Neural Activation By Bacterial Superantigens

Alexander W. Kusnecov, Ph.D. and Bruce S. Rabin, MD, Ph.D.

Department of Psychology, Rutgers University, New Brunswick NJ, USA and

Department of Pathology, University of Pittsburgh Medical Center, Pittsburgh, PA, USA.

Introduction

As should clearly emerge from this symposium there is little doubt that the nervous and immune systems share a mutually interactive system of communication. An integral aspect of this interaction involves behavioral repertoires that avoid, eliminate and cope with immunogenic pathogens. It is well known that increased sleep, reduced locomotion, anorexia, and fever occur in response to immune-derived cytokines and bacterial endotoxin (i.e. LPS). However, although these adjustments are adaptive, it is likely they occur at some behavioral expense. Clinically, neuropsychiatric and neuropsychological problems have arisen from the use of cytokine immunotherapy (eg., IL-2) ^{30; 31}. Therefore, while the potential physical benefits of cytokine immunotherapy provide cause for excitement, we are at a loss to describe ways to predict and circumvent the neurological impact of excessive cytokine exposure. Similarly concerns exist regarding critical care conditions (eg., bacterial septicemia and viral infections [eg., HIV]) where severe – at times morbid - physical and neural complications emerge from exuberant cytokine responses. It is highly conceivable, then, that neurobehavioral adjustments to infection and/or cytokine treatment may lead to additional demands on neurochemical resources supporting emotional and cognitive functions, subverting processes that reconstitute central imbalances and destabilizing mechanisms that promote mental equilibrium.

Experimentally, these questions require models that address the breadth of the immune system’s response capacity. Since adaptive immune function is ultimately steered by T cells secreting regulatory cytokines responses ¹⁹, and since these processes are involved in important neuroimmunopathological conditions (eg., encephalomyelitis) ³, the relationship between T cell activation and subsequent behavior deserves serious attention. In an attempt to broaden our understanding of the contribution that T cell cytokines might have on CNS function, we have been studying the protein antigen staphylococcal enterotoxin B (SEB), that in vivo significantly stimulates cytokine production by a subpopulation of T cells (CD4⁺/Vb 8⁺) via binding to the MHC Class II molecule on antigen presenting cells ^{2; 18 8-10 20}. As will be described below, research in in this laboratory has shown that mice challenged with SEB showed significant activation of the HPA axis ²⁶. More recent information shows this to occur through a CRH dependent mechanism, along with increased transcription of CRH mRNA in the PVN and central nucleus of the amygdala (ceA). The impact of these CNS alterations extended to behavioral changes that suggested the presence of increased emotionality under novel contextual conditions interepreted, on the basis of ACTH measurements, to be stressful. These results suggest a clear and testable hypothesis: activation of CD4⁺/Vb 8⁺ T cells with a protein antigen (SEB) results in increased anxiety behavior that is mediated by activated CRH neurons in the central nucleus of the amygdala (CeA). If accurate, this will highlight the potentially important emotional (and cognitive) costs to being immunologically challenged (i.e. during infection) and/or exposed to cytokines (eg., immunotherapy with IL-2). Indeed, there is growing evidence that emotional behavior may be modulated by immune processes ^{4; 5; 5; 16; 21; 27 21; 24; 25}, although there appears to be difficulty in dissociating some of the findings from illness effects ²¹.

Superantigens and Staphylococcal Enterotoxin B

Discrete administration of recombinant cytokines suggests ways in which the immune system may alter neural function, and as will be evident from other presentations in this symposium this approach continues to provide important and meaningful information. An alternative approach unitlizes endogenous cytokine inducers that address the role of CNS reactivity to immune activation within a more biological context. The LPS model has been prominent in this regard. However, the cytokines induced by LPS are largely macrophage-derived, and do not include those produced by T lymphocytes (eg., IL-2 and IFNg ), which are known to have effects on a variety of neural and behavioral functions ¹¹.

Recently, a class of protein antigens has been characterized that have "super" antigenic properties (and hence, are referred to as ‘superantigens’), in that they induce prononounced cytokine production by T cells in vivo, and subsequent to this, drive activated T cells into a lymphoproliferative phase that increases two-fold the initial percentage of the relevant T cell pool. This is quite a marked and impressive effect, since most benign protein antigens, such as ovalbumin and keyhole limpet hemocyanin, and various other immunogenic proteins do not stimulate sufficient numbers of T cells to allow for detectable levels of proliferation and cytokine production in vivo during the primary phase of immunological responding ^{14; 19; 29}. Furthermore, given the wide number of superantigenic molecules of both bacterial and viral origin, and the similarity of the immune effects observed to some autoimmune processes, some investigators have suggested that these molecules may serve not only as models for examining immune processes that take place, but also that surreptitious and stealth-like presence of microbial superantigenic proteins in the host may cause molecular mimicry that disrupts immune tolerance and initiates autoimmune disease ^{3; 15}.

To date, the most extensively studied superantigen is staphylococcal enterotoxin B (SEB), the immunological effects of which have been well characterized in the BALB/c mouse ^{2; 18 8-10}. If endogenous cytokines are the principal mediators of immunological effects on the CNS, it was hypthesized that SEB, as a potent inducer of TNFµ , IL-2 and other cytokines, should be able to exert neural effects. Indeed, our early studies showed this to be the case. Thus, it was found that challenge with SEB dose-dependently stimulated increased ACTH and corticosterone elevations in plasma ²⁶. This was fully T cell dependent, since inhibition of T cell function with cyclosporine A abrogated the HPA response. Depletion of macrophages had no effect, suggesting that any possible activation of these cells did not contribute in a significant way to neuroendocrine activation by SEB. Since these early studies, considerable information has been gathered on the effects of staphylococcal enterotoxin B (SEB) on CNS function: in particular, changes in central c-fos expression, CRH mRNA transcription, and neophobic behavior.

We should note that since these results are now in submission, and owing to the copyright uncertainty of internet symposia, we have chosen to limit actual presentation of some of this data.

GENERAL METHODOLOGY

Our studies have routinely utilized male BALB/cByJ mice (Jackson Laboratories, Bar Harbor) that arrive in the animal facility 5-6 weeks of age. A standard acclimation period of 10-14 days intervenes before experiments commence, and which were approved by the University of Pittsburgh Institutional Animal Care and Use Committee. Staphylococcal enterotoxin B (SEB) was purchased from Sigma (St Louis MO) and lipopolysaccharide (LPS: from E.coli 055:B5) from Difco (Detroit, Mn). Plasmids containing exonic segments of mouse IL-2 and CRH genes were provided by Dr. William Paul (National Institutes of Allergy and Infectious Diseases) and Dr. Audrey Seasholtz (University of Michigan, Ann Arbor), respectively. Animals (approximately 25 g in weight) were injected IP with varying doses of SEB or pyrogen free 0.9% Saline (Baxter Healthcare Corp., Deerfield Ill., USA). In some experiments, additional groups of animals were injected IP with 5 µg LPS. Sacrifice routinely occurred 2 hours after injection, and trunk blood was collected into chilled EDTA-coated vacutainer tubes (Beckton Dickinson, Rutherford, NJ), which is the peak time of the plasma corticosterone response to SEB. In some experiments animals were sacrificed after 4 and 6 hrs to assess splenic and neuronal IL-2 CRH mRNA levels by in situ hybridization using standard protocols optimized in our lab. In addition, using previously published procedures experiments examined c-Fos protein in brains of SEB and LPS challenged mice. All data were analyzed using analysis of variance (ANOVA). Significant main effects were analyzed further by a post hoc Tukey or Neuman-Keuls tests or Bonferroni t-test, while interaction effects were explored by contrast analysis of group means. Treatment effects or group differences were considered to be significant at p < 0.05.

RESULTS

Activation of specific limbic regions of the brain after immunological challenge with SEB. Our previous studies had revealed dose-dependent activation of the hypothalamic-pituitary-adrenal (HPA) axis following immunological challenge with SEB²⁶. The optimal dose for elevating plasma corticosterone concentrations was 50 m g SEB. Consistent with the effects of SEB challenge on elevated plasma ACTH and corticosterone, we reported increased c-Fos immunoreactivity in the paraventricular nucleus (PVN) of the hypothalamus (it should be noted that there was no increased c-Fos immunoreactivity in animals administered 25 m g SEB, a dose that does not elevate plasma corticosterone).

Since our interest has extended to an analysis of the behavioral effects of SEB challenge, we also assessed c-Fos immunoreactivity in brain regions known to mediate cognitive and emotional functions.

As expected, the hypothalamus proved to be a prominent region of activation. Thus, two hours after challenge with SEB, c-Fos was detected in the arcuate nucleus, suprachiasmatic nucleus, supraoptic nucleus, dorsomedial hypothalamus, lateral hypothalamus, and the paraventricular nucleus. In addition, c-Fos immunoreactivity was observed in the amygdala, bed nucleus of the stria terminalis, septum, cingulate gyrus and piriform cortex. Although there was some sporadic detection in the hippocampus, the number of c-Fos immunoreactive cells in the hippocampus did not exceed those observed in saline-injected controls. Since we did not perform an analysis of the brain stem, we do not know whether increased c-Fos immunoreactivity was detected in key hind brain regions involved in stress, such as the locus coeruleus and nucleus of the tractus solitarius.

It is sufficient to conclude from these findings that immunological challenge with SEB produces a significant excitation of neurons within limbic regions of the brain, and that this may serve to induce behaviors reflecting increased emotionality and/or anxiety. In view of the role the amygdala plays in memory for stressful or emotional events, cognitive processes may also be influenced. Indeed, studies with LPS, IL-1, and autoimmune mice have already suggested that emotional behavior may be modulated by immune processes ^{4; 5; 16; 21; 22; 27 5}, ^{21; 24; 25}.

Immunological challenge with SEB stimulates corticotropin releasing factor neurons in the brain. Since SEB challenge elevates plasma ACTH, and induces increased c-Fos immunoreactivity in the PVN and amygdala, we were particularly interested in determining whether these events coincide with activation of neurons synthesizing corticotropin releasing factor (CRH). This neuropeptide promotes numerous endocrine, autonomic, and behavioral functions in the brain. It is synthesized in large amounts in neurosecretory cells in the PVN, as well as in the central nucleus of the amygdala (ceA), where it is considered to play an important role in mediating fear and anxiety. Double-labelling immunocytochemistry revealed that some c-Fos expressing cells in the PVN and ceA also contained CRH (data not shown). In so far as c-Fos immunoreactivity reflects neuronal excitation, this observation suggests that immunoreactive neurons may have been stimulated to increase CRH gene transcription. Consequently, we tested whether CRH-containing neurons in the PVN and ceA displayed increased transcription of the CRH gene.

Animals were sacrificed at the same time of day 0, 2, 4 or 6 hr after challenge with 50 m g SEB. To ensure that T cell activation had occurred we determined by in situ hybridization the level of IL-2 mRNA in the spleen. In addition, plasma was assayed by commercial ELISA for circulating IL-2 protein. The level of mRNA for IL-2 was dramatically elevated in the spleens of SEB challenged mice at all time points. The spleens from Saline injected animals revealed no hybridization for IL-2 mRNA. Plasma IL-2 levels in SEB challenged animals reflected the IL-2 mRNA data (not shown; but see Shurin et al, 1997).

Examination of the brains from all animals did not reveal any hybridization for IL-2 mRNA. This suggests that within the time period examined, T cell activation with SEB is likely to be confined to lymphoid compartments, such as the spleen and lymph nodes.

The brains from these animals were assayed by in situ hybridization for CRH gene transcription. Consistent with other studies utilizing LPS, in vivo stimulation of T lymphocytes with SEB increased CRH mRNA in the PVN and ceA 4 and 6 hours after SEB challenge.

Table 1: Effect of SEB challenge on CRH mRNA levels (percent dpm of 0hr) in the PVN and central nucleus of the amygdala. * p < 0.05 compared to 0hr

Semiquantitative assessment of CRH mRNA at 2 hours after challenge did not result in a significant difference from that of Saline-injected control animals. These results support the information derived from double-labelling for c-Fos and CRH, indicating that CRH neurons are stimulated in SEB challenged animals. Since enhanced mRNA synthesis can reflect compensatory responses to synthesis and release of prestimulated levels of translated product, these results suggest possible release of CRH by neurosecretory cells of the PVN and neurons within the central nucleus of the Amygdala.

Whether increased CRH mRNA in the ceA is a reflection of increased synthesis and release remains to be determined.

Effects of Immunological challenge with SEB on Illness-related Behaviors

No Effect on Conditioned Taste Aversion. Given that immunological challenge with SEB achieves CNS alterations, such as those reported above, we asked whether these changes were secondary or at least associated with a generalized illness caused by elevated cytokine production. Therefore, we tested the ability of SEB to induce conditioned taste aversion or to disrupt consumption of a familiar drinking solution. The basic conditioned taste aversion paradigm involves pairing a novel taste solution with drug administration. If the drug has physiological effects which induce discomfort (as in illness), animals generally avoid or consume less of the taste solution when it is presented again on subsequent days. The conditioning stimulus was Prosobee, a commercially available infant liquid formula solution manufactured by Mead Johnson, and which we have found to be readily consumed by mice without the need for fluid deprivation.

There was no significant reduction in consumption of the CS as a result of having it explicitly paired with immunological changes induced by SEB. Furthermore, it was quite clear that a conditioned taste aversion could be induced if the CS was paired with the administration of 5 m g LPS. This has been reported previously by other laboratories ^{7; 13}, and is consistent with an extensive literature on the effects of LPS challenge on illness-like behavior.

No Effect on Disruption of Ongoing Ingestive Behavior. In other experiments animals were habituated for two days to a daily regimen of 1 hour exposure to a preferred drinking solution (viz., Prosobee). On the third day, they were either injected with saline or a 50 m g dose of SEB. Two hours later they were presented with the Prosobee solution. The results revealed no differences in the consumption of Prosobee solution between saline and SEB challenged mice. This suggests that at the time

when we have routinely observed elevated plasma IL-2 and TNF, and increased c-Fos immunoreactivity in the brain ²⁶, consumption of a palatable drinking solution is not disrupted. This result is also consistent with the conditioned taste aversion experiments which failed to demonstrate any aversive properties of SEB challenge. Furthermore, we have monitored overnight body weight changes in mice challenged with SEB, and found no reduction in body weight that would suggest the presence of illness.

As an interesting aside, and in regard to the behavioral effects of LPS, we have found that doses of LPS below 1 m g per mouse fail to induce conditioned taste aversion or disrupt consumption of a familiar drinking solution. This is in contrast to the same doses being capable of activating the HPA axis. Therefore, there may be a dose-related dissociation between the CNS effects of immunological stimuli, and any illness-like effects such immunological challenges may pose. In other words, CNS activation through an immunological challenge is not sufficient grounds for expecting illness-like behavioral changes, such as anorexia, lethargy, and reduced locomotion. However, given that CNS activation is clearly apparent, and this occurs (as in the case of SEB – see above) in areas serving cognitive/emotional behavior, other behavioral alterations may be expected, and that cannot be discerned using traditional procedures used to assess illness.

Effect of SEB Challenge on Taste Neophobia: Interaction with Isolation Stress

The results of the foregoing experiments indicated that immunological challenge with SEB did not affect ingestion of familiar solutions or solutions paired with administration of SEB. However, we had not tested for the effects of SEB on consumption of a novel drinking solution. When rats and mice are exposed to a solution with a novel taste, initial consumption is relatively low to that observed on subsequent exposures. This is interpreted as neophobic behavior – or fear of novelty – and dissipates with increased familiarity with the taste stimulus. In studies reported above, we had observed in SEB challenged mice increased c-Fos immunoreactivity and CRH gene transcription in the central nucleus of the amygdala. While CRH has been shown to suppress feeding behavior, the data reported above show that activation of CRH neurons is not associated with loss of appetite. Alternatively, it has been confirmed in rats and mice, that CRH is anxiogenic, and within the central nucleus of the amygdala, may mediate fear and/or anxiety ^{6; 12}. Therefore, we have initiated studies to determine whether the immunologically-induced central changes in CRH mRNA that we observed in SEB challenged mice are associated with neophobic behavior. The particular paradigm we used was consumption of a novel tasting fluid in a novel environment. This paradigm is the basis of many common tests of anxiety, especially those assessing anxiogenic and/or anxiolytic treatments ²⁸.

Experiment 1: SEB enhances taste neophobia. Previously unhandled, group-housed animals were challenged with 50 m g SEB or with an equivalent volume of Saline, and 2 hrs later removed from their cages and transferred to an individual holding cage equivalent in size, shape and bedding as the home cage. They were immediately exposed for 1 hr to a drinking bottle containing Prosobee liquid formula solution, to which they had no previous exposure. At the end of the 1 hr exposure, all animals were returned to their home cages and original cage mates. The results revealed a dramatic reduction (SEB: 1.73 ± 0.22, Saline: 0.81 ± 0.1, p < 0.05, N = 8 per group) in consumption of the novel tasting solution, and suggested that the CNS alterations in CRH mRNA after immunological challenge with SEB may indeed influence emotional reactivity.

Experiment 2: Enhancement of taste neophobia is dependent on the novelty of the appetitive context. Based on the procedure used to assess taste neophobia in Experiment 1, we asked whether the reduction in consumption of the novel tasting solution was solely affected by SEB challenge, or whether some influence was exerted by the temporary shift of housing from group to individual. Since research has shown isolation to be a stressor, we conducted an experiment to determine whether the relocation imposed on mice in our experiments was stressful. As expected, 15 and 30 minutes after placement into isolation, there was a significant elevation of plasma ACTH concentration (data not shown, in submission). Therefore, the possibility existed that immunologically challenged animals were sensitized to show enhanced neophobic behavior under environmental conditions that in and of themselves were novel and stressful.

A second experiment tested this hypothesis. Two groups (n=8 per group) received 5 days of habituation to relocation from the home cage to 1hr of isolated housing. Animals from two other groups (n=8 per group) that did not receive habituation to this procedure, were simply handled for 15-30 seconds and returned to the home cage with their cage mates. Two days after the final habituation session, animals in each condition (habituated or not habituated) were either challenged with 50 m g SEB or were administered Saline. Two hours after challenge, all animals were moved into isolated housing and exposed for 1 hr to the novel drinking solution, Prosobee.

The results revealed that SEB challenged animals that had been habituated to brief isolated housing did not drink significantly less than similarly habituated, saline-injected animals. However, SEB challenged animals that had not been habituated to the isolated housing revealed the same original result: significantly reduced consumption of the novel tasting solution. This suggests that immunologically challenged mice displayed taste neophobia only under contextual conditions that were in and of themselves novel and unpredictable. Similarly unhabituated animals that were challenged with Saline did not display enhanced taste neophobia, relative to habituated animals injected with Saline. This suggests that the novel housing conditions per se were insufficient to ward off exploratory behavior that included consumption of a novel tasting food substance. Furthermore, the failure of SEB to affect consumption in habituated animals suggested that under familiar contextual conditions immunologically mediated alterations in neuronal activity of the limbic system, and elevated CRH gene activity in particular, was not associated with augmented taste neophobia.

Experiment 3: Contextual novelty does not influence consumption of a familiar drinking stimulus in SEB challenged mice. The results of Experiment 2 suggested that the consummatory response to a novel solution was affected by an interaction between psychological stress (i.e. a combination of an unfamiliar environment and isolation) and immunological challenge with SEB. However, the psychological processes mediating reduced consumption of the novel taste solution remain unknown. It is possible that immunological challenge sensitizes animals to psychological stressors which then augment neophobic responses to discrete novel stimuli. Alternatively, the immunological challenge may sensitize stressors to alter basic motivational properties, such as appetite. Furthermore, since we did not monitor activity, we do not know whether SEB challenged animals may have been more lethargic when placed into isolation.

Therefore, we manipulated the novelty of the taste solution. On two separate days, two different groups of animals (n=16 per group) were preexposed in their home cage to either a bottle of Prosobee or Water. On the third day, animals in each group were either challenged with 50 m g SEB or an equal volume of Saline. Two hours later, all animals were moved into isolated housing as in previous experiments, and exposed for 1 hr to Prosobee solution. Therefore, for half of the animals Prosobee was a novel taste solution, whereas for all animals the relocation from group-housed conditions to isolation was a novel experience.

The results showed that SEB challenged animals that had previously experienced the taste of Prosobee, did not differ in their level of consumption from similarly experienced Saline-injected controls. However, if SEB-challenged animals were naïve to the taste of Prosobee, consumption was significantly reduced compared to similarly naïve Saline-injected control animals (SEB/NovTaste: 0.79 ± 0.11; Saline/NovTaste: 1.7 ± 0.2; SEB/FamTaste: 2.41 ± 0.21; Saline/FamTaste: 1.96 ± 0.14). Therefore, in spite of the novel (and presumably stressful) contextual conditions, consumption of the preferred drinking solution was affected only if it was novel. Therefore, taste novelty is an important property determining whether the interaction between exposure to a novel environment and SEB challenge induces a behavioral avoidance of the drinking solution in that novel environment. This renders unlikely the possibility that novel contextual conditions interact with immunological challenge to alter motivational processes subserving appetitive behavior, and more likely suggests the operation of augmented taste neophobia.

Conclusion

The results of the behavioral experiments, and the conclusion that psychological stress interacts with immunological challenge to promote enhanced gustatory neophobia, is consistent with the effects of SEB challenge on increased c-Fos immunoreactivity in various regions of the limbic system (eg., amygdala) as well as increased transcriptional activity of the CRH gene in the central nucleus of the amygdala. It remains to be determined whether CRH gene activation is important to the behavioral changes observed, and whether other potential neurochemical mediators may be involved. Neurons synthesizing CRH are highly responsive to immunological stimuli ^{1; 17; 23}, raising the possibility that fundamental alterations in mood and cognition may occur independently of or prior to more obvious signs of infectious illness. In apparent support of this, immunological cytokines [eg., interleukin-1 (IL-1)] and the bacterial endotoxin, lipopolysaccharide (LPS), have been shown to reduce exploratory behavior in novel or precarious test environments^{4; 5; 16; 21; 27}, in some cases dependent on central CRH receptors ⁵, but in others influenced by depressive illness-like effects on locomotion ²¹. The present findings appear to be in support of these data, although we have yet to determine the mechanisms through which T cell activation with SEB promotes the neophobic changes observed, and whether these are strictly dependent on the observed alterations in central CRH mRNA.

Acknowledgements

These studies could not have been conducted without the involvement of Nola Shanks, Ph.D., Galina Shurin, Ph.D., and Rumei Liang, M.D. who contributed significantly to the cognitive and technical aspects of this work.

Reference List