- Email: [email protected]
Stress hyperthermia: Physiological arguments that it is a fever
Physiology& Behavior,Vol. 49, pp. 1153-1157. ¢ Pergamon Press plc, 1991. Printed in the U.S.A.
0031-9384/91 $3.00 + .00
Stress Hyperthermia: Physiological Arguments That It Is a Fever E. B R I E S E 1 A N D M. C A B A N A C 2 Department of Physiology, Faculty of Medicine, Laval University, Qudbec G1K 7P4, Canada R e c e i v e d 16 July 1990 BRIESE, E. AND M. CABANAC. Stress hyperthermia: Physiological arguments that it is a fever. PHYSIOL BEHAV 49(6) 1153-1157, 1991.--The theory that stress (or emotional) rise in central temperature (To) in rats is a fever with an upward shift of the set-point temperature was tested with three experiments: 1) Measurement of tail skin temperature and T c during the emotional Tc rise; 2) Investigation of the effect of ambient temperature on the emotional Tc rise; and 3) The assessment of emotional Tc rise during daytime and nighttime. Skin vasomotor responses helped the increase of Tc toward a higher level and contributed to the regulation of central temperature at this new higher level. The cold environment did not diminish the emotional rise of central temperature as it would be expected in the case of a hyperthermia. However, at night emotional fever reached a higher level than during the daytime, suggesting that prostaglandin rise in T c is distinct from emotional or stress-induced hyperthermia. In conclusion, the experiments reported here confirm the hypothesis that the rise of T¢ induced by handling or disturbance of the rats is regulated, and is due to a shift of the set-point as occurs in fever. Hyperthermia
Fever
Set-point
Temperature regulation
GENTLE handling of rats, including measuring colonic temperature, moving the animals to a new environment, or disturbing them even without handling, produces a substantial increase in their central temperature (T¢) (2, 5, 22, 26, 27). This phenomenon is important because it occurs each time a rat is approached or manipulated: it contaminates all experimental work on nonanesthetized rats, which reminds one of Heisenberg's indeterminacy principle (17). Therefore, the phenomenon of the rise of T c by gentle handling of the animal must be taken into consideration when T c is measured. An important issue concerning this phenomenon is to know whether we are dealing with hyperthermia or fever. Hyperthermia, defined as a rise in central temperature above the set-point temperature (Tset), is simply due to an increase of the heat load 00,12). Thermoregulatory responses available are recruited in order to dissipate the extra heat and bring body temperature down to the lower level of Ts~r On the other hand, a rise in central temperature must be defined as a fever if the concomitant thermoregulatory responses are those the organism uses to increase the heat content of the body, which results in the rise of body temperature. The thermogenic and thermoconservative responses indicate then that the set-point has been shifted to a higher level and that the central temperature is now regulated at this higher level. Those who have addressed this issue believe that the rise in central temperature of rats by handling or disturbing the animals is a fever because sodium salicylate and indomethacin, two antipyretics, partially prevent the emotional rise of central temperature (4, 13, 22). It is generally assumed that
Emotion
Stress
Handling
Rat
antipyretics have no effect on normal body temperature (7, 19, 20, 25) or on hyperthermic elevation of T c (1, 8, 25) and that they only lower central temperature in febrile subjects. Consequently, the effect of these drugs on the emotional rise of Tc is given as evidence that the rise of T¢ is due to an upward shift of Tset, and is therefore a fever. The syllogism is as follows: if salicylate lowers T c only in febrile subjects and if the handlingproduced rise in T¢ is blocked by salicylate, then the handlingproduced rise in T~ is a fever. Yet the evidence is indirect. Another indirect argument, in favor of fever rather than hyperthermia, is the observation that antiserum against tumor necrosis factor increased the body temperature of rats subjected to mild stress in the same way as it does in lipopolysaccharide fever (15). Finally, when rats are housed at a low ambient temperature their stress-induced rise in body temperature does not differ from that seen when the animals are kept at a thermoneutral ambient temperature (14). Nevertheless, there are some arguments against the hypothesis of a resetting. Salicylate does lower the normal body temperature of rats (11,21). In addition, drugs other than prostaglandin antagonists block the emotional rise of T c. Naloxone, a specific narcotic antagonist, blocks the hyperthermia induced by handling (2, 18, 23) and it is not known whether the T~ elevation induced by endogenous opioids (16) is a fever or a hyperthermia. In the present work we explored the fever/hyperthermia question by three physiological experiments investigating: 1) the vasomotor response in the rat's tail during handling, 2) the influence of the ambient temperature, and 3) the influence of the circadian
1Invited Professor from the Department of Physiology, University of Los Andes, Apartado 109, M6rida 5101-A, Venezuela. 2Requests for reprints should be addressed to: M. Cabanac at the above address, EmaiI:CABANAC@LAVALVM1.
1153
1154
BRIESE AND ('ABANAC
cycle. The reasoning for these experiments is as follows. l) Vasomotor response: If the rise in T c when the rats are disturbed is a fever, then during the ascendent phase of the T c rise the coordinated vasomotor response should be a vasoconstriction, a thermoregulatory response aiming at heat conservation. On the other hand if the elevation of Tc is a hyperthermia the vasomotor response, as soon as the central temperature starts to rise, should be a vasodilatation. 2) Ambient temperature: If disturbing the rats produces an upward shift of their set-point temperature, as in fever, then the rise in T c must be independent of the ambient temperature. On the other hand, hyperthermia should be less pronounced in a cold than in a warm environment because a cold environment would increase heat loss without triggering the thermogenic and thermoconservative responses. 3) Circadian variations: Intracerebroventricular injections of prostaglandins of the E-series, substances mediating fever (9,24), increased T~ to a dose-dependent maximum which was independent of the circadian phase and of the initial T~. Only the change in T~, quantified by the difference between the highest temperature, reached and the initial one, was related to the initial T~. It was concluded (9) that the characteristic attribute of fever is that the regulated variable is the maximal temperature (peak or plateau) and not the difference between the upper level reached and the initial temperature. If this is true, and the emotional rise is a fever, then Tc should reach about the same plateau value in daytime as in nighttime. METHOD
Animals Twenty-seven male Wistar rats from Charles River Inc. Canada weighing between 385 and 425 g were used for the experiments. The rats were kept in animal quarters at a temperature of 24°C in individual standard cages. Food and water were available ad lib. Lights went on automatically at 0600 h, and off at 1800 h.
Su~e~ Seven rats were anesthetized with an intramuscular injection of ketamine hydrochloride (87 mg.kg -1) and xylazine hydrochloride (12.9 mg.kg-1). An incision was made on the midline of the cranium and the skull was exposed. Four 0-80 x 1/6" stainless steel screws were fixed to the skull and a trepanation of 1 mm in diameter was made 2 mm behind bregma, 1 mm lateral to the midline. An insulated copper-constantan thermocouple fastened into a 22 ga stainless steel tube, close welded at its distal end, was introduced through the trepanation into the brain at a depth of 7 mm. The thermocouple wire ended at the surface of the skull into male copper and constantan connectors fixed to the skull with acrylic cement.
Temperature Measurements Colonic temperature was measured with a copper-constantan thermocouple. The thermocouple wires were passed through a polyethylene 205 tube which was then filled with silicone rubber. The distal end of the probe was sealed and smoothed with heat conducting acrylic cement. Colonic temperature was measured at a depth of 70 ram. The rat was held at the proximal 1/3 of the tail; the lubricated probe was introduced into the colon, maintained in place until the reading was stable, and then removed. Intracranial temperature was measured with a light and flexible copper-constantan cable ending into female copper and
constantan connectors. At the time of the test session the female connectors were plugged into the male connectors on the rat~ head. The flexible thermocouple cable was loosely suspended t(~ a holder with a thin rubber band. The colonic or the intracranial thermocouple was connected to a digital thermometer (Bailey. BAT8). Skin temperature (Wtai0 was obtained from direct reading on an infrared thermometer for surface thermometry (Everest Interscience). This instrument concentrates the infrared radiations emitted by the point measured with a parabolic mirror, processes the captured energy and displays the surface temperature digitally. The exact locus whose temperature is read is indicated by a light spot emitted by the instrument. Surface temperature was measured at the base of the tail 20 mm distal from the limit between the hairy and glabrous skin. Small variations in the location of the point along the tail length where the temperature is measured can result in important variations in the measured temperature. Therefore, the precise point where Tta~ was measured was marked for each rat with indelible ink and the experimenter aimed the infrared thermometer's light spot on this mark.
Vasomotor Response Experiments The seven rats with implants were used in this experiment. At the time of the test session the rat was removed from its cage and placed in an open plastic transparent cage (0.18 m wide, 0.28 m long, 0.14 m high) on a table. Ttai|, and brain temperatures were measured every minute over 40 min.
Ambient Temperature In seven other rats the colonic temperature was taken every 10 rain for a total of 40 min, in the morning at 1000 h on separate days at 3 ambient temperatures of about 24, 19, and 10°C.
Circadian Variations Thirteen rats were used in this experiment. At the time of the test session each rat was taken out of its cage and placed on a table under an inverted standard wire mesh cage (0.18 m wide, 0.24 m long, 0.18 m high). Colonic temperature was measured every 9 min for 36 min. The sessions took place once in the morning at 0900 h, and once in the night at 2230 h.
Emotional Stress The stress situation was created by removing the rat from its home cage, placing it on a table, either under an inverted cage or into an open plastic transparent cage and having the animal subjected to the necessary manipulation for obtaining skin, brain, or colonic temperature. RESULTS
Experiment 1: Vasomotor Responses The results of the experiments when Tt, n was measured simultaneously with brain temperature are summarized in Fig. 1. Each point in Fig. 1 represents the mean value of the measurements taken in 7 rats. The curves of T¢ and TuaI have quite different shapes. T c tends to rise immediately and reaches a plateau value at minute 20. Three segments can be identified in the time course of Tual: 1) up to minute 12 T,,ii diminishes, indicating a vasoconstriction; 2) from minute 12 to minute 31 Tt~l rises sharply (vasodilatation); for the last 9 min the curve of Tmi ! is stable at about 31.5°(2. An important feature in Fig. 1 is the change of inflexion of the Tt~ curve which occurred between rain 9 and 12. This indicates the moment when vasoconstriction
EMOTIONAL FEVER
1155
39
38.5 A
E
ILl
W
n-
--z
38
3,3 tr uJ
o.
6" 0.....
37.5
32
rt"
N 37
QC ILl
31 m
W
k.-
n,"
z < rr an
W
I,IJ
n" LU a. z II
36.5
I.,-J
26:~
t~l
36 _J
27 ~
n 0
26
o 5
u 10
I 15
I 20
I 25
I 30
I 35
I 40
24
TIM E (rain)
25 0
10
20 MINUTES
30
40
FIG. 1. Mean surface tail temperature (bottom curve) and brain temperature (upper curve) in seven rats. In the morning, each rat was taken out of its cage, placed in an open transparent cage and the temperatures were measured every minute. Vertical line at rain 12 indicates the inflexion change of tail temperature curve when central temperature theoretically reached the set-point temperature.
changed to vasodilatation. This point is important because it may indicate the moment when T~ reached Tser The time course of the temperature elevation was not identical in all rats. As a result the mean curve of Fig. 1 is somewhat buffered. Figure 2 reproduces the results from a typical experiment with one rat. In this experiment the initial vasoconstriction was small but obvious. Vasoconstriction changed to vasodilatation after min 12 as indicated by the rise of Ttail. T~I and T c tended to oscillate in opposition of phase. The dashed lines in Fig. 2 indicate inflexion changes in Ttail. There are eight segments delimited by these dashed lines and in six of them when one temperature goes one way the other temperature goes in the opposite direction.
FIG. 2. Time course of surface tail temperature (bottom) and brain temperature (upper curve) of one rat during a 40-min session when the animal was taken out of its cage and placed in a plastic open cage. Dashed vertical lines indicate inflexion changes in tail-skin temperature. Arrows show the respective directions of T c and Trail. The dashed lines define eight intervals. In six of them T c and Tt~i~ go in opposite directions. In the last two, one temperature remains stable while the other goes up. p < 0 . 0 2 ) . The differences between daytime and nighttime coIonic temperature at min 18, 27, and 36 were assessed with a two-way ANOVA. The difference between the two plateaus was significant, F ( 1 , 7 2 ) = 17.6, p = 0 . 0 0 0 1 . DISCUSSION
Our results suggest that the emotional stress rise of T c in rats has the basic characteristics of a fever, i.e., it is associated with an upward shift of the Ts~t of the thermoregulatory system. The arguments are as follows. First, the skin vasomotor responses were coordinated so as to help with the Tc changes. Simultaneously with the rapid initial rise of T c, Tt=l decreased indicating vasoconstriction, a response that limits heat-loss and facilitates the rise of To. After an average time of 12 min the vasoconstriction gave way to vasodilatation, a heat dissipation response.
Experiment H: Ambient Temperature The results of this experiment are given in Table 1 and Fig. 3. The mean maximal T c reached by the seven rats was almost the same when the measurements were done at 19 and 24°C. The maximal mean T~ was 0.3°C higher at 10°C than at 19°(2. This difference was not significant according to a one-way ANOVA, F ( 2 , 1 2 ) = 3 . 6 6 , p = 0 . 0 5 7 5 . Differences between the plateaus shown in Fig. 3 considering the data from min 20, 30, and 40 were calculated by a two-way A N O V A and were not significant: F1(2,18 ) = 1.065, p = 0 . 3 6 5 4 ; F 2 ( 2 , 3 6 ) = 1.033, p=0.3663.
40
Ta: 10°C
Ta: 19°C
Ta: 24°C
E W
n-
39
i=1 n
:IE W
F-
38
z
9 Experiment 111: Circadian Cycle The results of the effect of time of day, with correspondingly different initial T¢, on the emotional stress in T~ is given in Fig. 4. Mean initial temperature in the morning was 37.3°(2 and at night 38.5°C and this difference was significant ( t = 11.79, p < 0 . 0 0 1 ) . Mean maximal temperature reached in the morning was 39.4°(2 and at night 39.70(2, a difference of 0.30(2 ( t = 3 . 8 9 ,
8 37
,
,
,
,
i
,
TIME
n
I
u
u
I
I
!
I
I
(10 min intervals)
FIG. 3. Mean colonic temperatures of seven rats and standard errors of the means. The temperatures were measured at lO-min intervals during the sessions of 40 rain at ambient temperatures of I0, 19, and 24°C.
1156
BRIESE ANI) ( AI~ANA("
TABLE I MEANS OF MAXIMALCOLONIC TEMPERATURES (°C) REACHED BY SEVEN RATS AT DISTINCT AMBIENT TEMPERATURES (°C~ Ambient temperature Mean colonic temperature Standard error
10 39.84 0.08
19 39.54 0.13
24 39.58 0.12
This should indicate that at this moment T~ had reached Tset (at least theoretically). From min 12 up to about min 31 Trail continued to increase while the ascending course of T~ gradually ended in a plateau at min 20. If the course of vasomotor responses is a criterion of the level of T~ relative to Tset, then after min 12 T~ was above Tset until about rain 31 when both T c and tail temperature stabilized. At first glance, from rain 12 to min 31 the animals appeared to be hyperthermic. However, nothing in the shape of T~ curve seems to constitute a signal for the sudden change from tail vasoconstriction to vasodilatation at min 12. This means that something else controls the time course of both curves, most likely the set-point mechanism. We speculate that the heat-dissipation response was triggered at min 12 because T c had overpassed Tset. The command to start and stop heat dissipation should be produced by a neural regulator which compares T~ and Tset. T~ integrates heat loss and heat gain. In
40
• Sn,o
LU
a:
39 Ill at. tU I'-0
38
8 37
I
I
I
I TIME
I (9 min
I
I
I
I
the present experiments we had only a window observation oi the time course of the heat dissipation response. From the strong vasoconstriction of the first 12 min we may deduce that heat production was insufficient although the animals were often seen shivering. From rain 12 to 31 heat gain was excessive since in.creasing vasodilatation was necessary to damp down the T,, increase and to maintain T~ at a stable level. Could this thermal state be qualified " h y p e r t h e r m i c " ? Probably not, because the oscillating Tt~ indicates that a regulato~ process was taking place. A new equilibrium was reached after about rain 3t up to the end of the experimental sessions. During this phase the stabilization of both Tta n and T~ as well as the reciprocal oscillations of both temperatures, as illustrated in Fig. 2, indicate that T, was regulated at this new higher level. The very existence of reciprocal oscillations is a strong argument in favor of regulation (6). A second physiological argument that the emotional rise in T~ is a fever is that the ambient temperature did not affect the elevation of T~. Most important is the fact that the cold environment did not diminish the T~ rise as would have occurred if the rise of T~ rise was a hyperthermia. Similar results were recently published (14). A controversial point is related to the results of the day/night experiment. The emotional rise of T~ reached a small (0.3°C) but significantly higher plateau during the night than during the day. It has been shown (9) that intracerebroventricular injections of prostaglandin induce a rise of T¢ regulated at the peak level which depends on the dose injected and is independent of the initial T~ and therefore of the time of day. Our results suggest that some part of emotional fever is not regulated according to the same paradigm. We conclude that the emotional elevation of T~ is similar to fever because the vasomotor responses contributed to increase T~ and to adjust and to maintain it at a new stable or oscillatory higher level. Another physiological argument in favor of fever is that the elevation of T~ due to handling was not diminished by a cool environment. The results suggest that thermogenic, thermoconservatory, and heat dissipation responses interplay during the emotional fever so as to raise T~ to a higher level. This further suggests that T~, is also shifted to a higher level.
I
ACKNOWLEDGEMENTS
intervals)
FIG. 4. Mean colonic temperatures and standard errors of the means of 13 rats. Measurements were taken every 9 min in sessions of 36-rain duration, once in the morning at 1000 h and once at night at 2230 h.
We wish to thank E. Satinoff for critically reading this manuscript. This work was supported by the Conseil de la Recherche en Sciences Naturelles et en GEnie (N. S. E. R. C. Canada) and Defense and Civil Institute of Environmental Medicine (D. C. I. E. M. Canada).
REFERENCES 1. Bass, D. E.; Jacobson, E. D. Effect of salicylate on work in the heat. Fed. Proc. 22:176; 1963. 2. BRisig, J.; Hdllt, V.; B~luerle, U.; Herz, A. Involvement of endorphins in emotional hyperthermia of rats. Life Sci. 23:2525-2532; 1978. 3. Briese, E. Hyperthermia in self-stimulating rats. Acta Physiol. Latinoam. 15:357-361; 1965. 4. Briese, E.; Cabanac, M. Emotional fever and salicylate. In: Sze16nyi, Z.; SzEkely, M., eds. Contributions to thermal physiology. Oxford: Pergamon Press; 1980:161-163. 5. Briese, E.; de Quijada, M. G. Colonic temperature of rats during handling. Acta Physiol. Latinoam. 20:97-102; 1970. 6. Cabanac, M.; Russek, M. REgulation et contrrle en biologie. Qurbec: Presses Univ. Laval; 1982. 7. Cranston, W. I.; Rawlins, M. D. Effects of intracerebral microinjections of sodium salicylate on temperature regulation in the rabbit. J. Physiol. (Lond.) 222:257-266; 1972.
8. Downey, J. A.; Darling, F. C. Effect of salicylates on elevation of body temperature during exercise. J. Appl. Physiol. 17:323-325; 1962. 9. Feng, J.; Price, M.; Cohen, J.; Satinoff, E. Prostaglandin fevers in rats: regulated change in body temperature or change in regulated body temperature? Am. J. Physiol. 257:R695-R699; 1989. 10. Gordon, C. J. A review of terms for regulated vs. forced neurochemical-induced changes in body temperature. Life Sci. 32:12851295; 1983. 11. Green, M. D.; Lomax, P. The effects of sodium salicylate on body temperature in the normothermic rat. Proc. West. Pharmacol. Soc. 16:257-261; 1973. 12. I.U.P.S. Commission for Thermal Physiology. Glossary of terms for thermal physiology. Pfltlgers Arch. 410:567-587; 1987. 13. Kluger, M. J.; O'Reilly, B.; Shope, T. R.; Vander, A. J. Further evidence that stress hypertherrrtia is a fever. Physiol. Behav. 39: 763-766; 1987.
EMOTIONAL FEVER
14. Long, N. C.; Vander, A. J.; Kluger, M. J. Stress-induced rise of body temperature in rats is the same in warm and cool environment. Physiol. Behav. 47:773-775; 1990. 15. Long, N. C.; Vander, A. J.; Kunkel, S. L.; Kluger, M. J. Antiserum against tumor necrosis factor increases stress hyperthermia in rats. Am. J. Physiol. 258:R591-R595; 1990. 16. Martin, G. E.; Bacino, C. B. Action of intracerebrally injected 13-endorphin on the rat's core temperature. Eur. J. Pharmacol. 59: 227-236; 1979. 17. Miles, G. H. Telemetering techniques for periodicity studies. Ann. NY Acad. Sci. 98:858-865; 1962. 18. Pae, Y. S. Hyperthermia in the rat from handling stress blocked by naloxone injected into the preoptic-anterior hypothalamus. Pharmacol. Biochem. Behav. 22:337-339; 1985. 19. Rawlins, M. D.; Rosendorff, C.; Cranston, W. I. The mechanism of action of antipyretics, In: Wolstenholme, G. E. W.; Birch, J., eds. Pyrogens and fever. Ciba Foundation Symposium. London, UK: Churchill Livingstone; 1971:175-191. 20. Rosendorff, C.; Cranston, W. I. Effects of salicylate on human temperature regulation. Clin. Sci. 35:81-91; 1968.
1157
21. Satinoff, E. Salicylate: Action on normal body temperature in rats. Science 176:532-533; 1972. 22. Singer, R.; Harker, C. T.; Vander, A. J.; Kluger, M. J. Hyperthermia induced by open-field stress is blocked by salicylate. Physiol. Behav. 36:1179-1182; 1986. 23. Stewart, J.; Eikelboom, R. Stress masks the hyperthermic effect of naloxone in rats. Life Sci. 25:1165-1172; 1979. 24. Stitt, J. T. Prostaglandin E as the neural mediator of the febrile response. Yale J. Biol. Med. 59:137-149; 1986. 25. Techtermann, F.; J6quier, E. Etude calorim6trique de l'effet du salicylate chez l'homme lors de l'hyperthennie induite par exposition au chaud ou par exercice musculaire. J. Physiol. (Pads) 74:615-621; 1978. 26, Vidal, C.; Suaudeau, C.; Jacob, J. Hyper- and hypothermia induced by non-noxious stress: effects of naloxone, diazepam and gammaacetylenic GABA. Life Sci. 33:587-590; 1983. 27, York, J. L.; Regan, S. G. Conditioned and un-conditioned influences on body temperature and ethanol hypothermia in laboratory rats. Pharmacol. Biochem. Behav. 17:119-124; 1982.
0031-9384/91 $3.00 + .00
Stress Hyperthermia: Physiological Arguments That It Is a Fever E. B R I E S E 1 A N D M. C A B A N A C 2 Department of Physiology, Faculty of Medicine, Laval University, Qudbec G1K 7P4, Canada R e c e i v e d 16 July 1990 BRIESE, E. AND M. CABANAC. Stress hyperthermia: Physiological arguments that it is a fever. PHYSIOL BEHAV 49(6) 1153-1157, 1991.--The theory that stress (or emotional) rise in central temperature (To) in rats is a fever with an upward shift of the set-point temperature was tested with three experiments: 1) Measurement of tail skin temperature and T c during the emotional Tc rise; 2) Investigation of the effect of ambient temperature on the emotional Tc rise; and 3) The assessment of emotional Tc rise during daytime and nighttime. Skin vasomotor responses helped the increase of Tc toward a higher level and contributed to the regulation of central temperature at this new higher level. The cold environment did not diminish the emotional rise of central temperature as it would be expected in the case of a hyperthermia. However, at night emotional fever reached a higher level than during the daytime, suggesting that prostaglandin rise in T c is distinct from emotional or stress-induced hyperthermia. In conclusion, the experiments reported here confirm the hypothesis that the rise of T¢ induced by handling or disturbance of the rats is regulated, and is due to a shift of the set-point as occurs in fever. Hyperthermia
Fever
Set-point
Temperature regulation
GENTLE handling of rats, including measuring colonic temperature, moving the animals to a new environment, or disturbing them even without handling, produces a substantial increase in their central temperature (T¢) (2, 5, 22, 26, 27). This phenomenon is important because it occurs each time a rat is approached or manipulated: it contaminates all experimental work on nonanesthetized rats, which reminds one of Heisenberg's indeterminacy principle (17). Therefore, the phenomenon of the rise of T c by gentle handling of the animal must be taken into consideration when T c is measured. An important issue concerning this phenomenon is to know whether we are dealing with hyperthermia or fever. Hyperthermia, defined as a rise in central temperature above the set-point temperature (Tset), is simply due to an increase of the heat load 00,12). Thermoregulatory responses available are recruited in order to dissipate the extra heat and bring body temperature down to the lower level of Ts~r On the other hand, a rise in central temperature must be defined as a fever if the concomitant thermoregulatory responses are those the organism uses to increase the heat content of the body, which results in the rise of body temperature. The thermogenic and thermoconservative responses indicate then that the set-point has been shifted to a higher level and that the central temperature is now regulated at this higher level. Those who have addressed this issue believe that the rise in central temperature of rats by handling or disturbing the animals is a fever because sodium salicylate and indomethacin, two antipyretics, partially prevent the emotional rise of central temperature (4, 13, 22). It is generally assumed that
Emotion
Stress
Handling
Rat
antipyretics have no effect on normal body temperature (7, 19, 20, 25) or on hyperthermic elevation of T c (1, 8, 25) and that they only lower central temperature in febrile subjects. Consequently, the effect of these drugs on the emotional rise of Tc is given as evidence that the rise of T¢ is due to an upward shift of Tset, and is therefore a fever. The syllogism is as follows: if salicylate lowers T c only in febrile subjects and if the handlingproduced rise in T¢ is blocked by salicylate, then the handlingproduced rise in T~ is a fever. Yet the evidence is indirect. Another indirect argument, in favor of fever rather than hyperthermia, is the observation that antiserum against tumor necrosis factor increased the body temperature of rats subjected to mild stress in the same way as it does in lipopolysaccharide fever (15). Finally, when rats are housed at a low ambient temperature their stress-induced rise in body temperature does not differ from that seen when the animals are kept at a thermoneutral ambient temperature (14). Nevertheless, there are some arguments against the hypothesis of a resetting. Salicylate does lower the normal body temperature of rats (11,21). In addition, drugs other than prostaglandin antagonists block the emotional rise of T c. Naloxone, a specific narcotic antagonist, blocks the hyperthermia induced by handling (2, 18, 23) and it is not known whether the T~ elevation induced by endogenous opioids (16) is a fever or a hyperthermia. In the present work we explored the fever/hyperthermia question by three physiological experiments investigating: 1) the vasomotor response in the rat's tail during handling, 2) the influence of the ambient temperature, and 3) the influence of the circadian
1Invited Professor from the Department of Physiology, University of Los Andes, Apartado 109, M6rida 5101-A, Venezuela. 2Requests for reprints should be addressed to: M. Cabanac at the above address, EmaiI:CABANAC@LAVALVM1.
1153
1154
BRIESE AND ('ABANAC
cycle. The reasoning for these experiments is as follows. l) Vasomotor response: If the rise in T c when the rats are disturbed is a fever, then during the ascendent phase of the T c rise the coordinated vasomotor response should be a vasoconstriction, a thermoregulatory response aiming at heat conservation. On the other hand if the elevation of Tc is a hyperthermia the vasomotor response, as soon as the central temperature starts to rise, should be a vasodilatation. 2) Ambient temperature: If disturbing the rats produces an upward shift of their set-point temperature, as in fever, then the rise in T c must be independent of the ambient temperature. On the other hand, hyperthermia should be less pronounced in a cold than in a warm environment because a cold environment would increase heat loss without triggering the thermogenic and thermoconservative responses. 3) Circadian variations: Intracerebroventricular injections of prostaglandins of the E-series, substances mediating fever (9,24), increased T~ to a dose-dependent maximum which was independent of the circadian phase and of the initial T~. Only the change in T~, quantified by the difference between the highest temperature, reached and the initial one, was related to the initial T~. It was concluded (9) that the characteristic attribute of fever is that the regulated variable is the maximal temperature (peak or plateau) and not the difference between the upper level reached and the initial temperature. If this is true, and the emotional rise is a fever, then Tc should reach about the same plateau value in daytime as in nighttime. METHOD
Animals Twenty-seven male Wistar rats from Charles River Inc. Canada weighing between 385 and 425 g were used for the experiments. The rats were kept in animal quarters at a temperature of 24°C in individual standard cages. Food and water were available ad lib. Lights went on automatically at 0600 h, and off at 1800 h.
Su~e~ Seven rats were anesthetized with an intramuscular injection of ketamine hydrochloride (87 mg.kg -1) and xylazine hydrochloride (12.9 mg.kg-1). An incision was made on the midline of the cranium and the skull was exposed. Four 0-80 x 1/6" stainless steel screws were fixed to the skull and a trepanation of 1 mm in diameter was made 2 mm behind bregma, 1 mm lateral to the midline. An insulated copper-constantan thermocouple fastened into a 22 ga stainless steel tube, close welded at its distal end, was introduced through the trepanation into the brain at a depth of 7 mm. The thermocouple wire ended at the surface of the skull into male copper and constantan connectors fixed to the skull with acrylic cement.
Temperature Measurements Colonic temperature was measured with a copper-constantan thermocouple. The thermocouple wires were passed through a polyethylene 205 tube which was then filled with silicone rubber. The distal end of the probe was sealed and smoothed with heat conducting acrylic cement. Colonic temperature was measured at a depth of 70 ram. The rat was held at the proximal 1/3 of the tail; the lubricated probe was introduced into the colon, maintained in place until the reading was stable, and then removed. Intracranial temperature was measured with a light and flexible copper-constantan cable ending into female copper and
constantan connectors. At the time of the test session the female connectors were plugged into the male connectors on the rat~ head. The flexible thermocouple cable was loosely suspended t(~ a holder with a thin rubber band. The colonic or the intracranial thermocouple was connected to a digital thermometer (Bailey. BAT8). Skin temperature (Wtai0 was obtained from direct reading on an infrared thermometer for surface thermometry (Everest Interscience). This instrument concentrates the infrared radiations emitted by the point measured with a parabolic mirror, processes the captured energy and displays the surface temperature digitally. The exact locus whose temperature is read is indicated by a light spot emitted by the instrument. Surface temperature was measured at the base of the tail 20 mm distal from the limit between the hairy and glabrous skin. Small variations in the location of the point along the tail length where the temperature is measured can result in important variations in the measured temperature. Therefore, the precise point where Tta~ was measured was marked for each rat with indelible ink and the experimenter aimed the infrared thermometer's light spot on this mark.
Vasomotor Response Experiments The seven rats with implants were used in this experiment. At the time of the test session the rat was removed from its cage and placed in an open plastic transparent cage (0.18 m wide, 0.28 m long, 0.14 m high) on a table. Ttai|, and brain temperatures were measured every minute over 40 min.
Ambient Temperature In seven other rats the colonic temperature was taken every 10 rain for a total of 40 min, in the morning at 1000 h on separate days at 3 ambient temperatures of about 24, 19, and 10°C.
Circadian Variations Thirteen rats were used in this experiment. At the time of the test session each rat was taken out of its cage and placed on a table under an inverted standard wire mesh cage (0.18 m wide, 0.24 m long, 0.18 m high). Colonic temperature was measured every 9 min for 36 min. The sessions took place once in the morning at 0900 h, and once in the night at 2230 h.
Emotional Stress The stress situation was created by removing the rat from its home cage, placing it on a table, either under an inverted cage or into an open plastic transparent cage and having the animal subjected to the necessary manipulation for obtaining skin, brain, or colonic temperature. RESULTS
Experiment 1: Vasomotor Responses The results of the experiments when Tt, n was measured simultaneously with brain temperature are summarized in Fig. 1. Each point in Fig. 1 represents the mean value of the measurements taken in 7 rats. The curves of T¢ and TuaI have quite different shapes. T c tends to rise immediately and reaches a plateau value at minute 20. Three segments can be identified in the time course of Tual: 1) up to minute 12 T,,ii diminishes, indicating a vasoconstriction; 2) from minute 12 to minute 31 Tt~l rises sharply (vasodilatation); for the last 9 min the curve of Tmi ! is stable at about 31.5°(2. An important feature in Fig. 1 is the change of inflexion of the Tt~ curve which occurred between rain 9 and 12. This indicates the moment when vasoconstriction
EMOTIONAL FEVER
1155
39
38.5 A
E
ILl
W
n-
--z
38
3,3 tr uJ
o.
6" 0.....
37.5
32
rt"
N 37
QC ILl
31 m
W
k.-
n,"
z < rr an
W
I,IJ
n" LU a. z II
36.5
I.,-J
26:~
t~l
36 _J
27 ~
n 0
26
o 5
u 10
I 15
I 20
I 25
I 30
I 35
I 40
24
TIM E (rain)
25 0
10
20 MINUTES
30
40
FIG. 1. Mean surface tail temperature (bottom curve) and brain temperature (upper curve) in seven rats. In the morning, each rat was taken out of its cage, placed in an open transparent cage and the temperatures were measured every minute. Vertical line at rain 12 indicates the inflexion change of tail temperature curve when central temperature theoretically reached the set-point temperature.
changed to vasodilatation. This point is important because it may indicate the moment when T~ reached Tser The time course of the temperature elevation was not identical in all rats. As a result the mean curve of Fig. 1 is somewhat buffered. Figure 2 reproduces the results from a typical experiment with one rat. In this experiment the initial vasoconstriction was small but obvious. Vasoconstriction changed to vasodilatation after min 12 as indicated by the rise of Ttail. T~I and T c tended to oscillate in opposition of phase. The dashed lines in Fig. 2 indicate inflexion changes in Ttail. There are eight segments delimited by these dashed lines and in six of them when one temperature goes one way the other temperature goes in the opposite direction.
FIG. 2. Time course of surface tail temperature (bottom) and brain temperature (upper curve) of one rat during a 40-min session when the animal was taken out of its cage and placed in a plastic open cage. Dashed vertical lines indicate inflexion changes in tail-skin temperature. Arrows show the respective directions of T c and Trail. The dashed lines define eight intervals. In six of them T c and Tt~i~ go in opposite directions. In the last two, one temperature remains stable while the other goes up. p < 0 . 0 2 ) . The differences between daytime and nighttime coIonic temperature at min 18, 27, and 36 were assessed with a two-way ANOVA. The difference between the two plateaus was significant, F ( 1 , 7 2 ) = 17.6, p = 0 . 0 0 0 1 . DISCUSSION
Our results suggest that the emotional stress rise of T c in rats has the basic characteristics of a fever, i.e., it is associated with an upward shift of the Ts~t of the thermoregulatory system. The arguments are as follows. First, the skin vasomotor responses were coordinated so as to help with the Tc changes. Simultaneously with the rapid initial rise of T c, Tt=l decreased indicating vasoconstriction, a response that limits heat-loss and facilitates the rise of To. After an average time of 12 min the vasoconstriction gave way to vasodilatation, a heat dissipation response.
Experiment H: Ambient Temperature The results of this experiment are given in Table 1 and Fig. 3. The mean maximal T c reached by the seven rats was almost the same when the measurements were done at 19 and 24°C. The maximal mean T~ was 0.3°C higher at 10°C than at 19°(2. This difference was not significant according to a one-way ANOVA, F ( 2 , 1 2 ) = 3 . 6 6 , p = 0 . 0 5 7 5 . Differences between the plateaus shown in Fig. 3 considering the data from min 20, 30, and 40 were calculated by a two-way A N O V A and were not significant: F1(2,18 ) = 1.065, p = 0 . 3 6 5 4 ; F 2 ( 2 , 3 6 ) = 1.033, p=0.3663.
40
Ta: 10°C
Ta: 19°C
Ta: 24°C
E W
n-
39
i=1 n
:IE W
F-
38
z
9 Experiment 111: Circadian Cycle The results of the effect of time of day, with correspondingly different initial T¢, on the emotional stress in T~ is given in Fig. 4. Mean initial temperature in the morning was 37.3°(2 and at night 38.5°C and this difference was significant ( t = 11.79, p < 0 . 0 0 1 ) . Mean maximal temperature reached in the morning was 39.4°(2 and at night 39.70(2, a difference of 0.30(2 ( t = 3 . 8 9 ,
8 37
,
,
,
,
i
,
TIME
n
I
u
u
I
I
!
I
I
(10 min intervals)
FIG. 3. Mean colonic temperatures of seven rats and standard errors of the means. The temperatures were measured at lO-min intervals during the sessions of 40 rain at ambient temperatures of I0, 19, and 24°C.
1156
BRIESE ANI) ( AI~ANA("
TABLE I MEANS OF MAXIMALCOLONIC TEMPERATURES (°C) REACHED BY SEVEN RATS AT DISTINCT AMBIENT TEMPERATURES (°C~ Ambient temperature Mean colonic temperature Standard error
10 39.84 0.08
19 39.54 0.13
24 39.58 0.12
This should indicate that at this moment T~ had reached Tset (at least theoretically). From min 12 up to about min 31 Trail continued to increase while the ascending course of T~ gradually ended in a plateau at min 20. If the course of vasomotor responses is a criterion of the level of T~ relative to Tset, then after min 12 T~ was above Tset until about rain 31 when both T c and tail temperature stabilized. At first glance, from rain 12 to min 31 the animals appeared to be hyperthermic. However, nothing in the shape of T~ curve seems to constitute a signal for the sudden change from tail vasoconstriction to vasodilatation at min 12. This means that something else controls the time course of both curves, most likely the set-point mechanism. We speculate that the heat-dissipation response was triggered at min 12 because T c had overpassed Tset. The command to start and stop heat dissipation should be produced by a neural regulator which compares T~ and Tset. T~ integrates heat loss and heat gain. In
40
• Sn,o
LU
a:
39 Ill at. tU I'-0
38
8 37
I
I
I
I TIME
I (9 min
I
I
I
I
the present experiments we had only a window observation oi the time course of the heat dissipation response. From the strong vasoconstriction of the first 12 min we may deduce that heat production was insufficient although the animals were often seen shivering. From rain 12 to 31 heat gain was excessive since in.creasing vasodilatation was necessary to damp down the T,, increase and to maintain T~ at a stable level. Could this thermal state be qualified " h y p e r t h e r m i c " ? Probably not, because the oscillating Tt~ indicates that a regulato~ process was taking place. A new equilibrium was reached after about rain 3t up to the end of the experimental sessions. During this phase the stabilization of both Tta n and T~ as well as the reciprocal oscillations of both temperatures, as illustrated in Fig. 2, indicate that T, was regulated at this new higher level. The very existence of reciprocal oscillations is a strong argument in favor of regulation (6). A second physiological argument that the emotional rise in T~ is a fever is that the ambient temperature did not affect the elevation of T~. Most important is the fact that the cold environment did not diminish the T~ rise as would have occurred if the rise of T~ rise was a hyperthermia. Similar results were recently published (14). A controversial point is related to the results of the day/night experiment. The emotional rise of T~ reached a small (0.3°C) but significantly higher plateau during the night than during the day. It has been shown (9) that intracerebroventricular injections of prostaglandin induce a rise of T¢ regulated at the peak level which depends on the dose injected and is independent of the initial T~ and therefore of the time of day. Our results suggest that some part of emotional fever is not regulated according to the same paradigm. We conclude that the emotional elevation of T~ is similar to fever because the vasomotor responses contributed to increase T~ and to adjust and to maintain it at a new stable or oscillatory higher level. Another physiological argument in favor of fever is that the elevation of T~ due to handling was not diminished by a cool environment. The results suggest that thermogenic, thermoconservatory, and heat dissipation responses interplay during the emotional fever so as to raise T~ to a higher level. This further suggests that T~, is also shifted to a higher level.
I
ACKNOWLEDGEMENTS
intervals)
FIG. 4. Mean colonic temperatures and standard errors of the means of 13 rats. Measurements were taken every 9 min in sessions of 36-rain duration, once in the morning at 1000 h and once at night at 2230 h.
We wish to thank E. Satinoff for critically reading this manuscript. This work was supported by the Conseil de la Recherche en Sciences Naturelles et en GEnie (N. S. E. R. C. Canada) and Defense and Civil Institute of Environmental Medicine (D. C. I. E. M. Canada).
REFERENCES 1. Bass, D. E.; Jacobson, E. D. Effect of salicylate on work in the heat. Fed. Proc. 22:176; 1963. 2. BRisig, J.; Hdllt, V.; B~luerle, U.; Herz, A. Involvement of endorphins in emotional hyperthermia of rats. Life Sci. 23:2525-2532; 1978. 3. Briese, E. Hyperthermia in self-stimulating rats. Acta Physiol. Latinoam. 15:357-361; 1965. 4. Briese, E.; Cabanac, M. Emotional fever and salicylate. In: Sze16nyi, Z.; SzEkely, M., eds. Contributions to thermal physiology. Oxford: Pergamon Press; 1980:161-163. 5. Briese, E.; de Quijada, M. G. Colonic temperature of rats during handling. Acta Physiol. Latinoam. 20:97-102; 1970. 6. Cabanac, M.; Russek, M. REgulation et contrrle en biologie. Qurbec: Presses Univ. Laval; 1982. 7. Cranston, W. I.; Rawlins, M. D. Effects of intracerebral microinjections of sodium salicylate on temperature regulation in the rabbit. J. Physiol. (Lond.) 222:257-266; 1972.
8. Downey, J. A.; Darling, F. C. Effect of salicylates on elevation of body temperature during exercise. J. Appl. Physiol. 17:323-325; 1962. 9. Feng, J.; Price, M.; Cohen, J.; Satinoff, E. Prostaglandin fevers in rats: regulated change in body temperature or change in regulated body temperature? Am. J. Physiol. 257:R695-R699; 1989. 10. Gordon, C. J. A review of terms for regulated vs. forced neurochemical-induced changes in body temperature. Life Sci. 32:12851295; 1983. 11. Green, M. D.; Lomax, P. The effects of sodium salicylate on body temperature in the normothermic rat. Proc. West. Pharmacol. Soc. 16:257-261; 1973. 12. I.U.P.S. Commission for Thermal Physiology. Glossary of terms for thermal physiology. Pfltlgers Arch. 410:567-587; 1987. 13. Kluger, M. J.; O'Reilly, B.; Shope, T. R.; Vander, A. J. Further evidence that stress hypertherrrtia is a fever. Physiol. Behav. 39: 763-766; 1987.
EMOTIONAL FEVER
14. Long, N. C.; Vander, A. J.; Kluger, M. J. Stress-induced rise of body temperature in rats is the same in warm and cool environment. Physiol. Behav. 47:773-775; 1990. 15. Long, N. C.; Vander, A. J.; Kunkel, S. L.; Kluger, M. J. Antiserum against tumor necrosis factor increases stress hyperthermia in rats. Am. J. Physiol. 258:R591-R595; 1990. 16. Martin, G. E.; Bacino, C. B. Action of intracerebrally injected 13-endorphin on the rat's core temperature. Eur. J. Pharmacol. 59: 227-236; 1979. 17. Miles, G. H. Telemetering techniques for periodicity studies. Ann. NY Acad. Sci. 98:858-865; 1962. 18. Pae, Y. S. Hyperthermia in the rat from handling stress blocked by naloxone injected into the preoptic-anterior hypothalamus. Pharmacol. Biochem. Behav. 22:337-339; 1985. 19. Rawlins, M. D.; Rosendorff, C.; Cranston, W. I. The mechanism of action of antipyretics, In: Wolstenholme, G. E. W.; Birch, J., eds. Pyrogens and fever. Ciba Foundation Symposium. London, UK: Churchill Livingstone; 1971:175-191. 20. Rosendorff, C.; Cranston, W. I. Effects of salicylate on human temperature regulation. Clin. Sci. 35:81-91; 1968.
1157
21. Satinoff, E. Salicylate: Action on normal body temperature in rats. Science 176:532-533; 1972. 22. Singer, R.; Harker, C. T.; Vander, A. J.; Kluger, M. J. Hyperthermia induced by open-field stress is blocked by salicylate. Physiol. Behav. 36:1179-1182; 1986. 23. Stewart, J.; Eikelboom, R. Stress masks the hyperthermic effect of naloxone in rats. Life Sci. 25:1165-1172; 1979. 24. Stitt, J. T. Prostaglandin E as the neural mediator of the febrile response. Yale J. Biol. Med. 59:137-149; 1986. 25. Techtermann, F.; J6quier, E. Etude calorim6trique de l'effet du salicylate chez l'homme lors de l'hyperthennie induite par exposition au chaud ou par exercice musculaire. J. Physiol. (Pads) 74:615-621; 1978. 26, Vidal, C.; Suaudeau, C.; Jacob, J. Hyper- and hypothermia induced by non-noxious stress: effects of naloxone, diazepam and gammaacetylenic GABA. Life Sci. 33:587-590; 1983. 27, York, J. L.; Regan, S. G. Conditioned and un-conditioned influences on body temperature and ethanol hypothermia in laboratory rats. Pharmacol. Biochem. Behav. 17:119-124; 1982.