We have demonstrated that after intraperitoneal lipopolysaccharide (LPS) injection, old rats mount fevers similar to those of young rats at an ambient temperature (Ta) of 31°C, but not at 21°C. The same is true for intraperitoneal or intravenous IL-1β administration. The underlying mechanism responsible for blunted fever in old rats may be a deficiency in communication between the periphery and the brain. Possibly, peripheral cytokine actions are altered in old rats, such that the signal that reaches the brain is diminished. Here, we hypothesized that at standard laboratory temperatures, not enough IL-1β is reaching the brain for fever to occur and that a warmer Ta would increase the influx of IL-1β into the brain, enabling old rats to generate fever. Young (3–5 mo) and old (23–29 mo) Long-Evans rats were maintained for 3 days at either Ta 21 or 31°C prior to intravenous injection with radiolabeled IL-1β to measure passage across the blood-brain barrier. Young rats showed similar influx of IL-1β into the brain at the two Tas, but old rats showed significant influx only at the warmer Ta. These data suggest that the lack of fever at a cool Ta may be due to a reduced influx of IL-1β into the brain.
- ambient temperature
some aged rodents (35, 47, 50), as well as elderly humans (9, 15, 37), show blunted fevers in response to illness, injury, and infection. Lack of fever in elderly humans can be a major problem. They may be infected yet show low or no fever, which can lead to a delay in diagnosis of the underlying cause and therefore a delay in treatment. However, there are two instances when old rats can generate fever equivalent to those of young rats: when the ambient temperature (Ta) is warm enough and if the pyrogen is administered directly into the brain.
We have previously shown that when given an opportunity to behaviorally thermoregulate, old rats choose to stay at a warmer Ta and then develop fevers as high as those of young rats (14). Further, when old rats are maintained at a warm Ta (31°C), their febrile responses to lipopolysaccharide (LPS) at the higher Ta are indistinguishable from those of their younger counterparts (38). This phenomenon is not specific to LPS. Similar results were seen with both intraperitoneal and intravenous administration of the proinflammatory cytokine IL-1β (10).
At ordinary laboratory Tas, old rats are able to develop fevers when the pyrogen is administered centrally. Previously, we showed that old rats develop fever equivalent to those of young rats after central administration of IL-1β at a Ta of 21–23°C (39). The same is true after central administration of PGE2, which is thought to be induced by IL-1β (42), and endogenous pyrogen (47). These results indicate that the central mechanisms of fever are functional in aged rats. Further, they demonstrate thermoeffector mechanisms in old rats are capable of raising Tb. Therefore, it may be that the underlying mechanism responsible for blunted fever in old rats is a deficiency in communication between the periphery and the brain. Possibly, peripheral cytokine actions are altered in old rats, resulting in diminished signaling to the brain.
One of the first cytokines involved in the fever response is the proinflammatory cytokine IL-1β, and there is evidence that it is actively transported across the blood-brain barrier (BBB) and into the brain (5, 6). Blocking IL-1β with IL-1 receptor antagonist (IL-1ra) or an anti-rat neutralizing IL-1β antibody results in a suppression of the fever response (24, 30). The majority of studies using IL-1β and IL-1R1 knockout mice, as well as IL-1β deletion mutants report reduced fever compared with wild-type mice after systemic injection of LPS, influenza virus, IL-1β, or turpentine (22, 27, 28, 33, 52).
Given the role of IL-1β in fever, a reduction in IL-1β signaling could result in blunted fevers in aged animals either by not inducing proinflammatory cytokines in the periphery and/or brain (e.g., IL-6, TNF-α), or by failing to induce the synthesis and release of PGE2 in the brain. Evidence for such diminished signaling was demonstrated by McLay et al. (34a), who showed that old mice had reduced influx of IL-1β into the brain compared with young mice at standard laboratory Tas. Because a pyrogenic signal must reach the brain for a fever to occur, this suggests that the brains of old rodents are not receiving the appropriate signal to initiate the autonomic responses that lead to fever.
In light of our Ta findings, along with the age-related reduction of IL-1β blood-to-brain influx found by McLay et al. (34a), we hypothesized that at cooler laboratory temperatures, not enough IL-1β is reaching the brain for fever to occur and that a warmer Ta would increase the influx of IL-1β into the brain, enabling old rats to generate fevers. Here, we report that young rats showed similar influx of IL-1β into the brain at the two Tas, but old rats showed significant influx, similar to the young rats, only at the warmer Ta.
MATERIALS AND METHODS
Animals and Surgery
Subjects were young (3–5 mo) and old (24–29 mo) male and female Long-Evans rats. Animals were housed individually in polycarbonate plastic cages (56 × 32 × 20 cm) with soft pellet bedding and maintained at Ta 23 ± 1°C on a 12:12-h light-dark cycle (lights on at 0700). Food and water were available ad libitum. Young rats were anesthetized with a solution of ketamine HCL (87 mg/kg body wt) and xylazine (13 mg/kg body wt) and implanted intraperitoneally with a battery-operated biotelemetry device (model VM; Mini-Mitter, Bend, OR). Because old rats are more sensitive to the effects of anesthesia than are young rats, old rats were anesthetized with 80% of this dose and similarly implanted with transmitters. At the time of transmitter implantation, rats were also implanted with jugular cannulas. Following surgery, rats were injected subcutaneously with buprenorphine (0.05 mg/kg). Cannulas were flushed with heparinized saline daily. Body temperature (Tb) was measured continuously from time of implantation until the end of the experiment. Rats were given 1 wk to recover from surgery before the beginning of the experiment. Rats were not put in the experiment unless they had reached their presurgery weight and their circadian temperature rhythms were normal. All procedures were approved by the University of Delaware Institutional Animal Care and Use Committee.
Young (n = 12, 6 male and 6 female) and old (n = 10, 5 male and 5 female) rats were housed at Ta 21 or 31°C for 3 days before injection. We have previously found that circadian temperature rhythms are briefly disrupted when the animals are moved to a new Ta but are normalized by day 3. On day 4, animals were given an intravenous injection containing 1 μg/kg 125I-IL-1β, and 131I-labeled albumin. Rats were decapitated at several time points between 1 and 60 min postinjection, allowing for multiple-time uptake analysis (discussed below). After decapitation, the brain was quickly removed and rinsed with saline, and trunk blood was collected. The brain was dissected over ice into six regions: hypothalamus, hippocampus, midbrain, cerebellum, pons-medulla, and cortex. Each region was placed in tared tubes and weighed. The level of radioactivity in plasma and brain regions was determined in a gamma counter. Values for whole brain were calculated by adding the weights and levels of radioactivity for all regions. At the time of injection, vaginal smears were obtained from the young female rats, whereby the tip of an eyedropper was filled with one to two drops of sterile saline, expelled into the vagina, and then collected and placed on a microscope slide. Slides were stained with cresyl violet and examined under a microscope, and the estrous stage was determined by the predominant cell type in the sample.
BSA was purchased from Sigma-Aldrich (St. Louis, MO). Albumin can leak into the circumventricular organs (CVOs) but is restricted from further progression by the ependymal barrier. If there is no significant influx of BSA, the presence of the BSA in tissues can then be attributed to blood spilled or leaked into the brain tissue (through CVOs) but outside the BBB. This volume can be subtracted from influx calculations as a correction for leakage and contamination. If the influx of the IL-1β is not greater than the influx of albumin, one can conclude that IL-1β is not crossing the BBB.
Rat recombinant, carrier-free IL-1β was purchased from R&D Systems (Minneapolis, MN). Labeled IL-1β was diluted in PBS with 1% BSA and injected intravenously at a dose of 1 μg/kg with counts/min (cpm) of about 1 × 107. Labeled IL-1β was combined with labeled BSA (∼1 × 107 cpm) in a single intravenous injection.
IL-1β was labeled with 125I and BSA with 131I by Amersham Biosciences to a specific activity of about 1,000–2,000 Ci/mmol.
Calculations for influx rates and single time-point blood-brain ratios were based on equations devised by Patlack et al. (37a) and Plotkin et al. (39a), respectively. The method has been fully described elsewhere (4, 18, 23).
Unidirectional influx constants (Ki), expressed in milliliters per gram per minute, were determined from the linear portion of the curve by the equations: Am/Cp(τ) = Ki (EXP time) + VI and EXP time = [∫0tCp(t)dτ]/Cp(t), where Am is the level of radioactivity in brain at time t, EXP time is the exposure time, Vi is the apparent volume of distribution in brain, Cp is the cpm/ml of plasma at time t, and τ is time.
Amount of radioactivity in brain.
The level of radioactivity in the brain in relation to the radioactivity in the initial injection was determined by the equation: R = cpmbr / [(cpmbl)(Wtbr)], where R is the brain level of labeled IL-1β or labeled albumin expressed on a per gram basis as a brain-blood ratio, cpmbr is the level of radioactivity in a sample of brain tissue, cpmbl is the level of radioactivity in a milliliter of plasma, and Wtbr is the weight in grams of the tissue sample. The % of labeled IL-1β in the brain was calculated by dividing the cpmbr per region by the total amount of radioactivity in the brain.
The amount of labeled IL-1β in the brain after correction for vascular content and leakage across the BBB (cpmbr−c) was determined by cpmbr−c = (RIL − RAlb) (Wtbr) (cpmbl−IL), where RIL is the brain-blood ratio for labled IL-1β, RAlb is the brain-blood ratio for labeled albumin, and cpmbl−IL is the level of radioactivity in a milliliter of plasma after intravenous injection of labeled IL-1β. The corrected labeled IL-1β levels can then be expressed as a percentage of total radioactivity injected by % of injection/region = 100 [(cpmbr−c)/(cpminj)], where cpminj is the total number of cpm of labeled IL-1β injected per animal.
Fifty microliters of plasma were diluted in 200 μl TCA to yield a 25% TCA concentration. The mixture was incubated on ice for 30 min and then centrifuged at 14,000 rpm for 15 min at 4°C. The supernatant was collected, and the pellet was rinsed with acetone. Both supernatant and pellet were counted separately in the gamma counter and summed to give total radioactive counts. Brains were homogenized and subjected to the same precipitation. The fraction of 125I label that precipitates with acid was plotted over time. The amount of radioactivity that precipitated was used as an indication of how much of the radioactive label was protein-bound (intact, labeled IL-1β). The rate of degradation was calculated by the slope of the line.
Tb was recorded at 1-min intervals throughout the experiment, averaged over 5 min, and analyzed using repeated-measures ANOVA. GraphPad Prism statistical software was used to estimate the linearity of the regression correlation. The slope of the regression line between tissue-plasma ratio and exposure time represents Ki (ml/g/min), the unidirectional influx rate reflecting blood-to-brain entry. Values for Ki were compared by ANOVA. Significance was set at P < 0.05. Transport was considered significant when the influx rate was significantly different from time 0 and significantly different from BSA. All data were analyzed for sex differences. None of the young female rats were in the estrous stage on the day of injection, and there were no differences between male and female rats in either age group. Therefore, the data from males and females were combined.
Tb After Labeled IL-1β at Ta 21 and 31°C
Young rats got similar fevers at both Ta 21 and 31°C, whereas old rats developed fever only at the warmer Ta (Fig. 1). Baseline Tb was similar across age and Ta (Fig. 1, insets). Tb responses of young and old rats to radiolabeled IL-1β were similar to those of rats injected with unlabeled IL-1β (data not shown).
Influx of IL-1β Into the Brain at Ta 21°C
At this lower Ta, young rats showed significant influx of IL-1β into the brain (Fig. 2, P < 0.001, compared with time 0 and BSA). BSA did not enter the brain in significant amounts in either age group. Old rats showed reduced influx of IL-1β into the brain compared with young rats (Fig. 2A, P < 0.01). Indeed, influx of IL-1β into the whole brain of old rats was not significantly different from time 0, nor did it differ from BSA.
Figure 2B shows influx rates (Ki) for different brain regions in young and old rats. Young rats showed significant influx into three brain regions (hypothalamus, P < 0.01, hippocampus, P < 0.05, and cerebellum, P < 0.05), whereas the cortex, pons-medulla, and midbrain showed no transport. Old rats showed no statistically significant influx into any brain region. Overall, age had a significant effect (P < 0.001).
Influx of IL-1β Into the Brain at Ta 31°C
IL-1β influx rates into whole brain were similar for young and old rats at this Ta. Figure 3A shows influx of IL-1β and BSA for young and old rats at Ta 31°C. Both groups showed significant influx of IL-1β into the brain (young, P < 0.01 and old, P < 0.01). BSA did not enter the brain in significant amounts in either age group.
Figure 3B shows the influx rates for the different brain regions for young and old rats at Ta 31°C. Young rats showed significant influx into two brain regions: the hypothalamus (P < 0.01) and the hippocampus (P < 0.01). Old rats showed significant influx into four brain regions: the hypothalamus (P < 0.05), the hippocampus (P < 0.01), the cerebellum (P < 0.05), and the pons-medulla (P < 0.05). Despite showing significant influx into more regions than the young rats, there were no differences between age groups at this Ta.
Ta 21 vs. 31°C
Whole brain influx rates for young rats at Ta 31°C were similar to those for young rats at Ta 21°C. Regionally, there were no differences in influx rates between the two Tas. Old rats at Ta 31°C showed much greater influx of IL-1β into whole brain (P < 0.001), hypothalamus (P < 0.001), hippocampus (P < 0.001), and pons-medulla (P < 0.001) than did old rats at Ta 21°C.
Percent of Labeled IL-1β in Brain Tissue
The percent of labeled IL-1β per gram of brain tissue is shown in Table 1. Means represent the average of percentages collapsed across time (from 10 to 50 min). The percent of labeled IL-1β in the brains of old rats at Ta 21°C was consistently low over time and lower than the percentage of IL-1β in all other groups (P < 0.05).
The amount of radioactivity that precipitated was used as an indication of how much of the radioactive label was protein-bound (intact, labeled IL-1β). The rate of degradation was calculated by the slope of the line. Figure 4 shows the degradation of label from IL-1β in the blood of young and old rats at Ta 21 and 31°C. The degradation of labeled IL-1β in the plasma of old rats at 21°C was much slower than in old rats at 31°C or young rats at either Ta (P < 0.05).
We can consider these experiments to be made up of four groups of rats: young and old at 21°C, and young and old at 31°C. If we do this, we find that three of these groups are indistinguishable from each other; the old rats at Ta 21°C are the odd group. The present results suggest that the difference in the fever response of young and old rats at the two Tas could be attributable to the amount of IL-1β that reaches the brain and/or the rate at which it enters the brain. Young rats at both Tas and old rats at Ta 31°C showed comparable influx of 125IL-1β into the brain. Only the old rats at Ta 21°C had significantly reduced influx compared with the other three groups. Indeed, this latter group showed no significant influx into any brain region. Thus, a cool Ta is a sufficient reason for lack of fever in the old rats, a potential mechanism being a reduced influx of IL-1β into the brain and a concomitant lack of fever response.
There are several possible explanations for these results. First, reduced influx of IL-1β into the brains of old rats may be a result of an increase in the production of the anti-inflammatory cytokine IL-1ra. IL-1β transport is related to the transport system for IL-1ra, which inhibits the transport of IL-1β into the brain (17). Levels of IL-1ra in response to peripheral administration of a pyrogen have not been measured in old rats, but possibly this cytokine is increased in old rats and inhibits the passage of IL-1β into the brain. Increased peripheral levels of IL-1ra could also inhibit IL-1β-induced production of other inflammatory cytokines in the periphery, thereby decreasing the pyrogenic signal to the brain.
A second possible explanation concerns changes in cerebral blood flow. Transport of IL-1 from blood to brain is most likely accomplished by a subset of capillaries in the microvascular bed (34). Decreased cerebral blood flow would mean less IL-1β present in the blood supply to the brain. The brain uses about 15% of cardiac output and accounts for 20–25% of whole body oxygen consumption. Previously, we demonstrated that oxygen consumption decreased about 15% in old rats after intraperitoneal LPS at 21°C (11). After LPS at Ta 31°C, old rats showed an increase in oxygen consumption accompanied by a fever that was similar to that of young rats. Since decreases in oxygen consumption are matched with decreases in cardiac output (7), the drop in oxygen consumption at Ta 21°C is likely to be accompanied by a decrease in cerebral blood flow. If so, it may contribute to the decrease in IL-1β influx into the brains of old rats. At Ta 31°C, the increase in oxygen consumption during fever may increase cerebral blood flow, leading to higher levels of IL-1β available for transport into the brain.
A third possibility is that the warmer Ta makes the BBB “leakier”. This is an appealing hypothesis since there are several studies showing that heat stress (Ta 40–45°C) alters the BBB such that relatively large molecules can enter the brain (44). However, exposure to such high temperatures is likely to cause irreparable damage to the BBB, whereas a Ta of 31°C is not (43). If Ta 31°C did significantly disrupt the BBB, we would expect to see increased influx of not only IL-1β, but BSA as well. We did not see this. There was no difference in BSA influx between the two Tas in young and old rats.
The lack of influx of BSA into the brain at Ta 31°C does not exclude the possibility that a more subtle form of “leakiness” is occurring at the BBB. Albumin is a very large molecule (roughly 66 kDa vs.17 kDa for IL-1β), and it is possible that at a warmer Ta, the BBB becomes more permeable to the passage of smaller molecules that are excluded from the brain at standard temperatures. To test this, it would be necessary to measure the passage of a substance that is much smaller than albumin. However, even if an increase in Ta did change the permeability of the BBB to small molecules, it would not necessarily mean that the transport system for IL-1 was also changed. Whether or not one sees age-related changes in BBB transport seems to depend entirely on which protein is being studied (3, 25, 45, 46, 48). It may be that the BBB in its entirety does not change with age and that the deficits are limited to specific transport systems.
McLay et al. (34a) also found reduced influx of IL-1β into the brains of aged mice at Ta 23°C and suggested that this might be due to increased levels of endogenous IL-1β already in the brains of old rodents. Because IL-1β is part of a saturable transport system, elevated levels of endogenous brain IL-1β would greatly reduce transport into the brain. Excess levels of IL-1β in the hippocampus of aged rats have been correlated with impaired LTP and neuronal death (31, 32). The addition of more IL-1β to the old brain through influx from the periphery would likely cause an even greater impairment. Viewed in this light, reduced influx of IL-1β into the brains of old rats would be a protective mechanism to keep brain levels of IL-1β as low as possible. Endogenous levels of IL-1β have not been measured when rats are maintained at a warmer Ta. Possibly, levels of IL-1β in the aged brain are lower at a higher Ta. If this were the case, the protective mechanism of reduced IL-1β influx would no longer be necessary.
In addition to cytokines crossing and/or acting at the level of the BBB, there is another pathway of fever in which peripheral cytokines bind to the vagus nerve, activating a neural pathway of fever. The role of the vagus in the generation of fever in aged rats has not been studied. However, in young animals, the role of the vagus in fever seems to depend greatly on dose and route of administration (21, 41). Severing the vagus nerve at the subdiaphragmatic level attenuates fever in response to low doses of peripherally administered LPS or IL-1β (16, 20, 21, 51) but fails to block fever when administered at higher doses (19, 21, 41). Although the vagus probably does not play a major part in fever at the dose and route of administration used here, there is evidence that Ta may influence the febrile response of vagotomized rats. Romanovsky et al. (41) found that differences between sham-operated and vagotomized young rats given LPS at Ta 25°C were eliminated when LPS was administered at Ta 31°C.
The beneficial aspects of a warm Ta are not limited to increased fever responses in old rats. The survival rate of herpes- and rabies-infected mice increased when mice were maintained at a warm Ta (2, 8). Mice infected with the protozoan parasite Trypanosoma cruzi demonstrate increased longevity, decreased parasitemia, and enhanced immune responses when maintained at 36°C (1, 12, 36). Survival of mice infected with Streptococcus pneumoniae increased as Ta increased from 19 to 34°C (29). Hoffman-Goetz and Keir (21a) found the survival rate of young and old mice given high doses of LPS was much higher for mice maintained at 30°C than for those maintained at 24°C. The effect was true in both young and old mice but was more pronounced in the aged mice.
Taken together, these studies support the idea that housing animals at a warm Ta can increase the ability of the immune system to fight off a pathogen. However, they give no indication of how the increased Ta facilitates this. It is generally accepted that the immune response is enhanced at febrile temperatures (13, 40), but simply maintaining an animal at a moderately warm Ta does not raise Tb on its own (11, 38). There is no reason to suppose that a relatively mild increase in Ta would cause a physiological change that may predispose the aged animal to develop fever and mount an effective immune response (26, 49). Whatever the mechanism, the phenomenon of a warmer Ta facilitating immune responses has the potential to profoundly impact treatment of infections in elderly humans. If a warm Ta can help humans fight infection, as it appears to do in other species, then this would be a relatively easy and economical means of aiding the immune response to potentially lethal infections.
Research was supported by National Institute of Mental Health Grant RO1-MH41138 to E. Satinoff.
Current address for J. B. Buchanan, University of Illinois Urbana-Champaign, 7 Animal Sciences Laboratory, 1207 W. Gregory Dr., Urbana, IL 61801 (e-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society