Clinical implications of the plasma EphA2 receptor level in critically ill patients with septic shock

The Eph/ephrin receptor ligand system is known to play a role in inflammation induced by infection, injury, and inflammatory diseases. The present study aimed to evaluate plasma EphA2 receptor levels in critically ill patients with sepsis. This study was a prospective cohort study evaluating samples and clinical data from the medical intensive care unit (MICU) of a 2000-bed university tertiary referral hospital in South Korea. Positive correlations of the plasma EphA2 receptor level with the acute physiology and chronic health evaluation (APACHE) II score and the sequential organ failure assessment (SOFA) score were observed. The area under the curve (AUC) for the plasma EphA2 receptor level on a receiver operating characteristic curve was 0.690 (95% confidence interval [CI], 0.608–0.764); the AUCs for the APACHE II score and SOFA scores were 0.659 (95% CI, 0.576–0.736) and 0.745 (95% CI, 0.666–0.814), respectively. A Cox proportional hazard model identified an association between an increased plasma EphA2 receptor level (>51.5 pg mL−1) and increased risk of 28-day mortality in the MICU (hazard ratio = 3.22, 95% CI, 1.709–6.049). An increased plasma EphA2 receptor level was associated with sepsis severity and 28-day mortality among sepsis patients.

SCIEntIfIC RepoRtS | 7: 17612 | DOI: 10.1038/s41598-017-17909- 7 We have found that the plasma EphA2 receptor level is elevated in sepsis patients, and therefore, we investigated the clinical implications of the plasma EphA2 receptor level with regard to sepsis. The primary outcome of the study was the association between the plasma EphA2 receptor level and all-cause 28-day mortality among sepsis patients. The secondary outcomes were an evaluation of the correlations between plasma EphA2 receptor levels and severity of sepsis as determined by the acute physiology and chronic health evaluation (APACE) II score and the sequential organ failure assessment (SOFA) score 23,24 . Because EphA2 receptor blocking antibodies are already undergoing clinical trials for cancer [25][26][27][28] , we also investigated the plasma EphA2 receptor level as a potential marker for therapeutic monitoring in sepsis patients.

Comparison of characteristics between 28-day survivors and 28-day non-survivors.
This study divided patients into two groups according to 28-day mortality: patients in the case group died within 28 days of admission to the MICU, whereas those in the control group survived longer than 28 days. The results of this inter-group comparison are shown in Table 2. The groups did not differ significantly with regard to sex,  median age, and median body mass index (BMI). Regarding clinical parameters, the groups did not differ in terms of the rate of positive blood cultures and median scores on the Charlson Comorbidity Index. However, the non-survivor group had higher rates of ARDS (32.6% vs. 12.7% for survivors, P = 0.005) and a sepsis status above severe (86.0% vs. 67.6% for survivors, P = 0.022). The median APACHE II score (29 vs. 24, P = 0.002) and median SOFA score (13 vs. 8, P < 0.001) were also significantly higher in the non-survivor group relative to the survivor group. Regarding laboratory parameters, no differences were observed between survivors and non-survivors in the median levels for C-reactive protein (CRP) (73.5 mg mL −1 vs. 97 mg mL −1 , P = 0.347) and procalcitonin (0.5 ng mL −1 vs. 1.8 ng mL −1 , P = 0.181) on the day of admission. However, the median plasma EphA2 receptor level was significantly higher in the non-survivor group than in the survivor group (78.5 pg mL −1 vs. 33.3 pg mL −1 , P < 0.001).
Association between the plasma EphA2 receptor level and severity or mortality. Positive correlations were observed between the plasma EphA2 receptor level and the APACHE II (r = 0.291, P < 0.001) and SOFA (r = 0.341, P < 0.001 Fig. 2) scores. Stratification of patients according to the IQRs for SOFA scores revealed an increasing trend in SOFA scores with increasing plasma EphA2 receptor levels (SOFA score <6 vs. SOFA score >13, P < 0.001; Fig. 3). The Area under the curve (AUC) for the plasma EphA2 receptor level on a receiver operating characteristic (ROC) curve was 0.690 (95% confidence interval [CI], 0.608-0.764); the AUCs for the APACHE II score and SOFA score were 0.659 (95% CI, 0.576-0.736) and 0.745 (95% CI, 0.666-0.814), respectively. In ROC curve comparisons, the AUC of the plasma EphA2 receptor level and SOFA score (0.690 vs. 0.745, P = 0.256) and APACHE II score (0.690 vs. 0.659, P = 0.605) did not differ significantly. However, the AUC of the SOFA score and APACHE II score differed significantly (0.745 vs. 0.659, P = 0.034; Fig. 4a). A cut-off value of >51.5 pg mL −1 for the plasma EphA2 receptor level was determined to predict 28-day mortality on the ROC (sensitivity, 62.8%; specificity, 71.6%). Patients were divided into two groups according to this plasma EphA2 receptor cut-off level of 51.5 pg mL −1 and subjected to a Cox proportional hazard model analysis involving several demographic and clinical characteristics and substituted plasma EphA2 receptor levels ( Table 3). In the univariate analysis, an increased plasma EphA2 receptor level (>51.5 pg mL −1 ; hazard ratio, 3.41; 95% CI, 1.750-6.636) was associated with 28-day mortality. In the multivariate analysis, sex, age, BMI, positive blood culture and sepsis status higher than severe were not found to be associated with 28-day mortality; in contrast, ARDS (hazard ratio, 2.02; 95% CI, 1.051-3.389) and an increased plasma EphA2 receptor level (>51.5 pg mL −1 ; hazard ratio, 3.22; 95% CI, 1.709-6.049) were associated with increased 28-day mortality in the MICU. Figure 4b shows the survival rates of the groups classified according to plasma EphA2 receptor levels; these survival rates differed significantly (P < 0.001).

Discussion
In the current study, we investigated potential associations of the plasma EphA2 receptor level with sepsis severity and 28 day-mortality among sepsis patients in the MICU. The main finding of this study was that the plasma EphA2 receptor level correlates with sepsis severity scores and increased mortality among sepsis patients. The APACHE II score and SOFA score are commonly used for initial severity assessments in sepsis patients 23,29 . The plasma EphA2 receptor level was comparable to these scores. The results of this analysis consistently indicated that an increased plasma EphA2 receptor level was related to the initial severity of sepsis. A higher SOFA score is associated with increased patient mortality 24 . In this study, plasma EphA2 receptor levels were evaluated in groups according to interquartile ranges of SOFA scores (Fig. 3). The result showed that a higher plasma EphA2 receptor level was associated with mortality in sepsis patients. The AUC of the ROC curve and results of the Cox hazard proportional model supported this association. Furthermore, we used a plasma EphA2 receptor cut-off value to discriminate mortality risk according to the Cox hazard proportional model. As a result, a plasma EphA2 receptor level >51.5 pg mL −1 was associated with a significant increase in mortality (hazard ratio, 3.22).
As mentioned previously, many studies have focused on the complex roles of the Eph receptor and ephrin ligand in malignancy 30,31 . Several studies have reported associations of EphA2 with multiple oncogenic signaling pathways such as MAP/ERK, phosphoinositide 3-kinase, E-cadherin, and integrin/FAK/paxillin 30,[32][33][34] . In the field of oncology, the EphA2 receptor has been studied as a therapeutic target and is being investigated in an ongoing clinical trial in patients with advanced malignancies [25][26][27][28] . Tumor development, progression, and therapeutic responses may be affected by inflammation 17 . Cancer and infectious disease have many similarities. For example, inhibitors of programmed cell death ligand 1, which were developed for patients with cancer, have also been actively investigated as therapeutic agents for chronic infection and/or sepsis 18 . Accordingly, it is noteworthy that the Eph2 receptor and ephrin ligand also affect inflammation through vascular endothelial injury 13,16,19 .  EphA2 and other inflammatory mediators, such as TNF-α and interferon (IFN)-γ, upregulate NF-κB, thus increasing intercellular adhesion molecule-1 expression and facilitating leukocyte migration and attachment 19,35 . Several emerging studies have shown a link between the EphA2 receptor and inflammation [36][37][38][39][40] . Animal models of lipopolysaccharide (LPS)-induced pneumonia have demonstrated an increase in the EphA2 receptor level after LPS exposure 37 . Another study reported the increased expression of EphA2 receptor with lung injury and found that EphA2 receptor blockade reduced lung injury and the passage of fluids and inflammatory cells 41 . EphA2 receptor signaling is associated with Src family kinases, mitogen activated protein kinase, p-21 activated kinase, post-synaptic density protein 95-dependent pathways, chemokine pathways, heterotrimeric G-protein pathways, and integrin-mediated pathways 19,22 . A recent study reported that the p-21 activated kinase was strongly associated with endothelial barrier in a sepsis murine model 22 . Given the results of this study and others, EphA2 receptor blockade may be a reasonable treatment target for reducing endothelial injury in sepsis patients. Unlike malignancy, however, sepsis does not involve a target lesion. Therefore, an appropriate blood marker with which to assess the therapeutic drug effect is needed if EphA2 receptor blocking is to be used as a therapeutic target. The plasma EphA2 receptor level may be useful to evaluate the effect of treatment in this setting.
The strengths of this study include the confirmed measurement of blood plasma EphA2 receptor levels and the identification of an association between human plasma EphA2 receptor levels and sepsis. These results provide clinical evidence for the use of EphA2 receptor as a therapeutic target in sepsis and provide a potential tool for monitoring responses to EphA2 receptor blocker treatment in sepsis patients.
However, our study has several limitations. First, this was a small study of plasma EphA2 receptor levels in sepsis patients at a single center without pediatric patients. However, this study is the first clinical study to use human blood samples to evaluate the relationship between plasma EphA2 receptor levels and sepsis. Previously, other studies investigated the relationship between the EphA2 receptor and inflammation. Most of these studies used in vitro or animal models 36,37,39,42 . However, no clinical proof of sepsis was provided. Second, the plasma EphA2 receptor level measurements have not been validated. The importance of proteases in regulating the Eph receptor and the ephrin ligand family has been shown 43 ; it is thought that the proteases that activate the Eph Kaplan-Meier survival analysis showed that the 28-day mortality of patients with plasma EphA2 receptor level ≥51.5 pg mL −1 was higher than that of patients with plasma EphA2 receptor level <51.5 pg mL −1 .  Table 3. Cox proportional hazard model of plasma EphA2 level and 28-day mortality. HR, hazard ratio; CI, confidence interval; BMI, body mass index; ARDS, acute respiratory distress syndrome.

Variable
receptor and the ephrin ligand family also degrade the Eph A2 receptor into the cleaved form, which in turn increases EphA2 receptors in the plasma. However, it is unclear whether the measured EphA2 receptor reflects the soluble or membrane-bound form due to limitation of the kit used in this study. Furthermore, in vitro and in vivo evidence that clearly supports an association between plasma EphA2 receptor signaling and sepsis is rare. Third, we did not analyze serial EphA2 receptor levels because we were unable to collect serial blood samples from the patients. Therefore, additional studies investigating the measurement of plasma EphA2 receptor levels in sepsis patients and evidence of an association between EphA2 receptor signaling and sepsis are needed.

Conclusion
An increased plasma EphA2 receptor level was associated with sepsis severity and 28-day mortality among sepsis patients. The plasma EphA2 receptor may be considered a potential therapeutic target in sepsis patients. The EphA2 receptor level may be used for treatment monitoring in sepsis patients treated with an EphA2 receptor blocker. Clinical data and blood sample collection. Data of all patients admitted to the MICU were collected from the hospital electronic medical records and their blood samples. The severity of each patient's condition was classified according to two different scoring systems: The SOFA and APACHE II scores were calculated 23,29 .

Methods
In addition, the Charlson Comorbidity Index was used to evaluate patients' comorbidities 45 . Clinical parameters such as the development of ARDS, blood culture positivity, 28-day mortality, and other demographic characteristics were evaluated. Venous blood samples were collected within 24 hours after admission to the MICU though central lines into tubes containing ethylenediaminetetraacetic acid. Simultaneously, we determined the CRP, procalcitonin, creatinine, and albumin levels, blood culture results, white blood cell and platelet counts, and APACHE II and SOFA scores. Plasma was prepared by centrifugation for 15 minutes at 800 × g and 4 °C. Supernatants from centrifuged blood were immediately aliquoted and stored at −80 °C until the analysis was performed. Plasma levels of the EphA2 receptor and other cytokines (IL-1β, IL-10, IL-18, IL-6, TNF-α, and interferon gamma induced protein-10) were measured using the Human Magnetic Luminex ® Screening Assay Kit (R&D Systems, Inc., Minneapolis, MN, USA). The kit detects all forms of the EphA2 receptor including the soluble, membrane-bound, phosphorylated and non-phosphorylated forms All samples and standards were assayed in duplicate using the Luminex 200 TM System (Merck Millipore, Darmstadt, Germany).
Ethical approval. The study protocol was submitted as an ICU cohort and approved by the Institutional Review Board (IRB) of Severance Hospital (IRB number: 4-2013-0585). All study procedures were performed in accordance with the relevant guidelines and regulations. In addition, this study was performed in compliance with the principles set forth in the Declaration of Helsinki.
Statistical analysis. Data are described as medians with IQRs. The chi-squared and Fisher's exact tests or the Mann-Whitney U test was used to assess differences between the two groups. The Kruskal-Wallis test was used to compare three or more groups of a qualitative parameter. Pearson correlation analyses were performed to estimate associations between the biomarkers and the APACHE II and SOFA scores. AUC analyses of ROC curves were performed to compare the plasma EphA2 receptor level, APACHE II score, and SOFA score. Subsequently, the 28-day survivor and non-survivor groups were analyzed using a Cox proportional hazard model with a plasma EphA2 receptor level cut-off value to predict 28-day mortality according to the ROC and several variables. The Kaplan-Meier method was used to report survival curves that were analyzed using the log-rank test. In all cases, a p-value of <0.05 was considered statistically significant. SPSS version 20 (IBM, Armonk, NY, USA) was used for statistical analyses; AUC analyses of ROC curves were performed using MedCalc software (version 16.4.3; MedCalc, Oostende, Belgium).