Serum Levels of the Interleukin-1 Receptor Family Member ST2 Predict Mortality and Clinical Outcome in Acute Myocardial Infarction
Background— Mechanically overloaded cardiomyocytes secrete a soluble interleukin-1 receptor family member called ST2. Serum levels of ST2 are associated with prognosis in nonischemic heart failure, but the predictive value of ST2 in patients with acute myocardial infarction is unknown.
Methods and Results— ST2 levels were measured in serum from 810 patients with acute myocardial infarction in the Thrombolysis In Myocardial Infarction (TIMI) 14 (362 patients) and Enoxaparin and TNK-tPA With or Without GPIIb/IIIa Inhibitor as Reperfusion Strategy in STEMI (ENTIRE)-TIMI 23 (448 patients) clinical trials. Baseline levels of ST2 were significantly higher in those patients who died (0.379 versus 0.233 ng/mL, P=0.0001) or developed new congestive heart failure (0.287 versus 0.233 ng/mL, P=0.009) by 30 days. In an analysis of outcomes at 30 days by ST2 quartiles, both death (P=0.001) and the combined death/heart failure end point (P=0.001) showed a significant graded association with levels of ST2; furthermore, in-hospital death (P=0.003) and death/heart failure (P=0.004) were also significantly associated with higher ST2 levels. In a logistic regression analysis that controlled for important clinical factors, increasing levels of ST2 remained associated with death at 30 days (P=0.047). ST2 levels rose during the first day after infarction and were maximal at 12 hours; ST2 levels at 12 hours were also independently associated with death at 30 days (P<0.001).
Conclusions— Serum levels of the interleukin-1 receptor family member ST2 predict mortality and heart failure in patients with acute myocardial infarction. These data suggest that ST2 may be a useful biomarker and that this novel inflammatory receptor may play a role in cardiac pathophysiology.
Received July 28, 2003; de novo received December 15, 2003; revision received February 9, 2004; accepted February 10, 2004.
The ST2 gene (also known as T1, IL1RL1, or Fit1) is a member of the interleukin-1 receptor host defense/inflammation family.1,2 The protein product of ST2 encodes a transmembrane receptor form (ST2L) and a truncated soluble receptor form (ST2) that can be detected in human serum.3,4 Although ST2 is an orphan receptor (ligand currently unknown), it is known that ST2 directs Th2 lymphocyte function.5 Genomic technology was used to identify ST2 as a gene markedly induced in mechanically overloaded cardiac myocytes.6,7 This suggests that ST2 is induced in conditions of myocardial overload such as myocardial infarction, when the remaining viable myocardium must bear more stress. Indeed, soluble ST2 levels are increased in the serum of patients 1 day after myocardial infarction.6 Furthermore, ST2 serum levels predict outcome in patients with heart failure, and a change in ST2 over time is also associated with prognosis.7 These data suggest that measurement of serum levels of ST2 may provide insight into the hemodynamic burden of the myocardium.
The relationship between this novel interleukin-1 receptor family member and clinical outcomes in acute myocardial infarction is undefined. Thus, the present study was designed to test the hypothesis that serum ST2 levels are associated with the risk of death or heart failure in patients presenting with ST-elevation myocardial infarction (STEMI). To do this, we measured ST2 levels in serum from patients with STEMI enrolled in the Thrombolysis in Myocardial Infarction (TIMI) 14 and Enoxaparin and TNK-tPA With or Without GPIIb/IIIa Inhibitor as Reperfusion Strategy in STEMI (ENTIRE)-TIMI 23 trials.
The TIMI 14 trial was a randomized, open-label, dose-ranging study of combination reperfusion therapy for patients with STEMI conducted between March 1997 and July 1998. Specifically, this study was an angiographic trial that evaluated several combinations of fibrinolytic with the glycoprotein IIb/IIIa antagonist abciximab.8,9 The ENTIRE-TIMI 23 trial was an open-label, dose-ranging, multicenter study conducted between February 2000 and September 2001 to evaluate enoxaparin as adjunctive antithrombin therapy with various forms of pharmacological reperfusion, including full-dose tenecteplase and half-dose tenecteplase plus abciximab.10 In both studies, patients were eligible for inclusion if they had a qualifying episode of ischemic discomfort of at least 30 minutes within 6 hours (ENTIRE) or 12 hours (TIMI 14) and exhibited at least 0.1-mV ST-segment elevation in 2 contiguous ECG leads. Exclusion criteria for both trials included increased risk of hemorrhage, severe renal insufficiency, and cardiogenic shock.
Serum samples collected at baseline and 1, 3, 12, and 24 hours after enrollment in TIMI 14 were evaluated. Serum samples from the ENTIRE-TIMI 23 trial were collected only at baseline. Serum was isolated within 60 minutes of sample collection and stored at −20°C or colder until shipped to the TIMI Biomarker Core Laboratory (Boston, Mass), where samples were maintained at −70°C. Soluble ST2 was measured by a sandwich double monoclonal antibody ELISA method (Medical & Biological Laboratories Co, Ltd). Serum samples or standards were incubated in microwells coated with anti-human ST2 antibody. After washing, peroxidase-conjugated anti-human ST2 antibody was added into the microwell and incubated. After a second washing, the peroxidase substrate was added, and the optical density at 450 nm was determined. High-sensitivity C-reactive protein (hs-CRP, Dade-Behring Inc), creatine kinase-MB isoenzyme (CK-MB), B-type natriuretic peptide (BNP, Shionoria BNP), and cardiac troponin I (cTnI; ACS:180, Bayer Diagnostics) were measured by previously described methods.11–13 Creatine kinase isoenzyme levels were measured locally at the enrolling site on admission, at 3 hours, and at 6- to 8-hour intervals for the first 24 hours. Because of sample availability limitations, BNP levels were measured in samples from ENTIRE-TIMI 23 but not TIMI 14.
Patients were divided into quartiles on the basis of their serum levels of ST2 at the time of enrollment. ST2 levels are described with the median and 25th and 75th percentiles. The association between baseline clinical characteristics and quartiles of ST2 were analyzed with the Kruskal-Wallis test for continuous variables and the χ2 test for categorical variables. Correlations between ST2 and other continuous baseline variables were studied with a nonparametric (Spearman) correlation coefficient. For evaluation of association with clinical outcomes, ST2 was compared between patients who met a study end point and those who did not with the Wilcoxon rank-sum test. Multivariate analysis of the association of ST2 (continuous variable) with outcomes was performed with logistic regression that included terms for established predictors of mortality in STEMI.14 Except where stated, results presented are for the combined TIMI 14 and ENTIRE-TIMI 23 study population.
Baseline ST2 and Clinical Variables
Most baseline clinical characteristics, including gender, age, weight, and extent of coronary artery disease, did not correlate with baseline ST2 levels (Table 1). Few patients in this population had either a prior history of or presented with clinical evidence of heart failure. Heart rate (P<0.0001) and systolic blood pressure (P=0.05) were associated with ST2 levels, consistent with the theory that ST2 is secreted by cardiac myocytes under biomechanical stress. Patients with the highest levels of ST2 were more likely to have presented with an anterior myocardial infarction (P=0.035) and to present later after chest pain onset (P<0.0001). The biomarkers cTnI, BNP, and CRP, which have all been shown to predict outcome after myocardial infarction,13,15,16 were correlated with ST2 by quartile analysis and, with the exception of BNP, showed statistically significant but quantitatively weak correlations when evaluated as continuous variables (Table 2).
ST2 and Clinical Outcomes
For the combined cohort of 810 patients, baseline ST2 was significantly associated with clinical outcomes at 30 days (Table 3). Specifically, levels of ST2 were significantly higher at presentation among patients who subsequently died (P=0.0001) or who developed new or worsening congestive heart failure (P=0.009) by 30 days after enrollment. Dichotomized at the median, elevated baseline levels of ST2 were indicative of higher mortality through 30 days of follow-up (log-rank P=0.0009; Figure 1). Moreover, in an analysis by quartiles of ST2, the risk of death (P=0.001) and the composite of death or congestive heart failure (P=0.001) increased in a graded, stepwise fashion with higher levels of ST2 (Table 4). This association between ST2 and clinical events was homogeneous between the 2 individual trials (TIMI 14 and ENTIRE-TIMI 23, P for heterogeneity=0.8).
Evolution of ST2 Serum Levels
Baseline ST2 levels analyzed by quartile were significantly correlated with the time to randomization (Tables 1 and 2⇑). This was not surprising, because on the basis of our prior studies, ST2 levels are anticipated to increase on the first day after coronary occlusion and to return to normal over the next 14 days.6 Among the TIMI 14 patients, analysis of serial measurements of serum ST2 in 228 patients revealed an increase with time, with most patients reaching a peak ST2 level at 12 hours (Figure 2).
After controlling for established clinical predictors in STEMI, including age, heart rate, systolic blood pressure, location of myocardial infarction, Killip class, and time from onset of chest pain, increasing levels of ST2 (per 1-ng/mL rise) remained an independent predictor of death at 30 days (OR 1.77, 95% CI 1.01 to 3.12; P=0.047). Data on ST2 and 3 established biomarkers (BNP, hs-CRP, and cTnI) were available in ENTIRE-TIMI 23. In this data set, mortality through 30 days showed a significant association with quartiles of cTnI (quartiles 1 through 4, respectively: 0%, 2.7%, 2.7%, and 7.9%; P=0.009) and BNP (quartiles 1 through 4, respectively: 0.6%, 1.4%, 0.9%, and 10.7%; P<0.001) but not hs-CRP (quartiles 1 through 4: 4.4%, 1.8%, 1.8%, and 5.4%; P=0.3). The association between ST2 and mortality was no longer significant when BNP and cTnI were added to the clinical model (limited to ENTIRE-TIMI 23). The predictive capacity of ST2 ascertained at later time points (3 and 12 hours in TIMI 14) was also evaluated and revealed a similar association between ST2 at 12 hours and mortality risk (OR 1.27, 95% CI 1.10 to 1.47; P=0.001). In multivariate analysis that included each of the clinical factors (age, heart rate, systolic blood pressure, location of myocardial infarction, Killip class, and time from onset of chest pain) and TIMI flow grade in the infarct-related artery, ST2 remained independently associated with mortality (P<0.01).
In the present study, we explored the potential role of serum measurement of a recently identified receptor of the interleukin-1 family in acute myocardial infarction. The soluble form of this receptor is rapidly secreted by cardiac myocytes when the cells are biomechanically overloaded; this suggests that the receptor may play a role in conditions in which the myocardium is suddenly overloaded, such as in myocardial infarction. To explore this, we measured serum ST2 levels at the time of presentation in a cohort of patients with ST-elevation myocardial infarction. The results demonstrate that levels of ST2 at the time of presentation with STEMI are associated with in-hospital and 30-day mortality. Furthermore, multivariate analysis demonstrated that ST2 level is independently associated with outcome after controlling for established clinical risk indicators in STEMI.
The significance of these data is twofold. Foremost, these data suggest that the interleukin-1 receptor family, which participates in host defense and differentiation of T cells,1 may participate in early events in acute myocardial infarction. These data support further investigation of this receptor as a potential novel target for modifying prognosis in patients with myocardial infarction. Secondarily, ST2 represents a novel biomarker that offers prognostic information in patients with acute myocardial infarction that is independent of clinical predictors. The present report extends our prior work demonstrating an association between ST2 and mortality among patients with nonischemic congestive heart failure,7 another condition of myocardial overload.
The mechanisms of induction and regulation of ST2 expression in acute myocardial infarction are not known. The mechanical overload may induce ST2, but in addition, proinflammatory cytokines from damaged tissues may activate neighboring cells to produce ST2. ST2 has been shown to regulate the expression of proinflammatory cytokines from macrophage, which may serve to prevent uncontrolled inflammatory reactions. It has also been shown that ST2 released in response to stress or injury can contribute toward the polarization of T helper cells to the Th2 phenotype. Myocardial infarction is associated with humoral (eg, cytokines and reactive oxygen species) and cell-mediated (eg, neutrophils and mononuclear cells) inflammatory reactions, which are a prerequisite for healing and scar formation. ST2 may contribute to this process, which can affect prognosis.
Although not excluded by the present study, it is unlikely that the relationship of ST2 and outcome after myocardial infarction is simply a reflection of the association of chronic elevations in inflammatory markers such as CRP and risk of myocardial infarction. ST2, like BNP, may be synthesized by cardiac myocytes themselves, and data from patients without apparent ischemic disease suggest that ST2 predicts prognosis in the absence of coronary artery disease. Furthermore, preliminary data suggest that ST2 levels in outpatients with stable coronary artery disease are unrelated to CRP levels (unpublished observations). Moreover, ST2 showed only a very modest (r=0.10) correlation with hs-CRP in the present data set. Although the present data support the complementary value of ST2 for risk assessment when added to a robust clinical model, ST2 did not contribute additional information to BNP and cTnI in the smaller data set limited to ENTIRE-TIMI 23. Notably, levels of ST2 were only weakly correlated with a measure of myonecrosis (cTnI) and were not associated with the plasma concentration of BNP, which suggests that ST2 reflects pathophysiology distinct from these established biomarkers. Additional studies are necessary to evaluate carefully the prognostic value of ST2 in conjunction with other available biomarkers. The study end points of TIMI 14 and ENTIRE-TIMI 23 were predetermined to be 30 days after enrollment, and thus longer-term follow-up was not available.
Although ST2 may be secreted by mechanically overloaded cardiac myocytes, many cells can secrete ST2. Thus, elevations in serum ST2 are not specific for acute myocardial infarction. In addition to nonischemic heart failure,7 patients with asthma17 or autoimmune diseases such as systemic lupus erythematosus18 may also have increased serum ST2 levels. Thus, ST2 measurement is unlikely to be useful in the initial diagnosis of myocardial infarction. However, ST2 levels correlate inversely with ejection fraction in patients with nonischemic heart failure,7 and additional studies with serial evaluations of ventricular function are necessary to determine whether the source of ST2 in patients with myocardial infarction is the overloaded myocardium.
These data lend validity to the concept that genomic technology can reveal a new potential pathophysiological pathway in a common disease. ST2 was identified initially through studies of the interleukin-1 family, but its role in myocardial disease was only recently suggested by genomic studies with DNA microarrays. Studies with DNA microarrays allow identification of potential new disease pathways, but this is only an initial step in understanding the role of the pathway. The association between ST2 and outcome in the present study now provides important data that support a role for ST2 in acute myocardial infarction. However, additional studies of the function of ST2 in myocardial infarction are needed. Furthermore, the ligand for the soluble and membrane ST2 receptors remains unidentified, and understanding the potentially competing roles of the membrane and soluble receptors will require identification of the ligand.
This study was supported in part by grants from the National Heart, Lung, and Blood Institute. Dr Shimpo was supported in part by Japan Heart Foundation/Bayer Yakuhin Research Grant Abroad, the Cell Science Research Foundation, and the Mochida Memorial Foundation for Medical and Pharmaceutical Research.
Brigham and Women’s Hospital has filed for a patent on ST2, with Dr Lee as inventor.
O’Neill L. The Toll/interleukin-1 receptor domain: a molecular switch for inflammation and host defence. Biochem Soc Trans. 2000; 28: 557–563.
Townsend MJ, Fallon PG, Matthews DJ, et al. T1/ST2-deficient mice demonstrate the importance of T1/ST2 in developing primary T helper cell type 2 responses. J Exp Med. 2000; 191: 1069–1076.
Walzl G, Matthews S, Kendall S, et al. Inhibition of T1/ST2 during respiratory syncytial virus infection prevents T helper cell type 2 (Th2)- but not Th1-driven immunopathology. J Exp Med. 2001; 193: 785–792.
Weinberg EO, Shimpo M, De Keulenaer GW, et al. Expression and regulation of ST2, an interleukin-1 receptor family member, in cardiomyocytes and myocardial infarction. Circulation. 2002; 106: 2961–2966.
Weinberg EO, Shimpo M, Hurwitz S, et al. Identification of serum soluble ST2 receptor as a novel heart failure biomarker. Circulation. 2003; 107: 721–726.
Antman EM, Giugliano RP, Gibson CM, et al. Abciximab facilitates the rate and extent of thrombolysis: results of the Thrombolysis In Myocardial Infarction (TIMI) 14 trial: the TIMI 14 Investigators. Circulation. 1999; 99: 2720–2732.
Antman EM, Gibson CM, de Lemos JA, et al. Combination reperfusion therapy with abciximab and reduced dose reteplase: results from TIMI 14: the Thrombolysis in Myocardial Infarction (TIMI) 14 Investigators. Eur Heart J. 2000; 21: 1944–1953.
Antman EM, Louwerenburg HW, Baars HF, et al. Enoxaparin as adjunctive antithrombin therapy for ST-elevation myocardial infarction: results of the ENTIRE-Thrombolysis in Myocardial Infarction (TIMI) 23 Trial. Circulation. 2002; 105: 1642–1649.
Morrow DA, Rifai N, Tanasijevic MJ, et al. Clinical efficacy of three assays for cardiac troponin I for risk stratification in acute coronary syndromes: a Thrombolysis In Myocardial Infarction (TIMI) 11B Substudy. Clin Chem. 2000; 46: 453–460.
Yasue H, Yoshimura M, Sumida H, et al. Localization and mechanism of secretion of B-type natriuretic peptide in comparison with those of A-type natriuretic peptide in normal subjects and patients with heart failure. Circulation. 1994; 90: 195–203.
Morrow DA, Antman EM, Charlesworth A, et al. TIMI risk score for ST-elevation myocardial infarction: a convenient, bedside, clinical score for risk assessment at presentation: an intravenous nPA for treatment of infarcting myocardium early II trial substudy. Circulation. 2000; 102: 2031–2037.