NUTRITION=

APPLIED
THE INTERNATIONAL
AND BASIC
JOURNAL
NUTRITIONAL
OF
SCIENCES

Volume 13 September 1997 Number 9

CURRENT CONCEPTS IN CLINICAL NUTRITION

Nutrition, Immunology, Neuroscience, and Behavior, Part VII

GUEST EDITOR:
CARLOS R. PLATA-SALAMAN, MD, DSC

From the School of Life and Health Sciences, University of Delaware, Newark, Delaware, USA

Cytskines and Cachexia

PATRICK MA’ITHYS, PHD AND ALFONS BILLIAU, MD, PHD

From The Rega Institute, Faculty of Medicine, Catholic University of Leuven, Belgium

Date accepted: 24 March 1997

ABSTRACT

Prolonged production of cytokines associated with cancer and chronic infections, and other long-term immune reactions
is increasingly recognized as a main causal factor of the often severe signs and symptoms that accompany these diseases:
weight loss, anorexia, and metabolic breakdown termed cachexia. The cytokine that initially was held responsible for
causing these changes was tumor necrosis factor (TNF). However, from various studies it has become clear that the action
of TNF can only be understood in the context of simultaneous presence of other cytokines, some of which have activities
that are at the least equally important as TNF in bringing about cachexia. This review summarizes the experimental evidence
for the involvement of cytokines in the pathogenesis of cachexia. Indirect evidence comes from the observation that cachexia
can be induced in animals by repeated injections of cytokines or by inoculation of cytokine-producing cells. Thus, cachexia
has been described in mice inoculated with tumor cells carrying and expressing genes for either TNF, interleukin-6 (IL-
6), leukemia inhibitory factor (LIF), ciliary neurotrophic factor (CNTF) and interferon-y (IFN-y ). More direct evidence
is provided by the observations that cachexia in experimental animal models can be mitigated by administration of specific
antagonists of cytokines. These latter type of studies revealed that cachexia can rarely, if ever, be attributed to one single
cytokine but rather to a set of cytokines that work in concert in cachexia. A pool of anticytokine antibodies or other
cytokine inhibitors might, therefore, be considered as a potential intervention for the treatment of cachectic patients, but
this approach may induce immunosuppression, and, therefore, danger exists that such treatment may benefit the infectious
agent or tumor. Nutrition 1997; 13:763-770. OElsevier Science Inc. 1997

Key words: cachexia, cytokines, cancer, infections, animal models

Correspondence to: Dr. P. Matthys, Laboratory of Immunobiology, The Rega Institute, Minderbroedersstraat 10, B-3000 Leuven, Belgium.

Studies in the authors’ laboratory are funded by the Concerted Actions initiative of the Regional Government of Flanders, the Inter-University
Attraction Pole Program of the Belgian Federal Government, and grants from the Fund for Scientific Research (FWO) of Flanders.

Nutrition 13:763-770, 1997

OElsevier Science Inc. 1997 0899-9007/97/$17.00
Printed in the USA. All rights reserved. PI1 SO899-9007(97)00185-S
764 CYTOKINES AND CACHEXIA

INTRODUCTION Involvement of the Cytokine Network in the Pathogenesis of

Cachexia
Involvement of Humoral Factors in the Pathogenesis of
Cachexia Cytokines are hormone-like proteins that fulfill a role as
chemicals messengers between cells in the body. The most
Cachexia is a syndrome of wasting associated with many salient characteristic of cytokines is that they can act as para-
forms of cancer and chronic diseases. The components of crine or autocrine rather than endocrine messengers. Moreover,
cachexia include weight loss, anorexia, depletion of both fat cells in the body are seldomly confronted with a single cyto-
and protein masses, muscle atrophy, gain in the proportion of kine, but rather with sets of several cytokines that interact syn-
body water, and a variety of metabolic changes.r3’ Although ergistically or antagonistically with each other. Thus, cytokines
investigators have used experimental infections to study the can be said to form local immune-regulatory networks, which
underlying mechanism, most of the available information has are essential for offering defense against noxious influences:
come from studies on cachexia associated with experimental “foreign” factors such as viruses, bacteria, parasites as well
tumors. Cancer-associated cachexia was initially considered to as “self’‘-derived factors such as cancer cells. However, in
be due to the metabolic requirements of the tumor.3 Tumors persistent and chronic diseases, the danger of overproduction
predominantly rely on anaerobic glycolysis and produce large as well as the endocrine spill-over is always present. When
amounts of lactate. This lactate has to be regenerated to glucose this happens, symptoms appear that are not directly due to the
in the liver via gluconeogenesis (Cori cycle), which is an en- infection or the tumor growth itself. Hence, despite being ac-
ergy-demanding process.’ Although abnormal gluconeogenesis tively involved in the removal of foreign elements from the
and lactate levels have been demonstrated in tumor-bearing body, cytokines are often the cause of the inflammatory symp-
animals and cancer patients,4 the energy demand of human toms associated with illness.
tumors rarely fully accounts for cachexia in cancer patients and
The concept that cytokines are involved in cachexia initially
can therefore not solely be responsible for wasting.1.5
arose from the observation that the metabolic derangements in
More recent studies have focused on the role of circulating
cancer patients are similar with those seen in infectious dis-
factors in the pathogenesis of cancer cachexia. Parabiotic trans-
eases. The idea also emerged from the observation that the
fer experiments between tumor-bearing and normal animals
previously isolated “cachectin” was found to be identical to
have provided evidence for the occurrence in the tumor-bearing
TNF, a cytokine that is produced by host immunocompetent
host of such circulating humoral factors responsible for many
cells. Research in several laboratories has then aimed at provid-
of the systemic metabolic disturbances that lead to a cachectic
ing evidence for the involvement of cytokines in the pathogene-
state.6.7
sis of cachexia associated with various inflammatory diseases.
Humoral Factors in Cancer Cachexia: Tumor or Host One line of research that provides indirect evidence for the
Origin? involvement of cytokines in cachexia comes from studies dem-
In the 198Os, many investigators had formulated the hypoth- onstrating increases in cytokine levels in the circulation of
esis that circulating factors that orchestrate the metabolic cachectic patients. Thus, multiple studies have reported in-
changes of cachexia are produced directly by tumor cells. Ki- creased blood cytokine levels in cancer cachexia,‘5*‘6 and in
tada et al.’ reported that a lipid-mobilizing factor was present the HIV wasting syndrome (reviewed in references 17 and 18).
in the serum of thymic lymphoma-bearing mice and in their Furthermore, the loss of body weight as observed in rheumatoid
tumor extracts. A lipolytic substance called toxohormone was cachexia,“*” has been linked to the release of cytokines, in
isolated from cell-free ascites fluid of mice with a sarcoma.g particular TNF. Increased circulating cytokine levels have also
Interestingly, out of the medium of a melanoma cell line been described in patients with inflammatory bowel disease
(SEKI), which induces severe cachexia in nude mice, lo a toxo- (reviewed in reference 21) , chronic liver disease” and chronic
hormone was purified and found to be identical to LIF.” Evi- obstructive pulmonary disease. 23It should be noted that numer-
dence for the production of lipolytic and proteolytic factors ous studies failed to demonstrate cytokine levels in pathologic
was found in mice bearing a colon adenocarcinoma (MAC16 conditions associated with cachexia. However, these negative
tumor I’). The lipolytic factor has recently been isolated and data do not eliminate the interference of cytokines in the patho-
appeared to be a 24 K proteoglycan that was also present in genesis of cachexia. Indeed, cytokines may act in a paracrine
urine of cachectic patients.13 Strassmann et a1.r4 reported that manner and in synergy which each other, at concentrations that
a similar colon adenocarcinoma secretes the cytokine IL-6 and are too insensitive to assess.
that this cytokine mediates cachexia in mice bearing such tu- Another line of research that incriminates cytokines in the
mors. Tumor-produced factors that cause anorexia include sero- pathogenesis of cachexia comes from in vitro experiments,
tonin (produced by gastrointestinal carcinoid cells) and demonstrating that cytokines can induce metabolic changes that
bombesin (produced by small cell lung cancer’). It should be are typically associated with cachexia. Thus, several cytokines
noted that some of these lipolytic tumor factors were purified have been found to exhibit an in vitro “cachectic state” by,
from ascites fluid or from serum of tumor-bearing animals and for instance, stimulating muscle proteolysis or by inhibiting
therefore may have been contaminated with host products. In lipoprotein lipase (LPL), an enzyme that is responsible for the
addition, other “tumor” factors have been identified as known movement of lipoprotein triglycerides from the circulation into
cytokines (IL-6, LIF) or substances (serotonin and bombesin) adipocytes. Whereas these type of observations are often used
that can be produced by host tissue as well. Therefore, the to isolate and identify new cytokines with a cachectic profile,
theory that circulating cachectic factors are only “tumor-de- convincing evidence for the involvement of cytokines in the
rived” substances was at least incomplete. Instead, because the pathogenesis of cachexia comes from in vivo experiments.
metabolic changes seen in cancer cachexia appear very similar Indirect evidence has been derived from studies in which
to those seen following infection (even with killed organisms or chronic administration of cytokines mimics cachexia. Similarly,
bacterial products), it was natural to suggest that host-derived cachexia has been described in mice inoculated with tumor
cytokines, in response to the tumor invasion, might be impli- cells carrying and expressing artificially implanted cytokine
cated in the pathogenesis of cancer cachexia. genes. More direct evidence for the contribution of cytokines
CYTOKINES AND CACHEXIA 765

in cachexia comes from the observation that metabolic alter- suppressed food and water intake, while peripheral administra-
ations in cachectic animals can be counteracted by administra- tion of an equal or higher dose had no such effect.
tion of specific antagonists like monoclonal antibodies that neu- From various other studies in which animals were given
tralize the endogenous cytokine activity. This review summa- repeated injections or a continuous infusion of TNF, it has
rizes the reports in which cytokines have been found to mimic become clear that in vivo administration of TNF is capable of
cachexia in vivo or to contribute to the wasting syndrome in triggering at least some of the clinical symptoms of cachexia.
experimental animal models of cachexia. The general properties However, in some of these studies, the cachectic symptoms
of each cytokine are cited but for more detailed information were found to be transient. Thus, mice or rats treated with daily
the reader is referred to specific reviews. injections of recombinant human TNF (rhuTNF) initially lost
weight but regained body weight after 2-4 d of treatment de-
TUMOR NECROSIS FACTOR spite continued treatment with cachectin.3”*33Failure to induce a
sustained weight loss appeared not to be due to the intermittent
More than a decade ago, a series of investigations, per- character of the injections, since continuous administration of
formed under supervision of Antony Cerami, resulted in the rhuTNF, with the use of mini-osmotic pumps, also failed to
isolation of a protein that was identified as the first humoral induce a long-term weight 10~s.~’Also, the mice became toler-
factor of cachexia. Rabbits that were infected with Trypano- ant to subsequent higher or escalating doses of rhuTNF. The
soma brucei exhibited severe weight loss and hypertriglyceride-
mechanism underlying tolerance to rhuTNF was not defined;
mia associated with an inhibition of LPL.24 A similar, suppres- humoral immune-mediated resistance was excluded.
sion of LPL could be observed in mice after injection of endo- Other studies in support of a role of TNF in cachexia in-
toxin, and could be transferred to normal mice.*’ Subsequently, volved the use of Chinese hamster ovary (CHO) cells, trans-
Cerami’s group isolated a factor from endotoxin-stimulated fected with the TNF gene.34 Implantation of these cells into the
macrophage cells, which inhibited LPL activity in cultured adi- limbs of nude mice resulted in continuous secreted serum levels
pocytes.26 In anticipation of its possible role in the pathogenesis of TNF. Animals receiving the parental CHO cells without the
of cachexia, the discoverers called this factor “cachectin.” TNF gene insert failed to produce detectable levels of serum
Later, the gene encoding cachectin was found to be identical TNF. Both groups of mice maintained normal body weight
with that encoding TNF-a, a cytokine associated with haem- during the first 3 wks. Thereafter, mice that had been injected
orrhagic necrosis of tumors in vivo and with cytotoxic effects with TNF-secreting tumor cells developed severe cachexia with
in vitro. TNF-a is a 17-kDa peptide that is structually and severe body-weight loss and reduced food intake. Autopsy of
functionally related to lymphotoxin or TNF-0. Although the these mice revealed losses of lean and adipose tissue. Such
monocyte/macrophage lineage is believed to be the main source symptoms were not observed in mice inoculated with the paren-
of TNF-a production in vivo, T cells, NK cells, and mast cells tal cells. In a subsequent study, Tracey et a1.35demonstrated
are also capable of producing TNF-a (for review, see reference that the effects associated with implantation of these TNF-
27). Microbial products are potent stimuli for the production secreting cells depend on their route of inoculation. Intracere-
of TNF-cu. Production is synergistically regulated by other cyto- bra1 injection of TNF-secreting cells resulted in body weight
kines, including IL-l and IFN-7. TNF-(Y has been implicated loss and anorexia, but depletion of lipid tissue was not more
in a number of diseases including shock reactions and autoim- pronounced than in pair-fed control animals. By contrast, TNF-
mune diseases. TNF-a is capable of inducing the production transfected cells inoculated in peripheral tissue triggered
of other cytokines such as IL-1 and IL-6. Several in vitro and cachexia including weight loss, depletion of lipid and protein
in vivo biological effects of TNF-(U resemble those of IL-1.27 stores, and anemia, without significant anorexia. Noteworthy,
In vivo studies have provided indirect evidence for the role both types of implantations resulted in equally elevated circulat-
of TNF-a in the pathogenesis of cachexia. Anorexia, weight ing TNF production. These data clearly indicate that TNF can
loss and depletion of whole body protein and lipid stores were reproduce the syndrome of tumor- or infection-associated
induced in rats by repeated sublethal injections of recombinant cachexia, but also that the net metabolic alterations may depend
human TNF-(Y for 1 wk.‘* Weight loss observed in TNF-a- on the site of production. In this context, it is interesting to
treated rats was not significantly different from that of pair-fed note that in keratin promotor-driven TNF-cu transgenic mice,
control animals, indicating that the reduced food intake might cachexia also developed but was accompanied with a graft-
account for the weight loss. However, body composition analy- versus-host disease-like response in the skin.36
sis revealed a significant depletion of body protein in cachectin- Evidence that more specifically delineates the role of endog-
treated rats but not in pair-fed control animals. Fat stores were enous TNF in the pathogenesis of cachexia comes from experi-
also diminished in TNF-a-treated rats but to a lesser degree ments in which the wasting syndrome associated with infections
than in the pair-fed controls. Clearly, the metabolic alterations or with tumor development was attenuated by treatment with
seen after TNF-a treatment resembled cachexia rather than antibodies neutralizing the activity of TNF. Mice bearing a
simple starvation. A similar conclusion was reached by Fong methylcholanthrene-induced fibrosarcoma developed signifi-
et alz9, who found a different pattern of redistribution of body cant anorexia, fat and protein loss within 2 wks post tumor cell
proteins in TNF-treated versus pair-fed or starved rats. Never- inoculation.37 Treatment of the mice with anti-TNF antibody
theless, Mahony and Tisdale3’ reported that weight loss seen significantly reduced hypophagia and the extent of fat and pro-
in TNF-treated mice is directly proportional to the decreased tein loss. Importantly, anti-TNF treatment also attenuated the
food and water intake. In this study, depletion of adipose tissue tumor growth and might, therefore, explain the anti-cachectic
was present in both TNF-injected and pair-fed control mice and property of the antibody. In a subsequent study using this meth-
muscle wasting could not be demonstated in TNF-treated mice. ylcholanthrene-induced sarcoma model, the anti-cachectic ac-
The authors concluded that the weight loss produced by TNF tivity of anti-TNF antibody was compared with that of anti-IL-
has to be seen as a result of reduced food and water intake. 1 receptor antibody. A combination of anti-TNF and anti-IL-l
Evidence to support a role of TNF in the anorexia was presented receptor antibodies had no additive effects on the inhibition of
by Plata-Salaman and colleagues,3’ who demonstrated that in- cancer cachexia, suggesting that the two antibodies acted
jections of TNF directly into the cerebral ventricles of rats through a common mechanism.38 Here again, both antibodies
766 CYTOKINES AND CACHEXIA

inhibited the tumor growth in vitro and in vivo, and this is an fat cells (reviewed in reference 44). Yet, IL-l is a potential
important finding that might explain their effects on cachexia. candidate to contribute to cachexia in vivo, but convincing data
The observed finding that anti-TNF antibody counteracted are scanty. Fong et al. ” found that IL-la was able to match
cachexia development in methylcholanthrene-induced sarcoma- TNF and LPS as inducers of anorexia and cachexia in rats. As
bearing mice was not confirmed in rats. Thus, body weight compared to pair-fed controls, IL-l-treated rats exhibited an
loss, food intake, and body composition of sarcoma tumor- accelerated peripheral protein loss while liver protein was pre-
bearing rats that were treated with anti-TNF antibody did not served, indicating that other mechanisms than hypophagia must
significantly differ from those of control antibody-treated rats.39 account for the observed body weight loss. Another study re-
The ability of the antibody to actually neutralize TNF in vivo ported a transient cachexia-like syndrome in rats after continu-
could not be questioned since a similar antibody treatment was ous infusion of murine IL-l using intraperitoneal implanted
found to be 100% protective against a lethal endotoxin chal- osmotic pumps. Animals infused with recombinant IL-l under-
lenge. Interestingly, in this study the anti-TNF antibodies failed went a marked weight loss and appetite loss as early as day 1
to affect the tumor growth. Taken together, in the methylcholan- but regained weight around day 3 post infusion and retained
threne-induced fibrosarcoma model, the action of TNF on normal body weight by day 1O.45The return of normal body
cachexia might be mediated by its growth-promoting effects weight was not due to osmotic pump failure, nor to the rapid
on the tumor. extent of fibrotic encapsulation of the pumps. More likely, toler-
Yoneda et a14’ studied cachexia in nude mice that were ance to IL-l appeared to be involved since rats failed to elicit
subcutaneously inoculated with a human squamous cell carci- cachexia after an implantation of a second pump. The role of
noma of the maxilla. Tumor growth was associated with severe TNF in this model was excluded since anti-TNF treatment
body weight loss in addition to splenomegaly, hypercalcemia, failed to inhibit weight loss. However, it is difficult to draw a
and leukocytosis. When cachectic animals were injected with firm conclusion from this study. Indeed, although eating activity
polyclonal neutralizing antibodies against TNF, at a time point was found to be significantly reduced in rats that were infused
when mice already displayed clear cachexia, further develop- with IL-l pumps, pair-fed controls were not included and body
ment of body weight loss was significantly damperred. In this composition was not analyed. The ability of IL-l to induce
model tumor size was not affected by the anti-TNF antibody anorexia is a well-recognized phenomenon.3’a46-52 The mecha-
treatment. nism is believed to be either an effect on hepatocytes, which
Loss of body weight and tissue protein has also been demon- subsequently affect the hypothalamic appetite center,47 or a
strated to occur in rats bearing a Yoshida ascites hepatoma.4’ direct effect on the central nervous system.3’~4g-52
Daily administration of anti-TNF antibodies to these rats coun- A significant contribution of IL-l to the initiation of
teracted protein degradation in gastrocnemius muscle, but failed cachexia at the level of the tumor was elegantly demonstrated
to prevent body weight loss. The total water content, although in a study using mice bearing a murine colon-26 adenocarci-
a source of variation in such a cachexia model, was not analyzed noma. Thus, intratumoral but not systemic injection of an IL-
in these rats. 1 receptor antagonist significantly attenuated the weight loss
The involvement of endogenous TNF in the pathogenesis and depletion of lean and fat tissue, without an apparent effect
of cachexia has also been studied in experimental models of on the tumor burden. In this model, IL-l-induced anorexia is
infectious diseases. Thus, BALBlc mice developed a wasting thought not to play a role since the cachexia progressed without
syndrome after infection with Trypanosoma cruzi .42 Cachexia a detectable loss of appetite. Strikingly, culturing of the C-26
was characterised by weight loss and decreases in lean carcass cells that were derived from the colon adenocarcinoma pro-
content. The weight and fat loss started to appear after the duced large quantaties of IL-6 after stimulation with IL- 1. The
parasitaemic phase of the infection. Early treatment of the mice IL-l-dependent production of IL-6 was also evident in vivo,
with monoclonal antibody against TNF significantly attenuated and a monoclonal antibody to IL-6 was able to suppress the
weight loss but was also associated with a decreased parasitae- development of cachexia associated with the colon tumor
mia. Since the onset and intensity of the body weight loss were growth.r4 Subsequently, the authors found that tumor-infiltrat-
found to be related to the parasitic load, a direct contribution ing macrophages produce IL-l, which in turn potentiate the
of TNF in bringing about cachexia is questionable. The obser- production of IL-6 by the tumor cells. Increased IL-6 release
vation that treatment of the mice at a late time point of the into the circulation then contributes to the wasting process of
disease, i.e., after parasitaemic phase failed to alter cachexia, the host.54
is in line with this conclusion. There is other evidence for a role of IL-6 in the pathogenesis
of cachexia. As with TNF and IL-l, IL-6 is a trigger of the
INTERLEUKIN 1 AND 6 acute phase response and is implicated in many diseases (for
Two forms of interleukin (IL-1 ) have been isolated, IL-lo review, see reference 55). As to metabolic activities, IL-6 is
and IL-lfi. A third member of the IL-l family is the IL-l able to inhibit the activity of LPL in adipose tissue in mice in
receptor antagonist (IL-lra). IL-la, IL-lo, and IL-lra can bind vivo.56 Likewise, muscle atrophy has been described to occur
to either of the two IL-l receptors, but IL-lra fails to transmit in IL-6 transgenic mice,57 and inoculation of mice with CHO
a signal, thus accounting for its antagonistic properties. tumor cells expressing IL-6 was found to be associated with
IL-l is primarily an inflammatory cytokine that has biologi- body weight loss and hypophagia.58 More evidence for a caus-
cal activities, which in many respects resemble those of TNF. ative role of endogenous IL-6 in the pathogenesis of cachexia
Being a product of mainly macrophages and endothelial cells, comes from the experiments performed by Strassmann and col-
it has been implicated in shock and autoimmune disorders. IL- leagues who reported that the IL-l-dependent IL-6 production
1 is known for its pyrogenic property and is a potent trigger was responsible for the adenocarcinoma-associated cachexia
of the acute-phase response (for review, see reference 43). Its (see above). These results were recently confirmed by several
ability to induce an autodestructive process in cartilaginous other laboratories,59-61 who used two subclones of the murine
tissue gave rise to the term “catabolin” at a time when the colon 26 adenocarcinoma cell line, one cachexigenic (clone
term interleukin-1 was not yet established. IL-l is also capable 20) and the other noncachexigenic (clone 5). In general, they
of inhibiting LPL activity and stimulating lipolysis in cultured found increased serum levels of biologically active IL-6 in mice
CYTOKINES AND CACHEXIA 767

carrying clone 20 but not in clone 5-bearing mice, suggesting cated in immunoregulatory and inflammatory activities (for a
that IL-6 might be involved in the syndrome of cachexia. Inter- recent update see reference 68). IFICy is produced by activated
estingly, in vitro both clones failed to express spontaneous T cells and natural killer cells, and is probably the most potent
mRNA of IL-6 whereas in vivo, such expression was selectively monocyte-macrophage activating factor. There is ample evi-
detected at the tumor site of clone 20.61 These results indicate dence for an either beneficial or detrimental role of endogenous
that clone 20 cells need to be triggered in vivo to produce IFN-7 in various diseases like infections, cancer, auto-immu-
substantial levels of IL-6. Furthermore, the injection of anti- nity, and shock reactions. With respect to the metabolic de-
IL-6 antibody partially inhibited weight loss caused by inocula- rangements associated with cachexia, IFN-y has been shown
tion of clone 20, indicating both that IL-6 is necessary for to inhibit the activity of LPL in an adipocyte cell line69 and in
the development of cachexia but also that other factors are viva.” In primary cultures of rat adipocytes, IFN-7 not only
involved.59,61 inhibits production of lipoprotein lipase, but also that of glyc-
erol phosphate dehydrogenase, another enzyme involved in li-
LEUKEMIA INHIBITORY FACTOR AND CILIARY
pogenesis.7’ Furthermore, IFN-y stimulates lipolysis in vitro
NEUROTROPHIC FACTOR and in vivo.72
Leukemia inhibitory factor (LIF), also designated as Several groups have found that IFN-y occupies a central role
HILDA or D-factor, is a cytokine with properties resembling in the pathogenesis of cachexia. We showed that a potentially
those of IL-6. For example, IL-6 and LIF induce macrophage lethal cachexia developed in nude mice inoculated with CHO cells
differentiation in a murine myeloid leukemic cell line and acute overexpressing the mouse IFN-y gene.73 In this model cachexia
phase protein synthesis in hepatocytes (reviewed in reference was characterized by loss of body weight, atrophy of fat stores and
55). Both cytokines share a subunit, glycoprotein 130 (gp130) reduced food intake. Experiments in which groups of nontumor-
in their receptors, which might explain their functional redun- bearing mice were pair-fed, indicated that body weight loss and
dancy. Mice inoculated with tumor cells of a hemopoietic cell fat loss were not due to the reduced food intake. The extent
line that over-expresses LIF were shown to develop a fatal of cachexia increased proportionally with the number of IFN=y-
syndrome characterized by weight loss, pancreatitis, thymus producing cells, and cachexia was blocked by pretreatment of
athrophy, and abnormalities in the adrenal cortex and calcium the mice with anti-IFN-y antibodies. In these studies it was also
metabolism.62 Mori et al.” isolated a factor from a human mela- demonstrated that cachexia developed when the non-lFN-y-pro-
noma cell line, which induced severe cachexia in nude mice, ducing cells were given together with exogenous II%-?, but not
and the factor appeared to be identical to LIF. Iseki et al.63 when either of these were given separately, suggesting that the
found expression of LIF mRNA in two out of five human cell tumor cells sensitize the host to the cachectic effects of IFN-y. A
lines that all had the ability to develop cancer cachexia when similar situation has been described to occur in mice bearing
inoculated in nude mice. The remaining three cell lines and Lewis lung tumors.‘4 Tumor development was associated with
their xenograft in nude mice failed to express LIF and IL-6, weight loss, anorexia and catabolism of fat and lean tissue. Weight
and the mechanism behind their potential to produce cachexia loss was significantly reduced by treatment with anti-IFN-y anti-
remains to be determined. body. The growth of the Lewis lung tumor was also antagonized
Ciliary neurotrophic factor (CNTF) is another cytokine that by the treatment. Since in untreated animals weight loss was found
shares structural and functional properties with IL-6 and LIF. to be increased with increasing tumor volume, the anti-cachectic
CNTF is implicated in the differentiation and survival of glial effect of anti-lFN-y antibody might have to be mediated by the
and neural cells, a property that gave rise to the term “neuro- effect on tumor growth. However, comparison of body weight
kine.” 55 Recent data have indicated that CNTF can induce changes in pairs of mice taken from control and anti-IFN-y anti-
severe damage in extra-neural tissues. Implantation of CNTF- body-treated groups and matched for tumor volume still revealed
secreting cells either subcutaneously or intraperitoneally, was significant reduction of body weight loss in the anti-IFN-y treated
shown to result in a rapid wasting syndrome characterized by group. Moreover, in mice treated with anti-IFN-y antibodies at a
weight loss, breakdown of fat tissue and skeletal muscle protein later time point, when the antibody failed to affect tumor size, the
and reduction of food and fluid intake.64.65Although pair-fed anti-cachectic effect of the antibody still prevailed. Treatment of
control experiments were not performed, the authors claimed tumor-bearing mice with IFN-y also resulted in a more pro-
that the observed anorexia was insufficient to explain the pro- nounced cachexia, without affecting the tumor volume. Body com-
found degree of body weight loss. In that study, cachexia was position analysis revealed that the anti-IFN-y antibody treatment
also induced by injections of recombinant rat or human CNTF significantly inhibited the loss of fat tissue and to a minor extent
three times per day. TNF, IL-6, and LIF serum levels in CNTF- the loss of lean tissue. Here again, experiments in which control
treated mice were below the detection limits suggesting that groups of mice were pair-fed indicated that the anorectic effect
CNTF-induced cachexia emerges independently of these cyto- of endogenous IFN-y was not responsible for cachexia.
kines. Although the gp130 receptor subunit of CNTF is shared Evidence to support a role for IFN-y in the pathogenesis of
with that of IL-6, recent data suggest that the cachectic proper- cancer cachexia has also been found by Langstein et a1.39These
ties of CNTF are distinct from those of IL-6.@,‘j7In these stud- investigators compared the effects of anti-IFN-y and anti-TNF
ies, it was found that administration of CNTF, but not of IL- antibodies on cachexia in rats bearing methylcholantrene-in-
6, was associated with profound anorexia and body weight duced tumors. Anti-IFN-y treatment clearly mitigated cachexia
loss, although both cytokines produced a similar acute phase as evident from decreased loss of body weight, increased food
response. Furthermore, Espat et a1.66demonstrated increased intake, and longer life span. In contrast, the anti-TNF treatment,
protein losses in CNTF-treated mice as compared to pair-fed although proven to be effective in a shock model, was unable
controls indicating that CNTF-induced cachexia was indepen- to alter the wasting syndrome. The authors concluded that in
dent of the associated anorexia. this model, endogenous production of IFN-7 is more important
than that of TNF in the pathogenesis of cachexia.
INTERFERON-y CONCLUSIONS
Interferon-y (IFN=y) was initially characterized by its abil- To date, studies with animal models incriminate several cy-
ity to “interfere’ ’ with virus replication, but has been impli- tokines such as TNF-n, IL-l, IL-6, LIF, CNTF, and IFN-y in
768 CYTOKINES AND CACHEXIA

the pathogenesis of cachexia. Continuous production of cyto- antibody treatment was sometimes associated with inhibition
kines has been shown to mimic at least some of the characteris- of tumor growth37,74or parasite proliferation.42 Since the degree
tics of cachexia, and administration of cytokine antagonists to of cachexia is related to tumor or parasitic load, additional
attenuate cachectic symptoms. However, in some of these stud- studies need to be done to delineate the true mechanism under-
ies, proper control by inclusion of pair-fed control groups and lying the anti-cachectic potential of the antibody. For example,
by analysis of body composition was not performed, and there- matching antibody-treated and control animals for tumor or
fore it is not always clear by which mechanism the cytokines parasitic load is one approach that can clarify whether the anti-
made their contribution to the wasting syndrome. Pair-fed con- cachectic effect of the antibody is due to its effect on tumor or
trols should be included to exclude the possibility that the wast- parasite development. The implication is that in some models
ing syndrome is simply due to reduced food intake. This type endogenous cytokines may cause cachexia in an indirect way
of controlled design has revealed that hypophagia, although by favoring tumor progression or parasite proliferation.
clearly present, was not causally involved in IFN-y-induced Our review emphasizes that cachexia, both in clinical or
cachexia. Thus, pair-feeding experiments indicated that IFN- experimental situations, has many different faces and back-
y-induced body weight loss and fat loss were not due to the grounds. In particular, different cytokines may be involved in
reduced food intake.73.74On the other hand, TNF- and IL-l- cachexia associated with different primary etiologies, Implicit
induced cachexia could be accounted for by reduced food in- assignment of the name “cachectin” to a single molecule iden-
take. 30~31*35*47
IL-6, finally, did not bring about anorexia although tical to TNF-a has indeed led to overemphasis on the role of
it did induce cachexia. These conclusions were drawn from TNF-a! in cachexia. In all probability cachexia develops as a
experiments in which cytokines were injected or produced in result of chronic overstimulation induced by several cytokines
the periphery. It should be noted, however, that various cyto- acting in concert and thereby affecting many different cells and
kines were found to induce anorexia when injected into the organ systems. Even in a simplified system in which cachexia
central nervous system, i.e., into the third cerebral ventricle is triggered by administration of one particular cytokine, a cas-
which is an area closely associated with the hypothalamic feed- cade of other cytokines targeting different organ systems is
ing and drinking region. Extensive work with respect to this probably involved. Thus, cachexia has to be seen as a result of
subject has been performed by Plata-Salarhan and collaborators. various cytokine-driven metabolic (lipolytic and cytolytic)
They demonstrated anorectic properties of interferons, TNF- and/or neuroendocrine (anorexia) interactions. Further clarifi-
(Y,IL-l/?, IL-lo, and chemokines including IL-8, MCP-1, and cation of the interactions can help to design strategies to avoid
RANTES.3’~50-52~75~76 The authors concluded that cytokines cachexia in patients suffering from cancer or chronic infections.
could directly participate in the regulation of feeding, at pico- One strategy may consist of administering antagonists against
gram-low nanogram range, while peripheral administration the cytokines, which are responsible for cachexia. The main
consistently required doses in the microgam range.2gB48~4g Out difficulty with this proposal strategy is that, with a few excep-
of the various categories of cytokines that were able to induce tions, cytokines are necessary for defense against the primary
anorexia, IL-lp seemed to be the most potent one.” cause of cachexia, i.e., infection and cancer. However, extreme
Another hiatus in some studies on experimental cachexia is cachexia by itself reduces natural defense, and, therefore, lesser
the lack of body composition analysis. Data on body composi- cachexia might not only ameliorate their discomfort but also
tion are needed to draw firm conclusions as to whether a given delay the breakdown of natural host defense mechanisms
cytokine affects fat or lean tissue. Studies in which body com- against the tumor or the infection.
position was performed revealed that IFN-7 and IL-6 affect fat A final consideration is that cytokines exert their influence
and muscle tissue, respectively. The ratio of dry to wet body on the metabolism of tissues through other categories of media-
weight should also be determined. Thus, acute body weight tors. One of these is leptin, also known as the obese gene
loss in mice given single injections of Staphylococcus entero- product, recently found to be increased in response to TNF and
toxin B” or anti-CD3 antibody” was found to be almost com- IL_ 1.79.80Another potential factor might be the 24 K proteogly-
pletely due to a loss of water (P. Matthys, unpublished data), can, a newly isolated cancer cachectic factor that has the ability
possibly as a result of severe diarrhea. to mimic cachexia in vivo.13 Elucidation of the mechanisms
A rather unexpected problem of interpretation was found to underlying cachexia is far from complete and remains a fasci-
arise in experiments with anticytokine antibodies in that such nating and promising field of investigation.
REFERENCES

1. Langstein HN, Norton JA. Mechanisms of cachexia. Nutr Cancer of a lipolytic factor (toxohormone-L) from cell-free fluid of ascites
1991;5:103 sarcoma 180. Cancer Res 1981;41:284
2. Norton JA, Peacock JL, Morrison SD. Cancer cachexia. Crit Rev 10. Kondo Y, Sato K, Ueyama Y, et al. Serum sialyltransferase and
Oncol Hematol 1987;7:289 liver catalase activity in cachectic nude mice bearing a human
3. Gold J. Proposed treatment of cancer by inhibition of gluconeogen- malignant melanoma. Cancer Res 1981;41:2912
esis. Oncology 1968;22:185 11. Mori M, Yamaguchi K, Honda S, et al. Cancer cachexia syndrome
4. Inculet RI, Peacock JL, Gorschboth CM, et al. Ciluconeogenesis in developed in nude mice bearing melanoma cells producing leuke-
the tumor-influenced rat hepatocyte: importance of tumor burden, mia-inhibitory factor. Cancer Res 1991;51:6656
lactate, insulin, and glucagon. J Nat1 Cancer Inst 1987;79:1039 12. Beck SA, Tisdale MJ. Production of lipolytic and proteolytic factors
5. Tisdale MJ. Cancer cachexia. Br J Cancer 1991;63:337 by a murine tumor-producing cachexia in the host. Cancer Res
6. Mordes JP, Rossini AA. A tumor-induced anorexia in the Wistar 1987;47:5919
rat. Science 1981;213:565 13. Todorov P, Cariuk P, McDevitt T, et al. Characterization of a
7. Norton JA, Moley JF, Green MW, Carson RE, Morrison SD. Para- cancer cachetic factor. Nature 1996;379:739
biotic transfer of cancer anorexialcachexia in male rats. Cancer 14. Strassmann G, Fong M, Kenney JS, et al. Evidence for the involve-
Res 1985;45:5547 ment of interleukin 6 in experimental cancer cachexia. J Clin Invest
8. Kitada S, Hays EF, Mead JF. A lipid mobilizing factor in serum 1992;89:1681
of tumor-bearing mice. Lipids 1980; 15:168 15. Tisdale MJ. Cancer cachexia. Anti-Cancer Drugs 1993;4:115
9. Masuno H, Yamasaki N, Okuda H. Purification and characterization 16. Shenouda G, Labateya M, Souhami L, et al. Elevated serum cyto-
CYTOKINES AND CACHEXIA 769

kines in patients with malignancy and weight loss. Int J Oncol 41. Costelli P, Carbo N, Tessitore L, et al. Tumor necrosis factor-a
1993;2:115 mediates changes in tissue protein turnover in a rat cancer cachexia
17. Coodley GO, Loveless MO, Merrill TM. The HIV wasting syn- model. J Clin Invest 1993;92:2783
drome: a review. J Acquired Immune Deficiency Synd 1994; 7:68 1 42. Truyens C, Torrico F, Angelo-Barrios A, et al. The cachexia associ-
18. Weinroth SE, Parenti DM, Simon GL. Wasting syndrome in AIDS: ated with Trypanosoma cruzi acute infection in mice is attenuated
pathophysiologic mechanisms and therapeutic approaches. Infect by anti-TNF-a, but not by anti-IL-6 or anti-IFN-y antibodies. Para-
Agents Dis 1995;4:76 site Immunol 1995; 17:561
19. Roubenoff R, Roubenoff RA, Ward LM, et al. Rheumatoid 43. Dinarello CA. Interleukin-1. In: A. Thomson, ed. The Cytokine
cachexia: depletion of lean body mass in rheumatoid arthritis. Possi- Handbook, 2nd ed. Academic Press, 1994:31
ble association with tumor necrosis factor. J Rheumatol 44. Grunfeld C, Feingold KR. The metabolic effects of tumor necrosis
1992; 19:1505 factor and other cytokines. Biotherapy 1991;3:143
20. Roubenoff R, Roubenoff RA, Cannon JG, et al. Rheumatoid 45. Ottemess IG, Golden HW, Brissette WH, et al. Effects of conti-
cachexia. Cytokine-driven hypermetabolism accompanying re- nously administered murine interleukin-la: tolerance development
duced body cell mass in chronic inflammation. J Clin Invest and granuloma formation. Infect Immun 1989;57:2742
1994;93:2379 46. Moldawer LL, Georgieff M, Lundholm K. Interleukin-1, tumor
21. Levine DS. Immune modulating therapies for idiopathic inflamma- necrosis factor-alpha (cachectin) and the pathogenesis of cancer
tory bowel diseases. Adv Pharmacol 1994;25:171 cachexia. Clin Physiol 1987;7:263
22. Tilg H. The role of cytokines in the pathophysiology of chronic 47. Hellerstein MK, Meydani SN, Meydani M, et al. Interleukin-1
liver diseases. Int J Clin Lab Res 1993;23:179 induced anorexia in the rat. Influence of prostaglandins. J Clin
23. Difrancia M, Barbier D, Mege JL, et al. Tumor necrosis factor-a Invest 1989; 84:228
levels and weight loss in chronic obstructive pulmonary disease. 48. Yang Z-J, Koseki M, Meguid MM, et al. Synergistic effect of
Am J Resp Crit Care Med 1993; 150:1453 rhTNF-a and rhIL-la in inducing anorexia in rats. Am J Physiol
24. Rouzer CA, Cerami A. Hypertriglyceridemia associated with Tryp- 1994;267:R1056
anosoma brucei infection in rabbits. Role of defective triglyceride 49. Uehara A, Sekiya C, Takasugi Y, et al. Anorexia induced by in-
removal. Mol Biochem Parasitol 1980;2:31 terleukin 1: involvement of corticotropin-releasing factor. Am J
25. Kawakami M, Cerami A. Studies of endotoxin-induced decrease Physiol 1989;257:R613
in lipoprotein lipase activity. J Exp Med 1981; 154:631 50. Plata-Salafian CR, Borkoksi JP. ChemokinesIintercrines and cen-
26. Beutler B, Mahoney J: Le Trang N, et al. Purification of cachectin, tral regulation of feeding. Am J Physiol 1994;266:R1711
a lipoprotein-lipase-suppressing hormone secreted by endotoxin- 51. Sonti G, Ilyin SE, Plata-Salaman CR. Anorexia induced by cytokine
induced RAW 264.7 cells. J Exp Med 1985; 161:984 interactions at pathophysiological concentrations. Am J Physiol
27. Tracey KJ. Tumor-necrosis factor-alpha. In: A. Thomson, ed. The 1996; 270:R1394
Cytokine Handbook, 2nd. ed. London: Academic Press, 1994:289 52. Plata-Salatian CR. Anorexia during acute and chronic disease. Nu-
28. Tracey KJ, Wei H, Manogue KR, et al. Cachectinltumor necrosis trition 1996; 12:69
factor induces cachexia, anemia, and inflammation. J Exp Med 53. Strassmann G, Masui Y, Chizzonite R, et al. Mechanisms of experi-
1988;167:1211 mental cancer cachexia. Local involvement of IL-l in colon-26
29. Fong Y, Moldawer LL, Marano M, et al. Cachectin/TNF or IL- tumor. J Immunol 1993; 150:2341
ILY induces cachexia with redistribution of body proteins. Am J 54. Strassmann G, Jacob CO, Evans R, et al. Interaction between mono-
Physiol 1989;256:R659 nuclear phagocytes and colon-26 carcinoma and its relevance to
30. Mahony SM, Tisdale MJ. Induction of weight loss and metabolic IL-6-mediated cancer cachexia. J Immunol 1992; 148:3674
alterations by human recombinant tumour necrosis factor. Br J 55. Hirano T. Interleukin-6. In: A. Thomson. ed. The Cytokine Hand-
Cancer 1988;58:345 book, 2nd ed. Academic Press, 1994:31
3 1. Plata-Salarfian CR, Oomura Y, Kai Y. Tumor necrosis factor and 56. Greenberg AS, Nordan RP, McIntosh J, et al. Interleukin 6 reduces
interleukin-lp: suppression of food intake by direct action in the lipoprotein lipase activity in adipose tissue of mice in vivo and in
central nervous system. Brain Res 1988;448: 106 3T3-Ll adipocytes: a possible role for innterleukin 6 in cancer
32. Sochkr SM, Friedman A, Martinez D. Recombinant human tumor cachexia. Cancer Res 1992;52:4113
necrosis factor induces acute reductions in food intake and body 57. Tsujinaka T, Fujita J, Ebisui C, et al. Interleukin 6 receptor antibody
weight in mice. J Exp Med 1988; 167: 1957 inhibits muscle atrophy and modulates proteolytic systems in in-
33. Stovroff MC, Fraker DL, Swedenborg JA, et al. Cachectinltumor terleukin 6 transgenic mice. J Clin Invest 1996;97:244
necrosis factor: a possible mediator of cancer anorexia in the rat. 58. Black K, Garrett R, Mundy GR. Chinese hamster ovary cells trans-
Cancer Res 1988;48:4567 fected with the murine interleukin-6 gene cause hypercalcemia as
34. Oliff A, Defeo-Jones D, Boyer M, et al. Tumors secreting human well as cachexia, leukocytosis and thrombocytosis in tumor-bearing
TNF-cachectin induce cachexia in mice. Cell 1987; 50:555 nude mice. Endocrinology 1991;28:2657
35. Tracey KJ, Morgello S, Koplin B, et al. Metabolic effects of 59. Fujimoto-Ouchi K, Tamura S, Mori K, et al. Establishment and
cachectin/tumor necrosis factor are modified by site of production. characterization of cachexia-inducing and -non-inducing clones of
J Clin Invest 1990;86:2014 murine colon 26 carcinoma. Int J Cancer 1995;61:522
36. Cheng J, Turksen K, Yu Q-C, et al. Cachexia and graft-vs-host- 60. Soda K, Kawakami M, Kashii A, et al. Manifestations of cancer
disease-type skin-changes in keratin promoter-driven TNFol cachexia induced by colon 26 adenocarcinoma are not fully ascrib-
transgenic mice. Genes Dev 1992;6: 144 able to interleukin-6. Int J Cancer 1995;62:332
37. Sherry BA, Gelin J, Fong Y, et al. Anti-cachectinltumor necrosis 61. Yasumoto K, Mukaida N, Harada A, et al. Molecular analysis of
factor-a antibodies attenuate development of cachexia in tumor the cytokine network involved in cachexia in colon 26 adenocarci-
models. FASEB J 1989;3:1956 noma-bearing mice. Cancer Res 1995;55:921
38. Gelin J, Moldawer LL, Lonnroth C, et al. Role of endogenous 62. Metcalf D, Gearing DP. Fatal syndrome in mice engrafted with
tumor necrosis factor (Y and interleukin 1 for experimental tumor cells producing high levels of the leukemia inhibitory factor. Proc
growth and the development of cancer cachexia. Cancer Res Nat1 Acad Sci USA 1989; 86:5948
1991;51:415 63. Iseki H, Kajimura N, Ohue C, et al. Cytokine production in five
39. Langstein HN, Doherty GM, Fraker DL, et al. The roles of y- tumor cell lines with activity to induce cancer cachexia syndrome
interferon and tumor necrosis factor (Yin an experimental rat model in nude mice. Jpn J Cancer Res 1995;86:562
of cancer cachexia. Cancer Res 1991;51:2302 64. Henderson JT, Seniuk NA, Richardson PM, et al. Systemic admin-
40. Yoneda T, Alsina MA, Chavez JB, et al. Evidence that tumor istration of ciliary neurotrophic factor induces cachexia in rodents.
necrosis factor plays a pathogenetic role in the paraneoplastic syn- J Clin Invest 1994;93:2632
dromes of cachexia, hypercalcemia, and leukocytes in a human 65. Henderson JT, Mullen BJM, Roder JC. Physiological effects of
tumor in nude mice. J Clin Invest 1991; 87:977 CNTF-induced wasting. Cytokine 1996; 8:784
770 CYTOKINES AND CACHEXIA

66. Espat NJ, Auffenberg T, Rosenberg JJ, et al. Ciliary neurotrophic inoculated with interferon-y-producing tumor cells. Int J Cancer
factor is catabolic and shares with IL-6 the capacity to induce an 1991;49:77
acute phase response. Am J Physiol 1996;271:R185 74. Matthys P, Heremans H, Opdenakker G, et al. Anti-interferon-y
67. Fantuzzi G, Benigni F, Sironi M, et al. Ciliary neurotrophic factor antibody treatment, growth of Lewis lung tumours in mice and
(CNTF) induces serum amyloid A, hypoglycaemia and anorexia, tumour-associated cachexia. Eur J Cancer 1991; 27: 182
and potentiates IL-l induced corticosterone and IL-6 production in 75. Plata-Salaman CR. Interferons and central regulation of feeding.
mice. Cytokine 1995;7: 150 Am J Physiol 1992;263:R1222
68. Billiau A. Interferon-y: biology and role in pathogenesis. Adv Im- 76. Plata-Saladran CR. Anorexia induced by activators of the signal
munol 1996;62:61 transducer of gp 130. NeuroReport 1996;7:841
69. Patton JS, Shepard HM, Wilking H, et al. Interferons and tumor 77. Matthys P, Mitera T, Heremans H, et al. Anti-interferon--y and anti-
necrosis factors have similar catabolic effects on 3T3 Ll cells. Proc IL-6 antibodies affect Staphylococcal enterotoxin B-induced weight
Nat1 Acad Sci USA 1986;83:8313 loss, hypoglycemia, and cytokine release in D-galactosamine-sensi-
70. Kurzrock R, Rohde MF, Quesada JR, et al. Recombinant y-inter- tized and unsensitized mice. Infect Immun 1994;63:1158
feron induces hypertriglyceridemia and inhibits postheparin lipase 78. Matthys P, Dillen C, Proost P, et al. Modification of the anti-
activity in cancer patients. J Exp Med 1986; 164: 1093 CD3-induced cytokine release syndrome by anti-IFN-y or anti-IL-
7 1. Gfegoire F, De Broux N, Hauser N, et al. Interferon-y and interleu- 6 antibody treatment: protective effects and biphasic changes in
kin-lp inhibit adipoconversion in cultured rodent preadipocytes. J blood cytokine levels. Eur J Immunol 1993;23:2209
Cell Physiol 1992; 151:300 79. Nabel GJ, Grunfeld C. Calories lost-another mediator of cancer
72. Memon RA, Feingold KR, Moser AH, et al. In vivo effects of cachexia? Nat Med 1996;2:397
interferon-a! and interferon-y on lipolysis and ketogenesis. Endocri- 80. Grunfeld C, Zhao C, Fuller J, et al. Endotoxin and cytokines induce
nology 1992; 131: 1695 expression of leptin, the ob gene product, in hamsters. A role for
73. Matthys P, Dijkmans R, Proost P, et al. Severe cachexia in mice leptin in the anorexia of infection. J Clin Invest 1996;97:2152

Part I appeared in Nutn’tion 1995; 11:39.

Part II appeared in Nutrition 1996; 12:69.
Part III appeared in Nutrition 1996; 12:303.
Part IV appeared in Nutrition 1997; 13: 1.
Part V appeared in Nutrition 1997; 13:403.
Part VI appeared in Nutrition 1997; 13:503.
Part VIII will appear in an upcoming issue of Nutrition.