Thrombin reprograms endothelial extracellular vesicles

In this issue of Blood Advances, Chatterjee et al¹ report that the phenotype of extracellular vesicles (EV) derived from endothelial cells in response to thrombin is dictated by the concentration of thrombin to which the cells are exposed. In response to lower concentrations of thrombin (50pM), endothelial cells released EV with anti-inflammatory properties mediated through their content of miR-409-5p, whereas cells exposed to higher thrombin concentrations (10nM) induced inflammatory responses, mediated through miR-155-5p (see figure).

The left panel depicts release of EV containing miR-409-5p by endothelial cells exposed to low concentrations of thrombin. These EV deliver miR-409-5p to other cells (monocytes), which blocks translation of USP7 and its potentiating effects on NF-kB–mediated inflammation. The right panel depicts higher concentrations of thrombin causing the release of EV containing miR-155-5p. Upon delivery to cells by EV, miR-155-5p promotes inflammatory responses by inhibiting translation of SOCS1, which regulates JAK-STAT signaling. TLR, toll-like receptor.

EV are membrane-bound nanoparticles that are released constitutively, although their release is often increased in response to cellular activation, injury, or apoptosis. They are roughly divided into 2 major subtypes, including exosomes (∼30-150 nm) that are derived from the cellular endosomal compartment, and microvesicles (∼100-1000 nm), derived by budding from the plasma membrane.² Microvesicles (previously referred to as microparticles) were originally studied for their prothrombotic properties,³ but with increased recognition of the physical and functional heterogeneity of EV, distinct roles for EV in mediating intercellular communication by delivering selectively packaged cargoes of proteins (receptors, adhesion molecules, signaling enzymes), lipids, and nucleic acids (particularly mRNA and microRNA [miRNA]) have been recognized, with the content of EV determined by the phenotypic state of the parental cell. In the vasculature, most EV are derived from platelets, although EV derived from essentially all cellular elements in contact with blood, including endothelial cells, may be detected and are often increased in settings of vascular inflammation.⁴ Endothelial cell–derived EV contribute to vascular homeostatic processes such as angiogenesis, inflammation, and hemostasis; however, under pathologic conditions, they may mediate endothelial dysfunction and augment thrombogenicity. Differences in number and content of circulating EV suggest their potential as biomarkers of various disorders, although their clinical utility in this role has been limited by lack of standardized approaches to measurement and analysis. However, the use of EV prepared ex vivo and containing specific cargo shows promise for therapeutic application, with several early phase clinical trials presently underway.

Building on previous studies that demonstrated that low-dose thrombin was associated with cytoprotective effects, whereas higher doses cause cytopathic effects, Chatterjee et al isolated EV from cultured endothelial cells exposed to thrombin at concentrations ranging from 5pM to 10 000pM. Concentration dependent increases in EV release were observed, although the size distribution and appearance by electron microscopy of EV isolated from endothelial cells treated with control vehicle, low, or high concentrations of thrombin did not differ, regardless of whether EV were isolated by ultracentrifugation or size exclusion chromatography. In contrast, the functional properties of isolated EV differed markedly depending on the concentration of thrombin their parental cells had been exposed to. Using an assay in which THP-1 cells or human peripheral blood mononuclear cells were exposed to lipopolysaccharide (LPS), Chatterjee et al found that the cells pretreated with EV derived from low thrombin–treated endothelial cells were anti-inflammatory, significantly reducing the secretion of tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and IL-6 by the monocytic cells. In contrast, EV derived from high thrombin–treated endothelial cells had no effect on LPS-induced cytokine release, and instead directly induced the secretion of inflammatory cytokines. Similar results were obtained when the barrier function of naïve endothelial cells was measured in response to LPS after pretreatment with EV stimulated by low or high thrombin.

To define the mechanisms by which EV altered the inflammatory response, the investigators profiled their miRNA content, based on previous studies demonstrating that miRNAs were often central mediators of EV effects. Upregulated miRNAs identified by sequencing were validated by quantitative polymerase chain reaction (qPCR), and their effects on LPS-induced inflammatory responses assessed. These studies demonstrated that miR-409-5p was the critical mediator of anti-inflammatory effects in low thrombin derived EV, whereas miR-155-5p mediated the proinflammatory effects of high thrombin derived EV. Importantly, the anti- and proinflammatory effects of EV containing these cytokines were confirmed in LPS-treated mice in vivo. Finally, using results of previous studies defining the targets of these miRNAs, the authors confirmed that the effects of miR-409-5p were mediated by binding to herpesvirus-associated ubiquitin-specific protease (USP7),⁵ which promotes inflammation by stabilizing NF-κB and TRAF-6 and preventing their ubiquitin-mediated degradation; thus inhibition of USP7 translation inhibited NF-κB signaling. miR-55-5p reduced the expression of SOCS1 (suppressor of cytokine signaling), a negative regulator of JAK-STAT signaling.⁶ Consistent with this effect, the augmentation of STAT1/3 activation by high thrombin induced EV were blocked by an anti–miR-155-5p.

This fascinating study provides new and important information on the proinflammatory and anti-inflammatory effects of thrombin; although these results are confined to endothelial cells, other cells, such as neurons, have demonstrated different responses to low- and high-dose thrombin, and it is likely that EV released by these cells may also reflect these cellular responses. It is interesting to speculate that the systemic effects of exosomal miR-409-5p and miR-155-5p, either derived from endothelial cells or loaded into synthetic liposomes, might suggest a novel therapeutic approach for inflammatory disorders such as sepsis.

Finally, although these studies were beautifully performed, they leave critical questions unanswered. Among these, perhaps the most perplexing is how do cells selectively package specific miRNA or other cargo in response to different concentrations of thrombin? The role of PAR-1,⁷ thrombomodulin,⁸ or perhaps other signaling proteins/receptors was not addressed in this report but should be examined in future studies. Regardless, this study provides the impetus for future work addressing these issues that will help us further understand inflammatory responses of the vasculature.

Conflict-of-interest disclosure: The authors declare no competing financial interests.