Introduction
Sepsis is still a leading cause of morbidity and mortality worldwide, despite modern intensive care medicine. Numerous biomarkers have been investigated in the diagnosis and clinical assessment of sepsis, but only few have found their way into clinical routine [
1]. Although good clinical scores for disease severity have been developed [
2‐
4], patient assessment on a daily basis remains a challenge and could be facilitated by molecular disease parameters. Evidence is growing that platelets play an active role in fighting infections and in innate immunity [
5‐
8], and thrombocytopenia is frequently observed in systemic infections. Previous studies have found that the severity of thrombocytopenia is associated with increased mortality rates for patients in intensive care units in general [
9‐
13]. Although this has been frequently interpreted as a consumption of platelets [
14], signs of cell death have also been observed in platelets. Just lately, our group was able to show that bacteria can directly activate the apoptotic pathway in platelets to induce platelet cell death
in vitro[
15]. Several other groups demonstrated that the septic milieu compromises mitochondrial function, which leads to dysfunction of blood cells, myocardium, and muscle wasting [
16‐
20]. In this context, platelet mitochondrial function was also affected in the sepsis patient, and mitochondrial dysfunction correlated with disease severity and poor outcome in different study designs [
21‐
24]. The focus of this work was to investigate platelet mitochondrial membrane potential (Mmp) with flow cytometry with regard to disease outcome and disease severity in patients with sepsis. Platelet Mmp was correlated with clinical disease-severity scores (APACHE II, SOFA, and SAPS II score) on admission and was compared with that in control patients. We reassessed platelet Mmp in patients with severe sepsis during the disease course and compared it in survivors and nonsurvivors. The protein Bcl-xL, which regulates apoptosis in platelets and is critical for cell survival [
25,
26], was investigated in severe sepsis with immunoblotting. Overall, in this study, we aimed to investigate whether platelet mitochondrial membrane depolarization could be a potential biomarker of platelets in sepsis. Our hypothesis was that mitochondrial membrane depolarization of platelets correlates with increased disease severity and worse clinical outcome.
Discussion
In this study, we demonstrated that platelet mitochondrial membrane depolarization correlates with disease severity (APACHE II, SOFA, and SAPS II score) and disease outcome in patients with sepsis. Mmp-Index was significantly decreased in a subgroup of patients with severe sepsis compared with sepsis patients without organ failure (nonsevere) and controls (Figure
2). Platelet counts in the severe-sepsis group were significantly reduced compared with the other groups, which implies that the increased apoptotic activity in the septic milieu reflects the loss of platelets and may explain the prognostic relevance of thrombocytopenia. Patients classified as sepsis without organ failure (nonsevere) showed no significant loss of mitochondrial membrane potential compared with control patients. This is likely because predicted mortality based on APACHE II, SOFA, and SAPS II scores in our patients with nonsevere sepsis was relatively low.
Overall, the link of thrombocytopenia as a predictor of clinical prognosis in patients in the ICU has been well established, although investigations on a molecular level are still rare. Besides a loss of platelets, hemostatic functions of platelets, such as aggregation, are impaired during sepsis, which also correlates with the severity of infection [
34]. During sepsis, platelets release prothrombotic particles, which can activate systemic coagulation and lead to lethal embolic events [
35]. Growing evidence suggests that platelets have the capacity to fight bacterial infections directly, and that they act as cells of innate immunity [
5,
7,
8]. We and others have been able to demonstrate that platelets directly interact with bacteria and release bactericidal proteins that inhibit bacterial growth [
6,
36‐
38]. Bacteria in return release potent exotoxins that drive platelets into cell death [
15]. It is thus not surprising that platelet numbers decline at a rate that reflects infectious disease activity and that thrombocytopenia has prognostic value, as suggested by previous work [
10]. Therefore, it is tempting to speculate that the loss of platelets is a result of direct bacterial action or the septic milieu rather than random clearance and consumption of platelets.
Mitochondria are the central target in the intrinsic apoptotic pathway leading to platelet apoptosis [
25], and mitochondrial damage constitutes an early indicator of cell death in platelets [
33]. Although closely linked to apoptosis [
25,
33,
39,
40], mitochondrial dysfunction is not an exclusive sign of cell death and can also occur with microcirculatory distress [
41,
42]. During clinical or experimental sepsis, mitochondrial dysfunction has also been observed in heart and skeletal muscle tissue [
18,
19], as well as in apoptosis of blood cells such as monocytes [
20] and lymphocytes [
43]. Based on experimental data from our own group and observations by others, we have reason to believe that mitochondrial dysfunction, platelet cell death, and thrombocytopenia are not coincidental in the septic milieu. We have previously shown that live bacteria (
E scherichia
coli, S taphylococcus
aureus) isolated from sepsis directly activate a calpain-dependent degradation pathway that eliminates the anti-apoptotic protein Bcl-xL [
15]. Bcl-xL is critical for platelet survival, and inhibition or neutralization of Bcl-xL induces platelet cell death. This process results in platelet mitochondrial membrane depolarization and programmed cell death of platelets
in vitro. We measured Bcl-xL protein expression in platelets from patients with sepsis on admission (day 1) and found decreased expression of Bcl-xL in patients with severe sepsis who had low Mmp-Index values compared with patients with nonsevere sepsis (Figure
4) in this study. We concluded that the septic milieu probably generates a multitude of factors, including toxins that damage platelets and reflect the degree of septic activity.
Other groups have described that bacterial factors of
S. aureus induce mitochondrial membrane depolarization and apoptosis in platelets [
44]. Although these observations support the hypothesis that mitochondrial dysfunction, apoptosis and thrombocytopenia are physiologically connected, we cannot prove the link in this complex
in vivo system, but they appear to be reasonable explanations of factors that contribute to the process of sepsis-induced thrombocytopenia. Undoubtedly, the etiology of thrombocytopenia in sepsis is multifactorial and includes microcirculatory and coagulation-associated effects.
A previous study first demonstrated a correlation of platelet mitochondrial membrane depolarization and clinical disease severity on admission in a diverse group of medical and surgical ICU patients with SIRS by a different approach [
24]. Before that, Sjövall and colleagues [
23] observed a substantial increase in platelet mitochondrial respiratory capacity in nonsurvivors compared with survivors of sepsis and demonstrated an association between mitochondrial dysfunction and clinical outcome in sepsis. They further detected mitochondrial damage by release of cytochrome
c from mitochondria. However, no correlation between mitochondrial respiration and standard clinical assessment scores for illness (APACHE II, SAPS, and SOFA Score) was found. Most likely, this is explained by a different approach to assess mitochondrial dysfunction, which probably differs from the loss of mitochondrial membrane potential in the apoptotic process.
In this study, we further aimed to investigate whether the degree of platelet mitochondrial membrane depolarization could also be correlated with disease severity during the disease course and outcome. We compared platelet mitochondrial membrane depolarization in a subgroup of patients with severe sepsis, including septic shock on admission and during clinical disease course, and compared Mmp-Index in survivors and nonsurvivors. On admission, initial Mmp-Index values in the severe-sepsis group were significantly decreased compared with controls, but no significant difference between survivors and nonsurvivors on admission was found (Figure
3A). Supporting these findings, no difference was noted regarding the clinical disease-severity scores APACHE II, SOFA, and SAPS II on admission between survivors and nonsurvivors in this study. Individual and group analysis of patients from the sepsis group illustrates that patients who survived sepsis showed recovery of Mmp-Index on follow-up, whereas no significant recovery in sepsis nonsurvivors was observed (Mmp-Index <0.5) (Figure
3A, B, and Additional file
1). It appears that follow-up Mmp-reads to identify clinical trends may be equally important to the absolute Mmp-values. Mmp-Index may prove a valuable tool to evaluate clinical recovery of the individual patient in addition to clinical parameters. The data suggest that markers of platelet apoptosis, such as mitochondrial membrane potential and Bcl-xL levels, may reflect infectious disease activity and disease severity in sepsis.
Conclusions
Platelet Mmp-Index proved to be a highly reproducible, easily accessible readout for mitochondrial integrity and showed good correlation with clinical disease severity during clinical course, correlated with clinical outcome, and normalized with recovery. In addition, we detected a proapoptotic phenotype of platelets based on Bcl-xL levels in patients with severe sepsis compared with controls. Further research on larger patient populations will, however, be necessary to establish markers of platelet apoptosis in daily routine. Nonetheless, results from this study and previous investigations seem to promise that markers of platelet apoptosis may assist in the assessment of disease severity, in risk stratification, and in management of patients with sepsis in the future.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
KG, experimental design, data collection and analysis, figure preparation, final approval of the manuscript; MA, patient enrollment, manuscript writing, critical revision, and final approval of the manuscript; RH, patient enrollment, data collection and analysis, final approval of the manuscript; PB, patient enrollment, data collection and analysis, final approval of the manuscript; TA, conception and design, critical revision, and final approval of the manuscript; AC, statistics and calculations, data analysis, final approval of the manuscript; HYS, conception and design, manuscript writing, final approval of the manuscript; SM, conception and design, critical revision, and final approval of the manuscript; BFK, conception and design, data collection and analysis, manuscript writing, and final approval of the manuscript. All authors read and approved the final manuscript.