Circulating microparticle levels are reduced in patients with ARDS

Acute respiratory distress syndrome (ARDS) remains common and carries significant morbidity and mortality [1] despite lung-protective ventilator strategies [2] and conservative fluid management [3]. One potential reason for the lack of proven clinical therapies is that patients are typically well into their clinical illness (48–72 h) before being enrolled in clinical trials, which may be beyond the time frame for successful therapeutic intervention. Compounding this problem is a lack of good predictors of development of ARDS in at-risk patients. One potential solution is the identification of biomarkers associated with development of ARDS in critically ill patients so that therapies can be started earlier in the most vulnerable patients. Although several biomarker studies of ARDS have been completed, no single biomarker has been predictive of development of ARDS [4]. In this study, we examined circulating microparticles (MPs) and coagulation activation in patients at risk for ARDS to determine if these biomarkers might be useful for identifying ARDS in at-risk patients and to determine if these biomarkers are associated with poor outcomes in patients who develop ARDS.

Activation of systemic coagulation is an early feature of ARDS [5] that contributes to microvascular thrombosis and organ dysfunction. MPs are central to activation and propagation of thrombosis through exposure of negatively charged phospholipids on their surface and delivery of tissue factor (TF) to the site of thrombosis [6‐10]. We previously observed that patients with ARDS had higher levels of MPs in the airspace than a critically ill control group of subjects with cardiogenic pulmonary edema and that higher levels were associated with worse clinical outcomes [11]. Further, we found that these MPs contain TF and have procoagulant activity. On the basis of these results showing an important role of MPs in coagulation and ARDS in the airspace, we hypothesized that circulating MPs would be higher in at-risk patients who develop ARDS than in those who do not develop ARDS. Further, we hypothesized that higher levels of circulating MPs would correlate with severity of ARDS, be associated with increased activation of coagulation, and predict poor clinical outcomes.

In this prospective study of patients at risk of developing ARDS, lower levels of circulating MPs were associated with the development of ARDS in critically ill patients at risk for ARDS. The association between lower plasma levels of MPs and the development of ARDS was independent of age, sepsis, and severity of illness. MP concentration differed by risk factor for ARDS and was stable between enrollment on ICU day 2 and ICU day 4. Interestingly, the inverse association between MP levels and development of ARDS was strongest in the subgroup of patients with sepsis.

The role of MPs in the development and evolution of ARDS is not well understood. Several in vitro studies have shown that MPs from different sources can induce inflammation and injury in lung endothelial cells. MPs from stimulated THP-1 monocytes in culture induce apoptosis in human pulmonary microvascular endothelial cells (hPMVECs) [22]. Similarly, platelet-derived MPs promote polymorphonuclear leukocyte-induced hPMVEC death [23]. In addition, MPs from apoptotic monocytes induce TF expression and platelet adhesion to human vascular endothelial cells [24]. In vivo, MPs have been shown to contribute significantly to experimental lung injury. Rats subjected to high tidal volume ventilation had increased circulating MPs after 4 h of ventilation [25]. Li et al. showed that circulating MPs isolated from the blood of lipopolysaccharide (LPS)-treated rats were sufficient to induce lung inflammation and cytokine production when injected either intravenously or intratracheally [26]. Similarly, mice and rats treated with MPs derived from cultured endothelial cells caused lung inflammation and permeability [27, 28]. On the basis of these studies demonstrating an injurious role for MPs in lung and vascular injury, we hypothesized that increased MP levels would be associated with increased risk of ARDS. However, we found instead that patients with lower circulating MPs were at increased risk of developing ARDS. This was a surprising observation and highlights the complexity of the impact of MPs on lung injury. Our findings are consistent with one previous clinical study of circulating MPs in patients with ARDS which showed that higher levels of endothelium-derived MPs were associated with reduced mortality in patients with ARDS [29]. Our data suggest that the effects of MPs may differ on the basis of specific clinical situations; for example, MPs were inversely associated with ARDS development only in patients with sepsis and not in those without sepsis. Further studies are necessary to fully understand the mechanisms that underlie these observations.

The reason for the contradiction between clinical studies of MPs in ARDS and experimental studies of MPs in culture or animal models is uncertain. One possibility is that there are subpopulations of MPs from some cellular sources that are protective. For example, MPs derived from human mesenchymal stem cells were protective against Escherichia coli-induced ALI when administered either intratracheally or intravenously [30, 31], whereas MPs from platelets, monocytes, or endothelial cells are injurious [22, 23, 27, 28, 32]. Second, it may be that bioactive MPs with exposed phospholipids generated in the setting of acute inflammation are sequestered in organs such as the lung rather than in the circulation. Whereas circulating MPs are lower in subjects with ARDS, MPs attached to the endothelium in the lung or other organs may be increased. Consistent with this hypothesis, MPs can be sequestered in developing thrombi through expression of adhesion molecules [33, 34]. Third, MP isolation and preparation for experimental use may alter MPs in such a way that they become proinflammatory. Supporting this concept, most studies using isolated MPs demonstrate injurious effects, whereas clinical studies of MPs show protective effects. In addition, our MP capture assay measures only MPs with phosphatidylserine on the external surface; this may be a subset of total MPs. Thus, different studies may measure different populations of MPs, depending on choice of measurement method. Fourth, in our clinical study, we compared MP levels between critically ill patients who develop ARDS and critically ill patients who do not develop ARDS, rather than comparing MP levels with those of healthy control subjects or a less ill hospitalized population. It may be that the degree of underlying illness is an important confounding variable that explains differences between human and animal studies of MPs. Finally, although we did not detect any significant changes in circulating MP concentration over 4 days in the ICU, it is possible that the timing of MP measurement is critical for evaluation of the pathogenic or prognostic roles of MPs. Measurement of MPs at the time of enrollment in the VALID study on the morning of ICU day 2 is already many hours into critical illness and often several days after initiation of symptoms. Perhaps MPs are higher in the very early stages of illness and have started to decrease by the time subjects are admitted to the ICU.

Another surprising finding was the only modest association between MP concentration and plasma procoagulant activity as measured by plasma recalcification time. Our group has previously shown that MPs in the airspace in ARDS have very high levels of TF and are highly procoagulant [11]. Thus, it was surprising that MPs were not more strongly associated with procoagulant activity in the circulation. The implication of this finding is twofold. First, it highlights an important difference between the circulation and the airspace and confirms that findings in one compartment cannot be extrapolated to another. The importance of compartmentalization of mediators has been an active and growing area of research. For example, loss of TF (the initiator of the extrinsic coagulation cascade) is protective in mice exposed to systemic LPS [35], but it was associated with worse lung injury when the LPS was delivered intratracheally [36]. Similarly, it is increasingly recognized that there are biologic and physiologic differences between ARDS with direct causes (pneumonia, aspiration) and ARDS with indirect causes (nonpulmonary sepsis, trauma) [37]. Second, the lack of association between MPs and coagulation markers suggests that it may be a noncoagulant subpopulation of MPs that regulates endothelial function and inflammation in the circulation.

Our study has some limitations. First, we measured MPs using a MP capture assay that measures the concentration of phosphatidylserine-positive MPs and does not provide information about MP cell of origin. In the study which showed that lower MP levels were associated with increased ARDS mortality, circulating MPs originated predominantly from leukocytes [29]. These MPs were characterized by flow cytometry rather than by capture assay and may have included more diverse populations of MPs. Second, although larger than samples in other clinical studies of MPs in ARDS, our sample size was limited to 90 patients with ARDS, which may limit detection of important associations of MPs and ARDS. The effects of MPs may differ depending on the overall clinical scenario. In addition, our study, although prospective, does not provide mechanistic information about the specific function of MPs. The unexpected inverse relationship between plasma MP levels and the development of ARDS will help to inform future mechanistic studies aimed at measuring MP sequestration in organs including the lung and the presence of circulating protective MPs. Whether circulating MPs affect development of organ failure outside the lung remains uncertain.

In summary, lower levels of circulating MPs early after ICU admission are independently associated with the development of ARDS, and this relationship is strongest in patients with sepsis. The present findings, in conjunction with a study by Guervilly et al. [29] showing a protective association between MP concentration and mortality in ARDS, suggest that the potential role of MPs in ARDS pathogenesis is complex and not well modeled in current experimental studies. Whether MP measurements could be used clinically to aid with diagnosis or prognosis of ARDS requires further study.