- Email: [email protected]
The neurobiology of aggression and rage: Role of cytokines
Brain, Behavior, and Immunity 20 (2006) 507–514 www.elsevier.com/locate/ybrbi
Invited minireview
The neurobiology of aggression and rage: Role of cytokines Steven S. Zalcman a,¤, Allan Siegel a,b a
b
Department of Psychiatry, UMDNJ-New Jersey Medical School, 183 South Orange Avenue, Newark, NJ 07103, USA Department of Neurology and Neurosciences, UMDNJ-New Jersey Medical School, 183 South Orange Avenue, Newark, NJ 07103, USA Received 11 April 2006; received in revised form 30 May 2006; accepted 30 May 2006 Available online 30 August 2006
Abstract Recent studies have suggested an important relationship linking cytokines, immunity and aggressive behavior. Clinical reports describe increasing levels of hostility, anger, and irritability in patients who receive cytokine immunotherapy, and there are reports of a positive correlation between cytokine levels and aggressive behavior in non-patient populations. On the basis of these reports and others describing the presence or actions of diVerent cytokines in regions of the brain associated with aggressive behavior, our laboratory embarked upon a program of research designed to identify and characterize the role of IL-1 and IL-2 in the hypothalamus and midbrain periaqueductal gray (PAG)—two regions functionally linked through reciprocal anatomical connections—in the regulation of feline defensive rage. A paradigm involved cytokine microinjections into either medial hypothalamus and elicitation of defensive rage behavior from the PAG or vice versa. These studies have revealed that both cytokines have potent eVects in modulating defensive rage behavior. With respect to IL-1, this cytokine facilitates defensive rage when microinjected into either the medial hypothalamus or PAG and these potentiating eVects are mediated through 5-HT2 receptors. In contrast, the eVects of IL-2 are dependent upon the anatomical locus. IL-2 microinjected into the medial hypothalamus suppresses defensive rage and this suppression is mediated through GABAA receptors, while microinjections of IL-2 in the PAG potentiate defensive rage, in which these eVects are mediated through NK-1 receptors. Present research is designed to further delineate the roles of cytokines in aggressive behavior and to begin to unravel the possible signaling pathways involved this process. © 2006 Elsevier Inc. All rights reserved. Keywords: Cytokines; Interleukin-1; Interleukin-2; Aggression; Defensive rage; Serotonin; GABA; NK-1
1. Introduction There is a consensus that the immune and central nervous systems (CNS) communicate in a bi-directional manner. A fundamental principle of psychoneuroimmunology is that this communication helps regulate an orchestrated immune response (Dantzer et al., 1999). This notably includes: (1) activation of neuroimmune feedback loops, and (2) stimulation of central neurochemical alterations that in turn underlie adaptive behavioral and physiological responses (i.e., the classic symptoms of sickness behavior). Cytokines released during the course of an immune
*
Corresponding author. Fax: +1 973 972 8305. E-mail address: [email protected] (S.S. Zalcman).
0889-1591/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbi.2006.05.002
response act as messengers that help modulate sickness behavior by inXuencing relevant neurotransmitter systems, and in some cases, by directly acting within the brain (Maier et al., 2001). Since cytokines also act as endogenous neuromodulators, it is possible to study their relationship to behavior independent of an ongoing immune response. In this context, it has been recently shown that cytokines are present in brain regions, such as the hypothalamus and midbrain periaqueductal gray (PAG), that are known to be associated with aggression and rage behavior (Siegel et al., 1999). Moreover, there is evidence that anger and hostility are increased in patients receiving repeated cytokine immunotherapy, supporting the view that cytokines may facilitate the expression of aggressive behavior. Accordingly, the following review will discuss research investigating the
508
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
relationship between cytokines, immunity, and aggressive behavior. This includes studies detailing the immunological consequences of an aggressive confrontation, the relationship between cytokines and aggressive behavior, links between social status and immune function and eVects of immune responding on the animal’s response to an aggressive challenge. We will also discuss our research program that uses an established model of feline aggressive behavior to directly determine the role of brain cytokines (notably interleukin (IL)-1 and IL-2) on the expression of defensive rage behavior, which is a close animal analog of a parallel form of human aggression. This approach allows us to identify speciWc brain sites wherein cytokines may modulate aggressive behavior, systematically determine the nature of their eVects, and identify speciWc cytokine-neurotransmitter receptor mechanisms that underlie these eVects. 2. Cytokines, immunity and aggressive behavior The relationship between cytokines, immunity, and aggressive behavior has been studied on many levels in human and infrahuman populations. In this section, we will present an overview of representative studies examining the aggression–immunity relationship. This includes: (a) human studies that evaluate anger/hostility in patients receiving repeated cytokine immunotherapy, or coordinate assessments of anger/hostility and endogenous cytokine or immune cell activity in non patient populations; and (b) animal studies examining immunological proWles of aggressive and submissive individuals, and the relationship between strain, social status and immunity. In recent years, an increasing amount of attention has focused on psychiatric abnormalities associated with cytokine administration (notably interferon (IFN)-2b and/or interleukin (IL)-2), which are used in the treatment of cancer, AIDS, and hepatitis C, among other disorders (e.g., Capuron et al., 2004). Studies have shown that cytokine immunotherapy results in a facilitation of aggressive behavior. SpeciWcally, measures of anger/hostility and irritability are increased in patients receiving cytokine immunotherapy (e.g., Kraus et al., 2003; McHuthison et al., 1998). It is important to consider that various factors may have contributed to these observations, including comorbid psychiatric abnormalities and processes related to the underlying disorder being treated. Indeed, patients also display depression, anxiety, and cognitive diYculties, among other psychiatric disturbances. Moreover, it is possible that an interaction between cytokine treatment and the disease process (e.g., chronic infection and subsequent alterations of cytokine activity) inXuences the expression of aggressive behavior. Support for this suggestion stems from a study by Kraus et al. (2003) who evaluated anger/hostility in patients with chronic hepatitis C infection before and after onset of IFN therapy. They found that anger/hostility was increased in patients before onset of cytokine treatment and that these increases were further augmented following onset of therapy. To be sure, one must temper one’s conclusions
concerning the relationship between cytokines and measures of aggressive behavior pending analyses of the possible contributions of other factors, including actual regimens of cytokine therapy, the use of other treatments, and certain demographic factors, among other factors. Studies using non-patient populations also suggest a link between aggressive behavior and cytokines. In these studies, measures of aggressive behavior are related to endogenous blood cytokine production. For example, Suarez et al. (2004), using healthy subjects, showed that Cook–Medley Hostility scores are increased coincident with an enhancement of LPS-stimulated production of proinXammatory cytokines by blood monocytes. Consistent with these Wndings, Kiecolt-Glaser and colleagues (2005) showed that higher levels of hostile marital interactions are associated with increased production of plasma proinXammatory cytokines. Taken together, these studies show that a positive association exists between cytokines, immune cell activity, and various measures of aggressive behavior in patient and non-patient populations. Increases in aggressive behavior have also been linked to enhancements of cytokine production and immune cell activity in animal studies. For example, Petitto et al. (1994) showed that IFN, IL-2 production, and T cell proliferation were higher in mice bred for high aggression (based on frequency of attacks after contact) than mice bred for low aggression. Inasmuch as these eVects were not related to diVerences in postweaning social experience or to gender diVerences, the authors suggested that a genetic linkage exists for genes associated with aggression and immunity. It is important to consider that aggressive confrontations can result in both immunoenhancing and immunosuppressive eVects. Avitsur and colleagues (2002) used a paired Wghting model to determine the eVects of Wghting on measures of splenic cell distribution and function. They found that six daily sessions resulted in an increase in monocytes and neutrophils, and a decrease in lymphocyte measures. These eVects occurred coincident with a state of glucocorticoid resistance in splenocytes. The authors suggested that the immune alterations associated with aggressive confrontations may be related, at least in part, to wound healing. Another model of social stress that has been used to examine the aggression–immune relationship is the social confrontation model or resident-intruder confrontation. In this model, an intruder is placed into an aggressive animal’s home cage. The intruder is typically attacked and defeated, and tends to show submissive or subdominant behavior. Stefanski and Ben-Eliyahu (1996) used this model to study the consequences of social confrontation on tumor retention. They found that rats receiving mammary tumor cells one hour into a seven hour confrontation had increased tumor retention, and that this eVect was blocked by a adrenergic antagonist. Studies have also examined the eVects of social status on measures of immunity. Typically, subdominant animals are immunosuppressed and show an
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
increased vulnerability to infection (e.g., Cohen et al., 1997). Finally, while a signiWcant amount of attention has focused on immunological consequences of aggressive encounters, there is some evidence that early life immune activation may induce long-lasting eVects on aggression. For example, Granger et al. (2001) showed that early life exposure to endotoxin reduced aggressive behavior in mice from a highaggressive line. Taken together, these studies demonstrate that exposure to an aggressive confrontation has marked eVects on various aspects of immune function, and that the type of confrontation and behavioral response inXuences the nature of the immunological changes. It is also important to consider that these changes may be due to a variety of factors, including wound healing, the stressfulness of the confrontation, the social status of the individual, and developmental factors. 3. Feline models of aggression and rage behavior Feline models have shed considerable light on the neuroanatomical substrates and neurotransmitter receptor mechanisms that regulate aggression and rage. As mentioned, we have used such models to directly determine the role of brain cytokines on forms of aggressive behavior, and to identify the cytokine-neurotransmitter receptor mechanisms that regulate these eVects. In the following section, we will provide an overview of these models. An expanding body of data indicates that aggressive behavior appears as a component of numerous clinical disorders associated with abnormal brain function, including aVective disorders, brain tumors, temporal lobe epilepsy, among other abnormalities (Bear, 1979). Such behavior usually falls into the category of defensive rage. Most notable is the aggression referred to as “episodic dyscontrol,” which is seen in patients with temporal lobe epilepsy or hypothalamic tumors. Typically, such patients display an explosive personality and may be physically or verbally assaultive in response to little or no provocation. Moreover, this disorder has been described in both adults and children (Nunn, 1986) lending credence to the idea that the same neural substrate underlies defensive-ragelike aggression over the course of brain development. Notably, this form of aggression is linked to structures associated with the limbic–hypothalamic–midbrain periaqueductal gray (PAG) axis (Siegel et al., 1999). Because the explosiveness of behavior and marked sympathoadrenal arousal in these disorders also occur in defensive rage behavior in the cat, it is our belief that this cat behavior is an appropriate model for study of the human disorders. Although fewer clinical studies have considered predatory forms of aggressive behavior, we likewise believe that it, too, is an appropriate model for study in the laboratory because it has been clearly identiWed in human studies (Vitiello et al., 1990). The discussion related below concerns the models of defensive rage and predatory attack behavior in the cat that have been studied extensively in our laboratory for many years.
509
Over the past three decades, the neural bases of defensive rage and predatory attack behavior, which can be elicited both under natural conditions and by electrical brain stimulation, have undergone systematic examination in our laboratory and elsewhere (see Siegel, 2005; Siegel et al., 1999). In attempting to understand the neural substrates of aggression, a crucial distinction must be made between two classes of neural structures. The Wrst class of structures, which include the hypothalamus and PAG, produce the expression of an attack response when stimulated at suprathreshold current levels. These structures, and the pathways arising from them, mediate, or carry the neural signals necessary for, the motor and autonomic aspects of aggression. Sites within these structures are termed, for example, “attack” sites, or more speciWcally, “defensive rage sites” or “predatory attack sites.” The second class comprises limbic structures whose stimulation does not produce an attack response; rather, stimulation of these structures increases or decreases the probability of occurrence of responses elicited by stimulation of an attack site. In other words, these structures modulate aggression and are termed “modulatory sites.” 3.1. Defensive rage behavior This form of aggressive behavior, initially described earlier in the last century, involves pronounced aVective signs such as vocalization, marked pupillary dilatation, retraction of the ears, arching of the back, and unsheathing of the claws (see Siegel, 2005; Siegel et al., 1999). Moreover, the cat will frequently modify its response by striking at a moving object. This response is ethologically signiWcant in that it occurs under natural conditions when the animal, its territory, or its kittens, are threatened by another animal of the same or diVerent species. Defensive rage behavior is elicited reliably by electrical stimulation of the medial preoptic-hypothalamus and by electrical or chemical stimulation of the PAG (Siegel et al., 1999). The hissing response is used as a measure of defensive rage because as a component of the stimulation-induced defensive rage response, hissing is always elicited on each trial, even in the absence of a threatening stimulus (see Siegel et al., 1999). The occurrence of hissing depends upon the parameters of stimulation and upon experimental manipulations such as paired trials of dual stimulation and drug administration. Because the latency of the hissing response is highly sensitive to such manipulations, their eVects can be clearly identiWed by analyzing the changes in latencies. Hence, for many years the latency for hissing has been a standard dependent variable used in our laboratory (see Siegel et al., 1999) as a measure of defensive rage. The threshold is less sensitive to experimental manipulations and is used as a dependent variable when appropriate. 3.2. Predatory attack behavior In nature, predatory attack behavior occurs as a motivated, stereotyped pattern of behavior directed at a prey
510
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
object. The cat initially stalks the prey (e.g., a rat) and then bites the back of its neck until the animal is killed (Siegel, 2005). Under experimental conditions, virtually the same response can be reproduced by electrical stimulation of the lateral perifornical hypothalamus (see Siegel et al., 1999), the ventral aspect of the midbrain PAG or caudal brainstem tegmentum. In contrast to defensive rage behavior, few autonomic signs other than mild pupillary dilatation are evident. 4. Neurochemistry of defensive rage and predatory attack in the cat–a summary Studies conducted in our laboratory over the past two decades have identiWed those classes of neurotransmitter receptors that facilitate or suppress the expression of defensive rage or predatory attack. These results of these studies are extensively reviewed in Siegel (2005) and Siegel et al. (1999), and are brieXy summarized here. 4.1. Excitatory neurotransmitter receptors Both dopamine and norepinephrine receptors have similar eVects upon neurons in the medial hypothalamus. SpeciWcally, activation of either norepinephrine 2-receptors or dopamine D2 receptors in the anterior medial hypothalamus facilitates defensive rage elicited from the ventromedial hypothalamus and dopamine D2 receptors further facilitate predatory attack as well. Other norepinephrine and dopamine receptors were not found to modulate aggression or rage. In contrast to the catecholamines, 5-HT2 and 5-HT1A receptors have diVerential eVects upon defensive rage behavior in which 5-HT2 receptors in the medial hypothalamus or PAG facilitate defensive rage, while 5-HT1A receptors suppress this response (see discussion below). The principal descending pathway from the medial hypothalamus over which defensive rage is expressed is one whose target is the PAG. The primary transmitter released from these endings is glutamate and its eVects are mediated through NMDA receptors. Two classes of neuropeptides that facilitate defensive rage have been identiWed and include substance P (SP), which acts through NK1 receptors in both the medial hypothalamus and PAG, and cholecystokinin (CCK), which acts through CCK-B receptors in the PAG. 4.2. Inhibitory neurotransmitter receptors Three classes of neurotransmitter receptors have been shown to have potent inhibitory eVects upon rage and/or aggression. Studies conducted on the role of GABA neurons have revealed that there exists reciprocal GABAergic inhibitory connections between the medial and lateral hypothalamus enables the lateral hypothalamus to suppress defensive rage from the medial hypothalamus and for the medial hypothalamus to suppress predatory attack
from the lateral hypothalamus and that these functions are mediated through GABAA receptors. Another class of receptors that suppresses defensive rage behavior is opioid peptides. SpeciWcally, enkephalinergic neurons arising from the central nucleus of amygdala, project to the PAG, and powerfully suppress defensive rage behavior at the level of the PAG. This inhibitory process is mediated through an opioid receptor mechanism within the PAG. As noted above, inhibition of defensive rage within both the medial hypothalamus and PAG are mediated through 5-HT1A receptors, a third class of inhibitory receptors. The presumed origin of serotonin Wbers that act through 5-HT receptors includes the pontine and midbrain raphe groups of neurons. 4.3. Cytokine eVects upon neurotransmitters associated with feline aggression It is well known that inXammatory cytokines are potent modulators of neurotransmitter activity in brain regions associated with mood, motivation, emotion, and behavioral responses to potentially threatening environmental stimuli (see Anisman et al., 2002; Dantzer et al., 1999; Dunn, 2001). Our studies have shown that IL-1 and IL-2 potently modulate defensive rage behavior. As will be discussed later in this review, we have shown that IL-1 in the medial hypothalamus facilitates defensive rage through a 5-HT2 receptor mechanism. Of further signiWcance is the dual eVect of IL-2 upon defensive rage. SpeciWcally, IL-2 in the medial hypothalamus inhibits defensive rage through a GABAA receptor mechanism, and potentiates this form of aggression through a substance P-NK-1 receptor mechanism in the PAG. Here, we brieXy summarize evidence linking IL-1 and IL-2 with these neurotransmitter systems that are associated with defensive rage and that have been shown to modulate cytokine-induced eVects upon this form of aggressive behavior. IL-1 mRNA and receptors are present in various brain regions, including hypothalamus (see Maier et al., 2001). As discussed earlier, 5-HT2 receptors in medial hypothalamus and PAG facilitate defensive rage behavior. It is thus of unique interest that IL-1 is a potent modulator of central 5HT activity. In rodents, systemic or intracebroventricular injections of IL-1 enhance the levels of tryptophan, hydroxyindolacetic acid (5-HIAA; the major metabolite of 5-HT), and stimulate 5-HT release in hypothalamus and extra-hypothalamic sites (see Anisman et al., 2002; Dantzer et al., 1999; Dunn, 2001). Further to the point, infusion of IL-1 into discrete hypothalamic nuclei likewise inXuences 5-HT activity. In particular, microinjections of IL-1 into the rat medial basal hypothalamus increase 5-HT release within this brain region (Shintani et al., 1993). Functionally, it is abundantly clear that IL-1 induces a broad range of behavioral alterations, and that some of these eVects are mediated centrally through 5-HT receptor mechanisms. For example, Imeri et al. (1999) showed that IL-1-induced sleep alterations are partly mediated through 5-HT2 receptors.
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
Collectively, these studies importantly show that IL-1 increases synaptic release of 5-HT, and that these interactions are evident in defensive rage sites in the hypothalamus. Of further importance is the evidence that certain behavioral eVects of IL-1 are mediated centrally through 5-HT2 receptor mechanisms. IL-2 mRNA and receptors have also been shown to be present in a variety of brain regions, including hypothalamus (see Anisman et al., 2002). IL-2 potently modulates central monoamine activity in hypothalamus and extrahypothalamic sites and induces behavioral changes linked with these systems (see Anisman et al., 2002). Inasmuch as IL-1 and IL-2 induce cytokine-speciWc alterations of central neurochemical activity and behavior (Zalcman et al., 1994, 1998), it might be expected that these cytokines would diVerentially modulate defensive rage behavior. Indeed, as mentioned, IL-2 exerts dual eVects upon defensive rage. In the medial hypothalamus IL-2 suppresses defensive rage; in the PAG it facilitates this form of aggression. The suppressive eVects of IL-2 are mediated through a GABAA receptor mechanism whereas the potentiating eVects occur through a substance P-NK-1 receptor mechanism. As noted above, GABAA receptors in the medial hypothalamus suppress feline defensive rage behavior. There is some evidence that IL-2 interacts with this neurotransmitter system. For example, it has been shown that IL-2 inhibits hippocampal acetylcholine release, and that this occurs through GABAA and GABAB receptor mechanisms (Seto et al., 1997). As previously discussed, substance P, acting through NK-1 receptors in the medial hypothalamus and PAG facilitates defensive rage behavior. Very little is known about the relationship between IL-2 and substance P-NK-1 receptors in the brain. It is thus of unique interest that intra-hippocampal infusion of IL-2 inXuences formaldehyde-induced substance P-like immunoreactivity (Wu et al., 1999). In the immune system, substance P induces an increase in IL-2 receptor expression (see Bhatt and Siegel, 2006). These Wndings suggest that IL-2 and substance P-NK1 receptors interact, and that the stimulation of one receptor modulates the expression of the other. 4.4. Cytokine modulation of defensive rage behavior As noted above, the relationships of a number of the major neurotransmitter systems upon defensive rage behavior and predatory attack in the cat have been identiWed. As well, studies have demonstrated the presence and/or actions of IL-1 and IL-2 in brain regions that regulate feline aggression (see Anisman et al., 2002; Dantzer et al., 1999; Dunn, 2001). Accordingly, a series of experiments were initiated in our laboratory involving regions that mediate defensive rage behavior. Namely, we aimed at identifying the roles of IL-1 and IL-2 in the hypothalamus and PAG elicited from either of these brain sites. Because the principal focus of these studies involved defensive rage behavior, a discussion of possible eVects of cytokines upon predatory attack is not included in the present review.
511
In the Wrst study, we sought to determine whether microinjections of IL-1 into the medial hypothalamus could modulate defensive rage behavior (Hassanain et al., 2003). The paradigm for this study and those that follow was based upon the neuroanatomical relationships that exist between the medial hypothalamus and the PAG. SpeciWcally, the downstream target of the medial hypothalamus mediating defensive rage behavior is the PAG (Siegel et al., 1999). Moreover, the PAG also projects back to the medial hypothalamus, completing a reciprocal feedback loop (Siegel et al., 1999). On the basis of this functional anatomical arrangement between the medial hypothalamus and PAG, a stimulating electrode was implanted into the PAG and a cannula-electrode was implanted into the medial hypothalamus from which defensive rage could be elicited from either site. In order to demonstrate the functional linkage between these two structures, a dual stimulation paradigm was employed to demonstrate that stimulation of the PAG facilitated the occurrence of defensive rage elicited from the medial hypothalamus and which further identiWed the speciWc sites from which this form of aggression are elicited. Then, IL-1 was microinjected into the behaviorally identiWed site via the cannula-electrode and the eVects of such microinjections upon defensive rage elicited from the PAG were observed. The results revealed a potent dose- and time-dependent facilitation of defensive rage (Hassanain et al., 2003). The kinetics of this eVect was of interest in that two separate peaks of maximal facilitation were noted, one occurring at 60 min, post-injection, and the second one occurring at 180 min, post-injection, suggesting the presence of an endogenous release of IL-1 with respect to the second peak. Moreover, these eVects were blocked by pretreatment with either an IL-1 type 1 receptor antagonist or with an IL-1 antibody. Facilitation is associated with 5-HT2 receptors in both hypothalamus and PAG. Therefore, we also determined whether pretreating the defensive rage site in the hypothalamus with a 5-HT2 antagonist would similarly block the potentiating eVects of IL-1. The results clearly indicated that this was case. A further analysis involving immunocytochemistry revealed the presence of 5-HT2 and IL-1 type I receptors in the medial hypothalamus which were co-localized, thus providing further support for the linkage of IL-1 and serotonin receptors (Hassanain et al., 2005). It should also be noted that our laboratory has recently obtained preliminary data indicating that activation of IL-1 receptors in the PAG likewise facilitate defensive rage behavior when elicited from the medial hypothalamus, suggesting that IL-1 receptors in both regions associated with defensive rage behavior mediate similar functions. Parenthetically, given the well-known roles of the PAG and hypothalamus in mediating responses to painful stimuli coupled with IL-1’s role in pain perception and sensitivity (Shavit et al., 2005), it seems reasonable to suggest that overlapping mechanisms may modulate defensive aggressive behavior and the response to painful stimuli. Such a link might be of adaptive value during an aggressive encounter.
512
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
Two important conclusions may be drawn from our observations: (1) Because we observed that relatively low (physiological) doses of IL-1 produced a facilitation of defensive rage, it is reasonable to suggest that these Wndings reXect an endogenous role for IL-1 in the modulation of defensive rage (Hassanain et al., 2003); and (2) that the potentiating eVects of IL-1 upon defensive rage behavior are mediated through IL-1 type I and 5-HT2 receptor mechanisms. Overall, this study represented the Wrst of its kind to show that a cytokine in the brain can modulate a form of aggressive behavior. Since IL-2 also modulates neurotransmitter activity in brain regions that regulate aggressive behavior in the cat, we also determined whether IL-2 could modulate defensive rage. Our Wndings showed that IL-2 modulates defensive rage with exquisite speciWcity. First, in contrast with the eVects of IL-1 upon defensive rage, microinjections of IL-2 into the medial hypothalamus induced a potent dosedependent suppression of this form of aggression (Bhatt et al., 2005). This eVect was blocked by pretreating the injection site with an antibody against the subunit of the IL-2 receptor or by a monoclonal antibody to IL-2. Inhibition of defensive rage is associated with GABAA receptors in the hypothalamus; we thus also showed that a GABAA receptor antagonist blocked the eVects of IL-2. We further showed the presence of IL-2 receptors ( subunit) and GABAA receptors in the medial hypothalamus, which provides anatomical support for the relationship between IL-2 and GABAA receptors. As indicated above, there is a reciprocal feedback loop and functional link between the medial hypothalamus and PAG that underlies defensive rage behavior. We thus determined whether microinjections of IL-2 into the PAG would also modulate defensive rage (Bhatt and Siegel, 2006). A striking Wnding was that, in contrast with its inhibitory eVects in the hypothalamus, IL-2 facilitated defensive rage when microinjected into the PAG, which diVers from the
uniform eVects associated with IL-1 in both hypothalamus and PAG. Hence, IL-2 produces opposite eVects on this form of feline aggression, which is dependent upon the site of its application. There were similarities and diVerences in the underlying mechanisms between the hypothalamus and PAG. As was the case in the hypothalamus, the behavioral eVects of IL-2 were blocked by pretreating the injection site in the PAG with an antibody against the subunit of the IL-2 receptor or by a monoclonal antibody to IL-2. However, the region-speciWc eVects of IL-2 were mediated by diVerent neurotransmitter receptor mechanisms. Facilitation of defensive rage in the PAG is mediated by substance P-NK1 receptors in the PAG. It is thus of unique interest that an NK1 receptor antagonist blocked the potentiating eVects of IL-2 in the PAG. Consistent with these Wndings are data obtained from our immunocytochemical-anatomical studies which provided evidence of the presence of IL-2 receptors in the region of the dorsal PAG from which defensive rage is typically elicited (Bhatt and Siegel, 2006). In summary, we have discovered that IL-1 and IL-2, acting through exquisite cytokine-, neurotransmitter-, and region-speciWc mechanisms, signiWcantly modulate defensive rage behavior. Fig. 1 illustrates cytokine modulation of defensive rage within the medial hypothalamus and PAG as well as the neurotransmitter receptor mechanisms that mediate these eVects. As discussed earlier, our electrical stimulation model of feline defensive rage is ethologically relevant since this behavior occurs under natural conditions when the animal, its territory, or its kittens, are threatened. As well, this form of aggressive behavior is a close animal analog of a parallel form of human aggression. Accordingly, we conclude that brain IL-1 and IL-2 act as important endogenous mediators of this form of aggressive behavior. Although there are links between immune responding and aggressive behavior, it should be underscored that we did not utilize a model of immune activation (e.g., endotoxin
Fig. 1. Schematic diagram depicting speciWcity of cytokine modulation of feline defensive rage behavior. Diagram illustrates reciprocal excitatory relationship between medial hypothalamus (MH) and periaqueductal gray (PAG) in mediating defensive rage behavior. The reciprocal inhibitory connections between medial and lateral hypothalamus are included to illustrate the relationship of these regions to predatory attack. Potent facilitation of defensive rage in MH is mediated through IL-1 type I and 5-HT2 receptor mechanisms. Preliminary data suggest the presence of a similar mechanism in the PAG. Suppression of defensive rage behavior in MH is mediated through IL-2 and GABAA receptor mechanisms. In contrast, IL-2 in the PAG powerfully potentiates defensive rage behavior. Such facilitation is mediated through IL-2 receptor and substance P-NK-1 receptor mechanisms.
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
challenge) or observe classic symptoms of sickness behavior in our studies. Nonetheless, it may be useful to employ one or more of the constructs that apply to our understanding of the role of cytokines (particularly IL-1) in mediating sickness behavior in clarifying cytokine modulation of defensive rage behavior. Concerning the facilitating eVects of IL-1 upon defensive rage behavior, one may postulate that an increase in defensive postures could help ward oV a potentially threatening confrontation, and thus be of adaptive value during an immune response. In this context, it is possible that defensive-aggressive behaviors involve processes that are associated with certain components of sickness behavior. Thus, one may view that IL-1, acting centrally through IL-1 type I and 5-HT2 receptor mechanisms, may reXect such a process and play an important role in mediating defensive rage behavior. Moreover, it would seem reasonable to suggest that during illness in humans, immune activation could result in increased anger/hostility, irritability and perhaps in other forms of defensive–aggressive behaviors. Because IL-2 can induce various eVects on a given behavior, it becomes more problematic in employing a sickness model to account for its eVects upon defensive rage behavior. It is thus of unique interest that the modulating eVects of IL-2 upon defensive rage are dependent upon the brain region into which it is microinjected. 5. Concluding remarks The relationship between cytokines, immunity, and aggressive behavior has been studied on many levels. This includes studies of the relationship between cytokines, immunity and anger/hostility in patient and non-patient populations, studies examining the immunological consequences of an aggressive confrontation, as well as links between social status, genetic factors and immunity. We suggest a new direction that uses a feline model to directly determine the role of cytokines on the expression of defensive rage behavior, which is a close animal analog of a parallel form of human aggression. This approach has allowed us to identify the role of IL-1 and IL-2 in the medial hypothalamus and PAG as powerful modulators of defensive rage behavior. Of further importance, we have shown that these cytokines modulate defensive rage in highly speciWc ways, and that they do so through selective cytokine-neurotransmitter receptor mechanisms. Acknowledgment This work was supported by NS 07941-34. References Anisman, H., Kokkinidis, L., Merali, Z., 2002. Further evidence for the depressive eVects of cytokines: anhedonia and neurochemical changes. Brain Behav. Immun. 16, 544–556. Avitsur, R., Stark, J.L., Dhabhar, F.S., Sheridan, J.F., 2002. Social stress alters splenocyte phenotype and function. J. Neuroimmunol. 132, 66–71.
513
Bear, D.M., 1979. Temporal lobe epilepsy—a syndrome of sensory-limbic hyperconnection. Cortex, 15357–15384. Bhatt, S., Zalcman, S., Hassanain, M., Siegel, A., 2005. Cytokine modulation of defensive rage behavior in the cat: role of GABAA and IL-2 receptors in the medial hypothalamus. Neuroscience 133, 17–28. Bhatt, S., Siegel, A., 2006. Potentiating role of interleukin 2 (IL-2) receptors in the midbrain periaqueductal gray (PAG) upon defensive rage behavior in the cat: role of neurokinin NK(1) receptors. Behav. Brain Res. 167, 251–260. Capuron, L., Ravaud, A., Miller, A.H., Dantzer, R., 2004. Baseline mood and psychosocial characteristics of patients developing depressive symptoms during interleukin-2 and/or interferon-alpha cancer therapy. Brain Behav. Immun. 18, 205–213. Cohen, S., Line, S., Manuck, S.B., Rabin, B.S., Heise, E.R., Kaplan, J.R., 1997. Chronic social stress, social status, and susceptibility to upper respiratory infections in nonhuman primates. Psychosom. Med. 59, 213–221. Dantzer, R., Aubert, A., Bluthe, R.M., Gheusi, G., Cremona, S., Laye, S., Konsman, J.P., Parent, P., Kelley, K.W., 1999. Mechanisms of the behavioural eVects of cytokines. Adv. Exp. Med. Biol. 461, 83–105. Dunn, A.J., 2001. EVects of cytokines and infection on brain neurochemistry. In: Ader, R., Felten, D.L., Cohen, N. (Eds.), Psychoneuroimmunology. Academic Press, pp. 649–666. Granger, D.A., Hood, K.E., Dreschel, N.A., Sergeant, E., Likos, A., 2001. Developmental eVects of early immune stress on aggressive, socially reactive, and inhibited behaviors. Dev. Psychopathol. 13, 599–610. Hassanain, M., Zalcman, S., Bhatt, S., Siegel, A., 2003. Interleukin-1 in the hypothalamus potentiates feline defensive rage: role of serotonin-2 receptors. Neuroscience 120, 227–233. Hassanain, M., Bhatt, S., Zalcman, S., Siegel, A., 2005. Potentiating role of interleukin-1 (IL-1) and IL-1 type 1 receptors in the medial hypothalamus in defensive rage behavior in the cat. Brain Res. 1048, 1–11. Imeri, L., Mancia, M., Opp, MR., 1999. Blockade of 5-hydroxytryptamine (serotonin)-2 receptors alters interleukin-1-induced changes in rat sleep. Neuroscience 92, 745–749. Kiecolt-Glaser, J.K., Loving, T.J., Stowell, J.R., Malarkey, W.B., Lemeshow, S., Dickinson, S.L., Glaser, R., 2005. Hostile marital interactions, proinXammatory cytokine production, and wound healing. Arch. Gen. Psychiatry 62, 1377–1384. Kraus, M.R., Schafer, A., Faller, H., Csef, H., Scheurlen, M., 2003. Psychiatric symptoms in patients with chronic hepatitis C receiving interferon alfa-2b therapy. J. Clin. Psychiatry 64, 708–714. Maier, S.F., Watkins, L.R., Nance, D.M., 2001. Multiple routes of action of interleukin-1 on the nervous system. In: Ader, R., Cohen, N., Felten, D.L. (Eds.), Psychoneuroimmunology. Academic Press, San Diego, pp. 563–583. McHuthison, J.G., Gordon, S.C., SchiV, E.R., et al., 1998. Interferon alfa2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C. N. Engl. J. Med. 339, 1485–1492. Nunn, K., 1986. The episodic dyscontrol syndrome in childhood. J. Child Psychol. Psychiat. 27, 439–446. Petitto, J.M., Gariepy, J.L., Gendreau, P.L., Rodriguiz, R., Lewis, M.H., Lysle, D.T., 1994. Difefrences in NK cell function in mice bred for high and low aggression: genetic linkage between complex behavioral and immunological traits? Brain Behav. Immun. 13, 175–186. Seto, D., Kar, S., Quirion, R., 1997. Evidence for direct and indirect mechanisms in the potent modulatory action of interleukin-2 on the release of acetylcholine in rat hippocampal slices. Br. J. Pharmacol. 120, 1151–1157. Shavit, Y., Wolf, G., Goshen, I., Livshits, D., Yirmiya, R., 2005. Interleukin-1 antagonizes morphine analgesia and underlies morphine tolerance. Pain 115, 50–59. Shintani, F., Kanba, S., Nakaki, T., Nibuya, M., Kinoshita, N., Suzuki, E., Yagi, G., Kato, R., Asai, M., 1993. Interleukin-1 beta augments release of norepinephrine, dopamine, and serotonin in the rat anterior hypothalamus. J. Neurosci. 13, 3574–3581. Siegel, A., Roeling, T.A.P., Gregg, T.R., Kruk, M.R., 1999. Neuropharmacology of brain-stimulation-evoked aggression. Neurosci. Biobehav. Rev. 23, 359–389.
514
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
Siegel, A., 2005. The Neurobiology of Aggression and Rage. CRC Press, Boca Raton, Fl. Stefanski, V., Ben-Eliyahu, S., 1996. Social confrontation and tumor metastasis in rats: defeat and beta-adrenergic mechanisms. Physiol. Behav. 60, 277–282. Suarez, E.C., Lewis, J.G., Krishnan, R.R., Young, K.H., 2004. Enhanced expression of cytokines and chemokines by blood monocytes to in vitro lipopolysaccharide stimulation are associated with hostility and severity of depressive symptoms in healthy women. Psychoneuroendocrinology 29, 1119–1128. Vitiello, B., Behar, D., Hunt, J., StoV, D., Ricciuti, A., 1990. Subtyping aggression in children and adolescents. J. Neuropsychiat. 2, 189–192.
Wu, X., Li, H.D., Li, X.C., Ruan, H.Z., Wang, J., 1999. EVects of intra-hippocampal injection of interleukin-2 on pain threshold and formaldehyde-induced substance P-like immunoreactivity in periaqueductal gray and spinal cord. Zhongguo Yao Li Xue Bao 20, 839–843. Zalcman, S., Green-Johnson, J.M., Murray, L.J.M., Nance, D.M., Dyck, D.G., Anisman, H., Greenberg, A.H., 1994. Cytokine-speciWc central monoamine alterations induced by interleukin (IL)-l, IL-2 and IL-6. Brain Res. 643, 40–49. Zalcman, S., Murray, L., Dyck, D.G., Greenberg, D.G., Nance, D.M., 1998. Interleukin (IL)-2 and IL-6 induce behavioral-activating eVects in mice. Brain Res. 811, 111–121.
Invited minireview
The neurobiology of aggression and rage: Role of cytokines Steven S. Zalcman a,¤, Allan Siegel a,b a
b
Department of Psychiatry, UMDNJ-New Jersey Medical School, 183 South Orange Avenue, Newark, NJ 07103, USA Department of Neurology and Neurosciences, UMDNJ-New Jersey Medical School, 183 South Orange Avenue, Newark, NJ 07103, USA Received 11 April 2006; received in revised form 30 May 2006; accepted 30 May 2006 Available online 30 August 2006
Abstract Recent studies have suggested an important relationship linking cytokines, immunity and aggressive behavior. Clinical reports describe increasing levels of hostility, anger, and irritability in patients who receive cytokine immunotherapy, and there are reports of a positive correlation between cytokine levels and aggressive behavior in non-patient populations. On the basis of these reports and others describing the presence or actions of diVerent cytokines in regions of the brain associated with aggressive behavior, our laboratory embarked upon a program of research designed to identify and characterize the role of IL-1 and IL-2 in the hypothalamus and midbrain periaqueductal gray (PAG)—two regions functionally linked through reciprocal anatomical connections—in the regulation of feline defensive rage. A paradigm involved cytokine microinjections into either medial hypothalamus and elicitation of defensive rage behavior from the PAG or vice versa. These studies have revealed that both cytokines have potent eVects in modulating defensive rage behavior. With respect to IL-1, this cytokine facilitates defensive rage when microinjected into either the medial hypothalamus or PAG and these potentiating eVects are mediated through 5-HT2 receptors. In contrast, the eVects of IL-2 are dependent upon the anatomical locus. IL-2 microinjected into the medial hypothalamus suppresses defensive rage and this suppression is mediated through GABAA receptors, while microinjections of IL-2 in the PAG potentiate defensive rage, in which these eVects are mediated through NK-1 receptors. Present research is designed to further delineate the roles of cytokines in aggressive behavior and to begin to unravel the possible signaling pathways involved this process. © 2006 Elsevier Inc. All rights reserved. Keywords: Cytokines; Interleukin-1; Interleukin-2; Aggression; Defensive rage; Serotonin; GABA; NK-1
1. Introduction There is a consensus that the immune and central nervous systems (CNS) communicate in a bi-directional manner. A fundamental principle of psychoneuroimmunology is that this communication helps regulate an orchestrated immune response (Dantzer et al., 1999). This notably includes: (1) activation of neuroimmune feedback loops, and (2) stimulation of central neurochemical alterations that in turn underlie adaptive behavioral and physiological responses (i.e., the classic symptoms of sickness behavior). Cytokines released during the course of an immune
*
Corresponding author. Fax: +1 973 972 8305. E-mail address: [email protected] (S.S. Zalcman).
0889-1591/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bbi.2006.05.002
response act as messengers that help modulate sickness behavior by inXuencing relevant neurotransmitter systems, and in some cases, by directly acting within the brain (Maier et al., 2001). Since cytokines also act as endogenous neuromodulators, it is possible to study their relationship to behavior independent of an ongoing immune response. In this context, it has been recently shown that cytokines are present in brain regions, such as the hypothalamus and midbrain periaqueductal gray (PAG), that are known to be associated with aggression and rage behavior (Siegel et al., 1999). Moreover, there is evidence that anger and hostility are increased in patients receiving repeated cytokine immunotherapy, supporting the view that cytokines may facilitate the expression of aggressive behavior. Accordingly, the following review will discuss research investigating the
508
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
relationship between cytokines, immunity, and aggressive behavior. This includes studies detailing the immunological consequences of an aggressive confrontation, the relationship between cytokines and aggressive behavior, links between social status and immune function and eVects of immune responding on the animal’s response to an aggressive challenge. We will also discuss our research program that uses an established model of feline aggressive behavior to directly determine the role of brain cytokines (notably interleukin (IL)-1 and IL-2) on the expression of defensive rage behavior, which is a close animal analog of a parallel form of human aggression. This approach allows us to identify speciWc brain sites wherein cytokines may modulate aggressive behavior, systematically determine the nature of their eVects, and identify speciWc cytokine-neurotransmitter receptor mechanisms that underlie these eVects. 2. Cytokines, immunity and aggressive behavior The relationship between cytokines, immunity, and aggressive behavior has been studied on many levels in human and infrahuman populations. In this section, we will present an overview of representative studies examining the aggression–immunity relationship. This includes: (a) human studies that evaluate anger/hostility in patients receiving repeated cytokine immunotherapy, or coordinate assessments of anger/hostility and endogenous cytokine or immune cell activity in non patient populations; and (b) animal studies examining immunological proWles of aggressive and submissive individuals, and the relationship between strain, social status and immunity. In recent years, an increasing amount of attention has focused on psychiatric abnormalities associated with cytokine administration (notably interferon (IFN)-2b and/or interleukin (IL)-2), which are used in the treatment of cancer, AIDS, and hepatitis C, among other disorders (e.g., Capuron et al., 2004). Studies have shown that cytokine immunotherapy results in a facilitation of aggressive behavior. SpeciWcally, measures of anger/hostility and irritability are increased in patients receiving cytokine immunotherapy (e.g., Kraus et al., 2003; McHuthison et al., 1998). It is important to consider that various factors may have contributed to these observations, including comorbid psychiatric abnormalities and processes related to the underlying disorder being treated. Indeed, patients also display depression, anxiety, and cognitive diYculties, among other psychiatric disturbances. Moreover, it is possible that an interaction between cytokine treatment and the disease process (e.g., chronic infection and subsequent alterations of cytokine activity) inXuences the expression of aggressive behavior. Support for this suggestion stems from a study by Kraus et al. (2003) who evaluated anger/hostility in patients with chronic hepatitis C infection before and after onset of IFN therapy. They found that anger/hostility was increased in patients before onset of cytokine treatment and that these increases were further augmented following onset of therapy. To be sure, one must temper one’s conclusions
concerning the relationship between cytokines and measures of aggressive behavior pending analyses of the possible contributions of other factors, including actual regimens of cytokine therapy, the use of other treatments, and certain demographic factors, among other factors. Studies using non-patient populations also suggest a link between aggressive behavior and cytokines. In these studies, measures of aggressive behavior are related to endogenous blood cytokine production. For example, Suarez et al. (2004), using healthy subjects, showed that Cook–Medley Hostility scores are increased coincident with an enhancement of LPS-stimulated production of proinXammatory cytokines by blood monocytes. Consistent with these Wndings, Kiecolt-Glaser and colleagues (2005) showed that higher levels of hostile marital interactions are associated with increased production of plasma proinXammatory cytokines. Taken together, these studies show that a positive association exists between cytokines, immune cell activity, and various measures of aggressive behavior in patient and non-patient populations. Increases in aggressive behavior have also been linked to enhancements of cytokine production and immune cell activity in animal studies. For example, Petitto et al. (1994) showed that IFN, IL-2 production, and T cell proliferation were higher in mice bred for high aggression (based on frequency of attacks after contact) than mice bred for low aggression. Inasmuch as these eVects were not related to diVerences in postweaning social experience or to gender diVerences, the authors suggested that a genetic linkage exists for genes associated with aggression and immunity. It is important to consider that aggressive confrontations can result in both immunoenhancing and immunosuppressive eVects. Avitsur and colleagues (2002) used a paired Wghting model to determine the eVects of Wghting on measures of splenic cell distribution and function. They found that six daily sessions resulted in an increase in monocytes and neutrophils, and a decrease in lymphocyte measures. These eVects occurred coincident with a state of glucocorticoid resistance in splenocytes. The authors suggested that the immune alterations associated with aggressive confrontations may be related, at least in part, to wound healing. Another model of social stress that has been used to examine the aggression–immune relationship is the social confrontation model or resident-intruder confrontation. In this model, an intruder is placed into an aggressive animal’s home cage. The intruder is typically attacked and defeated, and tends to show submissive or subdominant behavior. Stefanski and Ben-Eliyahu (1996) used this model to study the consequences of social confrontation on tumor retention. They found that rats receiving mammary tumor cells one hour into a seven hour confrontation had increased tumor retention, and that this eVect was blocked by a adrenergic antagonist. Studies have also examined the eVects of social status on measures of immunity. Typically, subdominant animals are immunosuppressed and show an
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
increased vulnerability to infection (e.g., Cohen et al., 1997). Finally, while a signiWcant amount of attention has focused on immunological consequences of aggressive encounters, there is some evidence that early life immune activation may induce long-lasting eVects on aggression. For example, Granger et al. (2001) showed that early life exposure to endotoxin reduced aggressive behavior in mice from a highaggressive line. Taken together, these studies demonstrate that exposure to an aggressive confrontation has marked eVects on various aspects of immune function, and that the type of confrontation and behavioral response inXuences the nature of the immunological changes. It is also important to consider that these changes may be due to a variety of factors, including wound healing, the stressfulness of the confrontation, the social status of the individual, and developmental factors. 3. Feline models of aggression and rage behavior Feline models have shed considerable light on the neuroanatomical substrates and neurotransmitter receptor mechanisms that regulate aggression and rage. As mentioned, we have used such models to directly determine the role of brain cytokines on forms of aggressive behavior, and to identify the cytokine-neurotransmitter receptor mechanisms that regulate these eVects. In the following section, we will provide an overview of these models. An expanding body of data indicates that aggressive behavior appears as a component of numerous clinical disorders associated with abnormal brain function, including aVective disorders, brain tumors, temporal lobe epilepsy, among other abnormalities (Bear, 1979). Such behavior usually falls into the category of defensive rage. Most notable is the aggression referred to as “episodic dyscontrol,” which is seen in patients with temporal lobe epilepsy or hypothalamic tumors. Typically, such patients display an explosive personality and may be physically or verbally assaultive in response to little or no provocation. Moreover, this disorder has been described in both adults and children (Nunn, 1986) lending credence to the idea that the same neural substrate underlies defensive-ragelike aggression over the course of brain development. Notably, this form of aggression is linked to structures associated with the limbic–hypothalamic–midbrain periaqueductal gray (PAG) axis (Siegel et al., 1999). Because the explosiveness of behavior and marked sympathoadrenal arousal in these disorders also occur in defensive rage behavior in the cat, it is our belief that this cat behavior is an appropriate model for study of the human disorders. Although fewer clinical studies have considered predatory forms of aggressive behavior, we likewise believe that it, too, is an appropriate model for study in the laboratory because it has been clearly identiWed in human studies (Vitiello et al., 1990). The discussion related below concerns the models of defensive rage and predatory attack behavior in the cat that have been studied extensively in our laboratory for many years.
509
Over the past three decades, the neural bases of defensive rage and predatory attack behavior, which can be elicited both under natural conditions and by electrical brain stimulation, have undergone systematic examination in our laboratory and elsewhere (see Siegel, 2005; Siegel et al., 1999). In attempting to understand the neural substrates of aggression, a crucial distinction must be made between two classes of neural structures. The Wrst class of structures, which include the hypothalamus and PAG, produce the expression of an attack response when stimulated at suprathreshold current levels. These structures, and the pathways arising from them, mediate, or carry the neural signals necessary for, the motor and autonomic aspects of aggression. Sites within these structures are termed, for example, “attack” sites, or more speciWcally, “defensive rage sites” or “predatory attack sites.” The second class comprises limbic structures whose stimulation does not produce an attack response; rather, stimulation of these structures increases or decreases the probability of occurrence of responses elicited by stimulation of an attack site. In other words, these structures modulate aggression and are termed “modulatory sites.” 3.1. Defensive rage behavior This form of aggressive behavior, initially described earlier in the last century, involves pronounced aVective signs such as vocalization, marked pupillary dilatation, retraction of the ears, arching of the back, and unsheathing of the claws (see Siegel, 2005; Siegel et al., 1999). Moreover, the cat will frequently modify its response by striking at a moving object. This response is ethologically signiWcant in that it occurs under natural conditions when the animal, its territory, or its kittens, are threatened by another animal of the same or diVerent species. Defensive rage behavior is elicited reliably by electrical stimulation of the medial preoptic-hypothalamus and by electrical or chemical stimulation of the PAG (Siegel et al., 1999). The hissing response is used as a measure of defensive rage because as a component of the stimulation-induced defensive rage response, hissing is always elicited on each trial, even in the absence of a threatening stimulus (see Siegel et al., 1999). The occurrence of hissing depends upon the parameters of stimulation and upon experimental manipulations such as paired trials of dual stimulation and drug administration. Because the latency of the hissing response is highly sensitive to such manipulations, their eVects can be clearly identiWed by analyzing the changes in latencies. Hence, for many years the latency for hissing has been a standard dependent variable used in our laboratory (see Siegel et al., 1999) as a measure of defensive rage. The threshold is less sensitive to experimental manipulations and is used as a dependent variable when appropriate. 3.2. Predatory attack behavior In nature, predatory attack behavior occurs as a motivated, stereotyped pattern of behavior directed at a prey
510
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
object. The cat initially stalks the prey (e.g., a rat) and then bites the back of its neck until the animal is killed (Siegel, 2005). Under experimental conditions, virtually the same response can be reproduced by electrical stimulation of the lateral perifornical hypothalamus (see Siegel et al., 1999), the ventral aspect of the midbrain PAG or caudal brainstem tegmentum. In contrast to defensive rage behavior, few autonomic signs other than mild pupillary dilatation are evident. 4. Neurochemistry of defensive rage and predatory attack in the cat–a summary Studies conducted in our laboratory over the past two decades have identiWed those classes of neurotransmitter receptors that facilitate or suppress the expression of defensive rage or predatory attack. These results of these studies are extensively reviewed in Siegel (2005) and Siegel et al. (1999), and are brieXy summarized here. 4.1. Excitatory neurotransmitter receptors Both dopamine and norepinephrine receptors have similar eVects upon neurons in the medial hypothalamus. SpeciWcally, activation of either norepinephrine 2-receptors or dopamine D2 receptors in the anterior medial hypothalamus facilitates defensive rage elicited from the ventromedial hypothalamus and dopamine D2 receptors further facilitate predatory attack as well. Other norepinephrine and dopamine receptors were not found to modulate aggression or rage. In contrast to the catecholamines, 5-HT2 and 5-HT1A receptors have diVerential eVects upon defensive rage behavior in which 5-HT2 receptors in the medial hypothalamus or PAG facilitate defensive rage, while 5-HT1A receptors suppress this response (see discussion below). The principal descending pathway from the medial hypothalamus over which defensive rage is expressed is one whose target is the PAG. The primary transmitter released from these endings is glutamate and its eVects are mediated through NMDA receptors. Two classes of neuropeptides that facilitate defensive rage have been identiWed and include substance P (SP), which acts through NK1 receptors in both the medial hypothalamus and PAG, and cholecystokinin (CCK), which acts through CCK-B receptors in the PAG. 4.2. Inhibitory neurotransmitter receptors Three classes of neurotransmitter receptors have been shown to have potent inhibitory eVects upon rage and/or aggression. Studies conducted on the role of GABA neurons have revealed that there exists reciprocal GABAergic inhibitory connections between the medial and lateral hypothalamus enables the lateral hypothalamus to suppress defensive rage from the medial hypothalamus and for the medial hypothalamus to suppress predatory attack
from the lateral hypothalamus and that these functions are mediated through GABAA receptors. Another class of receptors that suppresses defensive rage behavior is opioid peptides. SpeciWcally, enkephalinergic neurons arising from the central nucleus of amygdala, project to the PAG, and powerfully suppress defensive rage behavior at the level of the PAG. This inhibitory process is mediated through an opioid receptor mechanism within the PAG. As noted above, inhibition of defensive rage within both the medial hypothalamus and PAG are mediated through 5-HT1A receptors, a third class of inhibitory receptors. The presumed origin of serotonin Wbers that act through 5-HT receptors includes the pontine and midbrain raphe groups of neurons. 4.3. Cytokine eVects upon neurotransmitters associated with feline aggression It is well known that inXammatory cytokines are potent modulators of neurotransmitter activity in brain regions associated with mood, motivation, emotion, and behavioral responses to potentially threatening environmental stimuli (see Anisman et al., 2002; Dantzer et al., 1999; Dunn, 2001). Our studies have shown that IL-1 and IL-2 potently modulate defensive rage behavior. As will be discussed later in this review, we have shown that IL-1 in the medial hypothalamus facilitates defensive rage through a 5-HT2 receptor mechanism. Of further signiWcance is the dual eVect of IL-2 upon defensive rage. SpeciWcally, IL-2 in the medial hypothalamus inhibits defensive rage through a GABAA receptor mechanism, and potentiates this form of aggression through a substance P-NK-1 receptor mechanism in the PAG. Here, we brieXy summarize evidence linking IL-1 and IL-2 with these neurotransmitter systems that are associated with defensive rage and that have been shown to modulate cytokine-induced eVects upon this form of aggressive behavior. IL-1 mRNA and receptors are present in various brain regions, including hypothalamus (see Maier et al., 2001). As discussed earlier, 5-HT2 receptors in medial hypothalamus and PAG facilitate defensive rage behavior. It is thus of unique interest that IL-1 is a potent modulator of central 5HT activity. In rodents, systemic or intracebroventricular injections of IL-1 enhance the levels of tryptophan, hydroxyindolacetic acid (5-HIAA; the major metabolite of 5-HT), and stimulate 5-HT release in hypothalamus and extra-hypothalamic sites (see Anisman et al., 2002; Dantzer et al., 1999; Dunn, 2001). Further to the point, infusion of IL-1 into discrete hypothalamic nuclei likewise inXuences 5-HT activity. In particular, microinjections of IL-1 into the rat medial basal hypothalamus increase 5-HT release within this brain region (Shintani et al., 1993). Functionally, it is abundantly clear that IL-1 induces a broad range of behavioral alterations, and that some of these eVects are mediated centrally through 5-HT receptor mechanisms. For example, Imeri et al. (1999) showed that IL-1-induced sleep alterations are partly mediated through 5-HT2 receptors.
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
Collectively, these studies importantly show that IL-1 increases synaptic release of 5-HT, and that these interactions are evident in defensive rage sites in the hypothalamus. Of further importance is the evidence that certain behavioral eVects of IL-1 are mediated centrally through 5-HT2 receptor mechanisms. IL-2 mRNA and receptors have also been shown to be present in a variety of brain regions, including hypothalamus (see Anisman et al., 2002). IL-2 potently modulates central monoamine activity in hypothalamus and extrahypothalamic sites and induces behavioral changes linked with these systems (see Anisman et al., 2002). Inasmuch as IL-1 and IL-2 induce cytokine-speciWc alterations of central neurochemical activity and behavior (Zalcman et al., 1994, 1998), it might be expected that these cytokines would diVerentially modulate defensive rage behavior. Indeed, as mentioned, IL-2 exerts dual eVects upon defensive rage. In the medial hypothalamus IL-2 suppresses defensive rage; in the PAG it facilitates this form of aggression. The suppressive eVects of IL-2 are mediated through a GABAA receptor mechanism whereas the potentiating eVects occur through a substance P-NK-1 receptor mechanism. As noted above, GABAA receptors in the medial hypothalamus suppress feline defensive rage behavior. There is some evidence that IL-2 interacts with this neurotransmitter system. For example, it has been shown that IL-2 inhibits hippocampal acetylcholine release, and that this occurs through GABAA and GABAB receptor mechanisms (Seto et al., 1997). As previously discussed, substance P, acting through NK-1 receptors in the medial hypothalamus and PAG facilitates defensive rage behavior. Very little is known about the relationship between IL-2 and substance P-NK-1 receptors in the brain. It is thus of unique interest that intra-hippocampal infusion of IL-2 inXuences formaldehyde-induced substance P-like immunoreactivity (Wu et al., 1999). In the immune system, substance P induces an increase in IL-2 receptor expression (see Bhatt and Siegel, 2006). These Wndings suggest that IL-2 and substance P-NK1 receptors interact, and that the stimulation of one receptor modulates the expression of the other. 4.4. Cytokine modulation of defensive rage behavior As noted above, the relationships of a number of the major neurotransmitter systems upon defensive rage behavior and predatory attack in the cat have been identiWed. As well, studies have demonstrated the presence and/or actions of IL-1 and IL-2 in brain regions that regulate feline aggression (see Anisman et al., 2002; Dantzer et al., 1999; Dunn, 2001). Accordingly, a series of experiments were initiated in our laboratory involving regions that mediate defensive rage behavior. Namely, we aimed at identifying the roles of IL-1 and IL-2 in the hypothalamus and PAG elicited from either of these brain sites. Because the principal focus of these studies involved defensive rage behavior, a discussion of possible eVects of cytokines upon predatory attack is not included in the present review.
511
In the Wrst study, we sought to determine whether microinjections of IL-1 into the medial hypothalamus could modulate defensive rage behavior (Hassanain et al., 2003). The paradigm for this study and those that follow was based upon the neuroanatomical relationships that exist between the medial hypothalamus and the PAG. SpeciWcally, the downstream target of the medial hypothalamus mediating defensive rage behavior is the PAG (Siegel et al., 1999). Moreover, the PAG also projects back to the medial hypothalamus, completing a reciprocal feedback loop (Siegel et al., 1999). On the basis of this functional anatomical arrangement between the medial hypothalamus and PAG, a stimulating electrode was implanted into the PAG and a cannula-electrode was implanted into the medial hypothalamus from which defensive rage could be elicited from either site. In order to demonstrate the functional linkage between these two structures, a dual stimulation paradigm was employed to demonstrate that stimulation of the PAG facilitated the occurrence of defensive rage elicited from the medial hypothalamus and which further identiWed the speciWc sites from which this form of aggression are elicited. Then, IL-1 was microinjected into the behaviorally identiWed site via the cannula-electrode and the eVects of such microinjections upon defensive rage elicited from the PAG were observed. The results revealed a potent dose- and time-dependent facilitation of defensive rage (Hassanain et al., 2003). The kinetics of this eVect was of interest in that two separate peaks of maximal facilitation were noted, one occurring at 60 min, post-injection, and the second one occurring at 180 min, post-injection, suggesting the presence of an endogenous release of IL-1 with respect to the second peak. Moreover, these eVects were blocked by pretreatment with either an IL-1 type 1 receptor antagonist or with an IL-1 antibody. Facilitation is associated with 5-HT2 receptors in both hypothalamus and PAG. Therefore, we also determined whether pretreating the defensive rage site in the hypothalamus with a 5-HT2 antagonist would similarly block the potentiating eVects of IL-1. The results clearly indicated that this was case. A further analysis involving immunocytochemistry revealed the presence of 5-HT2 and IL-1 type I receptors in the medial hypothalamus which were co-localized, thus providing further support for the linkage of IL-1 and serotonin receptors (Hassanain et al., 2005). It should also be noted that our laboratory has recently obtained preliminary data indicating that activation of IL-1 receptors in the PAG likewise facilitate defensive rage behavior when elicited from the medial hypothalamus, suggesting that IL-1 receptors in both regions associated with defensive rage behavior mediate similar functions. Parenthetically, given the well-known roles of the PAG and hypothalamus in mediating responses to painful stimuli coupled with IL-1’s role in pain perception and sensitivity (Shavit et al., 2005), it seems reasonable to suggest that overlapping mechanisms may modulate defensive aggressive behavior and the response to painful stimuli. Such a link might be of adaptive value during an aggressive encounter.
512
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
Two important conclusions may be drawn from our observations: (1) Because we observed that relatively low (physiological) doses of IL-1 produced a facilitation of defensive rage, it is reasonable to suggest that these Wndings reXect an endogenous role for IL-1 in the modulation of defensive rage (Hassanain et al., 2003); and (2) that the potentiating eVects of IL-1 upon defensive rage behavior are mediated through IL-1 type I and 5-HT2 receptor mechanisms. Overall, this study represented the Wrst of its kind to show that a cytokine in the brain can modulate a form of aggressive behavior. Since IL-2 also modulates neurotransmitter activity in brain regions that regulate aggressive behavior in the cat, we also determined whether IL-2 could modulate defensive rage. Our Wndings showed that IL-2 modulates defensive rage with exquisite speciWcity. First, in contrast with the eVects of IL-1 upon defensive rage, microinjections of IL-2 into the medial hypothalamus induced a potent dosedependent suppression of this form of aggression (Bhatt et al., 2005). This eVect was blocked by pretreating the injection site with an antibody against the subunit of the IL-2 receptor or by a monoclonal antibody to IL-2. Inhibition of defensive rage is associated with GABAA receptors in the hypothalamus; we thus also showed that a GABAA receptor antagonist blocked the eVects of IL-2. We further showed the presence of IL-2 receptors ( subunit) and GABAA receptors in the medial hypothalamus, which provides anatomical support for the relationship between IL-2 and GABAA receptors. As indicated above, there is a reciprocal feedback loop and functional link between the medial hypothalamus and PAG that underlies defensive rage behavior. We thus determined whether microinjections of IL-2 into the PAG would also modulate defensive rage (Bhatt and Siegel, 2006). A striking Wnding was that, in contrast with its inhibitory eVects in the hypothalamus, IL-2 facilitated defensive rage when microinjected into the PAG, which diVers from the
uniform eVects associated with IL-1 in both hypothalamus and PAG. Hence, IL-2 produces opposite eVects on this form of feline aggression, which is dependent upon the site of its application. There were similarities and diVerences in the underlying mechanisms between the hypothalamus and PAG. As was the case in the hypothalamus, the behavioral eVects of IL-2 were blocked by pretreating the injection site in the PAG with an antibody against the subunit of the IL-2 receptor or by a monoclonal antibody to IL-2. However, the region-speciWc eVects of IL-2 were mediated by diVerent neurotransmitter receptor mechanisms. Facilitation of defensive rage in the PAG is mediated by substance P-NK1 receptors in the PAG. It is thus of unique interest that an NK1 receptor antagonist blocked the potentiating eVects of IL-2 in the PAG. Consistent with these Wndings are data obtained from our immunocytochemical-anatomical studies which provided evidence of the presence of IL-2 receptors in the region of the dorsal PAG from which defensive rage is typically elicited (Bhatt and Siegel, 2006). In summary, we have discovered that IL-1 and IL-2, acting through exquisite cytokine-, neurotransmitter-, and region-speciWc mechanisms, signiWcantly modulate defensive rage behavior. Fig. 1 illustrates cytokine modulation of defensive rage within the medial hypothalamus and PAG as well as the neurotransmitter receptor mechanisms that mediate these eVects. As discussed earlier, our electrical stimulation model of feline defensive rage is ethologically relevant since this behavior occurs under natural conditions when the animal, its territory, or its kittens, are threatened. As well, this form of aggressive behavior is a close animal analog of a parallel form of human aggression. Accordingly, we conclude that brain IL-1 and IL-2 act as important endogenous mediators of this form of aggressive behavior. Although there are links between immune responding and aggressive behavior, it should be underscored that we did not utilize a model of immune activation (e.g., endotoxin
Fig. 1. Schematic diagram depicting speciWcity of cytokine modulation of feline defensive rage behavior. Diagram illustrates reciprocal excitatory relationship between medial hypothalamus (MH) and periaqueductal gray (PAG) in mediating defensive rage behavior. The reciprocal inhibitory connections between medial and lateral hypothalamus are included to illustrate the relationship of these regions to predatory attack. Potent facilitation of defensive rage in MH is mediated through IL-1 type I and 5-HT2 receptor mechanisms. Preliminary data suggest the presence of a similar mechanism in the PAG. Suppression of defensive rage behavior in MH is mediated through IL-2 and GABAA receptor mechanisms. In contrast, IL-2 in the PAG powerfully potentiates defensive rage behavior. Such facilitation is mediated through IL-2 receptor and substance P-NK-1 receptor mechanisms.
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
challenge) or observe classic symptoms of sickness behavior in our studies. Nonetheless, it may be useful to employ one or more of the constructs that apply to our understanding of the role of cytokines (particularly IL-1) in mediating sickness behavior in clarifying cytokine modulation of defensive rage behavior. Concerning the facilitating eVects of IL-1 upon defensive rage behavior, one may postulate that an increase in defensive postures could help ward oV a potentially threatening confrontation, and thus be of adaptive value during an immune response. In this context, it is possible that defensive-aggressive behaviors involve processes that are associated with certain components of sickness behavior. Thus, one may view that IL-1, acting centrally through IL-1 type I and 5-HT2 receptor mechanisms, may reXect such a process and play an important role in mediating defensive rage behavior. Moreover, it would seem reasonable to suggest that during illness in humans, immune activation could result in increased anger/hostility, irritability and perhaps in other forms of defensive–aggressive behaviors. Because IL-2 can induce various eVects on a given behavior, it becomes more problematic in employing a sickness model to account for its eVects upon defensive rage behavior. It is thus of unique interest that the modulating eVects of IL-2 upon defensive rage are dependent upon the brain region into which it is microinjected. 5. Concluding remarks The relationship between cytokines, immunity, and aggressive behavior has been studied on many levels. This includes studies of the relationship between cytokines, immunity and anger/hostility in patient and non-patient populations, studies examining the immunological consequences of an aggressive confrontation, as well as links between social status, genetic factors and immunity. We suggest a new direction that uses a feline model to directly determine the role of cytokines on the expression of defensive rage behavior, which is a close animal analog of a parallel form of human aggression. This approach has allowed us to identify the role of IL-1 and IL-2 in the medial hypothalamus and PAG as powerful modulators of defensive rage behavior. Of further importance, we have shown that these cytokines modulate defensive rage in highly speciWc ways, and that they do so through selective cytokine-neurotransmitter receptor mechanisms. Acknowledgment This work was supported by NS 07941-34. References Anisman, H., Kokkinidis, L., Merali, Z., 2002. Further evidence for the depressive eVects of cytokines: anhedonia and neurochemical changes. Brain Behav. Immun. 16, 544–556. Avitsur, R., Stark, J.L., Dhabhar, F.S., Sheridan, J.F., 2002. Social stress alters splenocyte phenotype and function. J. Neuroimmunol. 132, 66–71.
513
Bear, D.M., 1979. Temporal lobe epilepsy—a syndrome of sensory-limbic hyperconnection. Cortex, 15357–15384. Bhatt, S., Zalcman, S., Hassanain, M., Siegel, A., 2005. Cytokine modulation of defensive rage behavior in the cat: role of GABAA and IL-2 receptors in the medial hypothalamus. Neuroscience 133, 17–28. Bhatt, S., Siegel, A., 2006. Potentiating role of interleukin 2 (IL-2) receptors in the midbrain periaqueductal gray (PAG) upon defensive rage behavior in the cat: role of neurokinin NK(1) receptors. Behav. Brain Res. 167, 251–260. Capuron, L., Ravaud, A., Miller, A.H., Dantzer, R., 2004. Baseline mood and psychosocial characteristics of patients developing depressive symptoms during interleukin-2 and/or interferon-alpha cancer therapy. Brain Behav. Immun. 18, 205–213. Cohen, S., Line, S., Manuck, S.B., Rabin, B.S., Heise, E.R., Kaplan, J.R., 1997. Chronic social stress, social status, and susceptibility to upper respiratory infections in nonhuman primates. Psychosom. Med. 59, 213–221. Dantzer, R., Aubert, A., Bluthe, R.M., Gheusi, G., Cremona, S., Laye, S., Konsman, J.P., Parent, P., Kelley, K.W., 1999. Mechanisms of the behavioural eVects of cytokines. Adv. Exp. Med. Biol. 461, 83–105. Dunn, A.J., 2001. EVects of cytokines and infection on brain neurochemistry. In: Ader, R., Felten, D.L., Cohen, N. (Eds.), Psychoneuroimmunology. Academic Press, pp. 649–666. Granger, D.A., Hood, K.E., Dreschel, N.A., Sergeant, E., Likos, A., 2001. Developmental eVects of early immune stress on aggressive, socially reactive, and inhibited behaviors. Dev. Psychopathol. 13, 599–610. Hassanain, M., Zalcman, S., Bhatt, S., Siegel, A., 2003. Interleukin-1 in the hypothalamus potentiates feline defensive rage: role of serotonin-2 receptors. Neuroscience 120, 227–233. Hassanain, M., Bhatt, S., Zalcman, S., Siegel, A., 2005. Potentiating role of interleukin-1 (IL-1) and IL-1 type 1 receptors in the medial hypothalamus in defensive rage behavior in the cat. Brain Res. 1048, 1–11. Imeri, L., Mancia, M., Opp, MR., 1999. Blockade of 5-hydroxytryptamine (serotonin)-2 receptors alters interleukin-1-induced changes in rat sleep. Neuroscience 92, 745–749. Kiecolt-Glaser, J.K., Loving, T.J., Stowell, J.R., Malarkey, W.B., Lemeshow, S., Dickinson, S.L., Glaser, R., 2005. Hostile marital interactions, proinXammatory cytokine production, and wound healing. Arch. Gen. Psychiatry 62, 1377–1384. Kraus, M.R., Schafer, A., Faller, H., Csef, H., Scheurlen, M., 2003. Psychiatric symptoms in patients with chronic hepatitis C receiving interferon alfa-2b therapy. J. Clin. Psychiatry 64, 708–714. Maier, S.F., Watkins, L.R., Nance, D.M., 2001. Multiple routes of action of interleukin-1 on the nervous system. In: Ader, R., Cohen, N., Felten, D.L. (Eds.), Psychoneuroimmunology. Academic Press, San Diego, pp. 563–583. McHuthison, J.G., Gordon, S.C., SchiV, E.R., et al., 1998. Interferon alfa2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C. N. Engl. J. Med. 339, 1485–1492. Nunn, K., 1986. The episodic dyscontrol syndrome in childhood. J. Child Psychol. Psychiat. 27, 439–446. Petitto, J.M., Gariepy, J.L., Gendreau, P.L., Rodriguiz, R., Lewis, M.H., Lysle, D.T., 1994. Difefrences in NK cell function in mice bred for high and low aggression: genetic linkage between complex behavioral and immunological traits? Brain Behav. Immun. 13, 175–186. Seto, D., Kar, S., Quirion, R., 1997. Evidence for direct and indirect mechanisms in the potent modulatory action of interleukin-2 on the release of acetylcholine in rat hippocampal slices. Br. J. Pharmacol. 120, 1151–1157. Shavit, Y., Wolf, G., Goshen, I., Livshits, D., Yirmiya, R., 2005. Interleukin-1 antagonizes morphine analgesia and underlies morphine tolerance. Pain 115, 50–59. Shintani, F., Kanba, S., Nakaki, T., Nibuya, M., Kinoshita, N., Suzuki, E., Yagi, G., Kato, R., Asai, M., 1993. Interleukin-1 beta augments release of norepinephrine, dopamine, and serotonin in the rat anterior hypothalamus. J. Neurosci. 13, 3574–3581. Siegel, A., Roeling, T.A.P., Gregg, T.R., Kruk, M.R., 1999. Neuropharmacology of brain-stimulation-evoked aggression. Neurosci. Biobehav. Rev. 23, 359–389.
514
S.S. Zalcman, A. Siegel / Brain, Behavior, and Immunity 20 (2006) 507–514
Siegel, A., 2005. The Neurobiology of Aggression and Rage. CRC Press, Boca Raton, Fl. Stefanski, V., Ben-Eliyahu, S., 1996. Social confrontation and tumor metastasis in rats: defeat and beta-adrenergic mechanisms. Physiol. Behav. 60, 277–282. Suarez, E.C., Lewis, J.G., Krishnan, R.R., Young, K.H., 2004. Enhanced expression of cytokines and chemokines by blood monocytes to in vitro lipopolysaccharide stimulation are associated with hostility and severity of depressive symptoms in healthy women. Psychoneuroendocrinology 29, 1119–1128. Vitiello, B., Behar, D., Hunt, J., StoV, D., Ricciuti, A., 1990. Subtyping aggression in children and adolescents. J. Neuropsychiat. 2, 189–192.
Wu, X., Li, H.D., Li, X.C., Ruan, H.Z., Wang, J., 1999. EVects of intra-hippocampal injection of interleukin-2 on pain threshold and formaldehyde-induced substance P-like immunoreactivity in periaqueductal gray and spinal cord. Zhongguo Yao Li Xue Bao 20, 839–843. Zalcman, S., Green-Johnson, J.M., Murray, L.J.M., Nance, D.M., Dyck, D.G., Anisman, H., Greenberg, A.H., 1994. Cytokine-speciWc central monoamine alterations induced by interleukin (IL)-l, IL-2 and IL-6. Brain Res. 643, 40–49. Zalcman, S., Murray, L., Dyck, D.G., Greenberg, D.G., Nance, D.M., 1998. Interleukin (IL)-2 and IL-6 induce behavioral-activating eVects in mice. Brain Res. 811, 111–121.