Fever, Temperature, and the Immune Response

DANIEL F. HANSON“ Department of Biology Washington University

Box 1229 One Brookings Drive

St . Louis, Missouri 63130

INTRODUCTION

Fever is an obvious sign of infection and has been used in diagnosis of disease for thousands of years.’ Fever as a biological trait is highly conserved in the vertebrates and, within a given vertebrate species, it is displayed by all the mem- bers of the species with only minor differences in As part of the “acute phase” response to infection, fever is temporally associated with host defense and this has led to speculation that it is a useful response of benefit to the infected host. Indeed, prior to modern antibiotic therapy, “fever therapy” was regarded as a useful, noninvasive treatment for some disease^.^.^ However, the actual biological role fever might play in the biology of infection has remained in dis- pute.zv6-’0 My laboratory has chosen to focus on how the temperature changes during fever regulate the immune responses of mouse lymphocytes and macro- phages.

CYTOKINES AS ENDOGENOUS PYROGENS

Our interest in the biological role of fever began the late 1970s when we were using conventional biochemical techniques to purify and characterize endogenous pyrogens (EP), the endogenous polypeptide mediators of fever made by phagocytic cells of the host in response to microbial products. Upon purification to homogene- ity, it was shown that rabbit endogenous pyrogens were in fact identical interleukin-1 (IL-lm, IL-l/3).”*’2 Although the concept of cytokine pleiotropy is now well e~tablished,’~ at that time the capacity of the same molecules for two such diverse functions as EP (fever by distal action on the brain) and IL-1 (local inflammation and T-cell costimulation) was quite remarkable to us. The coinci- dence of such diverse activities in a single molecular entity suggested to us that one activity might influence the other. In recent years the purification, molecular cloning and extensive characterization of the growing family of cytokines has led to an impressive roster of distinct molecules which possess intrinsic endogenous pyrogen activity. l4 These include IL- l a , IL- lp, tumor necrosis factor-a (TNF- a), TNF-P, interferon-a (IFN-a), IL-6 and macrophage inflammatory protein-1 (MIP-1). In addition, during animal studies and human clinical trials, the recombi-

a Tel.: (314) 935-7306; Fax: (314) 935-5125; E-mail: hanson@biodec.wustl.edu. 453

454 ANNALS NEW YORK ACADEMY OF SCIENCES

nant, purified cytokines IL-2 and IFN-y were also found to display pyrogenic activity, although this is now believed to result from their ability to induce one of the genuine EPs (above) either directly or in synergy with traces of bacterial products. These cytokines are released during the acute phase of infection when fever occurs, when host defense is most critical, and when both immediate inflam- matory defenses and more long-range lymphocytic immune responses are being actively engaged. Thus it seemed reasonable to us that the thermoregulatory properties of these molecules might have a large impact upon concomitant im- mune responses.

TEMPERATURE DEPENDENCE OF HOST DEFENSE

It has been proposed that the temperature changes associated with fever itself might increase host resistance to infection and the results of experiments both in viuo and in vitro suggest that this may be so.l5-l8 Indeed, if the physiological temperature changes associated with fever functioned to augment some fundamen- tal, critical aspect of host defense, this might serve to explain the complete conser- vation of the trait of fever in the vertebrates, animals whose host defense systems are more or less similar. However, it remains unclear which defensive functions are most critically affected by temperature mainly due to the diversity of experimental approaches to this question. The calculation of temperature coefficients or “Qlo values”19 allows comparison of temperature dependence data from different bio- logical systems and using different experimental temperature ranges. A review of the available literature demonstrated to us that most of the antigen-nonspecific defense systems (chemotaxis, phagocytosis, killing of ingested bacteria, comple- ment-dependent hemolysis) displayed rather low Qlo values on the order of 2-5, whereas systems involving the antigen-specific (or antigen-like) activation, prolif- eration and differentiation of lymphocytes appeared to display Qlo values on the order of 100-1000.17~20

THE CHOICE OF TEMPERATURE RANGE

The experimental temperature ranges employed in the work mentioned above have been diverse and unsystematic. Many studies have reflected the clinical definition of fever by comparing the efficiency of a host defense process at a “normal” core temperature of 37°C with its efficiency at a “febrile” core tempera- ture, usually 38.5-40°C. Another group of studies have used even higher “heat shock” temperatures, usually 41-45”C, but these clearly have little to say about the impact of normal fevers upon host defense. The great majority of fevers each of us experiences in the course of a lifetime are moderate in degree and self- limited in time, and the causative infection resolves itself without medical interven- tion. The dangerously high fevers often encountered in clinical management of severe or morbid infection are by far the biological exception. Such fevers and the temperatures accompanying them have little to say about the normal biology of fever or its evolutionary selection since they most probably represent the occasional failure of an immune system which is usually successful at host defense. Finally, those few studies which have employed temperatures below 37°C have usually employed temperatures well below the physiological range as probes of

HANSON IMMUNE RESPONSE 455

cell biology. But like the “heat shock” temperatures, such studies do not illuminate the role of normal body temperatures and normal fevers in routine host defense.

Unlike most studies in this field, work in our laboratory has utilized a range of temperatures which we define as “physiological” in the following way. Mammalian body temperature is not a single value but rather a series of highly organized thermal gradients whose slope and distribution are altered with changes in activity, environment and heaIth.’6*21*22 Indeed, approximately 50% of the mass of a healthy human being is normally below 37”C, the core temperature usually considered “normal.” Because the temperature gradient is somewhat variable throughout the body, there is no single thermal gradient normal to “health.” However, for experimental purposes, we have chosen to define such a “normal, healthy” gradi- ent as spanning 29°C (normal, thermoneutral skin temperature) to 37°C (normal, thermoneutral core temperature) (FIG. 1, curve #l). The choice of 29°C as a lower limit is a very conservative one based solely upon what is thermoneutral for a resting human being and does not take into account activity in colder environments. We have chosen to model fever as 39°C because this represents a modest, normal febrile core temperature. When fever is instituted, the thermal gradients of the body change in two ways. First and most obviously, the core temperature increases resulting in a proportionate increase in all temperatures along the theoretical gradient (FIG. 1, curve #2). Second, the gradient flattens due to behavioral selec- tion of warmer environments as shelter, clothing or piloerection as well as a reduction in convective cooling due to postural changes and torpor due to weakness and malaise during illness (FIG. 1, curve #3). Thus, during fever core temperatures may rise from 37 to 39”C, but skin temperatures may rise from 29°C toward 37°C as a sick animal remains sheltered in its nest or bed. With this view in mind, it is not obvious which thermal transitions mediated by fever might be beneficial to host defense. A large fraction of the immune system, including the spleen, gut, lung, liver, interior tissues and interior vasculature, can be considered to be at core temperature during health or fever (37OC-39”C). However, at least 50% of the body mass including the skin, peripheral tissues, peripheral vasculature and the peripheral lymph nodes are normally subject to a different set of lower temperature gradients. This half of the body increases its temperature upward during fever as well but the thermal transitions cover 29 to 37°C. Moreover, phagocytes and lymphocytes are constantly recirculating between these changing thermal compart- ments and must be able to function at a variety of different physiological tempera- tures. The immunological purpose of fever might be to heat the animal’s core. Alternatively, the core temperature increase might be an incidental consequence

FIGURE 1. Hypothetical thermal gradients in the mouse or human being. Curve #1 rep- resents the temperature gradient in health; curue #2 when the core temperature is ele- vated to fever without flattening of the gradi- ent; curue #3 is the flattened gradient during fever with behavioral selection of warmer environments with restricted activity due to malaise and torpor. Stippled area shows that peripheral tissues may undergo a greater overall temperature change than core tissues.

27 J . , . r I I 1 . I

0.0 0.1 0.4 0.6 0.8 1.0 1.1 1.4 1.6

456 ANNALS NEW YORK ACADEMY OF SCIENCES

to the need to warm the peripheral tissues to some critical level. Thus the “physio- logical temperature range we employ in our studies of the temperature dependence of murine immune responses is 29OC-39”C in increments of 2°C. This range includes the clinical definition of fever in human beings as well as temperatures found in the normal temperature gradients of humans and mice down to 29°C at the skin and specifically excluding very high febrile temperatures (40-42°C) and heat shock temperatures (42°C or higher) as these are not part of the normal biology of fever. Mice were chosen for our studies since they are inbred, their immune responses in uitro and in uiuo are well defined and they exhibit normal and febrile thermal gradients very similar to human beings.2’

THYMOCYTE ACTIVATION IS TEMPERATURE DEPENDENT, DNA SYNTHESIS AND CELL DOUBLING TIME ARE NOT

The presence of EP and IL-1 activities in the same molecule focused our initial studies on the effect of temperature on lymphocytes as reflected in the proliferative response of mouse thymocytes to IL-1 and the polyclonal antigen analog phyto- hemagglutinin (PHA).11,12%’7.24 The results of experiments of this kind are shown in TABLE 1. Neither the antigen analog PHA nor recombinant IL-1 causes significant proliferation, yet together they cause a potent proliferative response. Interestingly, this response is usually not increased by febrile core temperatures of 39°C. Instead, as the temperature is lowered into the range of the peripheral tissues, the response fails almost completely at 32-33°C. This suggested to us that fever’s impact upon the immune system might have more to do with peripheral tissue temperature changes than with the core temperature changes traditionally defined as “fever.” The average apparent Qlo value over 24 experiments with IL- 1 and PHA was 1 125 (range 77-6343). This is a figure which is enormously greater than those associated with simple or complex enzyme systems.20 IL-I also occurs in a membrane form prior to release from phagocytes and, indeed, most IL-1 activity may find expres- sion in this form during cell : cell interaction^.^' The thymocyte response to the membrane form of IL- 1 was also very temperature sensitive generating apparent Qlo values of 200-600. In contrast, thymocyte or T-cell proliferation to the T-cell growth factor IL-2 is not particularly sensitive to temperature (Qlo = 12, TABLE 1). This response reflects the growth of previously activated thymocytes utilizing exogenous IL-2 by dividing and synthesizing DNA. Our earlier studies on the temperature sensitivity of thymocyte DNA ~ y n t h e s i s ’ ~ . ~ ~ suggested that cell divi- sion and DNA synthesis are not remarkably sensitive to temperature change. Both the DNA synthesis and cell doubling times of two T-cell lines display Ql0 values of only 1.6-4.0 (TABLE 2), results which are similar to enzyme systems however complex they may be.2o Thymocyte responses to exogenous, saturating doses of IL-2 can be rendered very temperature sensitive by two means: the inclusion of either recombinant IL-1 or PHA. Such results argue that it is thymocyte activation for cell division that possesses one or more highly temperature sensitive events rather than cell division itself. In this context, two possible cell biological events could be strongly temperature dependent. The first is the endogenous thymocyte production of IL-2 or IL-7 in response to the synergistic action of IL-1 and PHA. These two growth factors are then used by activated thymocytes to generate the observed cell p r~ l i fe ra t ion .~~,~’ The second is the expression of receptors for IL-2 (IL-2R) or IL-7 triggered by exposure to PHA (as a polyclonal antigen analog) or by IL-1 .28 Recent experiments in our laboratory indicate that IL-2R expression

T * $ z E 5 3

Tri

tiate

d Th

ymid

ine

(CPM

) E

Q 10

0

TA

BL

E 1. T

empe

ratu

re S

ensi

tivity

of

Thy

moc

yte

Prol

ifera

tion

in R

espo

nse

to M

urin

e IL

-1 a

nd I

L-2

v)

w

IL-1

PH

A

32°C

37

°C

39°C

Rec

ombi

nant

5 U

/ml

+ 96

5 32

100

2509

3 11

06

IL-1

IL

-2

PHA

29

°C

33°C

37

°C

39°C

Q

io

Con

trol

RPC

mem

bran

e -

+ 19

8 608

595

516

4.0

Act

ivat

ed R

PC m

embr

ane'

-

+ 39

5 41

48

4268

1 52

097

348

Rec

ombi

nant

300

U/m

l -

+ 22

8 35

0 89

9 68

6 5.

6 -

50 U

/mld

-

830

2140

61

85

7170

12

R

ecom

bina

nt 3

00 U

/ml

50 U

/ml

-

2115

13

046

2452

66

2281

06

380

-

50 U

/ml

+ 35

20

4087

9 24

5226

16

2376

20

0

-

+"

398

644

78 1

2.6b

Phyt

ohem

aggl

utin

in 5

pg/

ml.

5 x

16

thym

ocyt

es s

timul

ated

with

1 x

lo

5 fix

ed r

esid

ent

perit

onea

l ce

lls s

timul

ated

24

hour

s w

ith 1

00 p

g/m

l he

at-k

illed

sta

phyl

ococ

ci

IL-2

aff

inity

pur

ified

fro

m E

L-4

sup.

* Qlo

calc

ulat

ed o

ver

32-3

7°C

or

29-3

7°C

inte

rval

s.

befo

re f

ixat

ion.

458 ANNALS NEW YORK ACADEMY OF SCIENCES

TABLE 2. Temperature Sensitivity of Proliferation and Cell Doubling Time in T-cell Lines

DNA Synthesis (CPM)" Cell Doubling Time (hr)

Cell Line 32°C 37°C Qlo 32°C 37°C Ql$ EL-4 14828 19036 1.6 26.4 14.4 3.4 CTLL-2' 8705 13794 2.5 22.5 11.2 4.0

DNA synthesis measured as incorporation of tritiated thymidine by 3000 EL-4 cells in

Qlo of cell doubling time calculated from 1/Td, where Td = cell doubling time in hrs. CTLL-2 cells growing in the presence of 100 U/ml IL-2.

a 22-hour pulse or by 2000 CTLL-2 cells in a 4-hour pulse.

is strongly temperature dependent (unpublished observations) and this may under- lie the observed temperature dependence of the proliferative responses of thymo- cyte cultures containing abundant, exogenous IL-2 plus activators like PHA or IL- 1.

EFFECTOR CREATION IS TEMPERATURE SENSITIVE BUT NOT EFFECTOR FUNCTION

Because thymocytes are not the mature T-cells involved in the immune re- sponses of the peripheral immune system, our next studies examined the effects of physiological temperature upon proliferation, effector differentiation and effector function in primary cell-mediated immune responses in ~ i t r o . ' ~ The proliferative responses of splenic T-cells to the polyclonal mitogen concanavalin A (Con A), to solid-phase monoclonal antibody (QCD3) against the CD3 component of the T-cell antigen receptor, or to irradiated allogenic spleen cells all displayed strong temperature sensitivity with calculated Qlo values of 50-500, whereas spleen cells previously stimulated with mitogens at a permissive temperature and then allowed to synthesize DNA at different temperatures displayed Qlo values of 1.8-2.4. We also began to examine the effects of temperature upon the kinetics of these responses. In theory, temperature could act to regulate the absolute extent of a response, it could act to delay a response at a lower temperature relative to a higher temperature but still achieve the same maximum only later, or both effects could occur. Thymocyte responses to IL-I + PHA had been found to display virtually no kinetic effect suggesting that temperature regulated the absolute extent of some critical components of the response. The same was true of spleen cell responses to mitogens like Con A (FIG. 2A). However, when we examined splenic proliferative responses to QCD3, to QCD3 plus another monoclonal antibody (QCD28) (which mimics the effect of costimulation of T-cells and provides a stronger signal than aCD3 alone), or to allogenic irradiated spleen cells we found that lower physiological temperatures both limited the extent of the response and delayed it as well. Proliferation due to QCD3 + QCD28 (FIG. 2B) was similar at 37 and 39°C while at 35°C the maximum was somewhat less and delayed one day. At 33°C the maximum was only 20% of that obtained at 37°C and delayed three days while there was no response below 33°C. The presence of kinetic effects makes Qlo values much less meaningful although they remain large in the early phases of these responses. In such responses temperature sensitivity is better

HANSON: IMMUNE RESPONSE

12 .

8 .

4 .

459

t

FIGUREZ. Kinetics of mouse splenic T-cell proliferative responses as a function of temperature. (A) 5 x lo4 DBAR cells/200 pl well stimulated with 1 pglml Con A. (B) 5 x lo4 Balb/c cells/200 p1 well stimu- lated with a combination of 2 pglml QCD3 and 2 pg/ml QCD28. Triplicate cultures pulsed for 4 hours on the indicated days. -0- = 29. -A- = 3 1 . -.- = 33.

from Han~on.*~ Reprinted by permission from the Journal of Immunology.)

-A- = 35; -0- = 37; -0- = 39. ((A)

1 2 3 4 5 6 7

Day

0 2 4 6 8 1 0 1 2 1 4

Day

described qualitatively by visual inspection or quantitatively by integrating such curves.

The reaction of murine splenic T-cells to irradiated allogenic cells is known as a mixed lymphocyte reaction (MLR), and it provokes a genuine effector response in the form of the differentiation of cytolytic T-lymphocytes (CTL) which will specifically lyse allogenic target cells of the same foreign histocompatibility geno- type. Quantitative analysis of the CTL content of such cultures revealed a profound effect of temperature upon CTL differentiation (FIG. 3A, apparent Qlo = 297). CTL differentiated equally well at 37 and 39°C but much less well at lower tempera- tures. At no temperature was there spontaneous formation of CTL without allo- genic stimulation nor was there differentiation of CTL against either self or against allospecificities other than the one used as the stimulator in the MLR. Kinetic analysis showed that temperature regulated both the extent and timing of the appearance of CTL in these cultures. In contrast, when CTL which had been formed at a permissive temperature (37°C) were allowed to lyse appropriate target cells at different temperatures, there was virtually no influence of temperature upon the effector function of lysis itself (FIG. 3B, apparent Qlo = 1.5). These data suggest that temperature restricts where new effector CTL can develop in the body but, once the effector CTL are formed, they can perform their lytic function wherever they are in the normal thermal gradient. Moreover, in no case was the “febrile” temperature of 39°C more stimulatory than the normal core temperature of 37”C, whereas slightly lower temperatures were markedly inhibitory. Together

460 ANNALS NEW YORK ACADEMY OF SCIENCES

1 10 100

FIGURE 3. Effects of physiological temper- ature upon the generation (A) and lytic effector function (B) of allospecific CTL. (A) Titration analysis at 37°C of CBA (H-2k) anti DBA/2 (H-2d) CTL populations gener- ated at different temperatures; target cells are P815 (H-2d) cells labeled with *‘Cr. (B) Titration analysis at different tempera- tures of CBA (H-2k) anti DBAR (H-2d) CTL generated at 37°C; target cells as above. -0- = 29; -.- = 33. -0- = 37. -0- = 39. ((A) from H a n ~ o n . ~ ~ Reprinted by permission from the Journal of Zmmu- nology.)

1 10 100

Effector : Target Ratio

with the thymocyte data, this finding suggests that the biological “purpose” of fever is to warm the peripheral tissues rather than the core.

Recent studies in our laboratory have focused on extending our findings with cell-mediated immunity to the other major limb of humoral immunity, antibody formation by B-lymphocytes (unpublished observations). This system has the advantage that mouse B-cells can be stimulated to proliferate and secrete antibody directly by bacterial lipopolysaccharide (LPS) or indirectly by stimulation of helper T-cells (Th). When B-cells were stimulated with LPS, they proliferated strongly with high apparent Qlo values in the first few days of the response. But with time even the lowest temperatures of 29 and 3 1°C mounted substantial proliferative responses. Likewise all cultures released substantial amounts of IgM and small amounts of IgG with only delay of such responses at the lower temperatures. Thus, with LPS as the B-cell stimulant, the effect of temperature is purely kinetic and probably isolates the effects of physiological temperatures upon the biology of the B-cell itself. Very different results are obtained when the B-cells respond through the agency of Th either as splenic Th stimulated with QCD3 + C D 2 8 or as cloned, antigen-specific, H-Zrestricted Th-hybridomas stimulated with specific antigen. In this form of the experiment virtually all the kinetic effect disappears. B-cells proliferate strongly at 37 and 3YC, about 50% as well at 35°C and very poorly at lower temperatures (Qlo = 100-500). The findings for antibody secretion in this version are similar with strong IgM and IgG at the higher permissive temperatures and catastrophic failure at the lower temperatures with no compen- sating late onset. Thus temperature can regulate both the extent and timing of B-cell responses but does so most profoundly when the B-cell is under the control of a Th which is the form of most antibody responses. These data suggest that it

HANSON: IMMUNE RESPONSE 461

is the helper T-cell which is the most temperature sensitive element in the lympho- cytic immune system. In contrast to antibody formation (above), previously acti- vated B-cells allowed to secrete antibody at different temperatures displayed Qlo values < 1.5 for secretion itself. Similarly, when antibody to sheep erythrocytes was allowed to fix complement and lyse the erythrocytes across the range of 29-39”C, lysis was nearly the same at all temperatures (Qlo slightly less than 1 .O) . Thus, as in cell-mediated immunity, temperature strongly regulates the initial activation of B-cell proliferative and antibody responses and the nature of this regulation can be either absolute or kinetic depending upon whether the response is under the control of a Th. Once formed, however, the antibody effectors can be secreted and can perform their effector function anywhere in the normal thermal gradient in a nearly temperature independent fashion. As found with cell-mediated immunity, B-cell responses were not greater or faster at 39°C (fever) relative to 37°C but failed rapidly at physiological temperatures only slightly below 37”C, findings which also suggest that the biological function of fever in the immune system is more directed at warming the periphery than the core.

MACROPHAGES AND HELPER T-CELLS DISPLAY DIFFERENT THERMAL RESPONSE PROFILES

The experiments discussed above provide evidence that physiological tempera- ture can profoundly regulate both the major limbs of lymphocytic immunity. But what are the elements of immune physiology which are so temperature dependent? The results of our studies with B-cell responses pointed to the Th as the critical cell for temperature sensitivity. Since this cell regulates most of the rest of the immune system by means of its production of cytokines, we propose that cytokine production by Th is one critical temperature-dependent event. The other major cellular source of cytokines in the immune system is the macrophage. Indeed, it is the macrophage which releases most of the EPs which cause fever as well as local inflammation, a condition which can be regarded as a localized form of fever (due to increased blood flow) save that it cannot increase beyond the prevailing core temperature of the host. Since macrophage secretion of pyrogeniciinflammatory cytokines is the source of temperature increase, it seemed unlikely that the secre- tion of these molecules should be grossly temperature dependent. Early experi- ments with rabbit macrophages stimulated with Staphylococci showed this to be the case” with Qlo values of 1 .O-4.6. Similar results have been obtained with mouse resident peritoneal macrophages and the murine macrophage cell line P388D1 stimulated with LPS or with Staphylococci. In fact, macrophages make both soluble and membrane IL-1 perfectly well at the lowest temperature of 29°C as well as at 33 and 37°C (TABLE 3). Most interesting is the finding that P388D1 macrophages make much less IL-1 at the febrile temperature of 39°C. This suggests that the pyrogenic action of IL-1 tends to shut off its own production once fever is achieved. Recent studies have shown a similar pattern for the production of TNF-a as well (unpublished observations).

In contrast to macrophages, cytokine production by helper T-cells displays a different pattern. In TABLE 4 is shown the pattern of temperature dependence of stimulated IL-2 production by the T-cell lymphoma EL-4. Very little of the cyto- kine is produced between 29 and 33°C. But between 33 and 37°C IL-2 production increases over tenfold while over the 37 to 39°C transition the increase is only 1.3- to 1.8-fold. The overall Qlo of this response is high (SO-lOOO), and we have

462 ANNALS NEW YORK ACADEMY OF SCIENCES

TABLE 3. Temperature Sensitivity of IL-I Production by P388D, Cells IL-I Units” per ml (Soluble) or per

lo6 Fixed Cells (Membrane)

Form of 1L-1 Stimulus 29°C 33°C 37°C 39°C Qlob

Soluble - 18.5 24.7 3.5 5.5 <1.0 Soluble LPS 86.4 155.7 64.5 43.4 11.0 Membrane - 38.0 64.1 10.8 7.5 <1.0 Membrane LPS 107.4 350.1 132.6 79.5 1.3

a 18 hour accumulation of IL-I assayed on C3H/HeJ thymocytes in the presence of 5 pg/ml PHA.

Qio values calculated over the 29-37°C interval. LPS stimulation at 10 pg/ml.

observed a similar temperature dependence in the production of the Th-derived cytokines IL-4 and IFN-y. This pattern is very similar to that observed in our early thymocyte experiments and suggests that the temperature dependence of IL-2 production is one of the central phenomena underlying the observed tempera- ture sensitivity of cell-mediated immune responses in uitro. As before, the lack of amplification observed at febrile temperature relative to 37°C suggests that fever influences the immune system mainly by increasing the temperature of peripheral tissues rather than increases in core temperature.

SUMMARY

Fever’s ability to manipulate the character and extent of physiological tempera- ture gradients correlates with the unusual influence different physiological temper- atures have upon model immune responses in uitro. This relationship may help to explain the remarkable evolutionary conservation of the febrile response to infection. A very restricted range of the upper physiological temperatures supports the activation of resting lymphocytes for proliferation and effector formation in the two major limbs of the immune system, cell-mediated immunity and humoral immunity. In contrast, once effectors are formed they can function in a fashion which is nearly independent of physiological temperature. This suggests that physi- ological temperature change acts to regulate the emergence of new immune re- sponses but does not restrict the activity of existing effector mechanisms once

TABLE 4. Temperature Sensitivity of IL-2 Production by EL-4 IL-2 Produced (Units/ml)”

Stimulus 29°C 33°C 37°C 39°C Qioh

None <0.3 <0.3 <0.3 <0.3 1 .o 10 ng/ml PMA 2.2 7.2 76.3 121 84-365 10 pg/ml Con A 0.3 0.6 10.2 13.1 82-1191

a 8 hour accumulation of IL-2 assayed on CTLL-2 cells. Qlo values have been calculated over the 29-37°C interval (lower number) or the 33-37°C

interval (higher number).

HANSON: IMMUNE RESPONSE 463

they have been formed. The differential sensitivity of these processes to different physiological temperatures suggests that fever’s biological purpose with respect to the immune system is the elimination of lower peripheral tissue temperatures rather than the elevation of core temperatures. However, further studies may reveal that some functions are amplified by the core temperature transitions while other functions are selectively regulated by peripheral tissue temperature transi- tions. The critical cell for the temperature dependence of immune responses seems to be the Th since its ability to produce cytokines is highly temperature dependent. In contrast, macrophages produce cytokines equally well at all temperatures except those of the febrile core, afeature which may serve to downregulate the production of endogenous pyrogens.

ACKNOWLEDGMENTS

The author wishes to thank Drs. Patrick Murphy and Anthony Defranco for invaluable support and discussion.

REFERENCES

1. 2. 3.

4. 5 .

ATKINS, E. 1982. Fever: its history, cause and function. Yale J. Biol. Med. 55: 283. KLUGER, M. J. 1978. The evolution and adaptive value of fever. Am. Sci. 66: 38. KLUGER, M. J. 1991. The adaptive value of fever. I n Fever: Basic Mechanisms and

SOLOMON, H. C. & I. KOPP. 1937. Fever therapy. N. Engl. J. Med. 217: 805. WAGNER-JAUREGG, J. 1965. The treatment of dementia paralytica by malaria inocula-

tion. In Nobel Lectures: Physiology of Medicine, 1922-1941. 159. Elsevier. New York.

BENNETT, I. L. & A. NICASTRI. 1960. Fever as a mechanism of resistance. Bacteriol. Rev. 24: 16.

KLUGER, M. J. 1986. Is fever beneficial? Yale J. Biol. Med. 59: 89. DUFF, G. W. 1986. Is fever beneficial to the host: a clinical perspective. Yale J. Biol.

BLATTEIS, C. M. 1986. Fever: is it beneficial? Yale J . Biol. Med. 59: 107. BANET, M. 1986. Fever in mammals: is it beneficial? Yale J. Biol. Med. 5 9 117. HANSON, D. F. & P. A. MURPHY. 1984. Demonstration of interleukin-1 activity in

apparently homogenous specimens of the PI 5 form of rabbit endogenous pyrogen. Infect. Immun. 45: 483.

MURPHY, P. A, , D. F. HANSON, Y.-N. G u o & D. E. ANGSTER. 1985. The effects of variations in pH and temperature on the activation of mouse thymocytes by both forms of interleukin-1. Yale J. Biol. Med. 58: 115.

Cytokines: coordinators of immune and inflammatory responses. Annu. Rev. Bio- chem. 59: 783.

DINARELLO, C. A. 1991. Endogenous pyrogens: the roleofcytokines in the pathogenesis of fever. I n Fever: Basic Mechanisms and Management. P. A. Mackowiak, Ed. 23. Raven Press. New York.

Management. P. A. Mackowiak, Ed. 105. Raven Press. New York.

6.

7. 8.

9. 10. 11.

Med. 59: 125.

12.

13. A U I , K . I., F. LEE, A. MIYAJIMA, S . MIYATAKE, N . ARAI & I. YOKOTA. 1990.

14.

15. 16.

ROBERTS, N. J., JR. 1979. Temperature and host defense. Microbiol. Rev. 43: 241. RODBARD, D., H. WACHSLIGHT-RODBARD & S. RODBARD. 1980. Temperature: acritical

factor determining localization and natural history of infectious, metabolic and immu- nological diseases. Perspect. Biol. Med. 23: 439.

HANSON, D. F., P. A. MURPHY, R. S. SILICANO & H. S. SHIN. 1983. Effect of temperature on the activation of thymocytes by interleukins 1 and 2. J. Immunol. 130 216.

17.

464 ANNALS NEW YORK ACADEMY OF SCIENCES

18.

19. 20.

21.

22.

23.

24.

25.

26.

27.

28.

ROBERTS, N. J., JR. 1991. The immunological consequences of fever. I n Fever: Basic Mechanisms and Management. P. A. Mackowiak, Ed. 125. Raven Press. New York.

GIESE, A. C. 1968. Cell Physiology. W. B. Saunders. Philadelphia. HANSON, D. F. 1982. Fever, monocyte endogenous pyrogen and immune responsive-

ness. Dissertation. The Johns Hopkins University. Baltimore, MD. HERRINGTON, L. P. 1940. The heat regulation of small laboratory animals at various

environmental temperatures. Am. J. Physiol. 129: 123. DUBOIS, E. F. 1948. Fever and the Regulation of Body Temperature. Charles C.

Thomas. Springfield. HANSON, D. F. 1993. Fever and the immune response: the effects of physiological

temperature on primary murine splenic T-cell responses in v i m . J. Immunol. 151: 436.

HANSON, D. F. & P. A. MURPHY. 1985. Temperature sensitivity of interleukin-depen- dent murine T-cell proliferation: Q2 mapping of the responses of peanut agglutinin negative thymocytes. J. Immunol. 135: 301 1.

KURT-JONES, E. A., D. I. BELLER, S. B. MIZEL & E. R. UNANUE. 1985. Identification of a membrane-associated interleukin 1 on macrophages. Proc. Natl. Acad. Sci. USA 8 2 1204.

HERBELIN, A., F. MACHAVOINE, E. SCHNEIDER, M. PAPIERNIK & M. DY. 1992. IL-7 is requisite for IL-1-induced thymocyte proliferation: involvement of IL-7 in the synergistic effects of granulocyte-macrophage colony-stimulating factor or tumor necrosis factor with IL-1. J. Immunol. 148: 99.

CONLON, P. J., C. S. HENNEY & S. GILLIS. 1982. Cytokine-dependent thymocyte responses: characterization of IL-1 and IL-2 target subpopulations and mechanism of action. J. Immunol. 128: 797.

LOWENTHAL, J. W., J. CEROTTINI & H. R. MACDONALD. 1986. Interleukin 1-dependent induction of both interleukin 2 secretion and interleukin 2 receptor expression by thymoma cells. J . Immunol. 137: 1226.