INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

A REGULATED RISE in body temperature (T_b) or fever is an adaptive response of the organism to infection, injury, or trauma aimed at facilitating host resistance and slowing the growth of the pathogen (22). Considerable evidence indicates that many circulating cytokines, such as interleukin (IL)-1, IL-6, tumor necrosis factor- (TNF-), and others, act as endogenous pyrogens and are responsible for the induction and maintenance of fever by raising the "set point" for T_b regulation (22, 23). On the basis of the extensive evidence indicating crucial roles for IL-1, IL-6, TNF-, and other cytokines in the development of the febrile response, we hypothesized that any endogenous factor involved in the mechanisms of cytokine production or clearance may also play an important role in fever. In the present study, we investigated the role in fever of ₂-macroglobulin (₂M), one of the major plasma proteinase inhibitors in mammals and a specific cytokine carrier that binds and possibly regulates metabolism of cytokines implicated in fever, including IL-1, IL-6, and TNF-.

₂M is a tetrameric glycoprotein (Mr ~750 kDa) present in human plasma at high concentrations (~2-4 mg/ml) and a proteinase inhibitor that binds proteinases from all major classes (5, 13, 34, 40). The major source of plasma ₂M is the hepatocyte (5, 34, 40); however, other cells including monocytes and macrophages synthesize and secrete ₂M (3, 11, 21). Proteinase cleavage of the sensitive peptide bonds in the ₂M "bait region" induces a conformational change in the ₂M molecule, which irreversibly traps the reacting proteinase (12, 34, 40). Conformational change of the ₂M also exposes binding sites for the ₂M receptor/low-density lipoprotein receptor-related protein (LRP) (2, 36), which is present on the surfaces of many different cell types, including hepatocytes and macrophages (30). After binding to LRP, ₂M-proteinase complexes rapidly undergo endocytosis, indicating that LRP is responsible for plasma clearance of conformationally modified ₂M (2, 13, 30, 36). ₂M in its native form does not bind LRP and has a prolonged half-life in circulation (34).

In humans, ₂M is constantly present in plasma, and the changes in plasma concentrations are moderate and rarely diagnostic for any disease [for discussion, see Umans et al. (37)]. There is evidence, however, that lipopolysaccharide (LPS) may increase as well as suppress (perhaps depending on the experimental conditions) production of ₂M by human monocytes and macrophages in vitro (3, 11, 21). These data indicate that during inflammation in humans, concentration of ₂M may change significantly on the tissue or cellular level, although alterations in plasma concentrations are negligible.

The ability to bind practically all cytokines, hormones, and growth factors is an intriguing feature of ₂M, and it indicates the potential role for this plasma protein in fever. Several studies demonstrated that ₂M is a specific cytokine carrier, which binds major pro- and anti-inflammatory cytokines, such as IL-1 (6, 7), IL-6 (29), TNF- (9, 20, 42, 44), and others. Many cytokines can bind both native and proteinase-activated forms of ₂M (9). Importantly, binding of cytokines to the protease-activated ₂M does not affect the interaction of protease-₂M-cytokine complex with LRP (25). However, the functional role of ₂M in regulation of cytokine metabolism in vivo is far from clear. As it was discussed by LaMarre et al. (25) and Crookston et al. (9), the function of ₂M as a cytokine-binding molecule is complicated due to the different conformational states of ₂M. By functioning as a cytokine carrier, ₂M in its native form may protect bound cytokines from proteolytic degradation and, therefore, lengthen the plasma half-life. Most of the cytokines bound to the native ₂M retain their biological activity [for review, see LaMarre et al. (25)]. On the other hand, only proteinase-activated ₂M is recognized by LRP, and cytokines bound to the proteinase-activated ₂M may be targeted to the cells expressing LRP for endocytosis and rapid clearance [for discussion, see LaMarre et al. (25)].

In the present study, the role of ₂M in fever and cytokine responses induced by LPS was investigated using ₂M gene knockout (₂M /) mice developed by Umans et al. (37). Murine ₂M is a close homolog of human ₂M. Similar to human ₂M, murine ₂M is a tetrameric glycoprotein with the molecular mass of ~720 kDa constantly present in plasma in a concentration of ~2 mg/ml. Murine ₂M also inhibits proteases from all known classes and undergoes identical human ₂M conformational changes upon reaction with protease. Although during experimental inflammation in mice moderate changes in plasma ₂M can be observed (1, 19), similar to human ₂M, murine ₂M is not an acute phase protein. Similarities between human and murine ₂Ms suggest that ₂M / mice represent an adequate animal model for determining the role of this plasma protease inhibitor in fever and cytokine responses during experimental inflammation.

We hypothesized that the role of ₂M in fever depends on whether this protease inhibitor facilitates or inhibits clearance of the "major" endogenous pyrogens, e.g., IL-1, IL-6, or TNF-. In this study, LPS-induced fevers in ₂M / mice were compared with those in wild-type (WT) control mice. To investigate possible involvement of cytokines, changes in plasma levels of IL-1, IL-6, and TNF- were determined during fever in ₂M / and ₂M WT mice. Because initial experiments showed that during fever plasma concentration of TNF-, but not of IL-1 or IL-6, was significantly lower in the ₂M / mice compared with their WT controls, LPS-induced changes in hepatic TNF- mRNA levels in knockout and WT mice were studied as well.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals

Surgery

T_b and Motor Activity Measurements

Body Weight and Food Intake Measurements

Induction of Fever

ELISA

RT-PCR

Experimental Design

Experiment 1. LPS-induced fever in ₂M gene knockout mice. Mice were assigned to one of four groups: ₂M / mice treated with either LPS (n = 16) or pyrogen-free saline (n = 11) and ₂M WT mice injected with either LPS (n = 20) or pyrogen-free saline (n = 14). All injections were performed at 0900. T_b and motor activity were monitored for 3 h before and 48 h after LPS or saline injections. Body weight and food intake were measured up to 4 days postinjection in six ₂M / mice and eight WT mice treated with LPS and in six ₂M / mice and eight WT mice injected with saline.

Experiment 2. Turpentine-induced fever in ₂M gene knockout mice. To test the ability of ₂M / mice to mount a normal thermogenic response, we compared fevers in ₂M / and WT mice developed during localized inflammation (induced by turpentine injection). Mice were assigned to one of four groups: ₂M / mice treated with either turpentine (n = 4) or pyrogen-free saline (n = 3) and ₂M WT mice injected with either turpentine (n = 5) or pyrogen-free saline (n = 4). Because of the limited availability of ₂M / mice for our studies, all ₂M / and WT mice were those previously used in experiment 1. To circumvent any effect of previous injections, only those mice previously injected with sterile saline were used. All injections were performed between 0900 and 1000. T_b was monitored for 48 h after turpentine or saline injections.

Experiment 3. LPS-induced changes in plasma levels of IL-1, IL-6, and TNF- in ₂M gene knockout mice. Plasma concentrations of cytokines were measured at 1.5, 4, and 27 h after injection of LPS because of the following reasons. In mice and rats, at 1.5 h after the LPS challenge, high plasma concentrations of both IL-6 and TNF- are observed, whereas TNF- level is at or near its peak (8, 41). In addition, results of experiment 1 showed that, at 1.5 h after LPS injections, T_b is significantly lower in the ₂M / mice compared with their WT controls. At 4 h after LPS treatment, both moderate increase in plasma concentration of IL-1 and still high plasma level of IL-6 are observed, whereas plasma TNF- concentration decreases to control levels (8). Because in some studies in mice fever was observed to last up to 30 h after LPS injections (26, 27), we also measured plasma levels of cytokines at 27 h after LPS challenge (at the middle of the next day after LPS injections). The design for experiment 1 was used. Mice were placed into one of four groups: LPS injected ₂M / (n = 5), saline injected ₂M / (n = 3), LPS injected WT (n = 6), or saline injected WT (n = 3). Mice were injected with either LPS (2.5 mg/kg) or saline between 0900 and 1000. Blood for cytokine analysis was collected from anesthetized (4% halothane in air mixture) ₂M and WT mice by cardiac puncture at 1.5, 4, and 27 h after the injections. Immediately after the collection, plasma was separated by centrifugation and stored at 20°C until assayed.

Experiment 4. LPS-induced changes in hepatic TNF- mRNA levels in ₂M gene knockout mice. There is evidence that, in response to LPS challenge, different hepatic cells are capable of TNF- expression, whereas Kupffer cells are the predominant source of circulating TNF- (8, 18, 28). Therefore, to compare LPS-induced production of TNF- in ₂M / and WT mice, hepatic TNF- mRNA levels were measured. Taking into account that the decay in hepatic TNF- mRNA level is slow [hepatic TNF- mRNA level only slightly decreases between 1 and 3 h after LPS challenge (33)], whereas half-life of TNF- protein in circulation is very short (4, 10), and because of the limited availability of the ₂M / mice for our studies, hepatic TNF- mRNA levels were measured at the same time as the plasma TNF- concentration, i.e., 1.5 h after injections of LPS or saline. The design for experiment 1 was used. Animals used in experiment 3 for blood collection were also used to harvest liver tissue. Immediately after blood collection, a portion of the liver was removed, immediately frozen in liquid nitrogen, and stored at 80°C until total RNA extraction and assay for TNF- mRNA levels.

Statistical Analysis

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Experiment 1. LPS-induced fever in ₂M gene knockout mice. The febrile response of ₂M / and WT mice is shown in Fig. 1. The injection procedure caused profound short-term, stress-induced rises in T_b. Interestingly, the magnitude of this initial stress-induced rise in T_b was significantly higher in WT mice compared with ₂M / mice, regardless of whether they were injected with LPS or saline (Fig. 1). Both groups of mice injected with saline returned to the normal level of T_b within 90 min with no difference between ₂M / (Fig. 1, ) and WT (Fig. 1, ) animals. As shown in Fig. 1, ₂M / mice developed short-lasted and markedly attenuated fever in response to LPS compared with their WT counterparts. ₂M WT mice responded to LPS with an ~0.9°C fever that began 90 min postinjection and lasted up to 7 h (Fig. 1, ). In ₂M / mice, LPS induced a short-term ~0.6°C rise in T_b, which returned to the values observed in ₂M / and WT mice treated with saline 4 h after the injection (Fig. 1, ). Thus febrile response to LPS in ₂M / mice was shorter by 3 h compared with fever in ₂M WT mice. There was no difference in T_b between ₂M / and WT mice treated with either LPS or saline 9-48 h after the injections (data not shown).

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Effect of lipopolysaccharide (LPS; 2.5 mg/kg) or pyrogen-free saline (equivalent volume) injected intraperitoneally on body temperature in 2-macroglobulin gene knockout (2M /) and wild-type (WT) mice. Data are presented as means ± SE. Numbers in parentheses indicate sample sizes. Arrowhead indicates time of LPS and saline injections.  Significant difference between 2M / and WT mice injected with LPS; between 2M / and WT mice injected with saline, P < 0.05. * Significant difference between LPS-injected 2M / and WT mice, P < 0.05.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Effect of LPS (2.5 mg/kg) or pyrogen-free saline (equivalent volume) injected intraperitoneally on change in body weight (A) and food intake (B) in 2M / and WT mice. Data are presented as means ± SE. Numbers in parentheses indicate sample sizes.

Experiment 2. Turpentine-induced fever in ₂M gene knockout mice. Local tissue injury following turpentine administration was accompanied by significant fever on the next day. There was no difference in turpentine-induced changes in T_b between ₂M / and WT mice (Fig. 3).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Effect of turpentine (20 µl) or pyrogen-free saline (20 µl) injected intramuscularly into the left hindlimb on body temperature in 2M / and WT mice. Data are presented as means ± SE. Numbers in parentheses indicate sample sizes. Arrowhead indicates time of LPS and saline injections. Between 22 and 31 h after the injections, body temperature in 2M / and WT mice treated with turpentine was significantly (P < 0.05) higher than in 2M / and WT mice injected with saline.

Experiment 3. LPS-induced changes in plasma levels of IL-1, IL-6, and TNF- in ₂M gene knockout mice. Plasma IL-1, IL-6, and TNF- levels were measured in ₂M / and WT mice at 1.5, 4, and 27 h following injection of LPS or saline (Fig. 4). At all time points tested following injection of saline, ₂M / and WT mice showed low plasma IL-1, IL-6, and TNF- concentrations that did not significantly differ between groups (Fig. 4). Although plasma levels of IL-1, IL-6, and TNF- in ₂M / and WT mice treated with saline were low, they were above the detection limits of the assays and were included in the statistical analysis. These baseline values of plasma IL-6 and TNF- concentrations are not adequately presented in Fig. 4, B and C, due to the large scale of the y-axis. The values of plasma TNF- concentrations in ₂M / and WT mice 1.5 h after saline injection are given in the text below, as it is important for the comparison with the values of the TNF- concentrations induced by LPS challenge.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Effect of LPS (2.5 mg/kg) or pyrogen-free saline (equivalent volume) injected intraperitoneally on plasma levels of interleukin (IL)-1 (A), IL-6 (B), and tumor necrosis factor- (TNF-; C) in 2M / and WT mice. Plasma IL-1, IL-6, and TNF- concentrations were measured at 1.5, 4, and 27 h following injection of LPS or saline. Data are presented as means ± SE. Numbers in parentheses indicate sample sizes. At 4 h after the injections, IL-1 levels in plasma of 2M / and WT mice treated with LPS were significantly (P < 0.05) higher than in plasma of 2M / and WT mice injected with saline (A). At 1.5 and 4 h after the injections, IL-6 levels in plasma of 2M / and WT mice treated with LPS were significantly (P < 0.05) higher than in plasma of 2M / and WT mice injected with saline (B). * Significant difference in plasma TNF- concentration between 2M / and WT mice injected with LPS, P < 0.05.

Experiment 4. LPS-induced changes in hepatic TNF- mRNA levels in ₂M gene knockout mice. TNF- mRNA levels in the liver were determined in ₂M / and WT mice at 1.5 h following injection of LPS or saline (when the difference in plasma concentration of TNF- between ₂M / and WT mice was observed). ₂M / and WT mice injected with saline showed low TNF- (Fig. 5) mRNA levels that did not significantly differ between groups. LPS led to a significant elevation in hepatic TNF- (Fig. 5) mRNA levels at 1.5 h postinjection. There was no significant difference in hepatic TNF- (Fig. 5) mRNA levels between ₂M / and WT mice 1.5 h after injection of LPS.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   A: detection of TNF- and -actin mRNAs by RT-PCR from RNA isolated from the liver of 2M / and WT mice 1.5 h after intraperitoneal injection of LPS (2.5 mg/kg) or pyrogen-free saline. B: quantification of the band intensities normalized for the housekeeping gene, -actin, showed significant increases of hepatic TNF- mRNA levels in 2M / and WT mice treated with LPS compared with their saline-injected counterparts. Samples of the hepatic RNA from 5 2M / mice injected with LPS, 3 2M / mice injected with saline, 6 WT mice injected with LPS, and 3 WT mice injected with saline were analyzed in 2 separate assays. Data are presented as means ± SE. * Significant difference between 2M / mice injected with LPS and saline; between WT mice injected with LPS and saline, P < 0.05.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In the present study, the role of ₂M in LPS-induced fever and cytokine responses was studied. ₂M / mice developed a short-term markedly attenuated fever in response to LPS, suggesting that this plasma protease inhibitor is essential for normal development of LPS-induced fever in mice. There was no difference in plasma levels of IL-1 and IL-6 between ₂M / and WT mice after injection of LPS. Plasma concentration of TNF- shortly after LPS challenge was significantly lower in the ₂M / mice compared with their WT counterparts. No difference in the hepatic TNF- mRNA levels between ₂M / and WT mice treated with LPS suggests augmented clearance of TNF- in ₂M / mice. No difference in fever between ₂M / and WT mice following turpentine injection suggests that ₂M / mice can increase their T_b in response to other stimuli to the same extent as the control animals, indicating that ₂M deficiency does not impair these animals' ability to mount an adequate thermogenic response. In view of the latter, an interesting observation is that the magnitude of the initial stress-induced rise in T_b (evoked by the injection procedure) was significantly lower in ₂M / mice compared with WT mice, regardless of whether they were injected with LPS or saline (Fig. 1). This observation may suggest that ₂M is also involved in the mechanisms of the development of stress-induced fever, and, to study the mechanisms of this involvement, the data have to be confirmed in "controlled" experiments, such as those in which animals are exposed to an identical novel environment.

Successful targeting of the ₂M gene and generation of the ₂M / mice (37, 38) provided a valuable tool to dissect the role of this protease inhibitor in febrile and cytokine responses induced by LPS. Unfortunately, this transgenic model is not ideal, because mouse plasma, unlike plasma of humans and other mammals, contains two different types of -macroglobulins: the tetrameric ₂M and the monomeric or single-chain murinoglobulins, the in vivo function of which is still unclear [for detailed discussion, see Umans et al. (37)]. Although levels of murinoglobulins in the plasma of adult nonpregnant ₂M / mice are unchanged compared with WT animals (37), we cannot completely exclude the possibility that, during development of the inflammatory response induced by either LPS or turpentine, murinoglobulins may, in part, functionally replace ₂M in the ₂M / mice. Therefore, as in any experiment involving gene knockout animals, caution should have been taken in the interpretation of the obtained data.

We reported recently that human ₂M induces moderate fever (~0.5°C) in mice when injected intravenously in amounts similar to or even smaller than those observed during the development of the systemic inflammatory response (14). Furthermore, it was shown that 1 h after intravenous injection, human ₂M induces moderate (smaller compared with LPS induced) but significant increase in plasma bioactivity of TNF- (14), suggesting that TNF- may mediate the pyrogenic effect of human ₂M in mice. It is possible that exogenous ₂M either inhibits rapid degradation of TNF-, which is constitutively produced under basal "nonpathological" conditions (28, 33), or stimulates TNF- synthesis, or both. Although the mechanisms of the increase in plasma TNF- level induced by intravenous injection of exogenous ₂M in mice have to be investigated, these findings coincide with the results of the present study, and, taken together, they suggest an important role for ₂M in the mechanisms of fever development.

The present data also correlate well with the observations of Umans et al. (37), who reported that ₂M / mice are resistant to the lethal effects of LPS. However, although LPS-induced fever was lower in ₂M / mice, the other "signs" of sickness syndrome, i.e, body weight loss, decrease in food intake, and decrease in motor activity, were identical to those in control WT mice (Fig. 2). Thus, in terms of induction of anorexia and lethargy, ₂M / mice are equally sensitive to LPS. These unexpected observations suggest that although playing an important role in the development of fever, ₂M probably is not involved in anorexia and lethargy induced by LPS in mice. In view of the results obtained in the present study, indicating that ₂M is likely to be involved in LPS-induced fever through the inhibition of TNF- clearance, these data are in agreement with the results obtained in the TNF double-receptor knockout mice, showing that TNF- does not mediate LPS-induced anorexia and lethargy (27). However, one should apply some caution in interpretation of our results. It is possible, due to inherent redundancy in hormone/cytokine action, that in both cases knockout mice compensate for removal of one gene by increased action of another in the regulation of some aspects of the acute phase response to LPS.

Although ₂M is not an acute phase protein in mice (unlike in rats and some other species), during experimental inflammation, moderate changes in plasma levels of murine ₂M do occur (1, 19). LPS in a dose of 20 µg/mouse (~1.0 mg/kg) induced a significant increase in plasma ₂M concentration 24 and 48 h after intraperitoneal injection (19). It is impossible to compare these results with the data obtained in our study, because different strains of mice and doses of LPS were used. However, the study of Isaac et al. (19) is important to this discussion as it shows that LPS can induce ₂M production in mice. Interesting data were obtained by LaMarre et al. (24), who showed that expression of LRP in murine macrophages can be markedly decreased by LPS. These data indicate that LPS is a natural regulator of the ₂M/LRP system: it can increase ₂M production and, at the same time, suppress the expression of the ₂M receptor.

We hypothesized initially that the effect of ₂M on the febrile response depends on whether this protease inhibitor facilitates or inhibits clearance of the "major" endogenous pyrogens, e.g., IL-1, IL-6, or TNF-. This hypothesis was supported by the vast amount of literature indicating that plasma ₂M is a broad-spectrum protease inhibitor and a cytokine-binding protein and carrier at the same time (5-7, 9, 13, 20, 25, 29, 42, 44). Extrapolation from in vitro studies suggested that depending on the conformational state of the ₂M molecule, its function in regulation of cytokine metabolism and, therefore, thermoregulatory febrile response could be different. As discussed by LaMarre et al. (25) and mentioned earlier in text, ₂M in its native form may protect bound cytokine from proteolytic degradation by functioning as a cytokine carrier and, therefore, lengthen its plasma half-life. On the other hand, proteinase-activated ₂M (that is recognized by LRP) may play an important role in the processes of cytokine clearance.

Results of the present study suggest an augmented rate of TNF- clearance in ₂M / mice. Several studies in vitro demonstrated binding of TNF- to native and protease-activated ₂M (9, 20, 42, 44). It has been shown that neither native nor protease-activated ₂M reduces biological activity of TNF- (44). Although TNF- binds native ₂M with lower affinity than protease-activated ₂M, we hypothesize that under most conditions and particularly during LPS-induced fever, native ₂M is more important in regulating TNF- clearance than protease-modified ₂M. First, our data suggest an augmented clearance of TNF- in ₂M / mice. If protease-activated ₂M played a significant role in LRP-mediated TNF- clearance in vivo, we would expect to observe the opposite (i.e., higher plasma concentration of TNF- in ₂M / mice compared with WT controls). Second, native ₂M is the predominant form of ₂M present in the plasma and in the extravascular microenvironments. In contrast, protease-activated ₂M under most conditions is present only at trace levels. We hypothesize that native ₂M, as a broad-spectrum protease inhibitor, protects bound TNF- from proteolytic degradation and, therefore, lengthens its plasma half-life. This conclusion is supported by the evidence that TNF- is efficiently destroyed by proteases released from activated polymorphonuclear neutrophils and that after proteolytic cleavage, TNF- fragments lack any TNF--like cytotoxic activity (31, 39).

Our data do not support the observations of Hochepied at al. (16), who showed an identical rate of clearance of injected TNF- in ₂M / and WT mice. Both data are difficult to reconcile, especially because the same mice were used in both studies. Presumably, the clearance mechanisms of injected TNF- are to some extent different from those of LPS-induced, endogenously produced TNF-. The latter could be significantly affected by other responses (physiological and humoral) induced by LPS and not mimicked by injection of TNF- alone.

The lack of a difference in plasma levels of IL-1 and IL-6 between ₂M / and WT mice after injection of LPS does not support the hypothesis that fever in ₂M / mice is attenuated due to decreased production or increased clearance of IL-1 or IL-6. These data were somewhat unexpected in view of the evidence obtained in experiments in vitro suggesting that ₂M is one of the major IL-1- and IL-6-binding plasma proteins (6, 7, 29). Data obtained in the present study reveal the limitations of a direct extrapolation from these in vitro studies and suggest that ₂M may not play a significant role in the regulation of IL-1 and IL-6 clearance during LPS-induced fever in mice.

Although the precise role of TNF- in fever is still unresolved, we propose that the putative mechanism of ₂M involvement in LPS-induced fever is through the inhibition of TNF- clearance. Indeed, depending on the experimental conditions and animal species (and probably also strains) used, this cytokine can act as an endogenous pyrogen as well as an endogenous antipyretic. Injection or infusion of TNF- appears to induce fever in several species (for review, see Refs. 22, 23, 35). When the actions of endogenous TNF- were blocked experimentally, this cytokine showed antipyretic properties in rats and mice and pyrogenic activity in rats, rabbits, guinea pigs, and humans (for recent reviews, see Refs. 15, 32, 35). The following arguments supporting the above hypothesis should be mentioned: 1) TNF- is a major proinflammatory cytokine, and its lower plasma levels indicate a smaller systemic inflammatory response. 2) We observed that intravenous injection of recombinant murine TNF- (1 µg) induces moderate (1°C) fever in C57BL/6J mice (A.V. Gourine and M.J. Kluger, unpublished observations) in mice of the same strain that was used in the present study. 3) It appears that, when TNF- is not involved in fever, ₂M is also without effect. Neither TNF- (27) nor ₂M is involved in fever induced by local inflammation (injection of turpentine). 4) Roth et al. (32) recently reported that in guinea pigs neutralization of TNF- by treating the animals with its type 1 soluble receptor significantly attenuates the second phase of the LPS-induced febrile response without affecting its initial phase. Interestingly, neutralization of TNF- by its type 1 soluble receptor affects the development of fever in guinea pigs in the same way as inactivation of the ₂M gene affects fever in mice, i.e., resulting in a marked attenuation of the late phase of the febrile response without markedly affecting the initial phase.

In conclusion, results of the present study suggest that ₂M is important for the normal development of LPS-induced fever in mice and that putative mechanism of ₂M involvement in LPS-induced fever is through the inhibition of TNF- clearance. We speculate that ₂M serves as an inhibitor of proteinases (e.g., elastase) responsible for rapid degradation of TNF- and, possibly, of other pyrogenic cytokines (but not IL-1 or IL-6).

Perspectives