Materials and Methods

Animals and inoculationsTwenty-one healthy, filovirus-seronegative, adult male cynomolgus macaques (Macaca fascicularis) and 4 rhesus macaques (Macaca mulatta) were used for this study. Macaques were inoculated intramuscularly with 1 mL of virus stock that contained 1000 pfu of EBOV (Zaire 1995 strain) that had been isolated from the serum of a human patient in 1995 and had been passaged 3 times in Vero cells [28]. Scheduled necropsies were performed at days 1 (3 cynomolgus macaques), 2 (3 cynomolgus macaques), 3 (1 rhesus and 4 cynomolgus macaques), 4 (1 rhesus and 4 cynomolgus macaques), 5 (1 rhesus and 4 cynomolgus macaques), 6 (3 cynomolgus macaques), and 9 (1 rhesus macaque) after infection. Longitudinal blood samples were analyzed by coagulation assays, as described below. Rhesus macaques were not used for any temporal comparisons; rather, this small cohort of macaques was used to detect any clear species-specific differences among the macaques that may warrant additional studies, while conserving primate resources

Ethical considerationsIn conducting this research, we adhered to the Institute of Laboratory Animal Resources, National Research Council’s Guide for the Care and Use of Laboratory Animals. Our animal facilities are fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. Human leukocytes used in this study were obtained from healthy donors, after receiving informed consent. These cells were obtained in accordance with a protocol (FY99-21) approved by the US Army Medical Research Institute of Infectious Diseases Office of Human Use and Ethics

Cell cultures and virusesElutriated monocytes were obtained from healthy human donors, isolated according to conventional procedures, and cultured or differentiated in 6-well plates [29]. Normal primary human umbilical vein endothelial cells (HUVECs), pooled from several donors, and primary human lung-derived microvascular endothelial cells (HMVEC-Ls) were obtained from Clonetics and were maintained according to the supplier’s recommendations. HUVECs and HMVEC-Ls were used at the second and third passages, respectively. Cultures were challenged with EBOV (Zaire 1995 strain) [28] at an MOI of 1.0. Culture medium from Vero cells represented a “virus-free stock” in all mock-infection experiments. Total RNA was extracted from infected/treated cells by use of Trizol (Life Technologies). Titration for infectious EBOV was determined as described elsewhere [3]

Coagulation testsPlasma levels of D-dimers were measured by an ELISA, according to the manufacturer’s directions (Diagnostica Stago). Plasma levels of protein C were determined by a chromatic hydrolysis assay (DiaPharma), according to the manufacturer’s recommendations. To control background for nonspecific hydrolysis, sterile water was substituted for the protein C activator. Plasma levels of tissue-type plasminogen activator antigen (tPA), urokinase-type plasminogen activator (uPA), and factor VIIa (FVIIa) were determined by ELISA (American Diagnostica), according to the manufacturer’s instructions

HistopathologyTissues were collected and immersion-fixed in 10% neutral-buffered formalin and processed for histopathology, immunohistochemistry, and transmission electron microscopy (TEM), as described elsewhere [3, 4]. Replicate sections of spleen, liver, and axillary lymph nodes were stained with phosphotungstic acid hematoxylin (PTAH), to demonstrate polymerized fibrin [4]

TEM of tissuesParaformaldehyde/glutaraldehyde-fixed tissues and peripheral blood mononuclear cells (PBMCs) for TEM were postfixed in 1% osmium tetroxide in 0.1 mol/L phosphate buffer, rinsed, stained with 0.5% uranyl acetate in ethanol, dehydrated in graded ethanol and propylene oxide, and embedded in Poly/Bed 812 resin (Polysciences). Ultrathin sections were cut, placed on 300-mesh nickel TEM grids, and stained with uranyl acetate and lead citrate

Immunoelectron microscopy (IEM) of cell culturesFor TF staining of cells, PHMs cultured in 6-well plates were incubated with dilutions (1:100) of an anti-TF murine monoclonal antibody against human TF (American Diagnostica) or an isotype-identical irrelevant murine antibody. PHMs were incubated with primary antibodies for 2 h, washed with Tris-buffered saline (TBS), and then incubated with dilutions (1:25) of a goat anti–mouse IgG conjugated to 10 nm of colloidal gold (Ted Pella) for 1 h. Cultures were washed in TBS, fixed in 2% glutaraldehyde in 0.1 mol/L phosphate buffer for 1 h, washed, postfixed in 1% osmium tetroxide, and processed for embedment in Poly/Bed 812 resin (Polysciences), as described above

IEM of culture fluids and plasmaFor TF staining of microparticles (MPs) in culture fluids or plasma, 1.5-mL aliquots of culture fluid or 0.5-mL aliquots of plasma were centrifuged at 16,000 g for 20 min. Pellets from culture fluid were resuspended in 50 μL of PBS, whereas pellets from plasma were recentrifuged in PBS (0.5 mL) and resuspended in 200 μL of PBS. Drops (5–10 μL) of the suspensions were applied to 300-mesh nickel TEM grids precoated with Formvar and carbon. Grids were immersed for 1 h in dilutions (1:100) of a murine monoclonal antibody against human TF (American Diagnostica) or in an isotype-identical irrelevant murine antibody in TBS containing 0.1% bovine serum albumin and 0.05% Tween-20. Grids were washed and transferred for 1 h to goat anti–mouse IgG labeled with 10-nm gold spheres (Ted Pella) diluted 1:25 in TBS. Grids were then washed by successive immersion in drops of TBS and PBS and were fixed in drops of 2% glutaraldehyde for 1 h. After rinsing in drops of PBS, grids were postfixed in drops of 1% osmium tetroxide, washed in distilled water, and negatively stained with 1% uranyl acetate

Flow cytometric analysisMPs were collected from plasma samples, as described above, resuspended in PBS containing a murine anti–TF antibody (American Diagnostica) or an isotype control, and incubated for 30 min. Samples were concentrated, washed in PBS, resuspended in PBS containing goat anti–mouse fluorescein isothiocyanate for 30 min, reconcentrated, rinsed in PBS, and resuspended in 500 μL of PBS. Samples were analyzed in a FACScan flow cytometer with CellQuest software (both from Becton Dickinson)

ImmunofluorescenceFormalin-fixed tissue sections were deparafinized and rehydrated, as described elsewhere [29]. Slides were placed in normal goat serum for 30 min, transferred to an anti–TF murine antibody (American Diagnostica), incubated overnight at 4°C, rinsed, incubated in goat anti–mouse alexa 594 (Molecular Probes) for 30 min, rinsed, and mounted in aqueous mounting medium containing 4′,6′-diamidino-2-phenylindole (DAPI; Vector Laboratories). The specificity of the anti–TF antibody for TF in formalin-fixed, paraffin-embedded tissues has been shown in other studies [30]. Cytospins or coverslips were rehydrated, rinsed in PBS, placed in normal goat serum for 30 min, transferred to the anti–TF antibody (American Diagnostica) or an EBOV antibody pool [26] for 1.5 h, incubated in goat anti–mouse alexa 594 (Molecular Probes) for 30 min, rinsed, and mounted in aqueous mounting medium containing DAPI. Negative controls included substitution of isotype-identical irrelevant murine antibodies for the anti-TF antibody and application of the anti-TF antibody to mock-infected cells and uninfected tissues

RNAse protection assaysPBMCs were separated from EDTA-treated peripheral blood collected from macaques before and after exposure to EBOV [29]. Custom Probe sets were obtained from Pharmingen. The Multiprobe RNAse Protection Assay was performed according to manufacturer’s directions (Pharmingen), with minor modifications [29]

Results

Development of coagulation abnormalitiesTo evaluate the role of the natural anticoagulant systems in the development of DIC during EBOV HF, we assessed several fibrinolytic pathways by use of plasma and/or serum samples temporally obtained from EBOV-infected macaques. Because the amount of available test material was limited, we focused our efforts on the most definitive assays, which had been validated in nonhuman primates and could readily be performed under biosafety level 4 containment. As an example of assay utility, preliminary studies showed that available assays for plasminogen activator inhibitor 1 were not compatible with cynomolgus or rhesus macaques

Our most striking finding was the rapid decrease in plasma levels of protein C (figure 1A), which occurred early during the course of disease and was evident in all macaques at day 2 after infection, and beyond. At days 2 and 3, levels decreased by >40% (2/4 macaques each day) versus preinfection levels. By days 4 and 5, levels decreased by >50% (3/4 macaques each day) versus preinfection values. Plasma levels of protein C rebounded at day 6, but only to 60% of preinfection levels (2/2 macaques). Levels of tPA increased sharply between days 3 and 4, from no detectable difference versus preinfection levels at day 3 to an ∼5.5-fold increase over preinfection levels at day 4; this elevation continued to increase through days 5 and 6 (14- and 18-fold over preinfection levels, respectively) (figure 1B). Levels of uPA increased in a similar manner, with elevations of ∼2-fold over preinfection levels on day 4, 4.5 fold-over preinfection levels on day 5, and nearly 8-fold over preinfection levels on day 6 (statistical significance was observed by day 4; P<.05). Levels of FVIIa decreased in most macaques during the course of disease, beginning with a >2-fold decrease at days 1 and 2 (2/3 and 3/4 macaques each day) and increasing to a >3-fold decrease by days 5 and 6 (2/4 and 2/2 macaques each day) (statistical significance was observed by day 5; P<.05). Development of fibrin-degradation products (D-dimers) showed rapid increases of 45-fold by day 4 and 55-fold by day 5 (figure 1C). Also of relevance in this study, thrombocytopenia developed as platelet counts decreased, from a preinfection mean of 412×10³ cells/mm³ to 154×10³ cells/mm³ on day 6 (figure 1D)

Evidence of fibrin in tissues: association with viral antigenTo further characterize the development of coagulation abnormalities during EBOV infections, the formation of fibrin deposition was analyzed longitudinally, by TEM and by histologic demonstration of polymerized fibrin, in tissue samples from EBOV-infected macaques. TEM performed at day 2 after infection showed rare deposits of fibrin in association with activated (increase in size of nucleus, numbers of lysosomes, and cytoplasmic flaps) tissue macrophages in the splenic marginal zone (1/3 macaques). By day 3, small focal deposits of fibrin were associated with activated tissue macrophages in red pulp and the marginal zone (2/4 macaques). PTAH-positive fibrin was evident by day 4; multifocal deposits were present in the sinusoids and marginal zones of spleen (figure 2A) and were less frequently observed in the cords of Bilroth (3/4 macaques). Fibrin was occasionally observed in hepatic sinusoids, usually closely associated with Kupffer cells (3/4 macaques), and multiple foci of PTAH-positive fibrin were detected in blood vessels of the renal medulla (2/4 macaques) (figure 2B). By day 5, the presence of fibrin deposits dramatically increased. Fibrin thrombi and fibrinocellular thrombi, composed of monocytes enmeshed in fibrin, were focally identified in vessels of most tissues evaluated. PTAH-positive fibrin was occasionally seen in hepatic sinusoids, usually closely associated with Kupffer cells (4/4 macaques); was frequently present in red pulp sinusoids, marginal zones of spleen, and the cords of Bilroth (4/4 macaques); and was especially prominent in blood vessels of the renal medulla (4/4 macaques). TEM findings at day 5 were similar. A high proportion of hepatic sinusoids were filled with fibrin deposits, fibrinocellular debris, and free virions. Deposits of fibrin and fibrinocellular debris in spleen were so heavy and widespread that normal architecture was almost completely lost (4/4 macaques); there was a strong demarcation of venous sinuses by fibrin deposits between the basement membrane and the cords of Bilroth. Macaques evaluated on day 6 showed similar patterns of splenic fibrin deposition, but with more abundant fibrin, and PTAH-positive fibrin thrombi were also present in pulmonary vessels (2/3 macaques) and renal glomeruli (1/3 macaques)

It is noteworthy that the presence of fibrin appeared to be closely associated with the presence of EBOV-infected cells. For example, PTAH staining did not show deposition of fibrin in the liver until day 4, the time when infected cells were readily detected by indirect fluorescent antibody analysis (IFA) (figure 2C and 2D). Likewise, the abundance of fibrin paralleled the level of EBOV antigen detected. Organ infectivity titers corroborated these findings. Specifically, by day 3 after infection, 2 of 4 macaques had infectious virus in the liver (virus titer [mean±SD, here and below], 2.40±2.77 log₁₀ pfu/g of tissue); by day 4, infectious virus was detected in livers of all 4 macaques (virus titer, 6.89±1.01 log₁₀ pfu/g of tissue). By day 2, infectious virus was detected in the spleen of 1 of 3 macaques (virus titer, 1.08±1.87 log₁₀ pfu/g of tissue); interestingly, the single macaque with infectious virus in the spleen was also the only macaque evaluated at day 2 with fibrin detected by TEM. By day 3, all macaques had infectious virus detected in their spleens (virus titer, 4.88±0.43 log₁₀ pfu/g of tissue); by day 4, the virus titer was 6.91±1.03 log₁₀ pfu/g of tissue

TEM also demonstrated a close association between fibrin deposits and EBOV-infected cells. The most remarkable TEM observation at day 4 was the appearance of fibrin mesh networks that completely encased many EBOV-infected lymphoid tissue macrophages (figure 3A), particularly in the subcapsular sinuses of inguinal lymph nodes (2/4 macaques). These fibrin mesh networks were also detected by TEM around EBOV-infected Kupffer cells and cells with morphology consistent with dendritic cells (DCs) or fibroblastic reticular cells. This fibrin encapsulation was observed only in cells with characteristic EBOV intracellular inclusion material and/or budding virions and proliferated membranes [31]; neighboring cells (neutrophils, lymphocytes, red blood cells, and macrophages) that did not exhibit morphological evidence of EBOV infection were not associated with these fibrin mesh networks (figure 3B)

Expression of TF mRNA and protein in PBMCs in vivo To explore the potential role of TF in inducing the coagulation cascade during EBOV infection, PBMCs were collected from EBOV-infected macaques, and mRNA transcripts were analyzed. Increased TF transcripts were observed in 3 of 4 macaques by day 3 after infection and in all macaques from day 4 and beyond (figure 4A). IFA for TF was also performed on cytospins prepared from available blood samples. IFA confirmed expression of TF and showed increased numbers of TF-positive cells during the course of the infection (figure 4B and 4C)

Expression of TF protein in tissuesImmunohistochemical analysis was performed on liver and spleen tissue from EBOV-infected macaques. Early-stage tissues, obtained at days 1 and 2, were compared with those obtained at mid-to-late stages of infection (days 4 and 5). Although antigen retrieval for TF proved to be problematic in tissues that had been stored in formalin for 4 weeks, clear increases in the number of cells expressing TF, particularly splenic macrophages, were seen at days 4 and 5. In addition to macrophages, other cells positive for TF antigen included polymorphonuclear neutrophils (PMNs) and endothelial cells. Interestingly, expression of TF in endothelial cells appeared to be focal and was not uniform (e.g., occasional vessels displayed intense staining, whereas many others showed mild-to-no staining). Circulating TF-positive cells were also observed in larger vessels (figure 4D). To confirm these findings, and to circumvent problems associated with storing tissues in formalin for extended periods of time, cytospins were prepared from dissociated spleens of available EBOV-infected macaques. IFA for TF on these samples confirmed increasing numbers of TF-positive cells during the disease course (figure 4E and 4F)

Production of membrane MPs in vivoMembrane MPs are shed from the plasma membrane of most eukaryotic cells undergoing activation or apoptosis [32–34]. To test the hypothesis that TF-expressing MPs are formed during EBOV infection, flow cytometry and IEM were performed on a cohort of plasma samples from EBOV-infected macaques. Flow cytometry demonstrated the development of a TF-positive MP population (figure 5A). IEM analysis of these samples confirmed that the MPs were fragments of cell membranes, of a variety of sizes (range, 0.1–1.0 μmol/L) and conformations, and that the labeling of TF on these MPs was very specific (figure 5B and 5C). TF-positive MPs were infrequently seen at day 1 (<1 TF-positive MP/field of 300 electron microscopy grid), but, by days 4 and 5, gold sphere–labeled MPs were readily detected in plasma of all macaques analyzed (>40 TF-positive MPs/field). In addition, the intensity of gold-sphere labeling of MP increased during the infection (e.g., at day 1, only 3–4 gold spheres were seen on MPs, whereas >10 gold spheres were usually seen on MPs by day 4)

Expression of TF mRNA and protein in vitroTo confirm expression of TF in mononuclear phagocytes, cultures of PHMs were established from 3 separate donors and infected with EBOV. Primary human endothelial cells, macrovascular HUVECs, and microvascular HMVEC-Ls were used for comparative purposes. Analysis of PHM and endothelial-cell gene response showed increased TF transcripts in PHMs as early as 1 h after infection (figure 6A); increased levels of TF transcripts were also seen in PHMs at 24 and 48 h. Increased levels of TF transcripts were not detected in HUVECs or HMVEC-Ls (data not shown), even though EBOV replicated in these cells, as evidenced by viral infectivity titration, which showed ∼6 log₁₀ pfu/mL of infectious EBOV in culture fluid. IFA confirmed TF protein expression in PHMs at 48 h (figure 6B–D). IEM also corroborated expression of TF and revealed that TF was primarily localized to PHM showing morphological evidence of EBOV infection [31]. TF-positive gold-sphere labeling was seen only rarely in PHMs that showed no morphologic evidence of infection. In many cases, TF was colocalized with areas of EBOV-induced proliferated membranes [31] (figure 6E and 6F). Occasional foci of TF were localized to nondescript areas along and just beneath the plasma membrane. Analysis of fluids from PHM cultures by IEM showed TF-positive MPs ranging in size from 0.1 μmol/L to nearly 1.0 μmol/L (figure 6G). Some of these MPs had attached viral particles and were interpreted to be portions of the EBOV-induced proliferated membranes seen in thin sections of EBOV-infected PHMs. It is noteworthy that the EBOV virions did not express TF, yet the attached membranes were TF positive by IEM

Discussion

The development of DIC is a characteristic clinical manifestation of EBOV infection in primates; however, the mechanisms that trigger coagulation abnormalities remain unknown. The present study provides the first insight into the pathogenesis of coagulation system dysregulation and suggests that the development of coagulation abnormalities may occur much earlier than previously thought. Although it is likely that the coagulopathy seen in EBOV HF is caused by multiple factors, particularly during the later stages of disease, the data presented here strongly implicate TF expression/release from EBOV-infected monocytes/macrophages as a key factor that induces the development of coagulation irregularities seen in EBOV infections

DIC is not a primary disorder but is a complication of an underlying disease, usually an infection, that triggers a massive activation of the clotting cascade. The major initiating factor is the release or expression of TF in response to endotoxins, various cytokines, or other cellular factors [6]. For example, in the case of dengue HF, Wills et al. [35] postulated that dengue infection may activate fibrinolysis by degrading fibrinogen directly, thus promoting secondary activation of procoagulant homeostatic mechanisms that include increases in plasma levels of TF. Here, we show that expression of TF was observed only in monocytes/macrophages with morphological evidence of EBOV replication, suggesting that expression of TF may be directly induced by EBOV infection. Given the large number of EBOV-infected monocyte/macrophages associated with primate infections [3, 25, 26], it is likely that the overproduction of TF is the primary initiating factor for the observed coagulopathy

In these tissues, the appearance of fibrin deposits, coinciding with detectable viremia and infectious virus, strongly supports a direct link with viral infection. The appearance of fibrin mesh networks that selectively encased EBOV-infected lymphoid tissue macrophages and Kupffer cells, by day 4, and DCs, by day 5, not only corroborate this observation but also suggest that viral infection resulted in up-regulation of a protein capable of triggering the coagulation cascade. RNA analysis, IFA, and TEM all show that expression of TF is up-regulated during EBOV infection and likely triggers activation of the coagulation system, producing the marked fibrin encasement of EBOV-infected cells

Analysis of macaque PBMC mRNA demonstrated increased levels of TF transcripts by day 3; IFA confirmed these findings. Because monocytes/macrophages are the only cells in the blood circulation compartment that express TF on their surface [15, 36] and because differential cell counts show that monocytes represented only 1%–6% of the total white blood cells of any of the 25 macaques in the present study, it is likely that the level of TF transcripts in monocytes is more marked than that in total PBMCs. Cells morphologically consistent with monocytes/macrophages or PMNs were positive for TF. TEM analysis of macaque tissues demonstrated fibrin formation only around EBOV-infected cells, including lymphoid monocytes/macrophages, Kupffer cells, and DCs. Fibrin formation was not observed around PMNs, which do not support EBOV replication. It is not surprising that PMNs may phagocytose and store TF protein released by monocytes/macrophages, but, unless they are induced to up-regulate mRNA transcripts and express TF protein on the cell surface, PMNs are not likely to trigger the coagulation system. The observation of TF by IFA and the lack of fibrin associated with these cell types by TEM is consistent with this view and also implicates monocytes/macrophages as key sources for TF protein during EBOV infection. Moreover, up-regulation of TF mRNA and protein in EBOV-infected monocytes/macrophages in vitro further corroborates this hypothesis

Levels of expression of TF may also be affected by the production of various cytokines and chemokines. Proinflammatory cytokines, such as interleukin (IL)–6, have been shown to effectively up-regulate expression of TF on monocytes [37, 38]. In EBOV-infected macaques, increased levels of IL-6 have been reported by day 4 after infection [29]. Splenocyte stimulation with TF also induces the selective release of the monocyte chemoattractant molecule macrophage inflammatory protein (MIP)–1α in vitro [39]. Increased levels of MIP-1α have been demonstrated in EBOV-infected macaques and cultures of infected monocytes/macrophages [29]. Taken together, these findings suggest that expression of TF can be amplified during the later stages of disease when EBOV-infected cells, which selectively express TF, release cytokines/chemokines that may either directly up-regulate levels of TF or attract additional preferred target cells to foci of infection, thereby exacerbating activation of the extrinsic coagulation pathway

The present study also implicates TF expressed on MPs, which may be generated by EBOV infection of monocytes/macrophages, as playing an important role in disease pathogenesis. We have demonstrated that TF-positive MPs are shed from EBOV-infected monocytes/macrophages in vitro; furthermore, we have shown increased numbers of TF-positive MPs in plasma of EBOV-infected macaques. The procoagulant potential of MPs is supported by clinical studies showing elevated levels of circulating MPs in patients with an increased risk for thromboembolic events (e.g., DIC) [33, 34]. Our observation of MPs during EBOV infection is not unprecedented. Recently, it was reported that patients with fulminant meningococcal sepsis had CD14- and TF-positive MPs and that these MPs also induced extreme thrombin generation in vitro [33]

The present study has suggested that protein C may be a critical component of the observed coagulation dysfunction in EBOV HF. The protein C system is one of the main anticoagulant mechanisms in blood [40]. Growing evidence also suggests that protein C has direct anti-inflammatory properties and modulating activity on cellular functions, likely by blocking NF-κB nuclear translocation [41–43]. The rapid decrease in plasma levels of protein C concomitant with progression of disease was a prominent finding in the present study. Reduced levels of protein C are found in the majority of patients with sepsis and are associated with an increased risk of death [44, 45]. Moreover, protein C deficiency appears to develop well before the onset of defined clinical variables of severe sepsis or septic shock [45, 46] and may be considered a prognostic indicator [47]. Clearly, in this primate model of EBOV, identifying protein C abnormalities (that so closely parallel protein C anomalies in human coagulopathies) offers an ideal target for chemotherapeutic interventions. Administration of activated protein C was protective in a baboon model of lethal Escherichia coli sepsis [48]. More recently, treatment with recombinant human activated protein C significantly reduced mortality in patients with severe sepsis [49]; furthermore, biological activity of activated protein C was demonstrated by finding greater decreases in levels of D-dimer and IL-6 in patients who received activated protein C than in those who received placebo

The selection of an appropriate nonhuman primate model of EBOV HF requires careful consideration of the species, sex, age, route of infection, dose of EBOV administered, and nature of the challenge virus itself. It is very difficult to compare results from the various published studies of EBOV HF in nonhuman primates because of considerable differences among these factors. Species-specific differences in the appearance of coagulopathies were reported in one study. Specifically, fibrin deposition in vessels was associated with EBOV infection of African green monkeys, whereas hemorrhages were a feature associated with EBOV disease in baboons [50]. Consistent with the previous findings shown for African green monkeys [26, 50] and rhesus macaques [3, 4], we observed fibrin in vessels of the macaques used in the present study. Although the present study primarily focused on cynomolgus macaques, we did not see species-specific differences in coagulopathy among the cynomolgus macaques and a small number of rhesus macaques that were primarily used as a pilot group to detect any clear species-specific differences

Identifying TF as an initiating aspect of DIC and dysregulation of protein C as a potentially exacerbating factor may be central to explaining the mechanisms that trigger the hemorrhagic diathesis of EBOV infection. The present study represents a first step in demonstrating that chemotherapeutic strategies aimed at controlling TF overexpression and dysregulation of the protein C system may ameliorate the effects of EBOV HF. Strategies that have proven to be successful for treating patients with similar clinical pictures, such as administering activated protein C to patients with severe sepsis, may also have utility in treating EBOV infections. It is also likely that combination therapy with multiple coagulation-modulating agents may ultimately prove to be the best approach for treating the inherent DIC. Rodent models of EBOV HF do not exhibit the hemorrhagic manifestations that characterize primate EBOV infections [4, 27], which may explain the failure of rodents to predict antiviral drug efficacy in primates [4, 28, 51]. The present study has suggested that antiviral strategies shown to control EBOV replication in rodents may be significantly improved by effective management of coagulopathy in primates