Sepsis remains a serious clinical problem, with an extraordinarily high morbidity and mortality. Several different strategies have been attempted to improve the survival in septic patients, but only the recent clinical trial with activated protein C has proven efficacious (1). There are several possibilities why these previous clinical trials with immunomodulator therapy failed (2–4). One potential reason is that all of the patients received the same therapy regardless of risk stratification. Patients with sepsis represent a heterogeneous population, and it is possible that some suffer from too much inflammation whereas others lack sufficient inflammation to effectively eradicate the infectious source.
Several clinical studies have demonstrated that the circulating plasma levels of interleukin (IL)-6 predict outcome (5,6). Specifically, patients with higher levels of IL-6 have a significantly increased mortality. Among the multiple different studies of the various cytokines, IL-6 has reproducibly been the best predictor of mortality. We have previously published in the cecal ligation and puncture model of sepsis that animals with the highest mortality have the highest levels of IL-6 (7,8). As the mortality decreases, the IL-6 levels decrease also. The plasma levels of IL-6 also quickly become elevated within a few hours of the surgical procedure inducing the sepsis. Other investigators have observed similar effects of higher levels of IL-6 in mice with greater mortality (9).
In this present study, we sought to determine whether the plasma levels of IL-6 observed 6 h after the initiation of sepsis could predict outcome. This study was designed to specifically test the hypothesis that mice, which die during the early stages of sepsis, have excessive inflammation as measured by plasma levels of IL-6. We additionally looked at other parameters, including body weight and the hematologic profile to determine other variables that could predict survival. Finally, based on these data, we attempted to perform a therapeutic intervention.
MATERIALS AND METHODS
Experimental animal model
Cecal ligation and puncture (CLP) was performed in adult female BALB/c mice (Harlan, Indianapolis, IN) using our previously defined protocol (7) modified after the original description (10). For these experiments, the mice were anesthetized with ketamine/xylazine (87 and 13 mg/kg), and a 2-cm incision was made in the lower abdomen. The cecum was exteriorized and ligated with 4-0 silk just below the ileocecal valve. The cecum was punctured twice with a 21-gauge needle and a small amount of stool extruded to ensure that the wounds were patent. The abdominal wall was closed with sutures, and the skin incision was closed with either surgical clips or wound glue (Nexaband, Veterinary Products Laboratories, Phoenix, AZ). One milliliter of warm normal saline was injected subcutaneously, and the animals were placed in a 37°C incubator for 30 min to help correct the hypothermia routinely induced by the ketamine/xylazine anesthesia. All mice were given antibiotic therapy consisting of 25 mg/kg imipenem given subcutaneously in 1 mL of D5W every 12 h that was continued for 5 days. The D5W was used to provide fluids because the mice did not eat or drink adequately for the first few days after surgery. In selected experiments, mice were randomized to have either resection of the necrotic cecum or antibiotic therapy alone. This procedure was completed within 9 h of the original CLP procedure. The initial CLP was performed with ketamine/xylazine anesthesia, but for the surgical resection, the mice were anesthetized with isoflurane. Isoflurane was used for the second surgical procedure because the mice became refractory to ketamine/xylazine. The abdomen entered through the original incision, the cecum was resected just below the suture used for ligation, and the abdominal cavity was flushed with 5 mL of sterile normal saline. The abdomen was then closed as in the original procedure, using wound glue for skin closure. The experiments in this article were performed in accordance with National Institutes of Health guidelines. All animal protocols were approved for use by the Institutional Committee on Care and Use of Animals.
Collection of specimens
Six hours after the onset of surgery, the distal tip of the tail was removed, and 20 μL of blood was obtained by drawing up the blood into a pipette tip that had been rinsed with EDTA (169 mM tripotassium salt). This blood was immediately diluted 1:10 in phospate-buffered saline with a 1:50 dilution of the EDTA, and the cells were removed by centrifugation. This material was used for measuring IL-6 in the rapid enzyme-linked immunosorbent assay (ELISA) described below. For the first 5 days, mice were weighed daily prior to the administration of the antibiotics. Twenty-four hours after the onset of surgery, an additional 20 μL of blood was used to perform a complete blood profile. The blood was collected using the EDTA rinsed pipette tip.
A complete blood profile was performed using the hemavet Mascot series 1500. This is an automated instrument that provides a differential of the peripheral blood constituents as well as a platelet count and red blood cell parameters (11,12).
The rapid ELISA was developed for detecting murine IL-6. This ELISA was based on commercially available matched antibody pairs (R & D Systems, Minneapolis, MN) and represents a modification of ELISAs routinely run in the laboratory (12,13). Briefly, ELISA plates are coated overnight with the monoclonal rat αIL-6. The next morning, all solutions are brought to a temperature of 37°C. The plates were blocked with 2% bovine serum albumin for 1 h. Next, standards and samples were added in an incubator for 20 min at 37°C with constant shaking. The plate was quickly washed three times with warm wash buffer. The detection antibody and streptavidin horseradish peroxidase are mixed together and added to the plates followed by a 20-min incubation. After washing, the chromogenic substrate TMB was added and the color developed at 37°C for about 5–10 min. Concentrations were calculated from the standard curve constructed with recombinant murine IL-6, also from R & D Systems.
Survival was analyzed by log rank analysis. Differences between the groups were compared with Student's t test (http://www.graphpad.com). Based on a preliminary subset of data, discriminate analysis was used to derive a linear combination of independent variables from the 24-h time point. This equation was y = 0.363 + (0.715 × weight change) + (0.566 × lymphocyte count) − (0.004 × platelet count). For the lymphocyte count and a platelet count, the exponent was not included. Those mice with negative numbers were predicted to live and those mice with positive numbers were predicted to die. Receiver operating characteristics curves were generated using NCSS 97 (http://www.ncss.com). Sensitivity was defined as the percentage of animals who died having plasma levels >2000 pg/mL. Specificity was defined as the animals who lived with an IL-6 level of less than 2000 pg/mL.
Several groups have identified that increasing the size of the puncture of the cecum will increase the lethality (7,9,14), but there are fewer studies that directly assessed an independent measure for the risk of subsequent mortality. A total of 79 mice were subjected to CLP with a 21-gauge needle and monitored for 21 days. Before the start of surgery, the groups of mice were very homogeneous because they were all from the same inbred strain (BALB/c), were from the same supplier, and were the same weight, age and sex. The overall survival during the 21-day interval was 56% (44/79, alive/total, Fig. 1). During the first 3 days, 18 mice died, yielding a 77% percent survival rate.
We next analyzed the plasma levels of IL-6 that were present 6 h after the initiation of sepsis because multiple studies have found higher levels of IL-6 in mice subjected to CLP (15–17). The mice were divided into two groups, those that were alive at 3 days and those that were dead. Mice that did not survive to 3 days had significantly greater concentrations of IL-6 in the circulation within 6 h of the surgical procedure (Fig. 2A). However, there was wide variation in these plasma levels, ranging from 81 pg/mL to nearly 10,000 pg/mL (Fig. 2B). Receiver operator characteristic curves were then drawn analyzing the status at day 3 (live or dead) to the plasma levels of IL-6. The total area under the curve was 0.834. Using a cut off level of 2000 pg/mL, the IL-6 measurement at 6 h had a sensitivity of 58% and a specificity of 97%. We next compared the survival of those mice with high plasma IL-6 to mice with low levels. Figure 3 demonstrates that the mice with high levels of IL-6 had significantly greater and more rapid mortality than those mice with an IL-6 level of less than 2000 pg/mL.
Mice were further analyzed at 24 h after the surgical procedure. At this time, the change in body weight and a hematologic profile were obtained. We did not obtain a hematologic profile at 6 h because previous results indicated that significant changes probably would not have developed by this time (8). There were significant differences in the weight change between survivors and nonsurvivors (Fig. 4). Those mice that died over the first 3 days because of sepsis had gained weight whereas the surviving mice lost weight. This weight loss occurred in spite of administration of 2 mL of fluid per day to each animal. Further investigations demonstrated that the mice became severely anorexic and failed to eat during the first few days after surgery (data not shown).
We also examined changes in the circulating numbers of neutrophils, lymphocytes, and platelets. Mice who survived beyond the 3-day time point had significantly more circulating neutrophils than the mice that died with some heterogeneity in the individual animal response (Fig. 5, A and B). Receiver operating characteristics showed that a neutrophil count of less than 0.95 × 106 cells/mL predicted outcome with a sensitivity of 81% and a specificity of 90%. Mice who survived beyond the 3-day period had significantly lower numbers of lymphocytes compared to the nonsurvivors (Fig. 6A). However, there was significant heterogeneity in the individual animal response (Fig. 6B). Platelet levels were lower in the animals that died during the first 3 days compared to those that survived (Fig. 7A), although there was substantial heterogeneity in this response also. Although there were differences between the groups, both survivors and non-survivors had significantly lower levels than normal mice (Table 1). There were no differences in the survivors and nonsurvivors in other hematologic parameters, including hemoglobin, hematocrit, red blood cell count, or monocytes.
A formula was derived that combined the weight change, lymphocyte, and platelet count to predict outcome. This formula is described in detail in the Materials and Methods section. Survival curves were generated based in this formula and showed that mice with the total score <0 are much more likely to survive (Fig. 8).
Both the formula and the IL-6 measurement predicted mortality; however, the IL-6 levels were available within 7 h and the other data were not available until 24 h after the initiation of sepsis. To initiate an early therapy, we designed an experiment to test the feasibility of initiating a therapeutic intervention on the basis of the IL-6 levels. Previous work has demonstrated that early resection of the necrotic cecum after CLP significantly improves survival (18–20). For our studies, CLP was performed and the IL-6 levels measured 6 h after surgery. As soon as these IL-6 levels became available, mice with IL-6 >2000 pg/mL had a second surgical procedure performed to resect the necrotic cecum and lavage the peritoneal cavity. A total of 14 mice were studied in this protocol, with six mice undergoing resection of the necrotic cecum and eight receiving antibiotic therapy alone. None of the mice that had undergone the surgical resection died whereas five of the eight mice treated with antibodies alone died over the next 21 days (Fig. 9). The improvement in survival was statistically significant (P = 0.02). Additionally, at autopsy none of the mice that had surgical resection of the cecum developed an abdominal abscess. In contrast, all of the other mice, regardless of IL-6 levels, developed an abdominal abscess.
All previous attempts to improve outcome in sepsis by modulating the inflammatory response have met with failure in the human sepsis trials, and the only new therapy that has improved outcome has been activated protein C (1). There have been multiple explanations advanced for these failures. One possibility is the heterogeneity in the patient populations that received treatment. The criteria used for patient enrollment are typically those of the systemic inflammatory response syndrome (21), in which the physiologic measurements cover a wide range of conditions not all of which will be lethal. Patient survival is predicated on the combination of the underlying condition of the patient, as well as the severity of the infectious challenge. In this present communication, we used an experimental animal model to precisely control the homogeneity of the population at risk before induction of sepsis. Specifically, in our experiments, all the mice were female, all the mice were approximately the same age and weight, and all the mice were genetically identical. In this system, the “patient population” is as homogeneous as possible, yet there is still heterogeneity in the response because about 45% of the mice died over 21 days. This heterogeneity in the response may be due to differences in the bacterial load within the cecum. Visual observation of time of surgery shows that some mice have a distended cecum with fecal material whereas others are relatively sparse. There may also be heterogeneity in the individual inflammatory response even though the experiments were conducted in inbred strain of mice.
Regardless of the cause of the heterogeneity, we attempted to determine whether there was a marker that could be used to predict outcome after CLP. We had previously published that animals with a high mortality rate have significantly elevated plasma levels of IL-6 (7,8). An important limitation to our study is that we only measured IL-6 at one time point and used this as the predictor of death. However, previous work has clearly demonstrated that in the CLP model, the mice with the highest mortality have elevated plasma levels of IL-6 over at least 24 h (7,11). Additionally, several studies in the recent literature have documented that those patients with sepsis who have high levels of plasma IL-6 are more likely to die. In a study with 242 septic patients, IL-6 was higher in patients with higher mortality (22). A study of 172 patients with severe melioidosis demonstrated that IL-6 is an independent predictor of survival (23). In a study of 180 patients who presented to the emergency department, the IL-6 levels on admission were predictors of progression to severe sepsis (24). This study has results similar to ours because the early time point IL-6 measurement was highly predictive of the clinical trajectory of the patient. It has also been proposed to use the plasma levels of IL-6 as a trigger to initiate treatment (25). In this study, patients were treated with a monoclonal antibody directed against tumor necrosis factor. Although the antibody did not improve survival in this study of 122 patients, there did appear to provide some benefit to those patients who had more than 1000 pg/mL of circulating IL-6 upon enrollment into the study. Additionally, one study found that the ratio of IL-6 to IL-10 was the best predictor of outcome (26).
The human studies have merely shown an association with elevated plasma levels of IL-6 and do not fully define a cause and effect relationship. The role of IL-6 in the pathogenesis of death and organ injury in sepsis has been examined by antibody inhibition studies or using IL-6 knockout mice. One of the earliest studies concluded that IL-6 was not important in a lipopolysaccharide model of sepsis in which inhibition of biological activity was achieved with monoclonal antibodies (27). Comparing IL-6 knockout mice to wild-type mice after CLP demonstrated no difference in survival (28). However, most investigators have found that IL-6 appears to exert a protective effect in animal models of sepsis. Experiments using a model of sepsis based on gavage with Escherichia coli demonstrated that inhibition of IL-6 improved both outcome and clearance of the bacteria (29). Feeding mice IL-6 helped to protect the mice from bacterial translocation afterhemorrhagic shock (30). IL-6 was considered beneficial in a neonatal model of group B streptococcal disease (31) and in a model of systemic E. coli infection (32). The high levels of IL-6 observed in mice that become septic after CLP are probably not directly injurious to the host and represent an protective mechanism.
The timing of treatment relative to the onset of sepsis has been extensively debated and has raised the possible reason why previous immunomodulating therapy for sepsis has failed. Some investigators believe that early treatment of sepsis is essential because the production of cytokines occurs quickly and during the initial phases of sepsis. These beliefs may have been based upon the endotoxin model of sepsis, which does indeed produce an explosive release of cytokines (11). However, in other animal models of sepsis, the release of cytokines is more prolonged compared to endotoxin (11). Studies in human patients with sepsis also demonstrate a more prolonged release of cytokines (33). Our data demonstrate that the plasma levels of IL-6 may be used as an accurate predictor of mortality. These levels offer the possibility of a diagnostic test for initiating therapy. In this case, surgical intervention in the animals with the high predicted mortality was able to offer 100% survival.
It should be stressed that IL-6 is not proposed as a mediator causing injury and death in sepsis but only as a marker for the level of inflammation. Other cytokines, such as tumor necrosis factor and IL-1, may be directly involved in damaging tissues and organs. Immunotherapy directed against these mediators has proven beneficial when both of the cytokines are inhibited at the same time (34). The present study would appear to indicate that those mice that died in the early time interval exhibited excessive inflammation as measured by increased levels of plasma IL-6. This group may benefit from decreasing the inflammatory response. In contrast, mice with low levels of IL-6 do not require immunosuppressive therapy to block the biological activities of the cytokines and actually may benefit from an augmentation of their inmate immune response.
Supported in part by NIH grant GM 44918.
1. Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez–Rodriguez A, Steingrub JS, Garber GE, Helterbrand JD, Ely EW, Fisher CJ, Jr., and Recombinant human protein CWEiSSsg: Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med
2. Abraham E: Why immunomodulatory therapies have not worked in sepsis. Intensive Care Med 25:556–566, 1999.
3. Abraham E, Marshall JC: Sepsis and mediator-directed therapy: rethinking the target populations. Mediator-directed therapy in sepsis: Rethinking the target populations. Toronto, Canada, 31 October–1 November 1998. Mol Med Today 5:56–58, 1999.
4. Kox WJ, Volk T, Kox SN, Volk HD: Immunomodulatory therapies in sepsis. Intensive Care Med 26:S124–S128, 2000.
5. Hack CE, De Groot ER, Felt-Bersma RJ, Nuijens JH, Strack Van ScR, Eerenberg-Belmer AJ, Thijs LG, Aarden LA: Increased plasma levels of interleukin-6 in sepsis [see comments]. Blood 74:1704–1710, 1989.
6. Waage A, Brandtzaeg P, Halstensen A, Kierulf P, Espevik T: The complex pattern of cytokines
in serum from patients with meningococcal septic shock. Association between interleukin 6, interleukin 1, and fatal outcome. J Exp Med 169:333–338, 1989.
7. Ebong S, Call D, Nemzek J, Bolgos G, Newcomb D, Remick D: Immunopathologic alterations in murine models of sepsis of increasing severity. Infection Immunity 67:6603–6610, 1999.
8. Ebong SJ, Call DR, Bolgos G, Newcomb DE, Granger JI, O'Reilly M, Remick DG: Immunopathologic responses to non-lethal sepsis. Shock 12:118–126, 1999.
9. Walley KR, Lukacs NW, Standiford TJ, Strieter RM, Kunkel SL: Balance of inflammatory cytokines
related to severity and mortality of murine sepsis. Infection Immunity 64:4733–4738, 1996.
10. Wichterman KA, Baue AE, Chaudry IH: Sepsis and septic shock—a review of laboratory models and a proposal. J Surg Res 29:189–201, 1980.
11. Remick DG, Newcomb DE, Bolgos GL, Call DR: Comparison of the mortality and inflammatory response of two models of sepsis: Lipopolysaccharide vs. cecal ligation and puncture. Shock 13:110–116, 2000.
12. Ebong SJ, Goyert SM, Nemzek JA, Kim J, Bolgos GL, Remick DG: Critical role of CD14 for production of proinflammatory cytokines
and cytokine inhibitors during sepsis with failure to alter morbidity or mortality. Infection Immunity 69:2099–2106, 2001.
13. Call DR, Nemzek JA, Ebong SJ, Bolgos GL, Newcomb DE, Remick DG: Ratio of local to systemic chemokine concentrations regulates neutrophil recruitment. Am J Pathol 158:715–721, 2001.
14. Lundblad R, Sandven P, Giercksky KE: The physical nature of a large bowel perforation predicts severity of the subsequent inflammatory response. Shock 3:455–461, 1995.
15. Villa P, Shaklee CL, Meazza C, Agnello D, Ghezzi P, Senaldi G: Granulocyte colony-stimulating factor and antibiotics in the prophylaxis of a murine model of polymicrobial peritonitis and sepsis. J Infect Dis 178:471–477, 1998.
16. Ertel W, Morrison MH, Wang P, Ba ZF, Ayala A, Chaudry IH: The complex pattern of cytokines
in sepsis. Association between prostaglandins, cachectin, and interleukins. Ann Surg 214:141–148, 1991.
17. Salkowski CA, Detore G, Franks A, Falk MC, Vogel SN: Pulmonary and hepatic gene expression following cecal ligation and puncture: Monophosphoryl lipid A prophylaxis attenuates sepsis-induced cytokine and chemokine expression and neutrophil infiltration. Infection Immunity 66:3569–3578, 1998.
18. Chaudry IH, Tabata Y, Schleck S, Baue AE: Effect of splenectomy on reticuloendothelial function and survival following sepsis. J Trauma 20:649–656, 1980.
19. Baker CC, Chaudry IH, Gaines HO, Baue AE: Evaluation of factors affecting mortality rate after sepsis in a murine cecal ligation and puncture model. Surgery 94:331–335, 1983.
20. Barke RA, Birklid S, Chapin RB, Roy S, Brady PS, Brady LJ: The effect of surgical treatment following peritoneal sepsis on hepatic gene expression. J Surg Res 60:101–106, 1996.
21. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, Schein RM, Sibbald WJ: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine [see comments]. Chest 101:1644–1655, 1992.
22. Oberhoffer M, Vogelsang H, Russwurm S, Hartung T, Reinhart K: Outcome prediction by traditional and new markers of inflammation
in patients with sepsis. Clin Chem Lab Med 37:363–368, 1999.
23. Simpson AJ, Smith MD, Weverling GJ, Suputtamongkol Y, Angus BJ, Chaowagul W, White NJ, van Deventer SJ, Prins JM: Prognostic value of cytokine concentrations (tumor necrosis factor-alpha, interleukin-6, and interleukin-10) and clinical parameters in severe melioidosis. J Infect Dis 181:621–625, 2000.
24. Terregino CA, Lopez BL, Karras DJ, Killian AJ, Arnold GK: Endogenous mediators in emergency department patients with presumed sepsis: are levels associated with progression to severe sepsis and death? [see comments]. Ann Emerg Med 35:26–34, 2000.
25. Reinhart K, Wiegand-Lohnert C, Grimminger F, Kaul M, Withington S, Treacher D, Eckart J, Willatts S, Bouza C, Krausch D, Stockenhuber F, Eiselstein J, Daum L, Kempeni J: Assessment of the safety and efficacy of the monoclonal anti-tumor necrosis factor antibody-fragment, MAK 195F, in patients with sepsis and septic shock: a multicenter, randomized, placebo-controlled, dose-ranging study [see comments] [published erratum appears in Crit Care Med
1996 Sep;24(9):1608]. Crit Care Med 24:733–742, 1996.
26. Taniguchi T, Koido Y, Aiboshi J, Yamashita T, Suzaki S Kurokawa A: Change in the ratio of interleukin-6 to interleukin-10 predicts a poor outcome in patients with systemic inflammatory response syndrome. Crit Care Med 27:1262–1264, 1999.
27. Libert C, Vink A, Coulie P, Brouckaert P, Everaerdt B, Van Snick J, Fiers W: Limited involvement of interleukin-6 in the pathogenesis of lethal septic shock as revealed by the effect of monoclonal antibodies against interleukin-6 or its receptor in various murine models. Eur J Immunol 22:2625–2630, 1992.
28. Leon LR, White AA, Kluger MJ: Role of IL-6 and TNF in thermoregulation and survival during sepsis in mice. Am J Physiol 275:R269–R277, 1998.
29. Gennari R, Alexander JW: Anti-interleukin-6 antibody treatment improves survival during gut-derived sepsis in a time-dependent manner by enhancing host defense. Crit Care Med 23:1945–1953, 1995.
30. Rollwagen FM, Li YY, Pacheco ND, Dick EJ, Kang YH: Microvascular effects of oral interleukin-6 on ischemia/reperfusion in the murine small intestine. Am J Pathol 156:1177–1182, 2000.
31. Mancuso G, Tomasello F, Migliardo M, Delfino D, Cochran J, Cook JA, Teti G: Beneficial effects of interleukin-6 in neonatal mouse models of group B streptococcal disease. Infection Immunity 62:4997–5002, 1994.
32. Dalrymple SA, Slattery R, Aud DM, Krishna M, Lucian LA, Murray R: Interleukin-6 is required for a protective immune response to systemic Escherichia coli
infection. Infection Immunity 64:3231–3235, 1996.
33. Endo S, Inada K, Yamada Y, Wakabayashi G, Ishikura H, Tanaka T, Sato S: Interleukin 18 (IL-18) levels in patients with sepsis. J Med 31:15–20, 2000.
34. Remick DG, Call DR, Ebong SJ, Newcomb DE, Nybom P, Nemzek JA, Bolgos GE: Combinantion immuntherapy with soluble tumor necrosis factor receptors plus interleukin 1 receptor antagonist decreases sepsis mortality. Crit Care Med 29:473–481, 2001.