Food deprivation alters thermoregulatory responses to lipopolysaccharide by enhancing cryogenic inflammatory signaling via prostaglandin D₂

Abstract

We tested the hypothesis that food deprivation alters body temperature (T_b) responses to bacterial LPS by enhancing inflammatory signaling that decreases T_b (cryogenic signaling) rather than by suppressing inflammatory signaling that increases T_b (febrigenic signaling). Free-feeding or food-deprived (24 h) rats received LPS at doses (500 and 2,500 μg/kg iv) that are high enough to activate both febrigenic and cryogenic signaling. At these doses, LPS caused fever in rats at an ambient temperature of 30°C, but produced hypothermia at an ambient temperature of 22°C. Whereas food deprivation had little effect on LPS fever, it enhanced LPS hypothermia, an effect that was particularly pronounced in rats injected with the higher LPS dose. Enhancement of hypothermia was not due to thermogenic incapacity, since food-deprived rats were fully capable of raising T_b in response to the thermogenic drug CL316,243 (1 mg/kg iv). Neither was enhancement of hypothermia associated with altered plasma levels of cytokines (TNF-α, IL-1β, and IL-6) or with reduced levels of an anti-inflammatory hormone (corticosterone). The levels of PGD₂ and PGE₂ during LPS hypothermia were augmented by food deprivation, although the ratio between them remained unchanged. Food deprivation, however, selectively enhanced the responsiveness of rats to the cryogenic action of PGD₂ (100 ng icv) without altering the responsiveness to febrigenic PGE₂ (100 ng icv). These findings support our hypothesis and indicate that cryogenic signaling via PGD₂ underlies enhancement of LPS hypothermia by food deprivation.

the ability of an organism to adjust its deep body temperature (T_b) in response to infection is an integral component of the host defense process. Humans (3, 11) and experimental animals (56, 65) respond to live bacteria or bacterial products with either fever (increased T_b) or hypothermia (decreased T_b). Fever prevails over hypothermia when systemic inflammation is less severe, whereas hypothermia prevails when systemic inflammation is more severe (1). Both fever and hypothermia seem to be regulated thermoregulatory responses brought about by brain-driven changes in thermoeffector activity (1, 2, 57). Fever is an energy-consuming strategy that enhances the bactericidal power of the immune system (4, 30). Hypothermia, an energy-conservation strategy (64), can have an adaptive value when inflammation is severe enough to compromise tissue perfusion (58) or when energy reserves would be threatened by the energetic cost of fever (69). Since the biological values of fever and hypothermia appear to be related to their different impacts on energy balance, it is important to understand how the state of energy balance plays a role in determining whether a host commits to fever or hypothermia.

Studies conducted in rabbits (28, 43, 72), guinea pigs (62), hamsters (6), and rats (23, 61) have shown that food deprivation attenuates the fever induced by bacterial LPS, and it can even turn fever into hypothermia. This switch from fever to hypothermia was thought to result from impairment in the ability of food-deprived animals to produce heat (62), but this idea has been challenged by recent studies (23, 61) reporting that food-deprived rats are fully capable of raising T_b in response to PGE₂, an essential mediator of fever (15, 34, 79). Interestingly, food-deprived rats can produce as much PGE₂ as free-feeding rats in response to LPS (23). The question then arises: how can food deprivation attenuate fever even when febrigenic signaling (at least via PGE₂) is not attenuated? We hypothesized that food deprivation alters thermoregulatory responses to LPS by enhancing inflammatory signaling that decreases T_b [so-called cryogenic signaling (24)] rather than by suppressing inflammatory signaling that increases T_b [febrigenic signaling (24)]. We recognize that the terms cryogenic and antipyretic are sometimes used interchangeably to refer to substances that attenuate fever regardless of the mechanism involved (29, 32, 73), but we prefer to use the term cryogenic to specifically define a substance that exerts a T_b-lowering effect, which can result either in attenuation of fever or in hypothermia (67, 69), and we reserve the term antipyretic to define a substance that does not exert a thermoregulatory effect of its own but can attenuate fever by interrupting febrigenic signaling (12, 36, 46).

The hypothesis raised above was tested in the present study. To distinguish febrigenic from cryogenic inflammatory signaling, we took advantage of the fact that LPS affects T_b differently depending on the ambient temperature (T_a). At neutral or supraneutral T_a (warm environment), rats challenged with LPS primarily develop fever, even when the dose of LPS is high (for a review, see Ref. 53). At subneutral T_a (cool environment), low doses of LPS cause fever, and moderate-to-high doses cause hypothermia followed by fever, with the magnitude of the hypothermia increasing along with the LPS dose (for a review, see Ref. 53). These observations suggest that the T_b of rats at neutral T_a reflects mostly the action of febrigenic signals, whereas T_b of rats at subneutral T_a reflects the combined actions of febrigenic and cryogenic signals. A plausible explanation for this T_a-dependence is that febrigenic signals alter heat loss effectors that govern T_b at neutral T_as, whereas both febrigenic and cryogenic signals alter heat production effectors that govern T_b at subneutral T_as (52). Investigation of thermoregulatory responses to LPS at neutral T_a and subneutral T_a is emerging as a powerful strategy to determine the extent of febrigenic vs. cryogenic signaling (66, 67). In the present study, this strategy was employed in conjunction with measurement of the levels of inflammatory mediators in free-feeding and food-deprived rats. We measured typical febrigenic mediators, IL-1β, IL-6, and PGE₂ (7, 35, 53), as well as mediators that have the potential to be cryogenic: TNF-α (5, 31, 37, 75) and PGD₂ (26, 78). We also measured corticosterone, a hormone that has anti-inflammatory actions capable of attenuating fever (antipyretic action) or hypothermia in systemic inflammation (12, 33, 49). Furthermore, we evaluated the responsiveness of free-feeding and food-deprived rats to the thermoregulatory actions of PGE₂ and PGD₂.

MATERIALS AND METHODS

Animals.

This study was conducted in adult male Wistar rats originated from Charles River Laboratories (Raleigh, NC). The colony room was maintained at a temperature of 22–24°C and a 12:12-h light-dark cycle (lights on at 07:00 AM). Rats were initially caged in pairs; after surgery, they were caged individually. Each rat was handled 5 min per day for 8 days; during a handling session, a rat was habituated to being dressed with an infusion harness that was used later in the experiments. The rats weighed 310 ± 18 g on the day prior to the experiments, i.e., just before the start of the 24-h food manipulation period. At the end of the food manipulation period, food-deprived rats had lost 6 ± 2 g (∼1.9% of body mass), whereas free-feeding rats had gained 2 ± 2 g (∼0.6% of body mass), which resulted in a difference of 8 g (∼2.5%) in the mass of free-feeding and food-deprived rats. A considerable portion of this difference is known to reflect differences in the mass of food and feces in the gastrointestinal tract (38). All protocols were approved by the Animal Care and Use Committee of the Stratton Veterans Affairs Medical Center.

Surgical preparation.

Seven to five days before an experiment, rats were implanted, as necessary, with an intravenous catheter (for administration of LPS or CL316,243, a thermogenic drug), an intracerebroventricular cannula (for administration of PGE₂ or PGD₂), and/or an intra-abdominal temperature datalogger (for recording of T_b). All procedures were performed under anesthesia with ketamine-xylazine-acepromazine (80:5:1 mg/kg ip) and antibiotic protection with enrofloxacin (2 mg/kg sc). During surgery, rats were maintained over a Delta phase isothermal pad (Braintree Scientific, Braintree, MA) warmed to 37°C.

For intravenous catheterization, a small longitudinal incision was made on the right ventral surface of the neck. The right jugular vein was exposed, freed from its surrounding connective tissue, and ligated. A 3-French polyurethane catheter (Instech Laboratories, Plymouth Meeting, PA) filled with heparinized (50 U/ml) saline was passed into the superior vena cava through the jugular vein and secured in place with ligatures. The distal end of the catheter was closed with a stainless steel plug, tunneled under the skin, and exteriorized at the nape. The intravenous catheters were flushed with heparinized saline every other day.

For intracerebroventricular cannulation, a rat was fixed to a stereotaxic apparatus (David Kopf, Tujunga, CA) with the incisor bar set at −3.3 mm. The skin was incised over the sagittal suture, the periosteum was excised, supporting microscrews were driven into the skull, and a steel guide cannula (Plastics One, Roanoke, VA) was implanted. The tip of the cannula was placed 0.5 mm dorsal to the right lateral ventricle by using the following stereotaxic coordinates: −0.5 mm from bregma, −1.5 mm from the midline, and 3.5 mm from the skull surface. The implanted cannula was attached to the supporting microscrews with acrylic cement.

For datalogger implantation, a midline laparotomy was performed and a spacer was used to keep the abdominal muscles separated. The rat was then positioned in lateral decubitus, and a SubCue miniature datalogger (Calgary, Alberta, Canada) was attached to the internal side of the dorsolateral abdominal wall with 4-0 silk sutures.

Food deprivation.

Food-deprivation was initiated at 11 am on the day before an experiment, and continued until the end of the experiment on the following day. During the food-deprivation period, a rat was placed in a cage equipped with an elevated metal grid floor to prevent coprophagy. Control rats had free access to food during the entire experiment, but were also placed on an elevated grid floor to ensure exposure to identical environmental conditions. Water was available ad libitum.

Experimental setup.

At 7 am on the day of the experiment, a rat in its cage was transferred to an environmental chamber (model NQ1; Environmental Growth Chambers, Chagrin Falls, OH), which contained fluorescent lamps that operated according to the light-dark cycle. In the chamber, T_a was maintained at 30.0 ± 0.1°C or 22.0 ± 0.1 °C; relative air humidity was always kept at 50 ± 5%. Whereas 30°C is a neutral T_a for rats in the chamber environment [as verified by the presence and variability of tail skin vasodilation (55)], 22°C is a subneutral T_a for rats in the same environment [as verified by the absence of tail skin vasodilation (55)]; see Supplemental Figure 1 (supplemental figures are available online at the American Journal of Physiology–Regulatory, Integrative and Comparative Physiology website). In a rat designated for intravenous drug administration, the preimplanted venous catheter was extended with a length of polyethylene-50 (PE-50) tubing. The extension was passed sequentially through a Covance infusion harness, which was worn by the rat, and a metal spring (Instech Solomon, Plymouth Meeting, PA), which protected the catheter extension from bites and scratches. The spring and extension, in turn, were connected to a swivel (Instech Solomon) mounted above the cage; the swivel prevented the extension from twisting as the rat moved about its cage. Another PE-50 extension, leaving the swivel, was passed to the exterior of the chamber through a wall port and connected to a 1-ml syringe. In a rat designated for intracerebroventricular drug administration, an injector needle (Plastics One) was fitted into the preimplanted guide cannula; the needle protruded 1 mm beyond the guide cannula to reach the lateral ventricle. The injector needle was coupled via a connector assembly (Plastics One) to a swivel (Instech) from which a PE-50 extension was passed to the exterior of the chamber and connected to a 10-μl Hamilton syringe. Both the intravenous and intracerebroventricular extensions were filled with the drugs of interest or their vehicles. This setup permits drug administration to be performed in a freely-moving, undisturbed rat.

With the infusion lines in place (usually around 8 AM), rats were allowed to become accustomed to the experimental setup for 3 h. Drug or its vehicle was administered at ∼11 am, i.e., 24 h after the initiation of the food manipulation period. If a rat was designated for measurement of T_b, it was left undisturbed for 360 min postinjection, then euthanized with pentobarbital sodium (100 mg/kg ip) for removal of the abdominal datalogger from which data were downloaded to a computer. If a rat was designated for blood collection, it was left undisturbed until the time point of interest (for details, see Results), at which the rat was anesthesized with ketamine-xylazine (8:0.5 mg/kg) via the extension of the venous catheter, and blood was collected by cardiac puncture.

Drug administration.

Systemic inflammation was induced by E. coli 0111:B4 LPS purchased from Sigma-Aldrich (St. Louis, MO). LPS was diluted to a final concentration of either 500 or 2,500 μg/ml in saline, and the resulting solution was bolus injected (1 ml/kg iv). The LPS doses administered were high enough to activate both febrigenic and cryogenic signaling (66, 67). Dosing of LPS to each rat was based on the body mass measured immediately before initiation of the 24-h food manipulation period, because a considerable portion of the mass lost by a rat over 24 h of food deprivation is simply a consequence of gastrointestinal emptying (38). Even if some actual body mass was lost during the 24-h food-deprivation period, it could not have exceeded the 2.5% difference in the mass of free-feeding and food-deprived rats (see Animals). Thus, any dosing difference between food-deprived and free-feeding rats is likely to be negligible in the present study.

CL316,243 (Sigma-Aldrich) was administered to compare the ability of food-deprived and free-feeding rats to increase their T_b in response to this thermogenic β₃-receptor agonist (13, 21). A solution of CL316,243 (1 mg/ml) in saline was bolus injected (1 ml/kg iv). As explained above, dosing was based on the body mass of the rats immediately before initiation of the food manipulation period.

PGD₂ and PGE₂ (Cayman, Ann Arbor, MI) were administered to compare the responsiveness of food-deprived and free-feeding rats to a cryogenic substance and a febrigenic substance, respectively. Each drug was dissolved to a final concentration of 20 ng/μl in saline containing 1% of ethanol. The solution was infused intracerebroventricularly at a rate of 1 μl/min for 5 min. Correct placement of the intracerebroventricular injector needle was confirmed by postmortem verification using Chicago Sky Blue (0.1%, 5 μl icv).

Assays.

Blood samples were placed into EDTA-coated Eppendorf tubes, the tubes were centrifuged (6,000 g, 10 min, 4°C), and the resulting plasma was stored at −80°C. The plasma levels of TNF-α, IL-1β, and IL-6 were determined by sandwich ELISA using kits from Thermo Scientific (manufactured by Pierce Biotechnology, Rockford, IL) and a SpectroMax Plus microplate spectrophotometer (Molecular Devices, Sunnyvale, CA). Plasma samples were assayed at a dilution of 1:200 in the TNF-α assay, 1:100 in the IL-6 assay, and 1:2 in the IL-1β assay. Assay ranges were 25.6–2,500 pg/ml for TNF-α and IL-1β, and 31–2,000 pg/ml for IL-6. All samples were run simultaneously, in duplicate.

Because cytokine production is modulated by the hypothamalo-pituitary-adrenal axis (77), we determined the plasma concentration of corticosterone, the final hormone of this axis. Corticosterone was assayed by competitive ELISA using a kit from Cayman Chemical (Ann Arbor, MI) and a SpectroMax Plus microplate spectrophotometer (Molecular Devices). Samples were assayed at a dilution of 1:100; the assay range was 41–8,000 pg/ml. All samples were run simultaneously, in duplicate.

Currently available antibody-based methods for prostaglandin analysis appear to be less accurate than chromatographic methods (22). Hence, we measured the plasma concentration of PGD₂ and PGE₂ by liquid chromatography-mass spectrometry according to the method of Masoodi and Nicolaou (40), with slight modifications. Each plasma sample (900 μl) was acidified to pH 3 with HCl (1 M) and subjected to solid-phase extraction as follows: after the sample was applied to a C18 Sep-Pak column (Waters, Milford, MA), the column was washed sequentially with water (2 ml) and 10% methanol in water (2 ml), then PGs were eluted from the column with methyl formate (2 ml). The methyl formate fraction was dried under a stream of nitrogen, and the resulting residue was reconstituted in 60 μl of methanol prior to liquid chromatography-mass spectrometry analysis. The analysis was performed using a Waters 600 liquid chromatography pump and injector coupled to a Waters Acquity tandem-quadrupole mass spectrometer. Prostaglandin separation was carried out in a 2.0 × 250 mm C18 Gemini column (Phenomenex, Torrance, CA) eluted isocratically with water-acetonitrile-formic acid (58:42:0.01) at a rate of 250 μl/min and temperature of 25°C; the injection volume was 20 μl. Detection was done by negative ion electrospray ionization tandem mass spectrometry using the following parameters: source temperature of 150°C, electrospray voltage of −3.6 kV, declustering potential of −21 V, and nitrogen gas collision at −16 eV. The method of detection was multiple reaction monitoring of the transition from the deprotonated molecules of m/z 351 to the most abundant fragment ion of m/z 271, which were present in the spectra of both PGD₂ and PGE₂; these PGs are geometrical isomers.

Calibration curves were made using standard methanolic solutions of PGD₂ and PGE₂ (both from Cayman). The retention time of PGD₂ was 8.4 min, whereas the retention time of PGE₂ was 7.3 min. For both compounds, the linear (r² = 0.999) quantification range extended from 0.5 ng/ml to at least 50 ng/ml. Recoveries from the extraction procedure were 89% and 90% for PGD₂ and PGE₂, respectively.

Data analysis.

The initial (basal) value of T_b of each rat was calculated by averaging T_b values recorded during the 30 min that preceded drug administration. Thermoregulatory responses to drugs were evaluated based on T_b changes in relation to initial T_b. Changes in TNF-α, IL-1β, IL-6, corticosterone, PGD₂, and PGE₂ were evaluated based on the absolute concentrations of these substances in plasma. Statistical comparisons were performed using Statistica Advanced 8.0 (StatSoft, Tulsa, OK). Repeated-measures ANOVA was employed to evaluate the effects of feeding status and treatment on consecutive measures of T_b, whereas factorial ANOVA was employed to evaluate the effects of feeding status and treatment on initial T_b or on plasma levels of inflammatory mediators. Post hoc analysis was conducted using the Fisher least significant difference test. The level of significance was set at P < 0.05. Data are reported as means ± SE.

RESULTS

Effects of food deprivation on initial T_b and thermogenic capacity.

At neutral T_a (30°C), there was no difference between the initial (basal) T_b values of food-deprived (24 h) and free-feeding rats (Fig. 1, left). However, at subneutral T_a (22°C), the initial T_b of food-deprived rats was ∼0.5°C lower than the initial T_b of free-feeding rats (Fig. 1, left); this difference was statistically significant (P < 0.0001). To investigate whether the lower initial T_b of food-deprived rats could reflect thermogenic incapacity, we assessed the ability of rats to elevate T_b in response to CL316,243 at a dose (1 mg/kg iv) known to activate thermogenesis (21). Both free-feeding and food-deprived rats exposed to subneutral T_a responded to CL316,243 with a transient rise in T_b (Fig. 1, right): T_b was statistically higher than initial T_b at 150–250 min postinjection (P < 0.04), and there was no statistical difference between the free-feeding and food-deprived groups. This result indicates that thermogenic capacity was not diminished by the food-deprivation regimen used in the present study.

Fig. 1.

Influences of food deprivation (24 h) on initial (basal) body temperature and thermogenic capacity. Left: summary of initial body temperatures of free-feeding and food-deprived rats exposed to an ambient temperature of 30°C (warm environment) or 22°C (cool environment); this summary includes values obtained in all thermophysiological experiments. Right: effect of a thermogenic β₃-receptor agonist (CL316,243; 1 mg/kg iv) on the body temperature of rats exposed to an ambient temperature of 22°C; within this experiment, initial body temperatures were 37.3 ± 0.1°C for free-feeding rats and 36.6 ± 0.3°C for food-deprived rats. The number of animals in each group is shown in parentheses. *Statistical differences between food-deprived and free-feeding rats; see text for more details about statistical results.

Effects of food deprivation on LPS-induced thermoregulatory responses.

We subsequently studied how food deprivation affects thermoregulatory responses to LPS in rats kept at different T_as. The results of this experiment are depicted in Fig. 2. Intravenous administration of saline (the LPS vehicle) caused no change in T_b, regardless of T_a (30°C or 22°C) or feeding status (free-feeding or food-deprived). On the other hand, intravenous LPS evoked thermoregulatory responses that were dependent on T_a and feeding status. At neutral T_a, both moderate (500 μg/kg) and high (2,500 μg/kg) doses of LPS produced fever. Statistical difference between febrile T_b and initial T_b was observed at 70–360 min following injection of the moderate LPS dose (P < 0.04) and at 40–360 min following injection of the high dose (P < 0.04). At the peak of fever, T_b was 2.0°C higher than initial T_b regardless of dose. The febrile responses to LPS were similar in free-feeding and food-deprived rats, with the exception of a mild attenuation in the fever of food-deprived rats at 100–160 min following the high LPS dose (P < 0.04).

Fig. 2.

Effects of intravenous LPS (doses indicated) or its vehicle (saline) on the body temperature of free-feeding or food-deprived (24 h) rats exposed to an ambient temperature of 30°C or 22°C. Within this experiment, initial body temperatures were 36.8 ± 0.1°C for free-feeding rats at 30°C; 36.7 ± 0.1°C for food-deprived rats at 30°C; 37.0 ± 0.1°C for free-feeding rats at 22°C; and 36.6 ± 0.1°C for food-deprived rats at 22°C. The number of animals in each group is shown in parentheses. *Statistical differences between food-deprived and free-feeding rats.

At subneutral T_a, the same doses of LPS elicited typical hypothermic responses. In free-feeding rats, hypothermic T_b was statistically different from initial T_b at 30–110 min following injection of the moderate LPS dose (P < 0.009) and at 50–100 min following injection of the high dose (P < 0.01); within these time windows, T_b fell by 1.5°C and 1.8°C, respectively. Compared with free-feeding rats, food-deprived rats responded to LPS with enhanced hypothermic responses. Enhancement of hypothermia was mild, although statistically significant (P < 0.04; 110–150 min), in food-deprived rats injected with the moderate LPS dose (500 μg/kg). However, it was very prominent in food-deprived rats injected with the high dose (2,500 μg/kg), in which hypothermia was significantly enhanced from 120 to 360 min postinjection (P < 0.01). By the end of this period, T_b had fallen by 2.4°C (Fig. 2). Of note, the hypothermic response of food-deprived rats to the high LPS dose (2,500 μg/kg) was significantly (P < 0.03) enhanced even compared with the response of free-feeding rats to a twofold higher dose (5,000 μg/kg); see Supplemental Figure 2. This finding rules out any possibility that small differences in LPS dosing due to a 2.5% difference in the mass of free-feeding and food-deprived rats might have accounted for the observed differences in LPS hypothermia.

Effects of food deprivation on the levels of inflammatory mediators.

To investigate whether enhancement of LPS hypothermia by food deprivation was associated with changes in the levels of inflammatory mediators, we studied the plasma levels of cytokines (TNF-α, IL-1β, IL-6) and PGD₂ and PGE₂ known to be involved in the thermoregulatory responses to LPS. Plasma samples were collected 100 min after administration of LPS (2,500 μg/kg iv) or saline (iv) in free-feeding or food-deprived rats exposed to a subneutral T_a of 22°C. This time point corresponds to the maximal rate of separation between the hypothermic responses to LPS in free-feeding and food-deprived rats (see Fig. 2). None of the inflammatory mediators was detectable in the plasma of saline-treated rats, whereas all of them surged in the plasma of LPS-treated rats (Figs. 3 and 4). The LPS-induced surges in plasma TNF-α, IL-1β, and IL-6 did not differ between free-feeding and food-deprived rats (Fig. 3). However, the LPS-induced surges in PGD₂ and PGE₂ were more than threefold higher in food-deprived rats than in free-feeding rats (Fig. 4; P < 0.004 for PGD₂; P < 0.02 for PGE₂). It is interesting though that food deprivation augmented the surges in PGD₂ and PGE₂ to a similar extent so that the PGD₂-to-PGE₂ ratio was not altered (Fig. 4).

Fig. 3.

Plasma levels of the cytokines TNF-α, IL-1β, and IL-6 at 100 min after administration of LPS (2,500 μg/kg iv) or saline (iv) in free-feeding or food-deprived (24 h) rats exposed to an ambient temperature of 22°C. The time of sample collection (100 min) corresponds to the maximal rate of separation between the hypothermic responses to LPS in free-feeding and food-deprived rats (see Fig. 2, bottom, right). The number of animals in each group is shown in parentheses. n.d., Nondetectable levels. There was no statistical difference between the levels found in food-deprived and free-feeding rats.

Fig. 4.

Plasma levels of PGD₂ and PGE₂, as well as the ratio between them, at 100 min after administration of LPS (2,500 μg/kg iv) or saline (iv) in free-feeding or food-deprived (24 h) rats exposed to an ambient temperature of 22°C. The number of animals in each group is shown in parentheses. N/A, not applicable. *Statistical differences between food-deprived and free-feeding rats.

In addition, plasma samples were assayed for corticosterone, a glucocorticoid hormone that is modulated by both food deprivation and LPS (16, 19) and can modulate thermoregulatory responses to LPS (12, 33, 49). In the absence of inflammation (saline treatment), plasma corticosterone tended to be higher in food-deprived rats (37 ± 15 ng/ml) than in free-feeding rats (28 ± 11 ng/ml), but this difference did not reach statistical significance. Compared with saline, LPS significantly (P < 0.0001) increased plasma corticosterone in both free-feeding and food-deprived rats (Fig. 5, left). Absolute corticosterone levels were higher in those LPS-treated rats that were food-deprived than in those that were free-feeding (P < 0.0001). However, the relative increase in corticosterone level (i.e., the ratio between levels in LPS- and saline-treated rats) did not differ between free-feeding and food-deprived rats (Fig. 5, right).

Fig. 5.

Plasma levels of corticosterone at 100 min after administration of LPS (2,500 μg/kg iv) or saline (iv) in free-feeding or food-deprived (24 h) rats exposed to an ambient temperature of 22°C. Relative changes in corticosterone are shown as the ratios between corticosterone levels in LPS-treated rats and in the corresponding saline-treated controls. The number of animals in each group is shown in parentheses. *Statistical differences between food-deprived and free-feeding rats.

Effects of food deprivation on the thermoregulatory responses to PGD₂ and PGE₂.

In addition to understanding how food deprivation alters the levels of inflammatory mediators, it is important to understand how it affects the responsiveness of the body to inflammatory mediators. Hence, we investigated whether food deprivation alters the ability of PGD₂ and PGE₂ to produce hypothermia and fever, respectively. Each substance or its vehicle was administered into the lateral cerebral ventricle at a dose (100 ng) known to affect T_b in rats (71, 78). The experiment was conducted in rats exposed to subneutral T_a (22°C), which permits the development of fever or hypothermia (53). The results of this experiment are shown in Fig. 6. Vehicle administration did not significantly alter the T_b of free-feeding or food-deprived rats. PGD₂ caused a modest (0.4°C) fall in the T_b of free-feeding rats, which was significantly different (P < 0.03) from initial T_b at 40–50 min postinjection. The hypothermic response to PGD₂ was largely exaggerated in food-deprived rats: not only was the response of the food-deprived rats longer-lasting than the response of free-feeding rats (P < 0.04; 60–170 min), but it was also more pronounced (0.8°C) in magnitude (P < 0.04). In contrast to PGD₂, PGE₂ induced fever. The fever induced by PGE₂ was virtually identical in free-feeding and food-deprived rats. The T_b of PGE₂-treated rats peaked 1.8°C above initial T_b, and it was statistically different from initial T_b at 20–50 min postinjection (P < 0.02).

Fig. 6.

Effects of intracerebroventricular PGD₂ (100 ng), PGE₂ (100 ng), or their vehicle (1% ethanol in saline) on the body temperature of free-feeding or food-deprived (24 h) rats exposed to an ambient temperature of 22°C. Within this experiment, initial body temperatures were 36.7 ± 0.1°C for the free-feeding rats and 36.3 ± 0.1°C for the food-deprived rats. The number of animals in each group is shown in parentheses. *Statistical differences between food-deprived and free-feeding rats.

DISCUSSION

The objective of the present study was to investigate how food deprivation alters febrigenic vs. cryogenic inflammatory signaling during systemic inflammation induced by LPS. We chose to deprive rats of food for no more than 24 h (prior to drug administration) to avoid the thermogenic incapacity known to disrupt thermoregulation following longer periods of food deprivation (20, 63). Indeed, food-deprived rats in the present study were fully able to increase T_b in response to a drug (CL316,243) that directly activates β₃-driven thermogenesis in brown fat. Interestingly, though, their basal T_b was slightly reduced at subneutral T_a (cool environment), but not at neutral T_a (warm environment). This observation may reflect the fact that food deprivation decreases the threshold for cold-induced thermogenesis even when it does not impair thermogenic capacity (60). The thermogenic threshold governs T_b at subneutral T_a, but it is of little importance for maintenance of T_b at neutral T_a (55). The T_b of rats at neutral T_a is governed by heat-loss effectors (55), which are unaffected by deprivation of food for as long as 3 days (60). To account for small (≤ 0.5°C) differences in basal T_b related to T_a and feeding status, we evaluated thermoregulatory responses to LPS based on the extent of T_b change in relation to initial (basal) T_b.

The present study was the first to evaluate whether food deprivation alters thermoregulatory responses to LPS in adult animals exposed to neutral T_a in which fever is the predominant thermoregulatory response and is virtually under the sole influence of febrigenic inflammatory signaling (68, 70). The results indicate that food deprivation has little or no effect on LPS fever at neutral T_a: the fever induced by the moderate LPS dose was not affected by food deprivation; the fever induced by the high dose was attenuated, but this attenuation was weak. Although these results do not rule out the possibility that food deprivation may weakly attenuate febrigenic signaling via a yet unidentified mechanism, they indicate that febrigenic signaling is not largely affected by food deprivation. To investigate whether food deprivation alters cryogenic signaling, we studied the thermoregulatory responses to LPS in rats exposed to subneutral T_a in which fever or hypothermia can occur as a reflection of the balance between febrigenic and cryogenic inflammatory signaling (67, 69, 73). Moderate and high doses of LPS caused hypothermia in rats at subneutral T_a, and the hypothermic response to either dose was enhanced by food deprivation. Enhancement of hypothermia by food deprivation was particularly pronounced in rats injected with the high LPS dose. Such an enhancement of hypothermia likely reflected an enhancement in cryogenic signaling, as our experiments at neutral T_a show that food deprivation does not exert a pronounced influence on febrigenic signaling.

The present study also provides insights into the mechanisms by which food deprivation enhances cryogenic signaling. Inflammatory mediators were assayed in samples collected at the time (100 min) corresponding to the maximal rate of separation between the hypothermic responses of free-feeding and food-deprived rats injected with LPS, when their cryogenic signaling is most likely to differ. The cytokines we assayed have different roles in thermoregulation. IL-1β and IL-6 appear to be exclusive mediators of fever (35), whereas TNF-α appears to be febrigenic at low doses but cryogenic at high doses (5). An increasing production of TNF-α, as systemic inflammation becomes more severe, is thought to promote a switch from fever to hypothermia (31, 37, 75). In the present study, the LPS-induced surges in TNF-α, IL-1β, and IL-6 were not altered by food deprivation, at least not at 100 min post-LPS. A study by Inoue et al. (23) agrees with the present results in that it demonstrated unaltered surges of IL-1β and IL-6 in the plasma of food-deprived rats injected with LPS, but it is not in line with the present results in that it reported a reduced surge of TNF-α in the plasma of the same animals. In yet another study, Faggioni et al. (16) demonstrated that food deprivation can augment the plasma TNF-α response to LPS. These discrepancies regarding TNF-α levels may be related to differences in the regimen of food deprivation and LPS dose. Compared with the present study, Inoue et al. (23) employed a longer (2-day) period of food deprivation and a lower dose of LPS, whereas Faggioni et al. (16) employed a longer (2-day) regimen of food deprivation in combination with a higher LPS dose. Regardless, the present results indicate that altered plasma levels of TNF-α are unlikely to be necessary for enhancement of cryogenic signaling by food deprivation, because plasma TNF-α was unaltered when the hypothermic response of food-deprived rats to LPS was enhanced at the highest rate. The present results, however, do not rule out possible alterations in other components of the TNF-α pathway, such as membrane-bound TNF-α, soluble TNF receptor antagonist, and cellular TNF receptors (14, 76).

Interestingly, plasma levels of PGE₂ and PGD₂ were augmented during the enhanced hypothermic response of food-deprived rats to LPS. This augmentation, at least in the case of PGE₂, may be the result of an increased synthesis in peripheral tissues, as it has been reported that food deprivation does not augment the synthesis of PGE₂ in the brain of LPS-injected rats (23). Regardless of its origin, circulating PGE₂ is fully capable of activating brain thermoregulatory circuitries to produce fever (39, 51, 54, 68). Circulating PGD₂ may also be able to activate brain thermoregulatory circuitries, as dense PGD₂ binding occurs in the preoptic region (80), a brain site that plays key roles in thermoregulation (45) and is accessible to circulating substances via the organum vasculosum of the lamina terminalis (8, 42, 59). Alternatively, PGD₂ receptors in leptomeninges may convey signals of circulating PGD₂ to the brain (50). Ueno et al. (78) were the first to report that PGD₂ causes hypothermia when microinjected intracerebroventricularly or into the preoptic area of conscious rats at subneutral T_a. Since then, the hypothermic effect of intracerebroventricular PGD₂ in rats has been reproduced in a study by Kandasamy and Hunt (26) and in the present study. On the contrary, Gao et al. (17) reported that PGD₂ administered into the cisterna magna does not cause hypothermia in rats and, in fact, causes a delayed fever. The basis for this conflicting observation is unknown, but could be related to an insufficient amount of intracisternal PGD₂ reaching the preoptic area or to the fact that Gao et al. (17) briefly anesthetized rats at the time of PGD₂ administration. Another consideration is that the spontaneous decomposition of PGD₂ in aqueous solutions is faster than the decomposition of many other prostaglandins (10). Presumably for this reason, we observed that PGD₂ solutions caused hypothermia only if stored no longer than 3 wk at −80°C.

Although food deprivation similarly enhanced the LPS-induced surges in PGD₂ and PGE₂, it selectively intensified the responsiveness of rats to the cryogenic action of PGD₂ without interfering with the responsiveness to febrigenic PGE₂. This is an important finding of the present study as it reveals that enhancement of cryogenic signaling via PGD₂ is associated with enhancement of LPS hypothermia by food deprivation. Regarding PGE₂, the augmented rise in its plasma level may have functional implications to prevent excessive falls in T_b. It should be noted, however, that any augmentation in PGE₂ levels by food deprivation does not seem to have functional implication for LPS fever at neutral T_a, as fever was not enhanced by food deprivation. There are at least two possible interpretations to this intriguing observation. On one hand, it is possible that febrigenic signaling reaches a ceiling at PGE₂ levels encountered in free-feeding rats so that it remains unchanged even when PGE₂ levels are further increased in food-deprived rats. On the other hand, it is possible that food deprivation does not enhance the LPS-induced surge in PGE₂ levels when rats are exposed to neutral T_a, even though it does so when rats are exposed to subneutral T_a. Future studies are necessary to investigate whether T_a has the ability to change the network of inflammatory mediators.

Another important finding of the present study is that enhancement of LPS hypothermia by food deprivation was not associated with impairment of the immune activation of corticosterone production. Such an impairment may occur following relatively long (at least 2 days) periods of food deprivation (16), and it would be expected to result in exaggeration of LPS hypothermia (33, 49), but it did not occur when rats in the present study were deprived of food for a single day. On the contrary, the regimen of food deprivation employed in the present study elevated the absolute concentration of corticosterone in the plasma of LPS-treated rats, even though the relative (LPS-induced) change in corticosterone was not altered. Nevertheless, the elevation in the absolute concentration of corticosterone was not strong enough to suppress LPS-induced surges in cytokines (which were unaltered by food deprivation), LPS-induced surges in prostaglandins (which were augmented by food deprivation), or LPS-induced hypothermia (which was enhanced by food deprivation).

In conclusion, the present study supports the hypothesis that food deprivation alters thermoregulatory responses to LPS mainly by enhancing cryogenic inflammatory signaling, and not by suppressing febrigenic signaling, at least not to a considerable extent. Enhancement of cryogenic signaling by food deprivation may involve PGD₂, as food deprivation not only enhanced the level of PGD₂ during the hypothermic response to LPS, but also increased the responsiveness of rats to the cryogenic (hypothermic) action of PGD₂.

Perspectives and Significance

The present study paves a new path for researching the mechanisms linking energy balance to cryogenic inflammatory signaling via PGD₂. Future studies will be necessary to determine which molecular alterations account for the enhanced responsiveness to cryogenic PGD₂ in food-deprived animals. Such alterations may occur at the level of PGD₂ receptor expression and signaling (9, 44), as well as at the level of PGD₂ transport across the blood-brain barrier (25, 27). Future studies will also be necessary to determine how feeding status interferes with each step of the PGD₂ biosynthetic pathway, which involves three sequential reactions catalyzed by phospholipase A₂, cyclooxygenase (also known as PGH synthase), and PGD synthase (47). Of these enzymes, cyclooxygenase isoforms have been shown to play important roles in LPS-induced fever and hypothermia, with fever being mediated by PGE₂ produced via cyclooxygenase-2 (15, 39, 41, 68) and hypothermia being mediated by a yet unidentified product of cyclooxygenase-1 (67). Such a product may be PGD₂ because of its cryogenic action (present study and Refs. 26, 78), although it should be noted that the direct involvement of PGD₂ in the mediation of LPS hypothermia has not yet been demonstrated. Inhibitors of PGD synthase and antagonists of PGD₂ receptors are becoming available, but their ability to block LPS hypothermia has never been tested. However, it is surprising that EDJ300520, a drug that preferentially inhibits the synthesis of PGD₂ over PGE₂ in macrophages, has been reported to attenuate LPS fever in rats (18). As EDJ300520 awaits a more detailed pharmacological characterization, the possibility that it attenuates fever via a mechanism other than inhibition of PGD₂ synthesis needs to be considered. Another possibility is that PGD₂ might have a dual role in mediating both fever and hypothermia. For instance, while peripherally acting PGD₂ may be febrigenic (as suggested in Ref. 18), centrally acting PGD₂ seems to be cryogenic (present study and Refs. 26, 78). It is worth mentioning that the relevance of the present study may extend beyond thermoregulation, since PGD₂ may be involved in other aspects of the pathophysiology of systemic inflammation, including sleep and pain (48, 74).

GRANTS

The study was supported by funds from the Albany College of Pharmacy and Health Sciences (to A. A. Steiner) and by a National Institute of Health Grant R15-AI072744 (to C. Feleder). X. Yao was the recipient of a scholarship from the China Scholarship Council.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).