| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
1
Laboratory of Integrative Biology (E.J.W., B.N.F., R.W.J.) and Laboratory of Immunophysiology (K.W.K.), Department of Animal Sciences, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801; Department of Biological Sciences (S.A.), Illinois State University, Normal, Illinois 61790; and Department of Veterinary PathoBiology (R.W.S., M.P.M.), University of Minnesota,,St. Paul, Minnesota 55108
Address all correspondence and requests for reprints to: Dr. Rodney W. Johnson, Laboratory of Integrative Biology, Department of Animal Sciences, University of Illinois, 1207 West Gregory Drive, Urbana, Illinois 61801. E-mail:rwjohn{at}ux6.uiuc.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(TNF-) concentrations. Sickness behaviors
were measured during 10-min tests at 0, 2, 4, 8, 12, and 24 h post
injection. The 5 and 50 µg/kg doses of LPS increased plasma
concentrations of cortisol and TNF-, while inducing anorexia,
hypersomnia, and fever. In contrast, although 0.5 µg/kg LPS induced
acute anorexia, hypersomnia, and fever, it did not increase plasma
TNF-; and the cortisol response was small and transient, suggesting
the behavioral system in pigs is more responsive to LPS than the
hypothalamic-pituitary-adrenal (HPA) axis. Because LPS-induced behavior
and activation of the HPA axis involve proinflammatory cytokines in the
brain, in a second study, unrestrained pigs with jugular catheters were
injected intracerebroventricularly (ICV) with recombinant porcine
TNF-. Vehicle or TNF- (0, 5, or 50 ng/kg) was injected ICV, and
plasma cortisol and behavior were determined as before. Pigs injected
ICV with 50 ng/kg TNF- showed anorexia, hypersomnia, and an abrupt
increase in plasma cortisol concentration. Whereas 5 ng/kg TNF- ICV
also induced marked sickness behavior, it failed to stimulate the HPA
axis, as indicated by plasma cortisol levels. That there was a distinct
difference in the magnitude of behavioral and endocrine responses to
LPS and TNF- suggests that different systems that are responsive to
inflammatory stimuli exhibit different sensitivities. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(TNF-). Synthesized by mononuclear myeloid
cells as a 26-kDa protein, membrane-bound TNF- undergoes processing
by a metalloproteinase to produce the secreted, biologically active
17-kDa form (1). The first suggestion that TNF- and other products
from activated leukocytes affect the central nervous system (CNS) and
alter behavior came from clinical trials, where recombinant cytokines
were injected peripherally to potentiate immune responses or treat
neoplastic disease (2, 3). These efforts were terminated when patients
receiving TNF- and interferon- developed severe flu-like
symptoms, including anorexia, fever, malaise, and lethargy. The
involvement of an immunogenic component, rather than a pathogenic
component, in the central effects of acute infection is now confirmed
by numerous independent studies in animals involving peripheral
administration of recombinant cytokines (4, 5).
Whether TNF- and other cytokines in peripheral blood directly access
the CNS has been the subject of considerable debate (6). Regardless, it
is clear that TNF- and its receptors are present in the CNS. For
instance, increased TNF- synthesis in brain has been observed in the
animal model for multiple sclerosis (i.e. experimental
allergic encephalomyelitis) (7), head injury or trauma (8), and after
peripheral injection of lipopolysaccharide (LPS) (9, 10). The cytokine
is likely produced by activated microglial cells that are present in
the CNS and, like macrophages, are derived from mononuclear myeloid
progenitors (11). Within the CNS, TNF- induces fever through a
direct action on hypothalamic neurons and by triggering the release of
IL-1 (12). It also induces anorexia (13) and is involved in the release
of ACTH by the pituitary (14, 15). Binding sites for TNF- have been
identified in the brainstem, cortex, cerebellum, thalamus, and basal
ganglia of the rat brain (16). Thus, there is considerable evidence
indicating that TNF- contributes to the behavioral and
neuroendocrine changes that characterize sickness.
Most of the studies concerning the behavioral and neuroendocrine
effects of proinflammatory cytokines have been conducted in rodents.
However, experimental limitations of rodent models and the growing
interest in a clinically relevant outbred animal model have prompted
the use of the pig as an alternative for many immunophysiological
investigations. Swine, which are similar to humans in size, physiology,
and dietary habits, are now used for studying, among other things: 1)
the suppressive effects of opiate drugs on humoral and cell-mediated
immune responses (17); 2) adult respiratory distress syndrome in sepsis
(18); and 3) organ transplantation between discordant species (19). The
pig also affords opportunities for studying the role of cytokines in
the pathogenesis of CNS infections and autoimmune diseases. For
example, Nam et al. (20) used cultured porcine astroglia to
demonstrate the rapid induction of PG F2 production in
response to recombinant human IL-1, whereas Megyeri et
al. (21) used a neonatal piglet model to show that intracisternal
injection of recombinant human TNF- increased the permeability of
the blood-brain barrier to sodium-fluorescein.
Although porcine macrophages produce the characteristic array of
proinflammatory cytokines in response to antigen in vitro
(22, 23), neither the secretion of TNF- in response to in
vivo challenge nor the behavioral and physiological effects of
TNF- in the pig brain have been adequately explored. In the present
study, we sought to address these issues and, in doing so, take
advantage of the pig model to simultaneously and chronologically
evaluate behavioral and physiological responses to inflammatory stimuli
administered either peripherally or centrally. We recently demonstrated
in the rat that the high sensitivity of the
hypothalamic-pituitary-adrenal (HPA) axis to LPS, and the resultant
increase in plasma corticosterone, are important mechanisms for
preventing profound behavioral disturbances in response to low-grade
immune stimulation (24, 25). In contrast, in the present study, by
monitoring the coincidental changes in behavior and plasma cortisol
after peripheral injection of LPS or central injection of TNF-, we
found evidence that suggests that the HPA axis in pigs is not as
responsive to inflammatory stimuli. Because glucocorticoids inhibit
LPS-induced sickness behavior (24, 25, 26), the present data suggest that
the relative insensitivity of the pig’s HPA axis to LPS and TNF-
contributed to the rather profound anorexia and hypersomnia that were
observed.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Surgical procedures
Anesthesia was induced with sodium thiamylal (20 mg/kg BW iv)
and maintained with halothane (2–5%) and oxygen (300 cc/min). Pigs
were administered penicillin (600,000 units im) postsurgically and were
allowed at least 7 days of recovery before beginning an experiment. All
procedures were approved by the University of Illinois Laboratory
Animal Care Advisory Committee.
Jugular catheterization.
The right jugular vein was
surgically exposed and a catheter (Tygon microbore tubing; id, 1.02 mm;
od, 1.78 mm; Fisher Scientific, Pittsburgh, PA) was advanced
posteriorly 15–20 cm to the vena cava and sutured in place. The distal
end of the catheter was directed sc to the dorsal neck region and
exteriorized between the scapulae for repeat blood sample collection.
Catheters were flushed daily with 5 ml sterile heparinized saline (4
U/ml) to prevent occlusion.
Intracerebroventricular (ICV) cannulation.
Pigs requiring a
central injection had a cannula stereotaxically placed into the left
lateral cerebral ventricle (ICV), as previously described for swine
(28). Briefly, the head of the pig was oriented in a large-animal
stereotaxic instrument so that the plane formed by the frontal and
parietal bones was parallel to the instrument table top. An 18-gauge
stainless-steel cannula was placed ICV according to predetermined
coordinates (anterior-posterior, +5.0 mm to the bregma; lateral, 5 mm;
horizontal, -14.0 mm to the dura mater). Two stainless-steel screws
and cranioplastic cement (Plastics One, Roanoke, VA) secured the
cannula. A back-flow of cerebrospinal fluid indicated the cannula was
ICV. To facilitate ICV injections, a catheter (Tygon microbore tubing;
id, 0.51 mm; od, 1.53 mm; Fisher Scientific) was extended sc from the
ICV cannula to the dorsal neck region and exteriorized.
Recombinant porcine TNF-
Porcine TNF- was cloned into the bacterial expression vector,
pGEX-2T (Pharmacia Biotech, Piscataway, NJ). Fusion protein, consisting
of cytokine and glutathione-S-transferase, was expressed in
Escherichia coli DH5. A 500-ml culture of recombinant
bacteria, amplified from a single colony, was grown in Superbroth with
100 µg/ml ampicillin to an OD600 of approximately 0.7 at
37 C in a 2-liter Fernbach flask. Fusion protein expression was induced
with 0.2 mM IPTG. After 4 h at 30 C, cells were
harvested by centrifugation at 400 x g for 15 min at 4
C and resuspended in 25 mM Tris HCl, pH 8.0, 5
mM EDTA, 50 µg/ml phenylmethylsulfonyl fluoride, and 25
µg/ml aprotinin, and frozen at -80 C. The insoluble fusion protein
aggregates were isolated by repeated rounds of sonic disruption and
centrifugation in the presence of 0.25% Triton X-100 and 0.25% Na
deoxycholate in 25 mM Tris HCl, pH 8.0, 5 mM
EDTA. Aggregates were solubilized and refolded in 8 M urea,
2 mM arginine, 5 mM EDTA, 0.005% Tween-20, 2
mM reduced glutathione, and 0.02 mM oxidized
glutathione. Refolded protein was dialyzed and concentrated on a 30-kDa
mw-cutoff membrane (S11Y30, Amicon, Beverly, MA) and passed over a
glutathione-sepharose 4B column (Pharmacia Biotech.) in 50
mM Tris HCl, 150 mM NaCl, pH 8.0 (Tris-saline).
TNF- was cleaved on the column with 0.5 µl/ml thrombin (Boehringer
Mannheim, Indianapolis, IN) in Tris-saline with 2.5 mM
CaCl2 and eluted with Tris-saline. Purity was assessed by
Coomassie blue stained SDS-polyacrylamide gels and activity by L929
bioassay (29). Porcine TNF- was stored frozen (-80 C) and later
diluted in sterile PBS for ICV injection.
Systemic responses
Body temperature (TB).
In some pigs after
jugular catheterization, temperature-sensitive radio transmitters
(Model VHF-T-1, Mini-mitter Co., Inc., Sunriver, OR) were implanted
adjacent to a jugular vein. TB was recorded at 20-min
intervals using an automated temperature-sensitive radio telemetry
system (DATACOL Data Acquisition System Version 5.0, Mini-Mitter Co.,
Inc.), as previously described (30).
Behavioral paradigm.
A food-motivated test was used to
assess the behavioral effects of LPS ip or TNF- ICV. A similar test
was used previously and was sensitive to LPS in both a time- and
dose-dependent fashion (30). In brief, motivation for food was
established by removing the food 12 h before administering
treatments. Food was returned for a 10-min test period, and behavior
was directly monitored by scan sampling at 30-sec intervals for the
entire period. Food intake and time spent standing, eating, and
somnolent (i.e. time spent in the drowsy state, as
determined by recumbency time with both eyes closed) was recorded.
Plasma TNF- and cortisol.
Serial blood samples were
collected into heparinized syringes (Sarstedt, Inc., Newton, NC) via an
indwelling jugular catheter and placed immediately on ice. Whole blood
samples were centrifuged (3500 x g for 15 min at 4 C),
and resultant plasma was stored at -80 C until assayed.
Total plasma TNF- was measured using a commercially available
enzyme-linked immunosorbent assay specific for porcine TNF-
(Endogen, Inc., Cambridge, MA). Plasma samples were assayed in
triplicate at either a 1:1 or 1:10 dilution. The assay has a lower
level of sensitivity of 10 pg/ml TNF-, and the intra- and interassay
coefficients of variation were less than 10% and less than 15%,
respectively.
Total plasma cortisol was measured using a commercially available 125I RIA kit (ICN Biomedicals, Inc., Costa Mesa, CA). Pooled plasma samples from pigs with high (>100 ng/ml) and low (<25 ng/ml) cortisol were used for validation. Recovery was validated using 20 µl of porcine plasma spiked with 0, 10, 100, or 1000 ng/ml of cortisol standards in a 25-µl reaction vol. Recovery levels ranged from 82–100%. Plasma samples also were serially diluted 1:2, 1:4, and 1:8 with the provided diluent to demonstrate parallelism with mean cortisol levels of 96.9, 81.6, and 71.4 ng/ml, respectively, after adjusting for dilution. Intra- and interassay variations were 7.2% and 9.3%, respectively. Sensitivity of the assay was 1.5 ng/ml.
Experimental protocols
Acute responses to peripheral LPS.
Twenty pigs, surgically
prepared with jugular catheters and temperature-sensitive radio
telemeters, were used in a completely randomized design to determine
the dose-effect relationship for LPS ip on food intake, behavior,
TB, plasma cortisol, and plasma TNF-. Pigs were fasted
12 h before treatments to enhance motivation for food. LPS from
Escherichia coli serotype K-235 (phenol extracted), which we
have used in pigs to induce sickness (30, 31), was purchased from Sigma
Chemical Co. (St. Louis, MO) and dissolved in sterile pyrogen-free
0.9% NaCl for ip injections. Saline or 0.5, 5.0, and 50 µg/kg LPS
was injected ip in a total vol of 2 ml at 0900 h (n = 5).
Serial blood samples were collected via indwelling jugular catheters
for plasma cortisol and TNF- determination. Pigs were subjected to
the behavioral test immediately before treatments were administered and
again at 2, 4, 8, 12, and 24 h post injection, and TB
was monitored at 20-min intervals throughout the experiment.
Acute responses to central TNF-.
Six pigs were surgically
prepared with iv and ICV cannulas to determine the central effects of
TNF- on food intake, behavior, and plasma cortisol. Pigs were fasted
12 h before treatments to enhance motivation for food. At
0900 h, they were subjected to the food-motivated test to
determine the baseline level of food intake and behavioral activity.
Immediately after the test, pigs were injected ICV with 100 µl PBS or
the same vol PBS containing 5 or 50 ng/kg recombinant porcine TNF-.
Pigs were subjected to the behavioral test again at 2, 4, 8, 12, and
24 h post injection. Serial blood samples were collected via
indwelling jugular catheters for determination of plasma cortisol
concentration. Each pig was assigned to each of the three treatments
(0, 5, or 50 ng/kg TNF-) after a Latin-square design, with pig and
day serving as blocking factors.
Statistical analysis
All data analyses were conducted using General Linear Model
procedures (32). Data were subjected to ANOVA to determine the
significance of main factors and main factor interactions. Cortisol and
temperature data were subjected to a two-factor repeated-measures
ANOVA. When ANOVA revealed a significant effect of dose or a dose
x time interaction, differences between treatment means were tested
using paired t tests. All data are presented as means
± SE.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Whereas saline-injected pigs
engaged in feeding behavior and consumed 175 g of food during the
10-min food-motivated test at 2 h, pigs receiving LPS (0.5, 5, or
50 µg/kg) did not eat. This same feeding pattern was evident at
4 h; but at 8 h, pigs that received 0.5 and 5 µg/kg LPS
consumed an amount of food similar to the saline control. However, food
intake for pigs injected ip with 50 µg/kg did not return to control
levels until 24 h (data not shown). In addition to anorexia,
hypersomnia was induced by LPS (Fig. 1). During the 10-min test
conducted 2 h post injection, pigs receiving saline, 0.5, 5, or 50
µg/kg LPS assumed a somnolent state for 0, 7, 8.5, and 8.75 min,
respectively. Somnolent behavior still was evident in LPS-injected pigs
at 4 h; but by 8 h, only those receiving the highest dose (50
µg/kg) showed hypersomnia.
|
). Two-way ANOVA of plasma cortisol concentration
revealed a significant effect of time (P < 0.001),
dose (P < 0.01), and the time x dose interaction
(P < 0.01). Before injection, plasma cortisol levels
were 25–35 ng/ml, which is consistent with nonstress concentrations
for pigs (28). The cortisol response to 0.5 µg/kg LPS, though
significantly different from saline control, was small and transient
compared with the response to the 5 and 50 µg/kg doses (Fig. 2). In
pigs receiving 0.5 µg/kg LPS, cortisol levels peaked at 1.5 h
post injection, remained elevated at 2 and 3 h, but returned to
baseline by 4 h. In contrast, the cortisol response to 5 and 50
µg/kg LPS was more abrupt and sustained. For example, pigs receiving
50 µg/kg LPS had increased cortisol within 1 h
(P < 0.05). Cortisol peaked at 3 h (199 ng/ml)
and remained high throughout the sampling period. As anticipated, 5
µg/kg LPS induced an intermediate response.
|
). However, the elevation in
TB was not dose-dependent. For example, at 4 h, the
change in TB for pigs injected with saline, 0.5, 5, and 50
µg/kg LPS was 0.229, 0.746, 1.283, and 0.836 C, respectively.
However, TB for pigs receiving 50 µg/kg LPS was elevated
longer. The presentation of food and heat increments of feeding had
obvious effects on TB. For instance, pigs that received
saline, consumed food when it was offered (Fig. 1
), and their
TB was increased above baseline after about 2.5 h;
pigs that received 5 µg/kg LPS did not eat (Fig. 1) and their
TB peaked at 4 h and returned to baseline by 8
h.
|
(Fig. 4). Two-way ANOVA of plasma TNF- concentration
revealed a significant effect of time (P < 0.001),
dose (P < 0.01), and the time x dose interaction
(P < 0.01). Consistent with a previous report (31),
before injection and after injection of saline or 0.5 µg/kg LPS (data
not shown), TNF- was undetectable by this assay, which has a lower
level of detection limit of 10 pg/ml. The TNF- response to 5 µg/kg
LPS was small and transient but was increased at 1.5 h compared
with saline (Fig. 4). The TNF- response to 50 µg/kg LPS was more
abrupt and sustained. Pigs receiving 50 µg/kg LPS had increased
TNF- at 1 h (P < 0.01); TNF- peaked at
1.5 h and was still elevated at 2 h.
|
in swine (5 and 50 ng/kg) markedly reduced food intake and
increased somnolence at 2 and 4 h post injection (Fig. 5). Time spent eating was proportional to food intake,
and therefore, data are not shown. To ensure that these effects were
caused by TNF- and not by contaminating endotoxin, the cytokine
preparation was heated to 90 C for 15 min. This treatment denatures
protein but has no effect on the biological activity of endotoxin.
Three separate pigs were injected ICV with 100 ng/kg heat-inactivated
TNF-, an amount 2-fold greater than the highest dose of biologically
active cytokine employed in the present study. The injectate containing
heat-denatured TNF- had no effect on either food intake or
somnolence (data no shown), indicating that endotoxin contamination was
minimal and not responsible for the change in behavior observed after
central administration of intact protein.
|
was accompanied by
increased activity of the HPA axis, as indicated by increased plasma
concentration of cortisol (Fig. 6). Two-way ANOVA of
plasma cortisol concentration revealed a main effect of dose
(P < 0.001), time (P < 0.01), and a
time x dose interaction (P < 0.01). Cortisol
levels for pigs receiving PBS ICV did not rise above baseline,
indicating that the injection procedures did not induce stress. Whereas
5 ng/kg TNF- induced anorexia and somnolence (Fig. 5), this dose did
not increase cortisol secretion (Fig. 6), suggesting that central
TNF- is more effective in inducing sickness than stimulating the
neuroendocrine system. The 50 ng/kg dose of TNF-, however, induced
an abrupt increase in plasma cortisol. In pigs receiving 50 ng/kg
TNF-, plasma cortisol peaked at 1.5 h and remained elevated for
4 h.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
in swine and confirms that this cytokine, when injected
directly into the CNS, is sufficient to induce anorexia, somnolence,
and secretion of cortisol. The important findings were that: 1) a low
dose of LPS administered ip, which induced marked sickness behavior and
fever, failed to increase plasma TNF- and induced only a small
transient increase in plasma cortisol; and 2) a low dose of TNF-
given ICV, which induced marked sickness behavior, did not increase
plasma cortisol. These results therefore suggest not only a
dissociation between the secretion of TNF- and the induction of
sickness but also between the behavioral and neuroendocrine effects of
TNF- in the brain. The change in behavior and neuroendocrine secretions associated with sickness are well-recognized components of the acute-phase response (5, 33). They are nonspecific to the invading pathogen and are attributed to cytokines released by activated macrophages. The HPA axis in the rat is one of the first systems to respond to LPS. The increase in plasma glucocorticoids is part of an important inhibitory mechanism that modulates immunological and inflammatory responses (34, 35). Indeed, it was recently reported that a low dose of LPS administered ip in rats maximally stimulated the HPA axis but failed to depress social behavior (24), a form of motivation used to quantify sickness (36). It was postulated, and later demonstrated, that the marked increase in plasma corticosterone in response to the low dose of LPS was responsible for precluding sickness behavior. Therefore, at least in the rat, the sensitivity of the HPA axis to inflammatory stimuli is important to prevent behavioral disturbances in response to low-grade immune stimulation.
An important attribute of the pig is that behavior, immune, and
neuroendocrine systems can be evaluated simultaneously and
chronologically in an integrated whole-animal model (28). In the
present study, the model afforded us the opportunity to characterize
the coincidental changes in behavior, plasma cortisol, and plasma
TNF- after peripheral injection of LPS. Pigs were injected ip with
four doses of LPS (i.e. 0, 0.5, 5, and 50 µg/kg) to
develop complete dose-response curves for several behavioral and
physiological effects (i.e. food intake, somnolence, fever,
and cortisol secretion). These doses are substantially less than those
typically used in rats and mice because, like human and nonhuman
primates, pigs are extremely sensitive to LPS (37). The same doses and
serotype of LPS previously were used in pigs to show that the induction
of sickness behavior by LPS involved a PG-dependent mechanism (30).
Consistent with that report, in the present study, pigs responded to ip
LPS with behavioral and physiological responses indicative of an acute
gram-negative bacterial infection. After ip administration of even the
smallest dose of LPS (0.5 µg/kg), pigs were anorectic, somnolent, and
febrile (Figs. 1 & 3). These responses were evident even after 4
h. The increase in plasma cortisol after the same low dose of LPS was
less dramatic (Fig. 2). Therefore, contrary to our previous study in
rats (which showed a marked increase in plasma corticosteroids but no
overt symptoms of sickness), after an ip injection of LPS, the pig
showed marked sickness behavior but only a small transient increase in
cortisol. The small but significant increase in cortisol may have had a
permissive effect on behavioral targets (i.e. the CNS),
therefore contributing to the marked behavioral response to LPS.
However, several studies indicate that adrenalectomized animals are
more sensitive to the behavioral effects of LPS or IL-1ß than intact
controls (24, 25, 26, 38). Alternatively, because corticosteroids have been
found to inhibit several effects of LPS, including fever (39, 40) and
sickness behavior (24, 25, 26, 38), the relative insensitivity of the
pig’s HPA axis to LPS may partially explain its noted sensitivity to
the behavioral effects of LPS.
In the present study, we also sought to measure TNF- in plasma of
pigs after ip injection of LPS. Although the increase in TNF after
injection of LPS is well described for mice and rats, to our best
knowledge, only two studies have reported plasma concentrations of TNF
in swine. Recently, we reported that pigs injected ip with LPS (5
µg/kg) had increased plasma concentrations of TNF-, IL-6, and
cortisol (31). Because the objective of that study was to determine the
relationship between plasma cytokines and alterations in macronutrient
metabolism, pigs were fasted, and behavior was not monitored. In
another study (41), pigs were anesthetized, and 5 µg/kg BW LPS was
continuously infused into the superior mesenteric artery over a 60-min
period. Again, TNF increased after intraarterial infusion of LPS, and
high plasma TNF was positively correlated with lethality caused by
endotoxic shock (41). In the present study, pigs were prepared with
indwelling jugular catheters, so that serial blood samples could be
obtained while simultaneously monitoring behavior. In addition, pigs
were implanted with radio telemeters for periodic undisturbed
monitoring of TB. The 50-µg/kg dose of LPS induced a
rapid increase in plasma TNF-, with levels being significantly
elevated at 1 h, peaking at 1.5 h, and still high (compared
with saline control) at 2 h (Fig. 4). Comparatively, 5 µg/kg
induced a moderate increase in TNF-, which was significant at
1.5 h only. Despite inducing fever and sickness behavior for more
than 4 h, the 0.5 µg/kg dose did not increase plasma TNF-.
This is consistent with an earlier report, where the same dose of the
LPS also failed to increase plasma concentration of TNF- (31). There
is, therefore, a dissociation between the induction of sickness
behavior by LPS and high circulating levels of TNF-. It is of note
that regardless of a purported link between TNF- and cachexia, there
is no clear correlation between weight loss and plasma TNF- in
cachectic patients (42).
Of course, that TNF- was not elevated in response to 0.5 µg/kg LPS
does not preclude the possibility that other cytokines, such as IL-1 or
IL-6, were. However, in a previous study, the same dose of LPS given ip
to pigs also failed to increase plasma IL-6 (31). A more plausible
explanation for the low dose of LPS inducing sickness without
increasing plasma TNF- is that cytokines produced in the periphery
act locally on sensory neurons, which when activated, induce cytokines
to be produced centrally. In support of this hypothesis, ip injection
of LPS increased messenger RNAs (mRNAs) encoding TNF-, IL-1, and
IL-6 in murine brain (10). Furthermore, subdiaphragmatic vagotomy
prevented the increased expression of cytokine mRNAs in mouse brain and
blocked the normal depression in social behavior (43) and anorexia (44)
caused by LPS. Of particular note is the finding that although vagotomy
blocked LPS-induced sickness behavior, it did not prevent the increase
in plasma IL-1 (43). Thus, increased cytokines in plasma do not ensure
sickness, nor does their absence preclude it. The behavioral response
seems dependent upon cytokines in the CNS.
To confirm this in pigs (and in doing so, take advantage of the pig
model to simultaneously and chronologically evaluate behavioral and
physiological responses to a cytokine administered directly into the
CNS), recombinant porcine TNF- was injected ICV into unrestrained
pigs. TNF- has not been measured in CSF of pigs, but the doses
injected ICV (i.e. 5 and 50 ng/kg) were projected to be
equal to or lower than that needed to achieve pathophysiologic
concentrations in the cerebrospinal fluid of animal models of disease
or humans with viral or bacterial meningitis (45). The doses also were
sufficiently low to ensure that centrally injected cytokine would not
diffuse into peripheral blood, although plasma was not assayed for
TNF- after ICV injection. The present results support the idea that
many neural effects of inflammatory challenge are mediated by
cytokines, like TNF-, acting directly in the CNS. When administered
into the CNS, porcine TNF-, at a dose as low as 5 ng/kg, induced a
transient but marked reduction in food intake and an increase in
somnolence. The anorectic properties of this cytokine are apparent,
because a reduction in food intake was evident, even though pigs were
subjected to a 12-h food withdrawal period before ICV injection. The
use of homologous cytokine was important because, e.g. in
mice and rats, recombinant human TNF- binds only the p55 receptor
and not the p75 receptor. This may explain why, in previous studies,
the rat was relatively insensitive to TNF- (i.e. human)
injected ICV (13, 45). Consistent with our finding that the pig is more
sensitive to the behavioral effects of LPS than to the neuroendocrine
effects of LPS, 5 ng/kg TNF- injected ICV reduced food intake and
behavioral activity without increasing plasma cortisol concentration.
Again, if the HPA axis is not as responsive to TNF- as our results
suggest, this may account for the profound behavioral effects of
central TNF-.
In summary, our results confirm that, in the pig, LPS induces the
secretion of TNF- and that this cytokine can act directly in the
brain to elicit behavioral and physiological effects associated with
inflammatory challenge. They also suggest a dissociation between the
behavioral and neuroendocrine effects of inflammatory stimuli in the
periphery and CNS. These data therefore indicate that the different
systems, which are responsive to inflammatory stimuli, exhibit
different sensitivities. It is suggested that the relative
insensitivity of the pig’s HPA axis to LPS and TNF- is, at least
partially, responsible for the marked behavioral effects that were
observed.
|
Footnotes |
|---|
Received November 5, 1996.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
processing by a metalloproteinase inhibitor. Nature 370:558–561[CrossRef][Medline]
expression by resident microglia and infiltrating
leukocytes in the central nervous system of mice with experimental
allergic encephalomyelitis. J Immunol 154:944–953[Abstract]
and IL-1 immunoreactivities in striatal
neurons after surgical injury to the hippocampus. Neuron 10:219–224
expression in the mouse brain after systemic
lipopolysaccharide administration. Proc Natl Acad Sci USA 91:11393–11397 in the brain. Biochem Biophys Res Comm 181:1532–1538[CrossRef][Medline]
.
Brain Res Bull 36:215–218[CrossRef][Medline]
constricts pial arterioles and increases blood-brain barrier
permeability in newborn piglets. Neurosci Lett 148:137–140[CrossRef][Medline]
, interleukin-6 and several blood metabolites in
swine following intraperitoneal injection of lipopolysaccharide. J Anim
Sci, in press
This article has been cited by other articles:
 |
|
 |
|
  J. L. Williams, J. E. Minton, J. A. Patterson, J. Marchant Forde, and S. D. Eicher Lairage during transport of eighteen-kilogram pigs has an impact on innate immunity and commensal bacteria diversity in the intestines J Anim Sci, May 1, 2008; 86(5): 1232 - 1244. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Moreau, P. Dubreuil, G. Molinaro, M. Chagnon, W. Muller-Esterl, Y. Lepage, F. Marceau, and A. Adam Expression of Metallopeptidases and Kinin Receptors in Swine Oropharyngeal Tissues: Effects of Angiotensin I-Converting Enzyme Inhibition and Inflammation J. Pharmacol. Exp. Ther., December 1, 2005; 315(3): 1065 - 1074. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Sakumoto, E. Kasuya, T. Komatsu, and T. Akita Central and peripheral concentrations of tumor necrosis factor-{alpha} in Chinese Meishan pigs stimulated with lipopolysaccharide J Anim Sci, May 1, 2003; 81(5): 1274 - 1280. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Webel, D. C. Mahan, R. W. Johnson, and D. H. Baker Pretreatment of Young Pigs with Vitamin E Attenuates the Elevation in Plasma Interleukin-6 and Cortisol Caused by a Challenge Dose of Lipopolysaccharide J. Nutr., October 1, 1998; 128(10): 1657 - 1660. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
