Introduction
The presence of circulating microparticles (MPs) in septic patients is well recognised [
1,
2] and is inducible by thrombin [
3], cytokines [
4], lipopolysaccharide (LPS) [
5] and collagen [
6]. Derived from cell membrane shedding as a result of activation or apoptosis, circulating MPs constitute a marker of vascular and systemic disease [
7]. Rearrangement of membrane phospholipids during MP release can result in increased phosphatidylserine availability with procoagulant activity. In patients with myocardial infarction and diabetes mellitus, elevated MP levels correlate with increased thromboembolic risk [
8,
9]. However, their functional role in the pathophysiology of sepsis remains unclear.
Elevated circulating MPs do not cause thrombosis in healthy individuals, principally due to the protective effects of the natural anticoagulant, activated protein C (APC) [
10]. APC is an anticoagulant [
11] with anti-inflammatory and anti-apoptotic properties [
12]. These beneficial effects may be explained by its binding to the endothelial protein C receptor (EPCR) with cleavage of endothelial protease activated receptor 1 (PAR1) [
13]. Although, the relative
in vivo usefulness of these effects are not yet known, recombinant human APC (rhAPC) is currently used to treat patients with sepsis [
14]. Its current use remains controversial because of reports of severe bleeding complications during rhAPC treatment [
15] and a second phase 3 trial is ongoing (PROWESS Shock) [
16].
We have previously demonstrated that APC can generate MPs
in vitro from EPCR-expressing cells, which retain anticoagulant and PAR1-dependent anti-inflammatory properties [
17].
In vivo demonstration of these APC-MP in septic patients during rhAPC infusion [
18] led us to hypothesize that such circulating MPs may retain their anti-inflammatory, and cytoprotective properties in these patients. An increased number of these MPs would thus translate into clinical benefits for the patient with severe sepsis.
Discussion
Our work is the first to demonstrate that APC can induce the generation and release of MPs
in vitro and
in vivo. A key finding from this study is that circulating MPs from patients during rhAPC treatment for severe sepsis possess APC-specific functional effects. The interaction between these MPs with endothelial cells can induce changes in gene expression, which translate into inhibiting apoptosis and reducing endothelial permeability. As these effects require PAR1 activation at the endothelial surface by APC in an EPCR-bound conformation [
18], these results provide functional confirmation to earlier microscopy evidence of the assembled EPCR-APC complex on
in vivo-derived MPs. Our findings also indicate that APC binding to EPCR within circulating MPs is stable and not displaceable by non-activated PC.
Whilst circulating MPs in sepsis and inflammatory conditions have been widely described, data on their function are limited. The majority highlight pro-inflammatory [
27,
28], pro-apoptotic [
29] and pro-coagulant [
30,
31] potential of circulating MPs. Whilst increased MP levels in disease states are generally associated with adverse outcomes [
32‐
34], Soriano reported that endothelial MP levels were higher in survivors of severe sepsis than in non-survivors [
35]. Our findings would support this with rhAPC treatment significantly increasing endothelial-derived CD13+, EPCR+ MPs with a clinical correlative trend towards improved outcome. Expression of the functional EPCR-APC complex may be relevant to this trend in survival benefit as such MPs could bypass the requirement for cellular co-localisation of EPCR and PAR1 to induce APC anti-inflammatory and cytoprotective signaling [
18]. As inflamed cell surfaces are EPCR deficient [
36], MPs could be a vehicle for APC to act at a distance on PAR1 in other cells. This would reduce endothelial apoptosis and permeability in distal vascular sites, as supported by increases in anti-apoptotic A20 and Bcl-x with concomitantly reduced pro-apoptotic Bax. This would also reduce endothelial permeability as KDR up-regulation has barrier protective functions [
18]. Our findings that APC-induced MPs are anticoagulant active suggest that these MPs might contribute to the bleeding risks associated with rhAPC treatment. No evidence of increased bleeding was, however, observed in the rhAPC-treated patient group.
Since thrombin can also cleave PAR1 to cause endothelial barrier disruption, the question as to how PAR1 mediates opposing effects has been debated and further clarified. Recently, APC-PAR1-dependent protection has been demonstrated
in vivo in reversing inflammation-induced vascular leakage [
37]. Such work in physiological systems, and also work in our own group [
18], suggest that a key molecular differential is the APC-induced interaction between EPCR and sphingosine 1-phosphate that is important to cytoskeleton stability [
38]. EPCR is co-localised with PAR1 in lipid rafts and APC facilitates their coupling to channel Gα
i-PAR1 protective signalling [
38]. Gα
i-PAR1 recruitment also modifies the inner leaflet of plasma membranes and this may be mechanistically linked to MP formation [
39]. With EPCR located within these glycosphingolipid-rich microdomains (rafts) and caveolae, this provides a likely pathway for calcium-regulated kinases and phosphatases, guanine nucleotide exchange factors and mitogen-activated protein kinase cassettes to form and release MP-EPCR [
40]. As to thrombomodulin (TM) expression on APC-induced MPs, our investigations do not show a consistent presence (data not shown). This is not unexpected given that the cellular distribution of TM is different from EPCR [
41] and MP generation by APC is EPCR-dependent but TM-independent.
A limitation of this study is in the numbers of rhAPC treated patients, especially with the relatively few deaths in the rhAPC treated group. To optimise the validity of our findings, we were careful to control for the rhAPC treated group with a group of patients who were equally eligible for rhAPC treatment but who were not treated due to clinician concern of bleeding risks. The platelet count in this group was significantly lower (Table
2) but should not have affected our MP findings or conclusions as platelets do not express EPCR. In addition, we have both inter- and intra-patient comparisons with sampling before and during rhAPC treatment. Through this and the comprehensive characterisation of both prognostic and functional aspects of circulating MPs, this study has taken a basic observational discovery along the translational pathway towards possible clinical relevance. It is important to note that our findings are limited to APC-induced MPs, which to our knowledge is not influenced by heparin or drugs that might have an endothelial effect. This is because the APC effect is specifically cell receptor-mediated, that is, via EPCR and PAR1. A limitation to clinical translation is that the data may not be applicable to the most seriously septic patients whose mortality is much higher than in the study patients.
Our findings are the first to highlight that there is an additional circulatory form of EPCR in human plasma, that is, MP-EPCR, which is different from soluble EPCR. Soluble EPCR is the cleaved, truncated form of EPCR (45kD) that cannot facilitate APC proteolytic anticoagulant activity [
42]. Another novel finding is that this provides the first evidence in humans that the APC-PAR1 pathway is physiologically relevant because MP-EPCR release is dependent on PAR1 activation by APC [
17]. Whilst the relative importance of free versus MP-associated APC remains to be clarified, free APC levels can be variable during treatment in patients with sepsis [
43]. This may be due to stability and clearance factors or that cell-bound APC is not measured. Conversely, our findings that MP-associated APC are stable both in measurable levels and activities would point to physiological and clinical relevance as bioactive effectors in rhAPC-treated patients. This conclusion comes with the caveat that their clinical relevance requires further exploration, especially in larger numbers of patients with septic shock and higher mortality.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MPC participated in the study design and coordination, carried out the molecular genetic studies and functional assays, and helped to draft the manuscript. VT carried out the MP number experiments and functional assays and helped to draft the manuscript. CD carried out the MP number experiments. IW and DW contributed patient samples. JT helped to draft the manuscript. CHT conceived of the study and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.