|
|
||||||||
C. R. Plata-Salamán is in the Division of Molecular Biology, School of Life and Health Sciences, University of Delaware, Newark, DE 197162590, USA.
| Abstract |
|---|
| Introduction |
|---|
|
|
|---|
| Feeding inhibition by cytokines |
|---|
|
|
|---|
, IL-1ß, the long-chain helical cytokine family that includes IL-6 subfamily members and leptin (OB protein), IL-8 (and other chemokines/intercrines), tumor necrosis factor (TNF)-
, and interferon (IFN)-
. Each of these anorexigenic cytokine categories uses a different transducing system. This indicates that cytokine-induced feeding inhibition represents a shared phenomenon. Data on feeding inhibition by cytokines have been obtained from animals and humans (12). The human data from immunotherapeutic protocols using cytokines are compelling. Feeding inhibition is one of the most common neurological manifestations observed during cytokine immunotherapy for viral diseases, cancer, or autoimmune processes.
Cytokines may also participate in the control of feeding in health and disease. Leptin is proposed to be an adiposity signal to regulate food intake and body weight during health. Proinflammatory cytokines are proposed to induce feeding inhibition during acute and chronic disease; feeding inhibition in disease is a component of the acute-phase response characterized by local and systemic reactions. A decrease in eating is a prominent clinical manifestation in disorders of all systems, and diseases with long-term feeding inhibition may be associated with the wasting syndrome (cachexia) (12). Feeding suppression during acute disease (e.g., infections) may have a beneficial effect by decreasing the availability of nutrients to pathogenic organisms. Animal studies have shown that force feeding during an acute infection can increase morbidity and mortality. Long-term feeding suppression during chronic disease, however, has deleterious effects and is associated with malnutrition and immunosuppression.
Dosage and route of administration.
Cytokine-induced feeding inhibition is observed when cytokines are administered in the periphery (e.g., intraperitoneally, intravenously, subcutaneously) or into the brain (e.g., hypothalamus, cerebral ventricles) (11). The magnitude of feeding inhibition depends on the type of cytokine, dosage, duration, and route of administration. Cytokines inhibit feeding when administered peripherally in the microgram range (12). An equivalent feeding inhibition is obtained when cytokines are administered into the brain in the picogram-low nanogram range (femtomolar-picomolar range) (12). This fact shows that low doses of cytokines administered into the brain suppress feeding by direct action in the central nervous system.
Chronic administration of cytokines (IL-1ß, TNF-
) via peripheral routes is accompanied by the development of tolerance to the feeding inhibitory effect. Tolerance is not observed when IL-1ß is injected repeatedly (every second day, i.e., when the feeding inhibitory effect of the preceding dose has subsided) (8). In fact, repeated intraperitoneal injection of IL-1ß results in sensitization to the anorectic effect of IL-1ß (8). Interestingly, tolerance does not develop to the chronic intracerebroventricular administration of IL-1ß within the time frame that tolerance fully develops during the peripheral administration. This evidence suggests that different systems are activated depending on the route of administration.
Specificity.
Cytokine-induced feeding inhibition can be blocked with the appropriate receptor antagonists, monoclonal antibodies, and other cytokine inhibitors (12). This evidence suggests specificity of cytokine action on feeding. It should also be noted that other neurological manifestations induced by specific cytokines, such as fever and sleep changes, can be dissociated from cytokine-induced feeding inhibition (7,12). In addition, the development of taste aversions has been reported to occur only when high doses of cytokines are administered (see Ref. 12 for original citations). However, because of endogenous synergistic cytokine interactions, it remains inconclusive whether the cytokine network induces feeding inhibition through taste aversions. Cytokines may affect macronutrient intake differentially (see Ref. 12). It has been reported that IL-1ß-treated animals ingest relatively the same quantity or a greater quantity of carbohydrate and significantly less protein, whereas the relative fat intake remains unchanged or decreases.
| Mechanisms of cytokine-induced feeding inhibition |
|---|
|
|
|---|
| Peripheral mechanisms |
|---|
|
|
|---|
Cytokines also induce the release of hormones that are considered as putative physiological satiety signals including cholecystokinin, glucagon, insulin, and leptin. Thus the feeding inhibition observed when cytokines are administered peripherally can be direct (i.e., mediated by the cytokine) or indirect (through other endogenous substances released by the cytokine). It should be noted that each cytokine induces an idiosyncratic endocrine profile that is also dependent on the dosage. A cytokine administered peripherally acts on peripheral organs but also on neural substrates.
| Peripheral-to-brain communication |
|---|
|
|
|---|
Other models propose that cytokines generate chemical mediators that can act on brain target sites. Endothelial cells of the cerebrovasculature have cytokine receptors, and data suggest that these receptors, when activated, could generate signals such as nitric oxide or prostaglandins that may modulate neural activities. Mediators from meningeal macrophages and perivascular microglia may also be involved (6).
Significant attention has been devoted to the peripheral
brain communication via afferent fibers (2, 3, 10). This model is based on previous gut-brain peptide work that demonstrated that afferent signaling through the autonomic nervous system was required to modulate various neural responses. For cytokines, this neural afferent signaling was first reported by Niijima, who found that administration of IL-1ß into the portal vein induced a dose-dependent increase in vagal afferent activity. More recent studies reported that total subdiaphragmatic vagotomy attenuated or blocked IL-1ß-induced effects when the cytokine was administered intraperitoneally (10). However, when IL-1ß was administered by other routes that include the intravenous, subcutaneous, or intracerebroventricularmajor routes relevant to disease conditionssimilar effects were observed in total subdiaphragmatic vagotomized and sham rats (3). In addition, studies using the selective vagal deafferentation procedure have shown that both deafferented and sham animals are equally responsive to IL-1ß- or bacterial lipopolysaccharide (an inducer of cytokines)-induced feeding suppression (15). As expected, deafferented animals did not respond to cholecystokinin, which requires intact vagal afferents to inhibit feeding (15). This evidence suggests that the model of vagal afferent signaling requires reevaluation.
Cytokines are also released within the brain (11). Cytokines can be released from immune system cells that cross a compromised blood-brain barrier during neurological diseases such as brain infections or brain tumors. There is also local production of cytokines within the brain. The sources of this intrinsic production are brain macrophages, endothelial cells of the cerebrovasculature, microglia, astrocytes, and possibly neuronal components (see Ref. 12 for original citations). Because of paracrine interactions, brain cells can interact to activate a complete brain immunological-inflammatory cascade in vivo (13).
| Brain mechanisms |
|---|
|
|
|---|
Hypothalamic neurophysiology.
Cytokines can inhibit feeding by modulating putative feeding-regulatory neurons. Neurons that are proposed to participate in the control of feeding respond to glucose. This response is stimulatory in the glucose-sensitive neurons of the hypothalamic ventromedial nucleus (VMN) but inhibitory in the glucose-sensitive neurons of the lateral hypothalamic area (LHA). Data show that cytokines act directly and specifically on these hypothalamic neurons (12). Direct application of IL-1ß specifically and reversibly excites the glucose-sensitive neurons in the VMN while suppressing the neuronal activity of the glucose-sensitive neurons in the LHA. Other cytokines (e.g., TNF-
, IFN-
) induce similar effects. IL-1ß and other cytokines modulated the majority of the glucose-sensitive neurons tested (which represent ~30% of the VMN or LHA neurons) but had no effect on the glucose-insensitive neurons in the VMN and LHA. This suggests specificity of neuronal action. Other hypothalamic nuclei (e.g., paraventricular nucleus) are also targets for cytokine action.
Cytokine-induced feeding inhibition and hypothalamic neurophysiology.
Computerized analysis of feeding discriminates among the different components of the meal pattern: meal size, meal duration, meal frequency, and intermeal intervals (see Ref. 12 for original citations). Thus meal pattern analysis provides a detailed description of the elements of eating, revealing the mode of action of manipulations that inhibit or enhance food intake. This allows differentiation between manipulations that modify meal initiation/frequency (integrated with hunger-associated mechanisms) and meal termination (integrated with satiety-associated mechanisms). Meal pattern analysis is also essential to determine brain sites for cytokine action. An action in the VMN (a site involved in the integrative control of meal termination) may be expressed by changes of meal size and meal duration, whereas an action in the LHA (a site involved in the integrative control of meal initiation) may be expressed by changes of meal frequency and prolongation of intermeal intervals. Thus this model predicts that excitation of VMN and inhibition of LHA glucose-sensitive neurons will result in feeding inhibition, by decreasing meal size and duration and meal frequency, respectively.
IL-1ß inhibits feeding by reducing meal size and meal duration when administered into the brain at doses that yield estimated pathophysiological concentrations in the cerebrospinal fluid (Table 1
). [IL-1ß pathophysiological concentrations in the cerebrospinal fluid refer to those observed in patients with human immunodeficiency virus (HIV) infection and bacterial meningitis.] Higher doses (suprapathophysiological range) also decrease meal frequency and prolong intermeal intervals (12). These effects by IL-1ß are blocked by the IL-1-receptor antagonist, a competitive inhibitor for the signaling receptor or IL-1 receptor type I. The meal pattern induced by IL-1ß (depending on the concentration) is in agreement with the neurophysiological effects on hypothalamic neurons, that is, excitation of glucose-sensitive neurons in the VMN and inhibition of LHA glucose-sensitive neurons. Data also show that the neurophysiological effects induced by IL-1ß and other cytokines involve the modulation of various neuronal ionic conductances including sodium-, potassium-, and calcium-channel currents.
|
Catecholamines, serotonin, and histamine are components of three important neurotransmitter systems involved in the control of feeding. Cytokines can modulate these systems. IL-1ß, for example, stimulates the release of catecholamines, serotonin, and histamine. This action of IL-1ß will inhibit feeding.
Cytokines also modulate neuropeptide systems. IL-1ß, for instance, induces the release of corticotropin-releasing factor, a potent feeding suppressant (see Ref. 11). In rodents, the intracerebroventricular administration of IL-1ß completely blocks the feeding-enhancing effect of neuropeptide Y (see Ref. 12).
Cytokine action on the hypothalamic chemistry has clinical relevance (4, 9, 12). A tumor-bearing rat model exhibits increased levels of IL-1 in the cerebrospinal fluid that correlate with the degree of feeding inhibition. In the same rat tumor model, the increase in hypothalamic serotonin concentrations is concomitant with the onset of the decrease in feeding (9); the hypothalamic concentration and release of neuropeptide Y is also reduced in tumor-bearing rats (4).
In the previous section, the direct action of IL-1ß on VMN neurons was considered. The intra-VMN administration of IL-1ß potently suppresses feeding, and IL-1ß modulates the glucose-sensitive neurons specifically. VMN neurons also exhibit an array of G protein
-subunit subclasses of which the G
o protein is predominant. G
o represents a transductional requirement for the modulation of normal feeding. IL-1ß-induced feeding inhibition is associated with a significant decrease of the G
o protein in the VMN. These effects by IL-1ß are blocked by the IL-1-receptor antagonist, and heat-inactivated IL-1ß has no effect. These findings are interesting. G protein-coupled receptors are proposed to participate in the control of feeding. Receptors coupled to G
o that respond to feeding-stimulatory signals include receptors for galanin, endogenous opioids, and possibly neuropeptide Y. IL-1ß has the ability to modulate mechanisms associated with a variety of these neuropeptides, and therefore, IL-1ß-induced modulation of G
o protein in the VMN may be involved in IL-1ß-induced feeding inhibition.
Data are also available on other transducing systems involved in cytokine-induced feeding inhibition. The full-length leptin receptor (predominantly expressed in the hypothalamus) is most related to glycoprotein 130 (gp130), the common signal transducer among receptors for members of the anorexigenic IL-6 subfamily. Thus gp130-related molecules could represent an interface mechanism for the control of feeding in health (via leptin) and disease (via IL-6 subfamily members) (12).
| Cytokine feedback systems |
|---|
|
|
|---|
or IL-1ß plus TNF-
plus IL-8 inhibits feeding synergistically. Computerized analysis shows that this activity depends on a synergistic decrease of meal size without affecting meal frequency when cytokines are used at estimated pathophysiological concentrations. These data show that the specificity on the meal pattern is maintained and is also consistent with the early satiety phenomenon observed during chronic disease or cytokine immunotherapy. Cytokines also induce the production of other cytokines (5). Experimental evidence of IL-1ß upregulating multiple cytokine components during IL-1ß-induced feeding inhibition has been obtained (13). This emphasizes two issues: 1) caution is essential when interpreting data obtained with individual cytokines; and 2) synergistic cytokine interactions may occur during disease conditions.
Cytokine balance.
Data show that the ratio between proinflammatory (e.g., IL-1ß, TNF-
) and anti-inflammatory [e.g., IL-1-receptor antagonist, transforming growth factor (TGF)-ß1] cytokines is relevant to the magnitude of feeding suppression. Studies with bacterial products (cytokine inducers) support this notion (Fig. 1
). Data obtained with an integrative behavioral-molecular in vivo approach show that bacterial lipopolysaccharide (from Gram-negative bacteria) and muramyl dipeptide (MDP; from Gram-positive bacteria) induce feeding inhibition and cytokine upregulation in brain regions including the hypothalamus. However, the profiles induced by lipopolysaccharide and MDP are different. Lipopolysaccharide is significantly more potent than MDP in suppressing feeding and in upregulating IL-1ß and TNF-
mRNAs in the brain (Fig. 1
). MDP, on the other hand, is more potent in upregulating IL-1- receptor antagonist and TGF-ß1 mRNAs. These data suggest that the magnitude of feeding suppression in response to bacterial products is associated with the balance between stimulatory (proinflammatory) and inhibitory (anti-inflammatory) cytokines. This balance may also have implications for the neuroinflammatory events associated with brain infections.
|
| Cytokines and long-term feeding inhibition |
|---|
|
|
|---|
, and TNF-
(see Ref. 12 for original citations). Data to support this notion have been obtained in studies on transgenic animals overexpressing cytokines, passive immunization against cytokines in cachectic rodent models, inoculation with cytokine-producing tumor cells, and transplants of malignant tumors (see Ref. 12). The general concept is that anorexigenic cytokines are upregulated during long-term feeding inhibition associated with cachexia. However, there has been controversy on the criterion to demonstrate cytokine involvement; that is, that cytokine concentrations should be elevated in the circulation of disease models associated with feeding inhibition and cachexia (see Ref. 12). Animal and human studies have reported inconsistent findings. One line of evidence found elevated levels of circulating cytokines in tumor-bearing rats and cytokine levels correlated with parameters of cachexia. Various human studies have also reported subpopulations of patients with elevated circulating cytokine concentrations associated with feeding inhibition and cachexia, including patients with various types of cancer and HIV-1 infection (see Ref. 12). Other reports, however, have not found increased circulating levels of cytokines.
To recapitulate the inconsistencies in the reported circulating levels of cytokines during chronic disease, an alternative model has been proposed. Cytokines have a short-half life, and paracrine/autocrine activities represent a predominant mode of cytokine action. It is known that organs (e.g., brain, liver, spleen) have the capability to produce cytokines through positive feedback systems. That is, once the network is activated within an organ, paracrine interactions can sustain cytokine production independently of cytokine concentrations in the circulation. This model is also consistent with the cascade pattern associated with cytokine production. Data to support this model have been obtained in animal models of disease.
Cytokines also induce metabolic changes and alterations in lipid, carbohydrate, and amino acid metabolism (see Ref. 12 for original citations). These cytokine-mediated metabolic alterations associated with long-term feeding inhibition are deleterious and can contribute to the wasting syndrome through lipolysis and skeletal muscle protein breakdown.
| Conclusions |
|---|
|
|
|---|
| Acknowledgments |
|---|
| References |
|---|
|
|
|---|
and TGF-ß1 mRNAs in obese (fa/fa) and lean (Fa/Fa) Zucker rats: implications to IL-1ß feedback systems and cytokine-cytokine interactions. J. Neurosci. Res. 49: 541550, 1997.[Medline]
This article has been cited by other articles:
![]() |
R. Rorato, A. M. Menezes, A. Giusti-Paiva, M. de Castro, J. Antunes-Rodrigues, and L. L. K. Elias Prostaglandin mediates endotoxaemia-induced hypophagia by activation of pro-opiomelanocortin and corticotrophin-releasing factor neurons in rats Exp Physiol, March 1, 2009; 94(3): 371 - 379. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Mattioli and M. Perfumi Rhodiola rosea L. extract reduces stress- and CRF-induced anorexia in rats J Psychopharmacol, September 1, 2007; 21(7): 742 - 750. [Abstract] [PDF] |
||||
![]() |
C. M POND Adipose tissue, the anatomists' Cinderella, goes to the ball at last, and meets some influential partners Postgrad. Med. J., November 1, 2000; 76(901): 671 - 673. [Full Text] |
||||
![]() |
D. O. McCarthy Cytokines and the Anorexia of Infection: Potential Mechanisms and Treatments Biol Res Nurs, April 1, 2000; 1(4): 287 - 298. [Abstract] [PDF] |
||||
![]() |
A. Inui Cancer Anorexia-Cachexia Syndrome: Are Neuropeptides the Key? Cancer Res., September 1, 1999; 59(18): 4493 - 4501. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |