Cell-derived microparticles: ‘Miniature envoys with many faces’


Yeon S. Ahn, School of Medicine, University of Miami, Miami, FL, USA.
Tel.: +1 305 243 6703; fax: +1 305 243 5957; e-mail: yahn@med.miami.edu

Endothelial cells (EC) normally present a non-adhesive and anticoagulant surface to flowing blood, but EC injury can provoke cell adhesion and initiation of coagulation with consequent thrombosis and inflammation [1,2]. Accordingly, detection of EC dysfunction is pivotal to the early diagnosis and prophylactic intervention of many prethrombotic conditions. Unfortunately, however, monitoring EC status is made difficult because of its inaccessibility. This situation has improved in the last few years, as it has become evident that EC, like other blood cells, shed membranous vesicles, termed endothelial microparticles (EMP), in amounts and phenotypes reflecting the states of the parent cell. They consist of somewhat heterogeneous species of microparticles (<1 μm) carrying many biomarkers of the parent cells, the analysis of which can therefore provide useful information on EC status. For example, it has been shown that EMP released upon activation exhibit distinctive antigenic phenotypes and functions compared to those released in apoptosis [3].

The number of papers on the value of EMP assay for assessing EC dysfunction has escalated during the last 5 years. There are some indications that EMP may be more sensitive than conventional tests in evaluating disease activity. In acute TTP, for example, EMP levels drop ahead of falling LDH or rising platelet counts [4,5]. In untreated hypertension, EMP correlated well with systolic and diastolic blood pressure, and appeared to be more sensitive than soluble markers of EC activation, sVAM-1, sICAM-1, and VWF [6]. (There is evidence that some soluble markers are in reality MP-bound (2).) In acute coronary syndromes (ACS), EMP were significantly elevated compared to stable disease or controls [7,8] and discriminated acute MI from unstable angina [8], suggesting that EMP levels reflect severity of vascular lesions. The latter was supported by finding EMP significantly higher in patients with high-risk angiographic lesions compared to those with low-risk lesions [9]. An EMP elevation was seen not only in an advanced stage of atherosclerosis such as ACS, but also in early stages such as metabolic syndrome [10] and diabetes mellitus [11]. A rapid rise of EMP was also demonstrated in healthy young volunteers following ingestion of a single high-fat meal, and the EMP correlated well with serum triglycerides [12]. EMP are also elevated in multiple sclerosis, an inflammatory disorder, during exacerbations but not remissions [13]. These findings support the concept that EMP assay can be useful in detecting even subtle disturbances of EC function.

However, a number of challenging problems remain in the field of MP investigations. Perhaps at the top of the list is the need for some kind of standards for comparing results among different laboratories, addressed in a recent forum in this journal [14]. On a different frontier is the problem of delineating the functional roles played by EMP (and other MP) in health and disease. There is growing consensus that EMP and other MP can act as diffusible messengers, transporting bioactive agents to initiate and mediate coagulation, inflammation, and cell–cell interactions. See Fig. 1. New observations on various roles of EMP is on the rise. For example, at the 2004 ASH meeting, two groups independently reported high levels of MP-associated tissue factor in cancer hypercoagulable states [15,16]. At the same meeting, Jy et al. [17] reported that EMP carry large-multimer VWF, which interacts with platelets to yield exceptionally stable platelet aggregates. Such observations amply support protean roles for EMP in hemostasis, thrombosis, and probably also inflammation, as EMP bind and activate leukocytes and enhance leukocyte transmigration [18,19]. The EMP are known to also carry an array of cytokines and other agents [1], some of which are indicated in Fig. 1, based on another recent review [20].

Figure 1.

Some functional agents associated with MP. Many other agents which are known to be membrane-associated and released from the parent cell are likely to also occur on MP, such as BAFF, fractalkine, MIP-1, RANTES, beta-thromboglobulin, CX3CR1, MMP-9 and others on platelets. A number of agents now regarded as soluble may in fact be MP-bound. Abbreviations: vWF, von Willebrand factor; β2GP1, beta-2 glycoprotein 1; TF, tissue factor; PF3, platelet factor 3 activity, due mainly to anionic phospholipids such as PS, phosphatidyl serine; TFPI, TF pathway inhibitor; EPCR, endothelial protein C receptor; PAF, platelet activating factor; PF4, platelet factor 4; TSP, thrombospondin; IL-1, interleukin-1; Fas/L, Fas and its ligand FasL; CD31, PECAM-1; CD40/40L, the latter being CD40 ligand; CD41, glycoprotein IIb or integrin aIIb; CD42, complex of GPIX, GPlb-alpha; CD51, vitronectin receptor, an integrin; CD54, ICAM-1; CD62E,P,L, selectins originally believed to specific for endothelium (E), platelets (P) and leukocytes (L); CD105, endoglin; CD106, VCAM-1; CD144, VE-cadherin or cadherin-5; AβPP, amyloid beta precursor protein.

New ideas

In the current issue, Hussein et al. [21] report some additional new findings regarding the generation and content of EMP. Most intriguing is their demonstration of caspase-3 in EMP from human umbilical vein endothelial cells (HUVEC) stimulated with interleukin-1 alpha (IL-1α). Moreover, not only do they find caspase-3 associated with EMP, but also with all MP in three lupus (SLE) patients, i.e. also from red cells (RMP), monocytes (LMP), and platelets (PMP), as determined by coexpression of glycophorin, CD14, and integrin β3, respectively. In the HUVEC experiments, EMP numbers measured by annexin V binding correlated with the number of apoptotic (detached) EC, leading them to conclude that EMP arise largely from detached (apoptotic) EC. They speculate that caspase-3 may be involved in membrane vesiculation and that shedding of caspase-3 in MP is one way EC minimize or escape apoptosis. They also propose the possibility that caspase-3 in MP may be delivered to other cells, inducing apoptosis. Consistent with this hypothesis are studies demonstrating that incubation of MP from ACS patients with EC in vitro induces EC dysfunction [22,23].

These observations are interesting, novel, and of potential importance in suggesting new roles of MP and mechanisms of their production. Most MP studies thus far have focused on their surface phenotypes and procoagulant functions. This is perhaps the first to gaze into the vesicle interior, using saponin permeation to reveal caspase-3 within all subtypes of MP examined. Before discussing the salient implications, however, some limitations of the study must be pointed out. First, they investigated only one agonist, IL-1α, which we suspect may be inducing predominantly apoptotic EC, as distinct from activated EC. Secondly, their detection of MP, both in vitro and in vivo, was by FITC-annexin V, which is specific for anionic phospholipids (PL) such as phosphatidylserine (PS); but we have found that PS-positive EMP are abundant only from apoptotic cells, whereas annexin V binding has low efficiency at detecting EMP from activated EC [2]. Thirdly, they examined MP from too few SLE patients to draw general conclusions about them, requiring confirmation in future studies.

Non-apoptotic function of caspase-3

Aside from the above caveats, the finding of caspase-3 in MP by Hussein et al. deserves comment. The central role of caspase-3 in apoptosis is well known, but less well known are its non-apoptotic functions. Caspase-3 and other caspases play roles in the enucleation and maturation of erythroid cells [24], differentiation of monocytes to macrophages [25], and in cell dispersal and migration [26]. Shcherbina and Remold-O'Donnell [27] demonstrated a caspase-3-like protease in platelets when activated by thrombin + collagen or calcium ionophore A23187, and its activation was associated with PS exposure, MP release, and cleavage of the cytoskeletal protein, moesin, yet had no effect on alpha granule release, shape change or aggregation. They employed the caspase inhibitor DVED-fmk as well as calpeptin, an inhibitor of calpain, which has also been implicated in vesicle shedding. They detected platelet procaspase-3 (32 kDa) but had difficulty finding the active product, caspase-3 (17 kDa), leading them to conclude that it must be fleeting or labile [27]. Wadhawan et al. [28] found ‘apoptosis-like’ changes in stored platelets, namely, release of cytochrome c from mitochondria, exposure of PS, and cleavage of procaspase-9, but wihout caspase-3 activation. On the other hand, Wolf et al. [29] concluded that calpain, rather than caspase, was key to the apoptotic changes in platelets, and that caspases might instead be involved in platelet senescence and production. Recent studies showed that the pan-caspase inhibitor zVAD-fmk decreased both ADP- and A23187-induced PS exposure, and decreased ADP-induced platelet aggregation, indicating a role of caspases in platelet activation [30], but this inhibitor also affects calpain. Caspase activation was also implicated in chronic platelet activation leading to hypercoagulable states in uremic patients [31]. Involvement of caspase in platelet storage lesion [28] raises the possibility of inhibiting it for improved platelet preservation in storage.

In light of Hussein et al.'s work [21], many of these discordant observations in platelets, at least, may be explained if caspase-3 quickly disappears because it is shed off with MP. Not only might caspase-3 be shed off with MP, but the possibility also exists that caspase-3 is central to the very process of vesiculation, as suggested by Hussein et al. [21] and by Shcherbina and O'Donnell [27].

Implications of caspase-3 in MP

The functional role of caspases in MP, if any, remains speculative, but it is tempting to wonder if they participate in the unknown mission or missions of MP: might they give the kiss of death to other cells, as Husein et al. suggest? It is easy to imagine them being phagocytosed, thus liberating active caspase-3 in phagocytes. (PS specifically signals phagocytosis via a PS receptor [32,33], although the specificity of this is recently questioned [34].) As platelets are shed from megakaryocytes, they may fairly be viewed as ‘giant’ MP and thus may share features with their ‘miniature’ cousins, the cell-derived MP. Indeed, we have shown that MP derived from platelets seem to function like surrogate platelets, substituting for platelets in thrombocytopenic patients [35]. Given that, the presence of caspase in MP may add another parallel between platelets and PMP. It remains to be seen if the functional role(s) of MP can be nearly as complex and delicate as platelets, of comparable importance and versatility. Whatever the final answer, it seems certain that cell-derived MP will prove to be more diverse, heterogenous, and active than we yet know –‘miniature envoys’ from many cell lines, swarming to their appointed sites in response to the appropriate alarm. They are well equipped for such a mission, though the details remain largely mysterious.

Our knowledge of MP is still nascent. Certainly, MP studies deserve further exploration, and we thank Hussein et al. for discovering some new clues to pursue.


We are indebted for the general support of the Wallace H Coulter Foundation.

Wenche Jy PhD, Lawrence L Horstman BS, and Joaquin J Jimenez MD at the Wallace H Coulter Platelet Lab have contributed to the commentary.