase of body events described above during the specific immune re-
temperature requires from a 7% to a 13% increase in sponse, the cytokines appear to lead to a glucocortiocid
caloric energy production or metabolism, depending on response by acting at the hypothalamus, thus initiating a
the species and circumstances. Furthermore, the produc- full hypothalamo-pituitary-adrenal response.
tion of acute phase proteins and all of the cellular pro- It is important to understand that the peripheral
liferation, production of cytokines, collagens, proteases, physiological action of glucocorticoids is to mobilize en-
and so forth, require a large supply of amino acid building ergy (Sapolsky, 1992). Glucocorticoids promote break-
blocks. An energy demand of this magnitude requires down of muscle protein into amino acids, facilitate the
changes that range from metabolism to behavior; this can conversion of amino acids and liver glycogen to glucose,
only be coordinated by the CNS. antagonize the action of insulin (insulin stimulates glucose
The cytokines, such as IL-1, IL-6, and TNF, may uptake into fat and muscle and suppresses liver glucose
again be key coordinating elements, both peripherally production), and enhance fat mobilization. In sum, glu-
and through their ability to communicate with the CNS. cocorticoids potentiate glucose increases while further
At the cellular level, IL-1 promotes the breakdown of breaking down protein. Here glucocorticoids and IL-1
muscle protein into amino acids (Baracos, Rodemann, operate in concert.
Dinarello, & Goldberg, 1983), a process that is responsible These actions produce the energy demanded by in-
for the muscle soreness experienced during infection. IL- flammation and the acute phase response. A final regu-
1 also increases the availability of glucose for metabolism latory factor is required. Recall that a number of poten-
by peripheral tissues and the release of fatty acids from tially destructive substances are released during inflam-
fat stores for similar use. mation and that there are positive feedback loops inherent
Recall that IL-1 and other cytokines produce a set in the biochemistry. Something must limit this process;
of behavioral changes that are similar to those seen after an argument can be made that it is glucocorticoids. Glu-
stress: decreases in activity, exploration, social interaction, cocorticoids do inhibit or oppose a large number of the
and food and water intake. Hyperalgesia was also dis- key mediators of inflammation. It should be obvious that
cussed. All of these are also considered to be part of the IL-1, IL-6, and TNF are the key orchestrators of inflam-
acute phase response. Somnolence and increased slow mation and the acute phase response, and glucocorticoids
wave sleep can be added to this set of changes. An ex- inhibit their synthesis at the genetic level (see Barnes &
amination of the list suggests that all of the behaviors are Adcock, 1993, for a review). In addition, glucocorticoids
designed to reduce unnecessary energy expenditure, so can interfere with the synthesis of receptors for these cy-
that available energy stores can be used to fight infection tokines, thereby interfering with both the substances and
or injury. Reductions in food and water intake might not the ability of cells to respond to them. Most or all of the
make sense in this context, but consider that organisms other mediators of inflammation that were omitted from
in which these systems evolved must forage to find food this review for simplicity are also inhibited by glucocor-
and water and that digestion is energy intensive. Indeed, ticoids (Barnes & Adcock). The conclusion is that glu-
IL-1 and TNF slow digestion. Hyperalgesia should operate cocorticoids exert a strong counterregulatory effect on
to reduce activity and direct behaviors such as licking to inflammatory processes. Consistent with this conclusion,
the site of inflammation. Increased sleep, particularly slow numerous experiments have demonstrated that removal
wave sleep, should reduce the brain's glucose demand, of the adrenal potentiates inflammation after infection
the brain being the body's major user of glucose. and can lead to septic shock (Barnes & Adcock).
In short, the intense energy demands of inflamma- Indeed, one might wonder how inflammation and
tion and the acute phase response may require a shift in the acute phase response ever proceed, given that glu-

December 1994 • American Psychologist 1013

 cocorticoids do rise. Obviously, the issue is one of balance It is possible to continue this line of argument by
between stimulatory and inhibitory forces. Glucocorti- speculating about evolutionary considerations. The innate
coids restrain; they do not prevent. In addition, gluco- immune response is very old in evolutionary terms,
corticoids do not rise until a number of hours have passed probably older than the fight-flight response. Macro-
and many of the early events have already taken place. phages are extremely primitive cells. After all, defense
The cytokines that stimulate glucocorticoids through an against pathogens and tissue damage repair is required
action at the brain have to accumulate in sufficient quan- by even simple organisms. However, fight-flight can come
tity to enter the circulation and ultimately alter neural into play only in an organism that has the sensory capacity
activity. Finally, many of the actions of glucocorticoids to detect predators or other dangers, the motor capacity
are genomic and therefore require substantial periods of to make organized movements directed away from the
time to occur. danger or to damage the predator, and the integrative
The Stress Response abilities to relate the two. Perhaps the fight-flight response
evolved out of the inflammatory-response-acute-phase
We are now in a position to speculate about why stress response machinery. Evolution works by using old parts
impacts on immunity. The first question should be "What for new purposes; organisms already had a system to de-
is a stress response, really?" It is really a fight-flight re- fend against damage and infection. Energy was still
sponse, a set of changes that mobilize the organism for needed, but it had to go to a different place, muscle and
energy production. That is, it involves a shunting of energy brain. A mechanism existed to produce energy and to
toward muscular exertion and high levels of brain energy reduce the energy demand made by inflammation and
use (Sapolsky, 1992). As in the innate immune response, the acute phase response by inhibiting them, thereby al-
energy stores are mobilized but motor function is en- lowing the energy to go to another area of demand such
hanced with increases in cardiac output due to increases as muscle. All that was required was to move the gluco-
in heart rate and contractile force and dilation of muscle corticoid response forward in time. So, perhaps all that
arterioles, increasing blood supply to exercising muscles. was needed was to initiate the hypothalamo-pituitary-
In addition, the sensory side is enhanced—vigilance is adrenal response from a new source. Remember that
induced, pupils are dilated for better distance vision, and macrophage-produced cytokines initiate the pituitary-
so forth. adrenal response by acting at the hypothalamus. So, in
Although both the innate immune and fight-flight the stress response it is initiated by neural input, rather
responses require energy mobilization, the energy must than by peripheral cytokines.
go to different places, and the behavioral requirements You might also note that stressors, or events that
are completely different for the two responses. In fact, elicit fight-flight, can bear a close resemblance to physical
they are roughly opposite. During a fight-flight emergency damage or events that produce physical damage. This
it would not be useful for energy to be used to produce concept goes back at least to Selye (1936), who viewed
inflammation and the acute phase response; energy needs inflammation as a prototypical stressor. Interestingly, IL-
to go to the muscles and brain. It would not be adaptive 1 and receptors for IL-1 are located in brain (Breder, Di-
to maintain fever, reduced activity, huddling, shivering, narello, & Saper, 1988; Takao, Tracey, Mitchell, &
and somnolence. Hyperalgesia would not be adaptive, in- DeSouza, 1990). Perhaps physically challenging fight-
asmuch as the organism would be likely to direct attention flight stimuli activate the hypothalamo-pituitary-adrenal
to sites of injury rather than engage in defense. Analgesia axis by activating brain IL-1. It is interesting to note that
would be desirable, and that is what stress produces (Kelly, physical restraint activates messenger RNA for IL-1 in
1986). In short, except for the fact that both produce brain (Minami et al., 1991). Physical stressors might even
energy mobilization, inflammation and the stress response be able to activate elements of the inflammatory response
generally have opposite effects. directly, thereby providing a pathway to brain. Perhaps
What this means is that during a fight-flight emer- more psychological stressors then evolved other pathways
gency it would be adaptive to produce energy and to in- to activate the hypothalamo-pituitary-adrenal axis.
hibit inflammation and the acute phase response, should A final consideration is that the specific immune
there be injury or infection during the encounter. More- response evolved out of and uses many of the components
over, if the encounter is extended with periods of respite of the innate immune system. Thus, stress-induced mod-
during which inflammation develops, it would be adaptive ulation of specific immunity might be a remnant of its
to reduce the innate immune response if the encounter effects on innate immunity. In fact, thinking of innate
starts again. There is something that does this—gluco- immunity and specific immunity as separate processes
corticoids. Therefore, during a fight-flight emergency it may be as misleading as thinking of cellular and humoral
would be useful to produce glucocorticoids quickly, rather immunity as separate.
than several hours after the initiating event. This would
produce energy and inhibit the inflammmatory response Conclusions
before it can develop, should there be injury. This is ex- The major theme of this article is that the immune system
actly what stressors do. Thus, when considering the innate and brain form a bidirectional interacting set of processes,
response, there is a clear argument for an adaptive func- each regulating the other. Psychological processes can in-
tion of stress-induced immunosuppression. fluence this network and in turn be modulated by it. We

1014 December 1994 • American Psychologist

 Besedovsky, H. Q, Sorkin, E., Felix, D., & Haas, H. (1977). Hypothalamic
hope that we have provided some insight into the adaptive changes during the immune response. European Journal of Immu-
reasons why these links might exist and be sensible. These nology, 7, 323-330.
links provide great promise in terms of understanding Besedovsky, H. Q, Sorkin, E., Keller, M., & Muller, J. (1975). Changes
health and disease, but as reviewed, a great deal of work in blood hormone levels during immune response. Proceedings of the
needs to be done before strong conclusions are warranted. Society for Experimental Biology and Medicine, 150, 466-470.
The issues involved are too important to allow sweeping Blalock, J. E., Smith, E. M., & Meyer, W. J. (1985). The pituitary-
adrenocortical axis and the immune system. Clinics in Endocrinology
conclusions at the present stage of knowledge. & Metabolism, 14, 1021-1038.
It is our feeling that the next few years will be an Bohus, B., & Koolhaas, J. M. (1991). Psychoimmunology of social factors
exciting time in PNI research. Because PNI is a new field, in rodents and other subprimate vertebrates. In R. Ader, D. L. Felden,
the existing knowledge is, of necessity, first-order knowl- & N. Cohen, (Eds.), Psychoneuroimmunology (2nd ed., pp. 807-831).
New York: Academic Press.
edge. Classical conditioning can modify immune pro- Bolles, R. C , & Fanselow, M. S. (1980). A preceptual-defensive-recu-
cesses, stress can alter immunity, and immune products perative model of fear and pain. Behavioral and Brain Sciences, 3,
can feed back and modulate behavior. However, the com- 291-323.
plexities, breadth, and richness of the interactions have Bonneau, R. H., Sheridan, J. F., Feng, N., & Glaser, R. (1991a). Stress-
yet to be elucidated. In addition, the details of the mech- induced suppression of herpes simplex virus (HSV)-specific cytotoxic
T lymphocyte and natural killer cell activity and enhancement of
anisms involved are largely unknown. In the next few acute pathogenesis following local HSV infection. Brain, Behavior,
years, work of increasing sophistication at the behavioral, and Immunity, 5, 170-192.
neural, and immunological levels should be accomplished. Bonneau, R. H., Sheridan, J. F., Feng, N., & Glaser, R. (1991b). Stress-
PNI is one of the new emerging interdisciplinary induced effects on cell mediated innate and adaptive memory com-
fields being driven by the growing realization that systems ponents of the murine immune response to herpes simplex virus in-
fection. Brain, Behavior, and Immunity, 5, 274-295.
cannot be understood in isolation. Simply studying im- Bovjberg, D. H., Redd, W. H., Maier, L. A., Holland, J. C , Lesko,
munology at the level of immune cells, neuroscience at L. M., Niedzwiecki, D., Rubin, S. E., & Hakes, T. B. (1990). Antic-
the level of neurons, and psychology at the level of be- ipatory immune suppression in women receiving cyclic chemotherapy
havior cannot capture the complex interactions between for ovarian cancer. Journal of Consulting and Clinical Psychology, 58,
levels. Living organisms are not composed of discon- 153-157.
Breder, C. D., Dinarello, C. A., & Saper, C. B. (1988). Interleukin-1
nected systems or processes. It is our conviction that pro- immunoreactive innervation of the human hypothalamus. Science,
gress waits at the interfaces between systems and levels. 240, 321-324.
Comprehension of health and disease in particular awaits Carlson, S. L., Felten, D. L., Livnat, S., & Felten, S. Y. (1987). Alterations
such analyses. of monoamines in specific central autonomic nuclei following im-
munization in mice. Brain, Behavior, and Immunity, 1, 52-64.
Carrol, B. J. (1978). Neuroendocrine function in psychiatric disorder.
REFERENCES In M. A. Lipton, A. DiMaschio, & K. F. Killam (Eds.), Psychophar-
macology: A generation of progress (pp. 487-497). New York: Raven
Ader, R. (1985). Conditioned immunopharmacologic effects in animals: Press.
Implications for a conditioning model of pharmacotherapy. In L. Coderre, T. J., Katz, J., Vaccarino, A. L., & Melzack, R. (1993). Con-
White, B. Tursky, & G. Schwartz (Eds.), Placebo: Theory, research, tribution of central neuroplasticity to pathological pain: Review of
and mechanisms (pp. 306-323). New York: Guilford. clinical and experimental evidence. Pain, 52, 259-285.
Ader, R., & Cohen, N. (1975). Behaviorally conditioned immuno- Coe, C. L., Rosenberg, L. T., & Levine, S. (1988). Effect of maternal
suppression. Psychosomatic Medicine, 37, 333-340. separation on the complement system and antibody response in infant
Ader, R., & Cohen, N. (1982). Behaviorally conditioned immuno- primates. International Journal of Neuroscience, 40, 289-302.
suppression and murine systemic lupus erythematosus. Science, 215,
Cohen, S., & Williamson, G. M. (1991). Stress and infectious disease
1534-1536.
in humans. Psychological Bulletin, 109, 5-24.
Ader, R., & Cohen, N. (1993). Psychoneuroimmunology: Conditioning
Croiset, G., Heijnen, C. J., Veldhuis, H. D., deWied, D., & Ballieux,
and stress. Annual Review of Psychology, 44, 53-85.
R. E. (1987). Modulation of the immune response by emotional stress.
Ader, R., Felten, D. L., & Cohen, N. (1991). Psychoneuroimmunology Life Sciences, 40, 775-782.
(2nd ed.). New York: Academic Press.
Cunnick, J. E., Lysle, D. T, Armfield, A., & Rabin, B. S. (1991). Stressor-
Banks, W. A., Kastin, A. J., & Durham, D. A. (1989). Bidirectional
induced changes in mitogenic activity are not associated with decreased
transport of interleukin-1 alpha across the blood-brain barrier. Brain
Research Bulletin, 23, 433-437. IL-2 production or changes in lymphocyte subsets. Clinical Immu-
nology & Immunopathology, 60, 419-429.
Baracos, V. E., Rodemann, H. P., Dinarello, C. A., & Goldberg, A. L.
(1983). Stimulation of muscle protein degradation and prostaglandin Dantzer, R., Bluthe, R. M., Kent, S., & Goodall, G. (1993). Behavioral
E2 release by leukocytic pyrogen (interleukin-1): A mechanism for effects of cytokines: An insight into mechanisms of sickness behavior.
the increased degradation of muscle proteins during fever. New England In E. B. DeSouza (Ed.), Neurobiology of cytokines (pp. 130-151). San
Journal of Medicine, 308, 553-558. Diego: Academic Press.
Barnes, P. J., & Adcock, I. (1993). Anti-inflammatory actions of steroids: Dorian, B. J., Garfinkel, P., Brown, G., Shore, A., Gladman, D., &
Molecular mechanism. Trends in Pharmacological Sciences, 14,436- Keystone, E. (1982). Aberrations in lymphocyte subpopulations and
441. function during psychological stress. Clinical and Experimental Im-
Baumann, H., & Gauldie, J. (1994). The acute phase response. Im- munology, 50, 132-139.
munology Today, 15, 74-81. Dunn, A. J. (1993). Infection as a stressor: A cytokine mediated activation
Ben-Eliyahu, S., Yirmiya, R., Liebeskind, J. C , Taylor, A. N., & Gale, of the hypothalamo-pituitary-adrenal area. In Corticotropin releasing
R. P. (1991). Stress increases metastatic spread of a mammary tumor factor (Ciba Foundation Symposium 172, pp. 226-243). West Sussex,
in rats: Evidence for mediation by the immune system. Brain, Be- England: Wiley.
havior, and Immunity, 5, 193-205. Dunn, A. J., Powell, M. L., & Small, P. A., Jr. (1989). Virus infection
Berkenbosch, F., VanOers, J., Del Rey, A., Tilders, F., & Besedovsky, H. as a stressor: Influenza virus elevates concentrations of corticosterone
(1987). Corticotropin-releasing factor-producing neurons in the rat and brain concentrations of MHPG and tryptophan. Physiology &
activated by interleukin-1. Science, 238, 524-526. Behavior, 45, 591-594.

December 1994 • American Psychologist 1015

 Esterling, B., & Rabin, B. S. (1987). Stress-induced alteration of T-lym- KJosterhalfen, W., & Klosterhalfen, S. (1983). Pavlovian conditioning
phocyte subsets and humoral immunity in mice. Behavioral Neuro- of immunosuppression modifies adjuvant arthritis in rats. Behavioral
science, 101, 115-119. Neuroscience, 97, 663-666.
Fawzy, I. F., Kemeny, M. E., Fawzy, N. W., Elashoff, R., Morton, D., Korneva, E. A. (1967). The effect of stimulating different mesencephalic
Cousins, N., & Fahey, J. L. (1990). A structured psychiatric inter- structures on protective immune response patterns. Sechenov Phys-
vention for cancer patients: II. Changes over time in immunological iological Journal of the USSR, 53, 42-50.
measures. Archives ofGeneral Psychiatry, 47, 313-319. Laudenslager, M. L., Held, P. E., Boccia, M., Reite, M. L., & Cohen,
Felten, D. L., Ackerman, K. D., Wiegand, S. J., & Felten, S. Y. (1987). J. J. (1990). Behavioral and immunological consequences of brief
Noradrenergic sympathetic innervation of the spleen: I. Nerve fibers mother-infant separation. Developmental Psychobiologv, 23, 247-264.
associate with lymphocytes and macrophages in specific compartments Livnat, S., Felten, S. Y, Carlson, S. K., Bellinger, D. L., & Felten, D. L.
of the splenic white pulp. Journal of Neuroscience Research 18, 28- (1985). Involvement of peripheral and central catecholamine systems
36. in neural-immune interactions. Journal of Neuroimmunology. 10 5-
Felten, S. Y., & Felten, D. L. (1991). Innervation of lymphoid tissue. In 13. ' ' '
R. Ader, D. L. Felten, & N. Cohen (Eds.), Psychoneuroimmunology Lysle, D. T, Cunnick, J. E., Kucinski, B. J., Fowler, H., & Rabin, B. S.
(2nd ed., pp. 27-71). San Diego: Academic Press. (1990). Characterization of immune alterations produced by a con-
Fleshner, M., Brohm. M. M., Laudenslager, M. L., Watkins, L. R., & ditioned aversive stimulus. Psychobiology. 18, 220-226.
Maier, S. F. (1993). Modulation of in vivo antibody response by a Lysle, D. T., Cunnick, 1. E., & Rabin, B. S. (1990). Stressor-induced
benzodiazepine inverse agonist (DMCM) administered centrally or alteration of lymphocyte proliferation in mice: Evidence for enhance-
peripherally. Physiology and Behavior, 54, 1149-1154. ment of mitogen responsiveness. Brain, Behavior, and Immunology,
Fleshner, M , Laudenslager, M. L., Simons, L., & Maier, S. F. (1989). 4, 269-277.
Reduced serum antibodies associated with social defeat in rats. Phys- Lysle, D. X, Lyte, M., Fowler, H., & Rabin, B. S. (1987). Shock-induced
iological Behavior, 45, 1183-1187. modulation of lymphocyte reactivity: Suppression, habituation, and
Fleshner, M., Watkins, L. R., Bellgrau, D., Laudenslager, M. L., & Maier, recovery. Life Sciences, 41, 1805-1814.
S. F. (1992). Specific changes in lymphocyte subpopulations: A mech- Macris, N. X, Schiavi, R. C , Camerino, M. S., & Stein, J. (1970). Effect
anism for stress-induced immunomodulation. Neuroimmunology, 41, of hypothalamic lesions on immune processes in the guinea pig.
131-142. American Journal of Physiology, 219, 1205.
Fredrikson, M., Furst, C. J., Lekander, M., Rotstein, S., Blomgren, H. Maier, S. F, & Laudenslager, M. L. (1988). Commentary: Inescapable
(1993). Trait anxiety and anticipatory immune reactions in women shock, shock controllability, and mitogen stimulated lymphocyte pro-
receiving adjuvant chemotherapy for breast cancer. Brain. Behavior, liferation. Brain, Behavior, and Immunity, 2, 87-91.
and Immunity, 7, 79-90. Maier, S. F., Wiertelak, E. P.. Martin, D., & Watkins, L. R. (1993).
Grochowitz, P. M., Schedlowski, M., Husband, A,. J., King, M. G., Interleukin-1 mediates the behavioral hyperalgesia produced by lith-
Hibberd, A. D., & Bowen, K. M. (1991). Behavioral conditioning ium chloride and endotoxin. Brain Research, 623, 321-326.
prolongs heart allograft survival in rats. Brain, Behavior, and Immunity, Mason, J. W. (1971). A re-evaluation of the concept of "non-specificity"
5, 349-356. in stress theory. Journal of Psychiatric Research. 8, 123-140.
Holsboer, F, Von Bardeleben, U., Gerken, A., Stalla, G. K., & Muller, Minami, M., Kuraishi, Y, Yamaguchi, X, Nakai, S., Hirai, Y, & Satoh,
O. A. (1984). Blunted corticotropin and normal cortisol response to M. (1991). Immobilization stress induces interleukin-1 beta mRNA
human corticotropin releasing factor in dperession. New England in the rat hypothalamus. Neuroscience Letters, 123, 254-256.
Journal of Medicine, ill, 1127-1137. Monjan, A. A., & Collector, M. I. (1977). Stress-induced modulation of
Irwin, M., Daniels, M., Bloom, E., Smith, T. L., & Weiner, H. (1987). the immune response. Science. 196. 307-308.
Life events, depressive symptoms, and immune function. American Mormede, P., Dantzer, R., Michaud, B., Kelly, K., & LeMoal, M. (1988).
Journal of Psychiatry, 144, 437. Influence of stressor predictability and behavioral control on lym-
Jiang, C. G., Morrow-Tesch, J. L., Beller, D. S., Levy, E. M., & Black, phocyte reactivity, antibody responses, and neuroendocrine activation
P. H. (1990). Immunosuppression in mice induced by cold water in rats. Physiology & Behavior. 43. 577-583.
stress. Brain, Behavior, and Immunity, 4, 278-291. Moynihan, J., Brenner, G., Koota, D., Breneman, S., Cohen, N., &
Katsuura, G., Arimura, A., Koves, K., & Gottchail, P. E. (1990). In- Ader, R. (1990). The effects of handling on antibody production, mi-
volvement of organum vasculosum of lamina terminalis and preoptic togen responses, spleen cell number, and lymphocyte subpopulations.
area in interleukin-1 beta induced ACTH release. American Journal Life Sciences. 46, 1937-1944.
of Physiology (Endocrinology and Metabolism), 258, E163-E171. Moynihan, J., Koota, D., Brenner, G., Cohen, N., & Ader, R. (1989).
Keller, S. E., Weiss, J. M., Schleifer, S. J., Miller, N. E., & Stein, M. Repeated intraperitoneal injections of saline attenuate the antibody
(1983). Stress-induced suppression of immunity in adrenaiectomized response to a subsequent intraperitoneal injection of antigen. Brain.
rats. Science, 221, 1301-1304. Behavior, and Immunity 3, 90-96.
Kelly, D. D. (1986). Stress-induced analgesia. Annals of the New York Nance, D. M., Rayson, D., & Carr, I. (1987). The effects of lesions in
Academy of Sciences, 46 7, the lateral septal and hippocampal areas on the humoral immune
Kemeny, M. E., Solomon, G. F, Morley, J. E., & Bennett, R. L. (1993). response of adult female rats. Brain, Behavior, and Immunity, 1, 292-
Psychoneuroimmunology. In C. B. Nemeroff(Ed.), A comprehensive 300.
textbook of neuroendocrinology (pp. 563-592). Boca Raton, FL: CRC Palmblad, J., Cantell, K., Strander, H., Froberg, J., Karlsson, C , Gron-
Press. strom, M., & Unger, P. (1976). Stressor exposure and immunological
Kent, S., Bluthe, R. M., Dantzer, R., Hardwick, A. J., Kelley, K. W., competence in man: Interferon producing capacity and phagocytosis.
Rothwell, N. J., & Vannice, J. L. (1992). Different receptor mechanisms Journal of Psychosomatic Research, 20, 193-203.
mediate the pyrogenic and behavioral effect of interleukin-1. Pro- Palmblad, J., Petrini, G., Wasserman, J., & Akerstedt. X (1979). Lym-
ceedings of the National Academy of Sciences, 89, 9117-9120. phocyte and granulocyte reactions during sleep deprivation. Psycho-
Kiecolt-Glaser, J. K., Fisher, L. D., Ogrocki, P., Stout, J. C , Speicher, somatic Medicine, 41, 273-285.
C. E., & Glaser, R. (1987). Marital quality, marital disruption, and Plaut, M. (1987). Lymphocyte hormone receptors. Annual Review of
immune function. Psychosomatic Medicine, 49, 13-25. Immunology. 5, 621-669.
Kiecolt-Glaser, J. K., Garner, W., Speicher, C , Penn, G. M., Holliday, Plotnikoff, N., Murgo, A., Faith, R., & Wybran, J. (1991). Stress and
J., & Glaser, R. (1984). Psychosocial modifiers of immunocompetence immunity. Boca Raton, FL: CRC Press.
in medical students. Psychosomatic Medicine, 46, 7-17. Rabin, B. S., Lyte, M., Epstein, L. H., & Caggiula, A. R. (1987). Alter-
Kiecolt-Glaser, J. K., Glaser, R., Shuttleworth, E. C , Dyer, C. S., Ogrocki, ation of immune competency by number of mice housed per cage.
P., & Speicher, C. E. (1987). Chronic stress and immunity in family Annals of the New York Academy of Sciences, 469, 492-500.
caregivers of Alzheimer's disease victims. Psyhcosomatic Medicine, Reinisch, C. I., & Litman, G. W. (1989). Evolutionary immunobiology.
49, 523-533. Immunology Today, 10. 278-293.

1016 December 1994 • American Psychologist

Rinner, I., Schauenstein, K., Mangge, H., & Porta, S. (1992). Opposite Solvason, H. B., Ghanta, V. K., & Hiramoto, R. N. (1988). Conditioned
effects of mild and severe stress on in vitro activation of rat peripheral augmentation of natural killer cell activity: Independence from no-
blood lymphocytes. Brain, Behavior, and Immunity, 6, 130-140. ciceptive effects and dependence on interferon-beta. Journal of Im-
Roszman, T. L., Cross, R. J., Brooks, W. H., & Markesbery, W. R. munology, 140, 661-665.
(1985). Neuroimmunomodulation: Effects of neural lesions on cellular Stanford, S. C , & Salmon, P. (1993). Stress: From synapse to syndrome.
immunity. In R. Guillemin, M. Cohn, & T. Melnechuk (Eds.), Neural London: Academic Press.
modulation of immunity (pp. 95-111). New York: Raven Press. Takao, T., Tracey, D. E., Mitchell, W. M., & DeSouza, E. B. (1990).
Rothwell, N. J. (1992). Metabolic responses to interlueukin-1. In N. J. Interleukin-1 receptors in mouse brain: Characterization and neuronal
Rothwell & R. D. Dantzer (Eds.), Interleukin-1 in the brain (pp. 115- localization. Endocrinology, 127, 3070-3078.
135). Oxford, England: Pergamon Press. Ursin, H., & Olff, M. (1993). The stress response. In S. C. Stanford &
Saphier, D. (1992). Electrophysiological studies of the effects of inter- P. Salmon (Eds.), Stress: From synapse to syndrome (pp. 4-24). Lon-
leukin-1 and a-interferon on the EEG and pituitary-adrenocortical
don: Academic Press.
activity. In J. J. Rothwell & R. D. Dantzer (Eds.), Interleukin-1 in the
Watkins, L. R., Wiertelak, E. P., Goehler, L., Mooney-Heiberger, K.,
brain (pp. 51-75). Oxford, England: Pergamon Press.
Sapolsky, R. M (1992). Stress: The aging brain and the mechanisms of Martinez, J., Furness, L., & Maier, S. F. (1994). Neurocircuitry of
neuron death. Cambridge, MA: MIT Press. illness-induced hyperalgesia. Brain Research, 639, 283-299
Schindler, R., Bhezzi, P., & Dinarello, C. A. (1991). IL-1 induces IL-1 Weiner, H. (1991). The dynamics of the organism: Implications of recent
IV; IFN suppresses IL-1 but not LPS-induced transcription of IL-1. biological thought for psychosomatic theory and research. In R. Ader,
Journal of Immunology, 144, 2216-2222. D. L. Felten, & N. Cohen (Eds.), Psychoneuroimmunology II (pp.
Schleifer, S. J., Keller, S. E., Camerino, M., Thornton, J. C , & Stein, 955-1013). New York: Academic Press.
M. (1983). Suppression of lymphocyte stimulation following bereave- Weiss, J. M., Sundar, S. K., Becker, K. J., & Cierpial, M. A. (1989).
ment. Journal of the American Medical Association, 250, 374. Behavioral and neural influences on cellular immune responses: Effects
Schleifer, S. J., Keller, S. E., Meyerson, A. T., Raskin, M. J., Davis, of stress and interleukin-1. Journal of Clinical Psychiatry, 50, 43-53.
K. L., & Stein, M. (1984). Lymphocyte function in major depressive Wiertelak, E. P., Smith, K. B., Furness, L., Mayr, X, Maier, S. F., &
disorder. Archives of General Psychiatry, 41, 484. Watkins, L. R, (1994). Acute and conditioned hyperalgesic responses
Selye, H. (1936). A syndrome produced by diverse noxious agents. Na- to illness. Pain, 56, 227-235.
ture, 38, 32-36. Zalcman, S., Minkiewicz-Janda, A., Richter, M., & Anisman, H. (1988).
Sheridan, J. E, Feng, N., Bonneau, R. H., Allen, C. M., Huneycutt, Critical periods associated with stressor effects on antibody titers and
B. S., & Glaser, R. (1991). Restraint stress differentially affects anti- on the plaque-forming cell response to sheep red blood cells. Brain,
viral cellular and humoral immune responses in mice. Journal of Behavior, and Immunity, 2, 254-266.
Neuroimmunology, 31, 245-255.
Zalcman, S., Richter, M., & Anisman, H. (1989). Alterations of immune
Sklar, L. S., & Anisman, H. (1981). Stress and cancer. Psychological
Bulletin, 89, 369-406. functioning following exposure to stressor-related cues. Brain, Be-
Smith, G. R., & McDaniels, S. M. (1983). Psychologically mediated havior, and Immunity, 3, 99-109.
effect on the delayed hypersensitivity reaction to tuberculin in humans. Zwilling, B. S., Brown, D., Christner, R., Faris, M., & Hilberger, M.,
Psychosomatic Medicine, 45, 65-70. McPeek, M., Epps, C. V., & Hartlaub, B. A. (1990). Differential effect
Solomon, G. (1969). Stress and antibody response in rats. Archives of of restraint stress on MHC class 11 expression by murine peritoneal
Allergy, 35, 97-104. macrophages. Brain, Behavior, and Immunity, 4, 330-338.

December 1994 • American Psychologist 1017