Not so long ago cytokines were thought to serve only as communication molecules between leukocytes, and other physiological systems were considered totally removed from immunologic reactions. However, the integrated view promoted by research in our laboratory is that the immune system engages in bi-directional communication with other physiologic systems rather than existing as a separate system that operates autonomously. To appreciate the idea that the immune system interacts with other systems, one only needs to recall the last time he or she was ill. In general, people with gram negative bacterial infections exhibit events that are mediated by the central nervous system (CNS)--anorexia, fever, hypersomnia, and lethargy to name a few. This means the immune system, which initially interacts with the pathogenic microorganism, conveys a message to the brain. Our research program has employed the best available technologies to understand the nature of this communication path and to explain the behavioral and physiological effects of cytokines at the whole animal, cellular, and molecular level. Some of our research is summarized below.

The Concept of Cytokines and Sickness. We have used a genetic approach to explore how the immune system interacts with the CNS to reduce food intake and induce sickness. The results from these studies demonstrate that an inbred strain of mice with a genetic defect in a receptor called CD14 is refractory to the behavioral and metabolic effects of Escherichia coli lipopolysaccharide (LPS) injected into the periphery (Physiol. Behav., 1997) or into the brain via an indwelling intracerebroventricular cannula (Brain Res., 1997). We showed that the refractoriness resulted from an inability to secrete cytokines like interleukin (IL)-1b. Therefore, when injected intraperitoneally or intracerebroventricularly with recombinant murine IL-1b, the mice carrying the genetic defect respond normally by reducing food intake, losing weight, and decreasing the amount of time engaged in social exploratory behavior. From a practical viewpoint, these data provide proof that cytokines and not LPS in the brain are intimately involved in regulating the behavioral and metabolic effects of infection by gram negative bacteria and, in doing so, demonstrate that cytokines in the brain represent a pragmatic target for manipulating LPS-induced anorexia. From a theoretical viewpoint, they demonstrate that the brain does not recognize inflammatory stimuli per se, but rather the cytokines they induce. Thus, cytokines are directly responsible for the motivational and cognitive deficits apparent in sick people and animals. This finding explains why different pathogens induce a similar repertoire of responses.

The Concept of Neuroimmunophysiology. A number of years ago we suggested that many of the behavioral and physiological manifestations of chronic debilitating diseases are caused by cytokines produced in the CNS, rather than in the periphery (Physiol. Behav., 1993ab). This hypothesis is based on a series of experiments involving intracerebroventricular injections of LPS and recombinant IL-1b and tumor necrosis factor-a (TNFa ). The important finding is that the anorexia, hypersomnia, fever, secretion of corticosterone, and hypercuppremia induced by stimulating the peripheral immune system with LPS can be mimicked by injecting very small amounts of LPS or cytokine into the CNS. That the effects of LPS are regulated by cytokines such as IL-1b , TNFa , and IL-6 suggested that a cytokine network--complete with cytokine-producing cells and receptors--is present in the CNS.

Hitherto, the metabolic abnormalities associated with chronic debilitating diseases such as acquired immune deficiency syndrome, multiple sclerosis, and certain lymphomas were thought to be caused by cytokines in the periphery. Our research demonstrating that symptoms indicative of cachexia could be induced by introducing inflammatory stimuli into the brain led to the idea that cytokines produced by microglia are, in part, responsible for the metabolic disturbances of infection. From this initial observation, we sought answers to a very fundamental question in neuroimmunophysiology: How do inflammatory stimuli localized in the brain induce peripheral endocrine and metabolic responses? We showed that nanogram levels of LPS injected into the rat brain induced lipolysis and increased plasma levels of a key metabolic cytokine, IL-6 (Am. J. Physiol., 1997). Because anorexia and cachexia are chronic progressive states, we used an osmotic pump and an intracerebroventricular cannula to chronically infuse IL-1b into the rat brain. The results indicated that chronic exposure of the CNS to IL-1b caused chronic anorexia, weight loss, and once again, high plasma levels of IL-6 (Brain Res., 1997). This was an important observation because reports from others had shown IL-6 to induce synthesis of hepatic acute phase proteins and degradation of fat and skeletal muscle. We went on to show that the induction of IL-6 and lipolysis was abrogated by blocking a -adrenergic receptors in the periphery (Am. J. Physiol., 1997). Therefore, increased sympathetic outflow and subsequent induction of IL-6 is one way inflammatory stimuli in the brain disrupt disparate metabolic systems in the periphery.

Having recognized that cytokines in the brain induce sickness behavior as well as peripheral endocrine and metabolic responses, we also sought to explore how this network is regulated. Stimulation of the hypothalamic-pituitary-adrenal axis and the resultant increase in plasma glucocorticoid levels is one of the first and most dramatic effects of LPS. It also is attributed to the cascade of cytokine synthesis and release and has been postulated to be part of a negative feedback mechanism that modulates immunological and inflammatory reactions, including the secretion of cytokines by macrophages. In 1996, we published results showing that glucocorticoids inhibit sickness behavior caused by LPS (Am. J. Physiol.). We hypothesized that glucocorticoids inhibit sickness behavior by reducing cytokine production in the brain. This hypothesis was supported by our finding that in the absence of adrenal glands, rats are more sensitive to the behavioral effects of centrally administered LPS (but not IL-1b ) than their sham-operated controls. The enhanced sensitivity caused by removal of the adrenal glands was blocked by implantation of a slow-release corticosterone pellet (Physiol. Behav., 1997). Since the microglial cell is the principal source of IL-1b (and other pro-inflammatory cytokines) in the brain, we proceeded to establish techniques for isolating and culturing microglia from the mouse brain. Using primary murine microglia cell cultures and a murine microglial cell line we provided evidence that in response to LPS, microglia express mRNA and protein for interleukin-1b converting enzyme (ICE), a cysteine protease that processes pro-IL-1b to its mature biologically active form (Mol. Brain Res., 1997). The synthetic glucocorticoid, dexamethasone, inhibits the expression of mRNA encoding ICE. Thus, glucocorticoids may regulate the production of IL-1b in the brain, in part, by inhibiting ICE.

In a subsequent study we showed that ICE is important for the characteristic anorectic response of mice to ICV LPS (Am. J. Physiol., 1999). Specifically, mice that were deficient in ICE (ICE^-/-) resisted the anorexia caused by ICV injection of LPS, but were sensitive to the anorectic properties of recombinant IL-1b . The typical anorectic response seen in wild-type (WT) mice after LPS was restored in ICE^-/- mice by ICV administration of the ICE analogue, cathepsin G. Conversely, anorexia induced by ICV injection of LPS in WT mice was blocked by prior ICV injection of the ICE antagonist, YVAD.CMK. Furthermore, in situ hybridization immunohistochemistry revealed intense expression of ICE mRNA in the hippocampus and dorsomedial hypothalamus of WT mice after ICV injection of LPS. Thus, ICE mRNA is expressed in brain after ICV injection of LPS and is important for induction of anorexia, presumably because it generates mature IL-1b . These results suggest that preventing generation of mature IL-1b can inhibit anorexia induced by LPS in the brain, and therefore, reveal ICE as a potential target for regulating food intake during brain inflammation.

The Concept of Brain Cytokines and Aging. Conditions that result in increased expression of cytokines in the CNS have profound influences on individuals' mental and physical health. We have suggested one condition that leads to increased expression of pro-inflammatory cytokines in the brain is senescence. Indeed, research in our lab confirmed that IL-6 is increasingly expressed in the brain of mice with age. We measured IL-6 in crude protein extracts from brains of juvenile (1-mo-old), adult (3-mo-old), and aged (24-mo-old) male BALB/c mice and found that the concentration of IL-6 increased with age. To extend this important observation we evaluated spontaneous IL-6 production by glial cells cultured from brains of neonate, adult, and aged mice. An age-associated increase in IL-6 mRNA and supernatant IL-6 concentration was evident, indicating glia from aged mice spontaneously express high levels of IL-6 relative to glia from adult and neonate mice. Flow cytometric analysis revealed that cultures established from aged brain compared to either adult or neonate brain comprised more microglia (i.e., MAC-1-positive cells). Furthermore, the proportion of microglia that was positive for IL-6 increased with age, whereas the proportion of astrocytes that were positive for IL-6 was not age-dependent. We interpreted these results to suggest that microglia are responsible for the increased IL-6 in the brain of aged mice. Current studies are investigating the effect of age on transcription factors that regulate the IL-6 gene. To support our research on aging, we have now established an in-house aging BALB/c mouse colony.

The Concept of Immunological Stress, Cytokines, and Growth. Sick or immune challenged (i.e., exposed to pathogens) animals neither eat well nor grow well. Although cytokines were thought to alter the balance between anabolic and catabolic processes in growing pigs, this hypothesis went untested because of the difficulties associated with measuring cytokines in plasma. Our laboratory established a bioassay for measuring IL-6 in porcine plasma and participated in the development of an ELISA specific for porcine TNFa . We had the idea to stimulate the pig's immune system with LPS and measure concurrently several pro-inflammatory cytokines and metabolites indicative of changes in macronutrient metabolism. The important finding was that the increase in IL-6, TNFa , and cortisol caused by LPS was followed by a substantial increase in plasma urea nitrogen, an indicator of protein degradation (J. Anim. Sci., 1997). Because the pigs were fasted throughout the study, we predicted that the increased plasma urea nitrogen was a result of increased skeletal muscle protein degradation. More recently, we have introduced a novel way to decrease the production of metabolically active cytokines in immune-stimulated pigs (J. Nutr., 1998), and have also shown that there is no advantage to increasing the dietary concentration of lysine for chicks whose food intake is depressed due to immune stimulation (J. Nutr., 1998).

To better study the relationship between disease and animal growth, we have now established an infectious disease model in swine. In our model, early-weaned pigs are housed in disease control chambers and inoculated with Mycoplasma hyopneumoniae. Mycoplasma pneumonia is prevalent in 99% of US swineherds and is characterized by a dry non-productive cough, lung lesions, and depressed growth. The ongoing studies are evaluating the relationship between inflammatory cytokines and protein accretion in pigs with mycoplasma pneumonia.

In addition to using neurosurgical techniques in small rodent species, our group used the intracerebroventricular cannula in swine to explore mechanisms of the neuroimmune axis. Hitherto the cost and/or unavailability of the large amounts of recombinant reagents needed for peripheral administration precluded studying the neuroimmune axis of large domestic food animals in a sophisticated fashion. In 1994, we published a paper in pigs showing that centrally administered corticotropin-releasing hormone increased plasma adrenocorticotropin, induced stress-associated behavior, and caused a rapid and profound suppression in lymphocyte proliferation (Endocrinology). This report introduced a new strategy for studying stress physiology in domestic animals. More recently, we have used the intracerebroventricular cannula in swine to demonstrate that recombinant porcine TNFa acts directly in the brain to reduce food intake, increase plasma cortisol, and induce behavioral symptoms of sickness (Endocrinology, 1997).

In a number of autoimmune, infectious and neoplastic diseases there is a decrease in motivation for food and a number of metabolic irregularities that precipitate degradation of body protein and fat. This shift towards negative energy balance is generally attributed to inflammatory cytokines, particularly TNFa . Plasma TNFa concentration, however, is poorly correlated with anorexia and cachexia. If TNFa is involved in anorexia and cachexia as evidence suggests, it likely acts in a paracrine fashion and thus, interacts with the CNS indirectly, via a mechanism that is yet unknown. The mechanism may be leptin.

Studies in our lab showed that LPS challenge in mice increased plasma leptin via a cytokine-dependent path (Endocrinology, 1998). Using in vitro cell culture techniques, we were able to show that TNFa induced leptin gene expression by acting directly on adipocytes. With TNF receptor knockout mice, we also showed that TNFa induced leptin secretion by activation of the p55 TNF receptor Am. J. Physiol., 2000). In current studies a luciferase reporter construct is being used to determine the transcription factor(s) activated by TNFa to induce leptin gene expression.

Return to top