Association of plasma exosomes with severity of organ failure and mortality in patients with sepsis

Abstract Current sepsis biomarkers may be helpful in determining organ failure and evaluating patient clinical course; however, direct molecular biomarkers to predict subsequent organ failure have not yet been discovered. Exosomes, a small population of extracellular vesicles, play an important role in the inflammatory response, coagulation process and cardiac dysfunction in sepsis. Nonetheless, the association of plasma exosome with severity and mortality of sepsis is not well known. Therefore, the overall levels of plasma exosome in sepsis patients were assessed and whether exosome levels were associated with organ failure and mortality was evaluated in the present study. Plasma level of exosomes was measured by ELISA. Among 220 patients with sepsis, 145 (66%) patients were diagnosed with septic shock. A trend of increased exosome levels in control, sepsis and septic shock groups was observed (204 µg/mL vs 525 µg/mL vs 802 µg/mL, P < 0.001). A positive linear relationship was observed between overall exosome levels and Sequential Organ Failure Assessment (SOFA) score in the study cohorts (r value = 0.47). When patients were divided into two groups according to best cut‐off level, a statistical difference in 28‐ and 90‐day mortality between patients with high and low plasma exosomes was observed. Elevated levels of plasma exosomes were associated with severity of organ failure and predictive of mortality in critically ill patients with sepsis.


| Study population
Critically ill adult patients (≥19 years of age) admitted to the medical intensive care unit (ICU) of Samsung Medical Center were considered eligible for inclusion in the registry. Exclusion criteria were as follows: (a) cognitive impairment, (b) inability to provide informed consent (or lack of an appropriate legal representative to do so), (c) ICU admission for simple procedure or postsurgical care, (d) transfer from other hospitals, (e) end-of-life decision or admission to facilitate comfort care, (f) haemoglobin <8 g/dL upon admission or persistent bleeding and (g) discharge within 24 hours of admission to the ICU.
Screening and enrolment were completed within 24 hours of ICU admission. A total of 220 patients registered between April 2014 and January 2019 were included in the analysis. In addition, a total of 20 healthy volunteers (≥19 years of age) donated 5 mL of blood samples for research purposes. Written informed consent was obtained from all healthy volunteers.

| Clinical data collection
A trained study coordinator used hospital records for each patient to prepare a standardized case report form. Clinical data consisting of patient demographics, reason for ICU admission, severity of illness scoring and laboratory data (PaO 2 /FiO 2 , lactic acid, C-reactive protein and procalcitonin) were obtained at the time of enrolment.

Illness severity was assessed using the Acute Physiology and Chronic
Health Evaluation II (APACHE II), 23 Simplified Acute Physiology Score 3 (SAPS 3) 24 and Sequential Organ Failure Assessment (SOFA) scores. 25 The primary outcome was 28-day mortality, and secondary outcomes were in-hospital mortality. Sepsis was defined according to the third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). 1 Since enrolment for the registry began in April 2014, patients enrolled before release of the new definition were reclassified.

| Measurement of plasma exosomes
In addition to clinical data, 19 mL of whole blood was drawn from each patient within 48 hours of study enrolment and healthy volunteers. To isolate plasma from whole blood, whole blood was centrifuged at 480 × g for 10 minutes at 4°C. Several plasma aliquots from each study participant were isolated and frozen at −80°C until further analysis.
ExoQuick exosome precipitation solution (System Biosciences, Palo Alto, CA, USA) was used to precipitate plasma exosomes. Briefly, plasma was thawed on ice and centrifuged at 1500 g for 10 minutes at 4°C. The supernatant was transferred to a new tube, and 2 µL of thrombin (System Bioscience, TMEXO-1) was added to 250 µL of plasma and incubated for 5 minutes at room temperature to remove fibrinogen. The plasma was then centrifuged at 10 000 rpm for 5 minutes, and the supernatant was collected. The plasma was then incubated with ExoQuick™ for 30 minutes at 4-5°C. The ExoQuick™/ plasma sample was then centrifuged twice at 1500 g for 30 and 5 minutes, respectively, in order to remove the supernatant. The pellet was resuspended in 200 µL of PBS. Then, the precipitated exosomes were used immediately.
To assess the size and morphology of exosomes, we performed transmission electron microscopy (TEM) at room temperature after the isolation of exosomes (80 kV, Phillip CM120; Figure S1).

| Statistical analysis
Data are presented as numbers (percentages) for categorical variables, and as median and interquartile range (IQR, 25th-75th percentiles) for continuous variables. Categorical variables were compared using the chi-square test or Fisher's exact test, and continuous variables were compared using the Mann-Whitney U test. Differences in plasma exosome levels in control, sepsis and septic shock groups were assessed with the Kruskal-Wallis test.
Associations between exosomes and severity of organ failure, as measured by SOFA score, were assessed using the linear regression.
We performed receiver operating characteristic (ROC) analysis to determine the predictive value of exosome level as a prognostic predictor of disease severity and calculated the optimal cut-off values of exosome level by Youden's index to analyse the relationship between exosome level and disease severity in our cohort. The optimal cut-off of plasma exosome levels for 28-day mortality prediction was assessed using Youden's index. 26 Patients were reclassified into two groups of high and low exosome levels based on the optimal cut-off level, and initial diagnosis, clinical status, severity of illness and 28-day mortality were compared between the groups. Kaplan-Meier equation was used to determine the 90-day mortality curves according to plasma exosome levels, which were then compared using the log-rank test.
All tests were two-sided, and a P-value <0.05 was considered statistically significant. All statistical analyses were performed using R version 3.5.3 (R Foundation for Statistical Computing, http://www.r-proje ct.org).

| Baseline characteristics
Baseline characteristics of all subjects are presented in Table 1.
The median age of patients was 67 years (IQR, 55-74 years), and 68% were male. During the study period, 145 (67%) patients presented with septic shock. The median Charlson comorbidity index of the patients was 2 points (1-3 points). A total of 91 (42%) patients required mechanical ventilation, and 166 (76%) patients needed vasopressor support. Significant differences were not observed in age, sex or comorbidities, in 28-day mortality and 90-day mortality between sepsis and septic shock patients (Table 1).

| Plasma exosome levels as a prognostic predictor of mortality and illness severity
The association between plasma exosome levels and severity of organ failure was assessed by linear regression method. There was a positive linear relationship between plasma exosome levels and SOFA scores in the study cohorts (r value = 0.47; 95% CI 0.36-0.56) ( Figure 2). To analyse the relationship between plasma exosome level and 28-day mortality in the study cohort, the optimal cut-off value of plasma exosomes was calculated using Youden's index. The best cut-off point was 809 μg/mL in the relationship between plasma exosome level and 28-day mortality. Patients were reclassified into two groups of high and low exosome levels based on the optimal cut-off level, and initial diagnosis, clinical status, severity of illness and 28day mortality were compared between the groups. The group with exosome levels higher than 809 μg/mL was significantly associated with septic shock, need for mechanical ventilation or vasopressor support, severity of illness defined by SAPS 3, APACHE II and SOFA scores, and 28-day mortality (Table 2). Furthermore, Kaplan-Meier survival estimation demonstrated a significant difference in 90-day survival between patients with high and low plasma exosome level (log-rank P = 0.004; hazard ratio for death 1.92; 95% CI, 1.21-3.05) ( Figure 3).
In the present study, the role of plasma exosome level as a potential diagnostic marker for septic shock was assessed and whether plasma exosomes could help predict the prognosis of patients with sepsis was evaluated. The result of our study indicated that the plasma exosome level was significantly higher in septic shock patients than that in sepsis patients and healthy volunteers. In addition, plasma exosome level was associated with severity of organ failure and predictive of mortality.
To the best of our knowledge, this is the first study to comprehensively evaluate the association between plasma exosome levels and severity and mortality of critically ill patients with sepsis or septic shock. The strength of our study lies in the quantification of overall exosomes in a relatively larger cohort of critically ill patients, whereas most of the preliminary studies were conducted in either animal models or focused on characterization of specific exosome contents or in small number of cohorts. As was evident in previous studies, exosomal contents such as miRNA or proteins from diverse origins provoke or prevent myocardial dysfunction, acute lung injury and acute kidney injury. 14,27-29 A murine model showed exosomal microRNA (miR-126) derived from endothelial progenitor cells was associated with microvascular dysfunction 28 ; however, the role of exosomal microRNA was not confirmed in humans. In addition, the specific plasma exosomal protein associated with progression of sepsis has been identified in sepsis patients, 30 but cannot be generalized to date. In another study, differences in exosomal microRNA expression in healthy volunteers, patients with sepsis and patients with septic shock were established using next-generation sequencing and quantitative reverse transcription-polymerase chain reaction. 31 Although three microRNAs were associated with mortality, the extraction and analysis method for microRNAs may have been too complicated for use as a routine biomarker, especially for critically ill patients who require rapid laboratory results. Finally, Real  (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) .021 Initial SOFA score 9 (6-11) 6 (4-9) 10 (8-12) <.001 Mortality 28-d mortality 39 (18) 12 (16) 27 (19) .712 In-hospital mortality 53 (24) 17 (23) 36 (25) .868 90-d mortality 73 (33) 23 (31) 50 (35) .651 Note: Data are presented as median (interquartile range) or number (%). Compared to previous studies, our data clearly showed that the overall level of plasma exosome in sepsis patients was associated with organ failure and mortality, suggesting its potential as a biomarker for assessing severity and predicting mortality of sepsis.
However, additional multicenter studies are required to confirm its efficacy and reliability as a biomarker for sepsis. Furthermore, considering the fact that there is no validated single marker for diagnosis and prognostication of sepsis, it seems necessary for the future investigations to evaluate the competency of exosome not only as a single biomarker compared to the known markers or scoring systems but also the efficacy of combining exosome with other biomarkers or incorporating it into existing scoring systems.
Of note, pan-exosome antibody of CD9 was used in our study to capture exosomes in whole, therefore, not specific contents of exosomes such as protein, lipid, mRNA or miRNA but plasma level of  Previous studies demonstrated that exosomes are implicated in the development of organ dysfunction through cell-to-cell communication or organ crosstalk in sepsis. [34][35][36][37] This finding suggests that exosomes may play a role in organ failure and sepsis without evident changes in the level of specific contents. In this regard, we tried to analyse the relationship between overall exosome and severity and mortality of sepsis. However, further studies are warranted to determine the role of specific exosomal contents as well as mechanism of action of exosomes in sepsis.
The present study had several limitations. First, the study was conducted at a single referral centre, which may limit the generalizability of the data. Second, subsets of exosomal miRNAs and proteins were not analysed as potential biomarkers in the study.
Previous studies have illustrated that various exosomal contents influence the pathogenesis of sepsis and host response in diverse ways. Although overall quantitation of exosome may be an advan-

| CON CLUS IONS
In conclusion, the study results indicate that plasma exosome levels were associated with severity of organ failure and predictive mortality in critically ill patients with sepsis.

CO N FLI C T S O F I NTE R E S T
The authors disclose that they do not have any conflicts of interest.

E TH I C S A PPROVA L A N D CO N S E NT TO PA RTI CI PATE
The study was approved by the institutional review board of Samsung Medical Center (IRB no. 2013-12-033). Written informed consent was obtained from patients or their legally authorized representative prior to enrolment.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data set used in this study is available from the corresponding author on reasonable request.

F I G U R E 3
Kaplan-Meier survival estimation of patients with high and low plasma exosome level (log-rank P = 0.004)