INTRODUCTION

TOP ABSTRACT INTRODUCTION METHOD RESULTS DISCUSSION REFERENCES

PREVIOUS STUDIES HAVE ESTABLISHED that the nonspecific host response to inflammation is mediated largely by the release of proinflammatory cytokines such as interleukin-1 (IL-1), tumor necrosis factor- (TNF-), and interleukin-6 (IL-6) (7). However, because of redundancy in the actions of these three cytokines, the relative contributions of TNF-, IL-1, and IL-6 in mediating the inflammatory response in vivo remains unclear. Two previous reports have demonstrated that blocking IL-1 binding to its type I receptor (8, 22) or inhibiting IL-6 with monoclonal antibodies (21) leads to a marked attenuation of the inflammatory response to a turpentine abscess. Similar results have been seen with IL-1, type I IL-1 receptor, and IL-6 knockout mice (4, 12, 13). However, a protective effect with blockade of the IL-1 type II receptor (IL-1RII) (22) has not been demonstrated. The collective findings of these studies suggest that the anorexia, weight loss, and the hepatic acute phase response to a turpentine abscess are mediated at least in part by an endogenous IL-6 response that is initiated by IL-1 binding to its type I receptor (24).

In contrast, the effect of TNF- blockade on the magnitude of the inflammatory response to a turpentine abscess remains controversial. Leon et al. (14) and Gershenwald et al. (8) were unable to demonstrate any differences in the inflammatory response to a turpentine abscess in TNF receptor knockout mice or in wild-type mice treated systemically with a TNF- antibody, respectively. In contrast, Luheshi et al. (15) and Cooper et al. (2) reported that they could block the plasma IL-6 and febrile responses with anti-TNF- antibodies. These studies suggest that the contribution of TNF- to specific components of the acute phase response to a turpentine abscess remain unresolved.

IL-1 signaling is antagonized by IL-1ra, a third member of the IL-1 superfamily, which binds to the type I IL-1 receptor with similar affinity to IL-1 but does not transduce a signal (3). Exogenous administration of IL-1ra reduces the pathological consequences of IL-1 in gram-negative bacteremic shock, rheumatoid arthritis, and ischemic injury (1, 6, 18). Studies from IL-1ra knockout and overexpressing mice have demonstrated that an endogenous IL-1ra response plays a critical role in protecting against lethality to endotoxemic shock and collagen-induced arthritis (10, 17). Overexpression of IL-1ra also increases sensitivity to listeriosis (10, 17). Yet little information is available concerning the role played by endogenously produced IL-1ra in the acute phase responses to an acute inflammatory insult.

In the present report, we evaluated the individual contribution of IL-1 and TNF- signaling and IL-1ra on the host nonspecific inflammatory response to a turpentine abscess. Transgenic mice lacking either a functional IL-1 type I receptor (IL-1RI^/), TNF p55 receptor (TNFR-I^/), both functional receptors (IL-1RI^//TNFR-I^/), or IL-1ra protein (IL-1ra^/), or mice overexpressing IL-1ra protein (IL-1ra^tg) were employed in a turpentine abscess model of acute inflammation. The present study demonstrates that the host nonspecific inflammatory response (anorexia, weight loss, hepatic acute phase protein, and plasma IL-6) to a turpentine abscess is mediated by IL-1 binding to its type I receptor and is antagonized by the endogenous expression of IL-1ra. This is the first in vivo report using transgenic animals establishing that an endogenous IL-1ra response regulates not only the magnitude of the inflammatory response to a turpentine abscess, but also its outcome. In contrast, TNF signaling only plays a role in the inflammatory response to a turpentine abscess in the absence of IL-1 signaling.

	METHOD

TOP ABSTRACT INTRODUCTION METHOD RESULTS DISCUSSION REFERENCES

A total of 191 B6×129, C57BL/6, or B6×CBA mice of mixed sex (equally distributed between male and female) weighing 20-27 g was housed in a temperature-controlled room with alternating 12:12-h light-dark cycles. Creation of the TNFR-I^/ mice has been previously described by Pfeffer and colleagues (23). The IL-1RI^/ mice were generated by interrupting the continuity of the p80 gene with the insertion of the neo gene into the germ line (13). To generate the double knockout (IL-1RI^//TNFR-I^/) mice, TNFR-I^/ and IL-1RI^/ mice were crossed and sister-brother F₁ progeny were mated to select for the double knockout genotype. All three strains of mice were maintained on a B6×129 background by sister-brother homozygote matings. Wild-type B6×129 mice were originally obtained from heterozygote crossings, and homozygous wild-type mice were propagated by sister-brother homozygote matings.

IL-1ra knockout mice were obtained on a C57BL/6 background (10). Because the knockout phenotype is associated with reduced numbers of offspring and growth, heterozygote crossings were performed. For these studies, heterozygote and wild-type mice were employed in the studies as controls. To assure that the starting weights among the animals were similar, the animals were body weight matched. IL-1ra overexpressors were obtained on a B6×CBA background and were also maintained by heterozygote crossings.

IL-1RI^/ and IL-1RI^//TNFR-I^/ mice were obtained from Hoffmann-LaRoche (Nutley, NJ). These animals were allowed to equilibrate 7-10 days, during which time the animals were provided free access to chow from 1700 to 0900 daily. The TNFR-I^/, wild-type B6×129, IL-1ra^/, wild-type C57BL/6, IL-1ra^tg, and B6×CBA animals were bred and maintained at the University of Florida College of Medicine Health Science Center Animal Resources Department. In all cases, body weight and food intake were monitored daily at 0900. The protocol was approved by the Institutional Care and Use Committee of the University of Florida, and the laboratory adheres to the Guiding Principles of Laboratory Animal Care as promulgated by the American Physiological Society.

Experimental designs. To evaluate the role that IL-1 and TNF signaling plays in the acute phase response, two separate experiments were conducted. The need to divide the study into two experiments was based on the limited availability of the TNFR-I^/ and IL-1RI^//TNFR-I^/ knockout mice. In the first experiment, 15 IL-1RI^/, 10 TNFR-I^/, and 13 wild-type B6×129 mice were studied. In the second experiment, 17 IL-1RI^/, 15 IL-1RI^//TNFR-I^/, and 13 B6×129 wild-type mice were used. On day 0 of the experiment, two to four mice in each group were killed by cervical dislocation; the remaining mice received bilateral injections of 100 µl sc of steam-distilled turpentine into each hindlimb. On postturpentine day 1, all of the mice were bled from their retroorbital plexus (50 µl). Mice were killed by cervical dislocation on postturpentine day 3 and day 5. At death, venous blood was obtained by cardiac puncture and collected via a heparinized syringe, and plasma was stored at 70°C until assayed. Body weight and food intake were recorded daily starting 3 days before the study until the day of death.

To evaluate the role of IL-1ra in the acute phase response, two additional experiments were performed. In the first experiment, 36 wild-type C57BL/6, 18 IL-1ra^+/ heterozygote, and 18 IL-1ra^/ homozygote knockout mice equally distributed among males and females were studied. Six IL-1ra^+/, six IL-1ra^/, and twelve wild-type C57BL/6 mice were killed on day 0 and served as healthy controls. The remaining 12 IL-1ra knockout and heterozygote, and 24 wild-type mice received bilateral injections of 100 µl sc of steam-distilled turpentine. On postturpentine day 1, all of the mice were bled from their retroorbital plexus (50 µl), whereas one-half of the remaining mice in each group were killed by cervical dislocation on postturpentine day 3 and day 5.

In the second study, 18 B6×CBA wild-type and 18 B6×CBA IL-1ra overexpressing mice (IL-1ra^tg) of mixed sex were studied. Six in each group were killed on day 0 to serve as controls, while the remainder received a turpentine abscess. All of the mice were bled on day 1 from their retroorbital plexus and one-half of the remaining animals in each group were killed on day 3 and day 5.

Analytic assays. Plasma IL-6 levels were measured by ELISA using commercially available antibodies (Endogen, Boston, MA). Recombinant murine IL-6 was used as standard, and sensitivity of the assay was between 32 and 100 pg/ml. The murine acute phase protein, serum amyloid P (SAP), was measured by immunoelectrophoresis (8). Five microliters of plasma samples diluted 1:3 to 1:10 were added to wells in a 1% agarose gel containing 0.3% of rabbit anti-murine amyloid P antiserum (Calbiochem, San Diego, CA). Gels were blotted, dried, and stained in 0.2% Coomassie brilliant blue. The quantities of acute phase proteins were calculated by comparison of the heights of the peaks with those of a commercial murine standard (Calbiochem). Murine serum amyloid A (SAA) was determined using a commercial ELISA (BioSource International, Camarillo, CA). The sensitivity of the assay is 20 ng/ml. Total seromucoid was determined chemically from the differential solubility of heavily glycosylated serum proteins in 1.2 M perchlorate-2% phosphotungstic acid (9). The seromucoid fraction is comprised of >90% ₁-acid glycoprotein.

Statistics. Data are presented as means ± SE. Data were analyzed using a commercially available statistics package on a Pentium-based personal computer (SigmaStat, Jandel Scientific, Santa Clara, CA). There were no statistical differences in the sex distribution among the groups, and no effort was made to examine specific sex-related changes in the response to the turpentine abscess. Differences in food intake and body weight among groups of mice after a turpentine abscess were analyzed by two-way ANOVA (time in days after the abscess vs. genetic background). Similarly, changes in plasma IL-6 and hepatic acute phase reactant concentrations were compared by two-way ANOVA (days 1, 3, and 5 vs. genetic background). Post hoc analysis was performed using the all pairwise multiple comparison procedure, Student-Newman-Keuls method. Survival was evaluated statistically using a log-rank test comparing Kaplan-Meier survival curves. In all cases, statistical significance was determined at the 95% confidence interval.

	RESULTS

TOP ABSTRACT INTRODUCTION METHOD RESULTS DISCUSSION REFERENCES

IL-1 and TNF receptor signaling studies. Data from the two experiments have been combined for brevity. In both experiments, wild-type B6×129 and IL-1RI^/ mice were studied, and results obtained from the two experiments were very similar. Subcutaneous administration of turpentine in wild-type B6×129 mice resulted in decreased food intake, moderate weight loss, and an increase in the concentrations of several acute-phase response proteins and IL-6 (Figs. 1-3). In contrast, when IL-1RI^/ mice were injected with turpentine, they were significantly (P < 0.05) less anorexic and their weight loss was significantly (P < 0.05) diminished compared with wild-type mice or TNFR-I^/ mice (Fig. 1) on days 1 and 2 (both P < 0.05). In actuality, the IL-1RI^/ knockout mice appeared to gain weight on the first postturpentine day, despite no significant change in food intake. Surprisingly, TNFR-I^/knockout mice had a statistically significant greater degree of anorexia (P < 0.05) and early weight loss (P < 0.05) on days 1 and 2 than wild-type mice. The animals lacking both functional TNFR-I and IL-1RI receptors had an intermediate food intake and body weight change response similar to wild-type animals. It should be noted that we did not analyze data from male and female mice separately, although during the course of the study, we did not note any obvious differences in the magnitude of the responses between male and female animals.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Changes in food intake (A) and body weight (B) in mice administered turpentine subcutaneously. Values represent means ± SE. , interleukin-1 type I receptor (IL-1RI/) mice; , tumor necrosis factor type I receptor (TNFR-I/) mice; , both IL-1 type I receptor and TNF type I receptor (IL-1RI//TNFR-I/); and , wild-type B6×B129 mice. * Differences in food intake and body weight are statistically significant compared with wild-type control mice on that day.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Changes in plasma IL-6 (A) and amyloid A (B) levels in mice with turpentine abscess. * Plasma IL-6 and amyloid A concentrations were statistically different than wild-type mice at same time point.  Statistical difference between IL-1RI//TNFR-I/ and IL-1RI/ mice at same time point. IL-1RI/ mice had a markedly diminished IL-6 and amyloid A response compared with wild-type animals at all 3 time points (both P < 0.01). TNFR-I/ mice had a significantly increased IL-6 response on day 1, whereas the plasma amyloid A response was generally unaffected compared with wild-type animals (with the exception of day 3 when it was reduced). IL-6 (on day 3 and 5) and amyloid A responses (on day 5) were further diminished in IL-1RI//TNFR-I/ mice compared with IL-1RI/ mice (both P < 0.05).

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Changes in amyloid P (A) and seromucoid (B) concentrations in mice with turpentine abscess. IL-1RI/ mice had a markedly diminished response compared with wild-type animals by 2-way ANOVA (both P < 0.05). Amyloid P concentrations were significantly reduced in TNFR-I/ mice compared with wild-type mice, whereas seromucoid levels were unaffected. Both seromucoid and amyloid P concentrations were further diminished in IL-1RI//TNFR-I/ mice compared with p80/ mice (both P < 0.05). * Plasma amyloid P and seromucoid concentrations were statistically different than wild-type mice at same time point.  Statistical difference between IL-1RI//TNFR-I/ and IL-1RI/ mice at same time point.

Several components of the hepatic acute phase response were similarly attenuated. Plasma IL-6 levels in IL-1RI^/ knockout mice were significantly diminished compared with wild-type mice at all time points (P < 0.05) (Fig. 2A). In the same manner, SAA concentrations were also reduced postturpentine injection in the IL-1RI^/ mice on days 1 and 3 (P < 0.05), whereas amyloid P concentrations were reduced on days 3 and 5 (P < 0.05), when matched to the wild-type mice (Figs. 2 and 3). Seromucoid concentrations were reduced at all time points in the IL-1RI^/ mice (P < 0.05) vs. the wild-type controls.

Interestingly, TNFR-I^/ knockout mice exhibited a variable acute phase protein response. Plasma IL-6 concentrations rose more in the TNFR-I^/ knockout mice compared with the wild-type mice only on the first postturpentine day (Fig. 2). Seromucoid concentrations did not vary between TNFR-I^/ knockout and wild-type mice, whereas amyloid P concentrations were reduced only on day 5 (P < 0.05) and amyloid A concentrations were reduced on day 3 (P < 0.05).

Mice lacking both functional IL-1RI and TNFR-I receptors exhibited further reductions in their serum acute phase responses compared with the IL-1RI^/ mice. This was most apparent in plasma IL-6 concentrations, which were significantly reduced (all P < 0.05) on days 3 and 5. Some more modest further reductions were also seen in amyloid A, amyloid P, and seromucoid concentrations in the TNFR-I knockout mice, but in most cases, they did not reach statistical significance.

Studies in IL-1ra knockout and overexpressing mice. Before study, all of the animals appeared to be healthy. IL-1ra^/ knockout mice show some runting during growth and development (10), but in the present study, young adult mice of similar body weights were employed. IL-1ra^/ knockout mice had basal IL-6 concentrations significantly higher than either heterozygote (IL-1ra^+/) or wild-type mice on a similar background (179 ± 90 vs. <32 vs. <32 pg/ml, P < 0.05). Similarly, basal amyloid P concentrations were also significantly increased in IL-1ra^/knockout mice compared with wild-type animals, indicating these animals had an ongoing acute phase protein response (170 ± 65 vs. 72 ± 26 vs. 35 ± 14 µg/dl, P < 0.01). However, food intake in the 3- to 5-day period before a turpentine abscess did not reveal a significant difference in basal food intake. Baseline plasma IL-6 and acute phase protein levels were not different in mice overexpressing IL-1ra than they were in their wild-type controls (Fig. 4).

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Changes in plasma IL-6 (A) and amyloid P (B) concentrations in IL-1 receptor antagonist (IL-1ra) knockout and overexpressing mice after a turpentine abscess. Values represent means ± SE. IL-6 concentrations were significantly increased in IL-1ra/ (days 1 and 3) and significantly reduced in IL-1ra protein (IL-1ratg) mice (days 1 and 3) compared with their background controls (both P < 0.05). However, amyloid P concentrations were only reduced in IL-1ra overexpressing mice (P < 0.05 on days 1 and 5). * Plasma IL-6 and amyloid P concentrations were statistically different than wild-type mice at same time point.

Induction of a turpentine abscess produced anorexia and weight loss in all of the groups of mice. Surprisingly, 50% of the IL-1ra knockout mice (6 of 12) died spontaneously by day 5, whereas there was no mortality in any of the other experimental groups. Two IL-1ra knockout mice died within the first 24 h, another three mice died between days 1 and 2, and a final knockout mouse died between days 2 and 3. Because of this mortality, it was difficult to perform statistical analyses on biochemical analyses, food intake, and body wt changes after day 3 as the number of surviving IL-1ra knockout animals was small (n = 3) and potentially not reflective of the entire group.

Between wild-type B6×CBA mice and similar animals overexpressing IL-1ra, there was no significant difference in the amount of weight loss or food intake (Fig. 5). There was a trend toward greater weight loss and reduced food intake in homozygous IL-1ra knockout mice compared with either the heterozygote or wild-type C57BL/6 mice (Fig. 6), although the differences were modest.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Changes in food intake (A) and body weight (B) in IL-1ra overexpressing mice administered turpentine subcutaneously. Values represent means ± SE. , IL-1ratg mice; , wild-type B6×CBA mice.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Changes in food intake (A) and body weight (B) in IL-1ra knockout mice administered turpentine subcutaneously. Values represent means ± SE. Data are only presented for first 3 days, as mortality in the IL-1ra/ mice was 50% after that time point, and values from remaining mice would not be representative. , IL-1ra/ mice; , IL-1ra+/ mice; and , wild-type B6×B129 mice.

The plasma IL-6 response varied significantly depending upon the degree of IL-1ra expression (Fig. 4). Peak IL-6 concentrations were eight times higher in IL-1ra knockout mice on day 1 (P < 0.01) than in either the heterozygote or wild-type animals. Similarly, IL-6 concentrations were reduced by 40% in the IL-1ra overexpressing mice compared with wild-type animals on days 1 and 3 (P < 0.05).

In contrast, the SAP response was unaffected by the lack of IL-1ra expression, whereas overexpression of IL-1ra reduced the peak amyloid P response by 30% (P < 0.05 on days 1 and 5) (Fig. 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHOD RESULTS DISCUSSION REFERENCES

By using genetically engineered mice lacking either functional IL-1RI or TNFR-I receptors and mice lacking IL-1ra protein or mice overexpressing IL-1ra protein, the present study clearly demonstrates that the host nonspecific inflammatory response to a turpentine abscess is mediated principally by IL-1 binding to its type I receptor. The current studies extend these findings by 1) demonstrating that TNF signaling via its type I receptor plays no role in the innate immune response to a turpentine abscess in the presence of IL-1 receptor signaling and 2) showing that the magnitude of the nonspecific inflammatory response results from the balance between IL-1 signaling and IL-1ra expression. In the latter case, the absence of IL-1ra expression is associated not only with an exaggerated inflammatory response, but also with significant mortality.

Previous studies have demonstrated that either blocking IL-1 binding to its type I receptor or blocking IL-6 with monoclonal antibodies attenuates the nonspecific inflammatory response and that blocking TNF- or the type II IL-1 receptor does not confer protection (8, 14, 22). By employing IL-1RI^/, TNFR-I^/, or double receptor knockout mice, we can now further demonstrate that endogenously produced IL-1 plays the significant role in the nonspecific inflammatory response to a turpentine abscess. Coupled with the earlier findings of Fantuzzi et al. (5) and Kopf et al. (12), who demonstrated an attenuated inflammatory response to a turpentine abscess in IL-1 and IL-6 knockout mice, respectively, it can be concluded that many of the components of the host response to a turpentine abscess are mediated via IL-1 signaling through the IL-1RI and the subsequent production of IL-6.

We can also confirm that in mice lacking a functional TNFR-I, the anorexia, weight loss, and acute phase response secondary to a turpentine abscess were generally unaffected, and the anorexia, weight loss, and plasma IL-6 responses may actually have been modestly augmented. These results confirm similar findings on body weight and food intake in TNFR-I, TNFR-II double receptor knockout mice reported earlier by Leon et al. (14), and, importantly, can extend their observations to other components of the hepatic acute phase protein response. Thus we can confirm that in a turpentine abscess model of acute inflammation, TNF- signaling contributes little to the inflammatory responses.

A novel and unexpected observation of the present study was that despite no significant reduction in any of the nonspecific host responses in mice lacking a functional TNFR-I receptor, mice lacking both a functional TNFR-I and IL-1RI receptor had a trend toward a more attenuated hepatic acute phase protein response than similar animals lacking only a functional IL-1RI. A ready explanation for this observation is lacking, given the failure of TNFR-I knockout mice to show any consistent attenuation in the acute phase response. However, there is considerable evidence to suggest that IL-1 and TNF- potentiate each other's tissue responses (20, 25). Administration of TNF- and IL-1 together, in doses insufficient to produce pathological responses when given alone, produce significant pathology. Therefore, it may be that endogenous production of TNF- in this model is in itself insufficient to elicit a response in the presence of IL-1 signaling but may be sufficient to elicit some response in its absence.

The findings further show that IL-1 signaling of the acute phase response is balanced by expression of IL-1ra. Mice lacking a functional IL-1ra exhibited some runting during the prestudy period, a finding reported earlier by Hirsch and colleagues (10). More interestingly, these mice exhibited biochemical evidence of a nonspecific inflammatory response despite their housing in a specific pathogen-free environment. IL-1ra^/ knockout mice had increased basal levels of IL-6 and evidence of an hepatic acute phase protein response as reflected by increased amyloid P concentrations. It is tempting to speculate that the runting and increased basal IL-6 and amyloid P production in these mice are indicative of a functional role for IL-1ra in antagonizing some basal IL-1 production in the otherwise healthy mouse. This would imply that the plasma IL-1ra concentrations (0.5-1 ng/ml) found in a normal mouse play a critical role in normal growth and homeostasis, and are essential for minimizing an acute phase response in the absence of an identifiable inflammatory locus.

Mice lacking IL-1ra experienced a significantly exaggerated plasma IL-6 response after a turpentine abscess. In fact, 50% of the IL-1ra knockout mice expired within 5 days after a turpentine abscess. Although a detailed necropsy and histological examination of the expired mice were not performed, the animals appeared to have died from shock and/or an exaggerated inflammatory response. It should be noted that the doses of turpentine used in the present study (100 µl) are 10 times greater than the concentrations used in other reports (2, 15).

The findings in this regard are similar to the findings reported originally by Hirsch et al. (10) who observed that IL-1ra knockout mice were more susceptible to endotoxin-induced lethality than were wild-type mice. Horai et al. (11) recently observed that cox-2 expression was increased in the diencephalon of IL-1ra knockout mice after a turpentine abscess, whereas it was reduced in IL-1 knockout mice, suggesting that central nervous system prostanoid production is also regulated by this balance between IL-1 and IL-1ra expression.

In contrast, overexpression of IL-1ra resulted in an attenuation of the plasma IL-6 and acute phase response but no further improvements in either food intake or body weight. These findings are therefore consistent with earlier studies demonstrating that systemic administration of IL-1ra can attenuate both the plasma IL-6 and febrile responses to turpentine (15). However, Lundkvist and colleagues (16) noted that overexpression of IL-1ra in the central nervous system did not reduce the plasma IL-6 response to endotoxin, suggesting that IL-6 synthesis is more dependent on peripheral and not central overexpression of IL-1ra.

The present studies would further suggest that the hepatic acute phase response is more sensitive to the balance between IL-1 signaling and IL-1ra expression than is either food intake or body weight. Consistent with this conclusion is our earlier observation that lower quantities of exogenously administered IL-1 are required to elicit an acute phase response than are necessary to alter food intake and body weight (19). In conclusion, these studies demonstrate that activation of the acute phase response to a turpentine abscess is dependent primarily on IL-1RI signaling, and it is the balance between endogenous IL-1ra and IL-1 expression that determines the magnitude of the inflammatory response. Inflammation-induced increases in IL-1ra expression serve to modulate the magnitude of the inflammatory response and apparently play a critical role in its successful outcome.