Platelet-like particles released from Ebola virus-infected megakaryocytic cells behave like virus-like particles

TL;DR

Ebola virus can cause platelet-like particles to carry viral components, which may help the virus spread in the body and cause disease.

Contribution

Ebola virus-infected megakaryocytes release platelet-like particles containing viral proteins and genetic material that can produce new virus.

Findings

Platelet-like particles from infected cells contain EBOV proteins and genetic material.

Viral mRNA and genome RNA are synthesized in these particles, enabling virus production in recipient cells.

The viral glycoprotein on PLPs helps them enter other cells.

Abstract

Ebola virus (EBOV) is likely a zoonotic and re-emerging virus that causes severe outbreaks of Ebola virus disease. The virus spreads to various tissues during the late stage of infection and has been detected in immune-privileged sites of survivors. However, the mechanism of how EBOV disseminates throughout the body is not completely elucidated. In this study, by using a biologically contained EBOVΔVP30 system, we demonstrate that a megakaryocytic-like MEG-01 cell line that stably expresses VP30 (MEG-01 VP30 cells) is susceptible to EBOVΔVP30 infection and that MEG-01 VP30 cells exposed to EBOVΔVP30 produce platelet-like particles (PLPs) that contain EBOV proteins and viral genetic material. We further found that the viral envelope glycoprotein is expressed on the surface of the produced PLPs and contributes to PLP internalization into recipient cells. In addition, viral mRNA and genome RNA are actively synthesized in these PLPs, which may lead to progeny EBOV production from recipient cells that internalize the PLPs. Taken together, our data provide new insights into the potential role of platelets in the widespread dissemination of EBOV and the pathogenesis of Ebola virus disease.

At the late stage of Ebola virus (EBOV) disease, the virus spreads to various tissues, including immune-privileged sites such as the brain, eyes, testes, and bone marrow. Even after recovery, some survivors experience relapses, and transmission of virus through breast milk or semen may occur. While EBOV is thought to disseminate throughout the body via the bloodstream or lymphatic system and through interaction with nearby cells, growing evidence shows that platelets act as important immune mediators. Platelets not only communicate with innate immune cells and endothelial cells but also interact with pathogens, including viruses. These observations led us to ask whether platelets could act as carriers of viral components (i.e., viral proteins, mRNA, and genomic RNA) capable of supporting further virus production. Using a biologically contained EBOVΔVP30 system, we show that megakaryocytes, the cells that produce platelets, can release platelet-like particles (PLPs) that contain EBOV proteins and genetic material after exposure to EBOVΔVP30. Furthermore, within these PLPs, viral mRNA and genomic RNA are actively synthesized, which can drive active progeny virus production in recipient cells that internalize them. Our study reveals a potential role for platelets during EBOV infection.

Figures

Susceptibility of megakaryocytic MEG-01 VP30 cells to EBOVΔVP30.A. Percentage of GFP-positive MEG-01 VP30 cells. PMA-treated or untreated cells were exposed to EBOVΔVP30-GFP at an MOI of 5 and GFP-positive cells were quantified by flow cytometry 2 days after exposure. Data are representative of two independent experiments. B. EBOVΔVP30 titers from MEG-01 VP30 cells. PMA-treated or untreated cells were exposed to EBOVΔVP30-GFP at an MOI of 1. Supernatants were collected daily, and viral titers were determined using Vero VP30 cells. Data are presented as means ± SD of three independent experiments. The dotted lines indicate the lower limit of detection (20 ffu/ml). Statistical significance was assessed by using the multiple unpaired t-test. *p < 0.05, p < 0.01.

Examination of platelet-like particles (PLPs) released from EBOVΔVP30-exposed megakaryocytic cells.A. Detection of EBOV proteins by western blot from PLPs released form EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were collected on day 4 post-exposure from PMA-treated MEG-01 WT or VP30 cells exposed to EBOVΔVP30 at an MOI of 5. The indicated proteins were analyzed by immunoblotting. Data are representative of two independent experiments. B. Relative amount of EBOV gRNA in PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were prepared as described in (A). EBOV gRNA was quantified by RT-qPCR using the indicated genome-specific primer pairs and normalized to the gRNA in PLPs from MEG-01 WT cells. Data are presented as means ± SD from two independent experiments performed in triplicate. C. Localization of EBOV GP in PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were prepared as described in (A). The platelet marker CD41 (magenta) and EBOV GP (green) were visualized by immunofluorescence microscopy using specific antibodies, along with the corresponding bright-field image. Scale bars, 5 μm. D. Interaction between EBOV NP/VP35 (top panel) and NP/VP40 (bottom panel) in PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were prepared as described in (A). NP/VP35 and NP/VP40 complexes (green) were visualized by using a proximity ligation assay (PLA) with specific antibodies and overlaid on the corresponding bright-field image. Scale bars, 5 μm.

Synthesis of EBOV mRNA and gRNA in PLPs from MEG-01 VP30 cells.Relative expression levels of EBOV mRNA (A) and gRNA (B) in PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were collected on day 4 post-exposure from PMA-treated MEG-01 VP30 cells exposed to EBOVΔVP30-GFP at an MOI of 5. EBOV mRNA and gRNA were quantified by RT-qPCR using the indicated specific primer pairs and normalized to the amount in the day 0 samples. Data are presented as means ± SD of three independent experiments performed in triplicate.

Internalization of PLPs by recipient cells.A,B. Expression levels of GP and CD41-mCherry in MEG-01 cells stably expressing CD41-mCherry with or without EBOV GP (A), and in PLPs released from these stable cell lines (B). Protein levels were analyzed by immunoblotting using the indicated antibodies. C. Percent of mCherry-positive cells that internalized PLPs containing CD41-mCherry or CD41-mCherry/GP. Huh7 VP30 cells were co-incubated with the indicated PLPs for 1, 3, or 5 h, followed by quantification of mCherry-positive cells by flow cytometry. Data are presented as means ± SD of three independent experiments. Statistical significance was assessed by use of a two-way ANOVA followed by Turkey’s multiple comparisons test. *p < 0.05, p < 0.0001.

PLP-mediated transfer of viral components to recipient cells.Number of GFP-positive cells following co-incubation with PLPs released from EBOVΔVP30-exposed MEG-01 cells (A). PLPs were collected from PMA-treated MEG-01 WT and VP30 cells exposed to EBOVΔVP30-GFP. Huh7 VP30 cells (2 x 105 cells) were co-incubated with the indicated PLPs (2 x 106) for two days. As a control, cells were cultured in the final wash supernatant (B). GFP-positive cells were quantified by flow cytometry. Data are representative of two independent experiments. C-E. EBOVΔVP30 titers in three different VP30-expressing cell types co-cultured with PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were prepared as described in (A). Huh7 VP30 cells (2 x 105 cells, C), HUVEC VP30 cells (2 x 105 cells, D), and PMA-differentiated THP-1 VP30 cells (2 x 105 cells, E) were co-cultured with the indicated PLPs (2 x 106). Virus titers on days 2, 4, and 6 were determined using Vero VP30 cells. Data are presented as means ± SD of three independent experiments. The dotted lines indicate the lower limit of detection (1.3 log10 ffu/ml). Statistical significance was assessed by use of a one-way ANOVA followed by Turkey’s multiple comparisons test. *p < 0.001, p < 0.0001.

Full text

Introduction

Ebola virus (EBOV) is an enveloped, filamentous virus that belongs to the Filoviridae family of negative-sense, single stranded RNA viruses. In humans and non-human primates, EBOV causes a severe disease characterized by systemic inflammation, coagulopathy and, in some cases, hemorrhagic manifestations. Although there were relatively few reports of hemorrhagic complications during the 2013–2016 Ebola virus disease outbreak in Western Africa, mild signs of a bleeding diathesis (e.g., conjunctival bleeding, petechiae, gastrointestinal bleeding, and continued oozing at venipuncture site) were commonly observed [1–4]. Although pathogenesis of hemorrhagic manifestations remains uncertain, some clinical signs (e.g., petechiae) suggest a link with platelet abnormalities including thrombocytopenia. In addition, the high case fatality rate of Ebola virus disease is associated with widespread multi-organ dysfunction across the hepatic, cardiovascular, pulmonary, renal, central nervous, and gastrointestinal systems, potentially driven by microvascular thrombosis and impaired tissue perfusion [5,6]. Chaotic inflammatory responses triggered by infected immune cells producing cytokines and tissue factors can lead to endothelial dysfunction and consumptive coagulopathy, although evidence for disseminated intravascular coagulation in Ebola virus disease remains controversial [5–7]. In severe cases, vascular endothelial dysfunction resulting in vascular leakage, third spacing, and hypovolemic shock can occur during the end stage of infection unless appropriate supportive care, treatment, or therapeutics are given [8,9].

Platelets are small anucleate cells produced from megakaryocytes that primarily reside and develop in the bone marrow. The recent finding of megakaryocytes in the lung has led to the suggestion that the lungs may be another site of platelet production [10,11] while under particular conditions, such as sepsis or physiological stress, megakaryocytes in the liver and spleen can contribute to extramedullary platelet production [12,13]. Various cell surface molecules on platelets are inherited from the precursor megakaryocytes [14–17] together with mRNA and intracellular proteins associated with transcription and post-transcription [18]. Using these cell surface molecules, platelets play vital roles not only in maintaining homeostasis and initiating thrombosis [19,20] but also in recognizing invasion of pathogens including DNA and RNA viruses such as dengue virus (DENV), influenza A virus (FLUAV), human immunodeficiency virus type 1 (HIV-1), Hepatitis C virus, and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [21]. Recognition of these pathogens results in platelet activation, which induces the release of various inflammatory mediators as well as the elimination of the viruses by phagocytosis [22,23]. Furthermore, platelets have the potential to serve as carriers or reservoirs for infectious viral pathogens that bind to or are internalized into platelets (e.g., DENV [24,25], FLUAV [26], HIV-1 [27,28], SARS-CoV-2 [29]). As a result, the virus particles could potentially be delivered to cells in other tissues or organs through the bloodstream via platelets [21]. Of note, several known cell surface molecules on platelets have been identified as attachment factors facilitating EBOV entry or inducing cytokine responses [e.g., DC-SIGN (CD209), TAM receptors, TLR2, TLR4, and FcγRII (FCGR2) [14–17]].

Studies of EBOV-infected patients and experimentally infected animals have shown that EBOV has broad cellular tropism, infecting many types of cells in various tissues (e.g., kidney, liver, spleen, and heart), especially late in infection [5,30]. EBOV has also been found in the respiratory tract [31] and viral antigen has been detected in relatively immune-privileged sanctuaries such as the brain, eyes, testes, and bone marrow [32–36]. In an EBOV-infected nonhuman primate model, viral antigen-positive megakaryocytes were detected in the bone marrow on day 5 post-infection; however, it remains unclear whether this reflects direct infection or megakaryocyte emperipolesis [33]. In survivors of the Ebola virus disease outbreak in Western and Central Africa, both EBOV RNA and infectious virus have been detected in body fluids including cerebrospinal fluid [37,38], ocular fluid [39], breast milk [40–44], and semen [45–49] long after clearance of acute viremia. In addition, survivors suffer from long-term sequelae of infection [50–59] and disease relapse [37,38,60,61] and they are more likely to become a source of new infection [62–64]. In contrast, in the early stages of infection, the primary target cells are endothelial cells and immune cells such as dendritic cells, monocytes, and macrophages that accumulate around mucus membranes or nonintact skin (e.g., sites of blood vessel injury) where platelets also gather. After infection of these target cells, newly produced progeny viruses are presumed to be disseminated to the entire body by traveling through lymphatic vessels or the bloodstream and by interacting with surrounding cells such as monocytes exhibiting macrophage-differentiation signatures [34,65]. However, there are other possibilities that may explain how the virus travels through the body to access other tissues and organs including immune-privileged sites. Here, we investigate the potential role of platelets as a carrier for the viral genome and the role of megakaryocytes as producers of virus-like platelet particles.

Results

Susceptibility of megakaryocytic cells to EBOVΔVP30

EBOV-specific antigens have been detected in the bone marrow and lungs of EBOV-infected patients, but it is unclear whether EBOV is present within megakaryocytes through direct infection, which can reside in these two anatomical locations [31]. Therefore, we first investigated the susceptibility of MEG-01 cells, a megakaryocytic-like cell capable of producing platelet-like particles (PLPs), which are structurally and functionally similar to primary platelets, after differentiation by treatment with phorbol 12-myristate 13-acetate (PMA) [66–71]. To examine the infectivity of this cell line, we used a biologically contained EBOV (EBOVΔVP30), which is replication-competent and phenotypically similar to wild-type EBOV in cells stably expressing VP30, but can be handled under BSL-2 containment [72,73]. To support the replication of EBOVΔVP30, MEG-01 cells were generated to stably express EBOV VP30 (MEG-01 VP30 cells), a viral protein essential for EBOV replication [72]. Untreated MEG-01 VP30 cells or cells pre-treated with PMA for 24 h were exposed to EBOVΔVP30 that expresses GFP instead of VP30 (EBOVΔVP30-GFP) at a multiplicity of infection (MOI) of 5. Two days after exposure, the number of GFP-positive cells was analyzed by flow cytometry. The percentage of GFP-positive cells in the PMA-untreated cells was 1.12%, indicating undifferentiated MEG-01 VP30 cells did not efficiently support EBOVΔVP30 infection (Fig 1A, top panels). However, after treatment with PMA, 20% of the MEG-01 VP30 cells were GFP-positive (Fig 1A, bottom panels), suggesting that this megakaryocyte-like cell line is susceptible to EBOVΔVP30 infection.

*Susceptibility of megakaryocytic MEG-01 VP30 cells to EBOVΔVP30.A. Percentage of GFP-positive MEG-01 VP30 cells. PMA-treated or untreated cells were exposed to EBOVΔVP30-GFP at an MOI of 5 and GFP-positive cells were quantified by flow cytometry 2 days after exposure. Data are representative of two independent experiments. B. EBOVΔVP30 titers from MEG-01 VP30 cells. PMA-treated or untreated cells were exposed to EBOVΔVP30-GFP at an MOI of 1. Supernatants were collected daily, and viral titers were determined using Vero VP30 cells. Data are presented as means ± SD of three independent experiments. The dotted lines indicate the lower limit of detection (20 ffu/ml). Statistical significance was assessed by using the multiple unpaired t-test. *p < 0.05, *p < 0.01.

Next, we assessed the ability of MEG-01 cells to produce infectious viruses. Cells (untreated or PMA-treated) were exposed to EBOVΔVP30-GFP at an MOI of 1. Cell culture supernatants were harvested every 24 hours for six days, and virus titers were determined on Vero cells that stably express VP30 (Vero VP30 cells). In PMA-treated MEG-01 VP30 cells, virus titers were significantly higher than those in untreated cells at the later time points; however, peak titers remained low, not exceeding 2,000 focus-forming units (ffu)/ml on day 6 after exposure (Fig 1B).

Having observed production of infectious virus, albeit at low levels, from MEG-1 VP30 cells, we next sought to confirm the presence of viral proteins and genome (g)RNA in these cells. In PMA-treated MEG-1 VP30 cells, viral proteins were detected by western blot analysis at 2 days after exposure at levels similar to viral protein expression in infected Vero VP30 cells; in exposed PMA-treated wild-type MEG-01 cells (that do not express EBOV VP30), no newly synthesized viral proteins were detected (S1A Fig). The presence of viral gRNA in cells was analyzed by RT-qPCR using specific primer pairs for the NP gene and the genomic trailer region, which are located at the 3’ and 5’ ends of the EBOV genome, respectively. In MEG-1 VP30 cells, there was an increase in the amounts of viral gRNA specific for the NP gene and trailer region of the genome from day 2 to day 3 after exposure compared to exposed wild-type MEG-01 cells (S1B Fig). Collectively, these observations demonstrate that the megakaryocyte-like cell line can support EBOVΔVP30 replication, although the amount of replicative intermediate produced is limited.

Examination of platelet-like particles (PLPs) released from EBOVΔVP30-exposed megakaryocytic cells

After treatment with PMA, MEG-01 cells produce small (2–4 µm in diameter) PLPs that are released into the cell culture supernatant [70,71,74]. Because MEG-01 VP30 cells can support limited EBOVΔVP30 replication, we next examined whether PLPs released from infected MEG-01 VP30 cells contain viral proteins and viral gRNA.

PMA-treated wild-type or VP30-expressing MEG-01 cells were exposed to EBOVΔVP30 at an MOI of 5, and four days after exposure, cell culture supernatants were harvested. Given their smaller size (about 2 µm) compared to the larger MEG-01 cells (30–40 µm; [68]), the PLPs were isolated from the cell culture supernatant by 5-µm filtration and then pelleted by centrifugation [74].

From the PLPs isolated from the EBOVΔVP30-exposed wild-type MEG-01 cells, the platelet surface marker CD41 was detected by western blotting, but viral proteins were not detected (Fig 2A). In contrast, CD41 and EBOV proteins were detected from PLPs isolated from the exposed MEG-01 VP30 cells (Fig 2A). By RT-qPCR, there was at least a 200-fold increase in viral gRNA (both NP gene and trailer region) detected in the PLPs isolated from the exposed MEG-01 VP30 cells compared to those isolated from the exposed wild-type cells (Fig 2B).

Examination of platelet-like particles (PLPs) released from EBOVΔVP30-exposed megakaryocytic cells.A. Detection of EBOV proteins by western blot from PLPs released form EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were collected on day 4 post-exposure from PMA-treated MEG-01 WT or VP30 cells exposed to EBOVΔVP30 at an MOI of 5. The indicated proteins were analyzed by immunoblotting. Data are representative of two independent experiments. B. Relative amount of EBOV gRNA in PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were prepared as described in (A). EBOV gRNA was quantified by RT-qPCR using the indicated genome-specific primer pairs and normalized to the gRNA in PLPs from MEG-01 WT cells. Data are presented as means ± SD from two independent experiments performed in triplicate. C. Localization of EBOV GP in PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were prepared as described in (A). The platelet marker CD41 (magenta) and EBOV GP (green) were visualized by immunofluorescence microscopy using specific antibodies, along with the corresponding bright-field image. Scale bars, 5 μm. D. Interaction between EBOV NP/VP35 (top panel) and NP/VP40 (bottom panel) in PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were prepared as described in (A). NP/VP35 and NP/VP40 complexes (green) were visualized by using a proximity ligation assay (PLA) with specific antibodies and overlaid on the corresponding bright-field image. Scale bars, 5 μm.

Next, we attempted to detect the presence of viral proteins on the surface of the PLPs. Using immunofluorescent microscopy, we detected CD41 on PLPs isolated from both EBOVΔVP30-exposed wild-type MEG-01 cells and MEG-01 VP30 cells (Fig 2C, top and bottom panels). However, the EBOV surface glycoprotein (GP) was only detected on the surface of PLPs isolated from the exposed MEG-01 VP30 cells and co-localized with CD41 (Fig 2C, bottom panel).

To further investigate viral proteins within the PLPs, we used a proximity ligation assay (PLA) that enables visual detection of protein interactions within cells [73]. First, we visualized the interaction of EBOV NP and VP35, which is important not only for viral transcription and RNA replication, but also for forming nucleocapsids consisting of NP-RNA complexes [75,76]. The direct interaction between NP and VP35 was detected within PLPs isolated from EBOVΔVP30-exposed MEG-01 VP30 cells (Fig 2D, top panel). Likewise, a direct interaction of NP with VP40 was also detected within the PLPs isolated from the exposed MEG-01 VP30 cells (Fig 2D, bottom panel). Taken together, these data suggest that EBOVΔVP30-infected megakaryocyte-like cells can produce PLPs that carry both viral protein and gRNA.

PLPs from exposed MEG-01 VP30 cells synthesize EBOV mRNA and gRNA

Our finding that viral proteins and gRNA are present within PLPs released from MEG-01 VP30 cells exposed to EBOVΔVP30 (Fig 2), together with reports that platelets can undergo transcription and translation using inherited mRNA and proteins from megakaryocytes [18], prompted us to investigate whether PLPs per se support transcription, translation, and replication leading to progeny virus production. We prepared PLPs from EBOVΔVP30-exposed MEG-01 VP30 cells and monitored changes in viral mRNA and gRNA expression over three days. Compared to the basal mRNA level on day 0, NP mRNA expression was slightly increased on day 1 (by around three-fold) and increased by more than 10-fold on day 3 (Fig 3A); gRNA expression comparably increased over this time period (Fig 3B). In a parallel experiment, we assessed the virus titers in the supernatants of the PLPs, but there was no infectious virus production (S2 Fig).

Synthesis of EBOV mRNA and gRNA in PLPs from MEG-01 VP30 cells.Relative expression levels of EBOV mRNA (A) and gRNA (B) in PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were collected on day 4 post-exposure from PMA-treated MEG-01 VP30 cells exposed to EBOVΔVP30-GFP at an MOI of 5. EBOV mRNA and gRNA were quantified by RT-qPCR using the indicated specific primer pairs and normalized to the amount in the day 0 samples. Data are presented as means ± SD of three independent experiments performed in triplicate.

EBOVΔVP30 particles are not internalized into isolated PLPs

It has been reported that some virus particles (i.e., HIV-1) on the surface of PLPs are engulfed by the PLPs, enabling the transfer of infectious virus to recipient cells in this manner [27,28,77]. Therefore, we examined whether PLPs from MEG-01 cells could transfer EBOV particles by the same mechanism.

PLPs isolated from PMA-treated wild-type MEG-01 cells or MEG-01 VP30 cells were incubated with EBOVΔVP30-GFP at an MOI of 10 for 90 minutes to allow for virus attachment and potential engulfment by the PLPs. After this incubation, the PLPs were treated with or without 0.1% trypsin to remove uninternalized virus particles and then washed with PBS four times. The washed PLPs (2 x 10^6^) were then co-cultured with Huh7 VP30 cells (2 x 10^5^ cells) for 2 days, and the number of GFP-expressing cells was measured by flow cytometry.

Co-culturing of Huh7 VP30 cells with trypsin-untreated PLPs resulted in over 50% GFP-positive cells, which was similar between PLPs produced from wild-type and VP30-expressing cells (S3 Fig). In contrast, when isolated PLPs were washed with a trypsin solution, the number of GFP-positive cells decreased to 0.89%–1.55%, indicating that most virus particles remained on the surface of the PLPs without being internalized. These results suggest that EBOVΔVP30 particles on the surface of the PLPs are most likely not internalized into the platelets.

Internalization of PLPs by recipient cells

EBOV GP on the virion surface facilitates both virus attachment and entry into permissive cell types. Given that PLPs released from exposed MEG-01 VP30 cells have EBOV GP on their surface (Fig 2), we next assessed the efficiency with which recipient cells internalized PLPs. To visualize internalization of the PLPs, we fused the platelet marker CD41 with the fluorescent reporter gene mCherry (CD41-mCherry). MEG-01 cell lines stably expressing CD41-mCherry only or with EBOV GP were generated (Fig 4A). Isolated PLPs produced from these cell lines were confirmed to express CD41-mCherry or EBOV GP by western blot analysis (Fig 4B). PLPs (2 x10^6^) produced from wild-type, CD41-mCherry, or CD41-mCherry/GP MEG-01 cells were isolated and then co-cultured with Huh7 VP30 cells (2 x 10^5^ cells).

*Internalization of PLPs by recipient cells.A,B. Expression levels of GP and CD41-mCherry in MEG-01 cells stably expressing CD41-mCherry with or without EBOV GP (A), and in PLPs released from these stable cell lines (B). Protein levels were analyzed by immunoblotting using the indicated antibodies. C. Percent of mCherry-positive cells that internalized PLPs containing CD41-mCherry or CD41-mCherry/GP. Huh7 VP30 cells were co-incubated with the indicated PLPs for 1, 3, or 5 h, followed by quantification of mCherry-positive cells by flow cytometry. Data are presented as means ± SD of three independent experiments. Statistical significance was assessed by use of a two-way ANOVA followed by Turkey’s multiple comparisons test. *p < 0.05, ***p < 0.0001.

Because EBOV particles can be internalized within 10 min of attaching to the cell surface, and viral membrane fusion with the endosomal membrane begins 3–4 h after internalization [78], we incubated the different CD41-mCherry PLPs (with or without EBOV GP) and Huh7 VP30 cells together for 1, 3, or 5 h. Unbound and uninternalized PLPs were removed by treatment with 0.25% trypsin, and cells were harvested. The number of cells with internalized PLPs (mCherry positive) was measured by flow cytometry. There was an approximate 2-fold increase in mCherry-positive Huh7 VP30 cells that were incubated with PLPs bearing EBOV GP (PLPs CD41-mCherry/GP) compared to the PLPs without EBOV GP (PLPs CD41-mCherry) throughout the time course, even at the earliest time point (Figs 4C and S4 Fig). This increase in uptake of PLPs bearing EBOV GP suggests the tropism of the PLPs may be affected by the interaction with a cellular receptor.

PLP-mediated transfer of viral components to recipient cells

We demonstrated that PLPs harbor both viral proteins and gRNA, and that PLPs with the EBOV GP can be internalized into recipient cells. We next assessed whether the internalized PLPs can result in the production of progeny viruses in the recipient cell. PLPs were isolated from either PMA-treated, wild-type MEG-01 cells or PMA-treated, MEG-01 VP30 cells that were exposed to EBOVΔVP30-GFP. The isolated PLPs (2 x 10^6^) were co-cultured with Huh7 VP30 cells (2 x 10^5^ cells), and two days later, GFP expression was assessed in the recipient Huh7 VP30 cells by flow cytometry.

When Huh7 VP30 cells were co-cultured with PLPs isolated from PMA-treated, EBOVΔVP30-GFP-exposed MEG-01 VP30 cells, 6.4% of the cells were GFP-positive, compared to <0.1% GFP-positive cells when Huh7 VP30 cells were co-cultured with PLPs from wild-type MEG-01 cells (Fig 5A, top panels). After the MEG-01 cells were exposed, they were washed extensively (four of times with PBS) to remove unbound EBOVΔVP30-GFP. However, there is a possibility that unbound residual virus from the initial exposure of the MEG-01 cells could be a source of virus for the recipient Huh7 VP30 cells. Therefore, the final wash of the exposed MEG-01 cells was used as the inoculum for the recipient Huh7 VP30 cells. However, the amount of GFP-positive Huh7 VP30 cells was < 1% when the final wash was used as the inoculum (Fig 5B), indicating that the source of infection of the recipient cells was the purified PLPs isolated from the exposed MEG-01 VP30 cells.

*PLP-mediated transfer of viral components to recipient cells.Number of GFP-positive cells following co-incubation with PLPs released from EBOVΔVP30-exposed MEG-01 cells (A). PLPs were collected from PMA-treated MEG-01 WT and VP30 cells exposed to EBOVΔVP30-GFP. Huh7 VP30 cells (2 x 105 cells) were co-incubated with the indicated PLPs (2 x 106) for two days. As a control, cells were cultured in the final wash supernatant (B). GFP-positive cells were quantified by flow cytometry. Data are representative of two independent experiments. C-E. EBOVΔVP30 titers in three different VP30-expressing cell types co-cultured with PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PLPs were prepared as described in (A). Huh7 VP30 cells (2 x 105 cells, C), HUVEC VP30 cells (2 x 105 cells, D), and PMA-differentiated THP-1 VP30 cells (2 x 105 cells, E) were co-cultured with the indicated PLPs (2 x 106). Virus titers on days 2, 4, and 6 were determined using Vero VP30 cells. Data are presented as means ± SD of three independent experiments. The dotted lines indicate the lower limit of detection (1.3 log10 ffu/ml). Statistical significance was assessed by use of a one-way ANOVA followed by Turkey’s multiple comparisons test. ***p < 0.001, ***p < 0.0001.

Next, to examine the production of infectious progeny viruses from the cells that internalized the PLPs, we monitored virus titers in the cell culture supernatant of Huh7 VP30 cells (2 x 10^5^ cells) co-cultured with purified PLPs (2 x 10^6^). Virus titers from the recipient cells co-cultured with purified and washed PLPs isolated from exposed MEG-01 VP30 cells increased 1,000-fold, started at 10^5^ ffu/ml on day 2 after co-culturing and reaching almost 10^8^ ffu/ml on day 6 (Fig 5C). An increase in virus titers was also observed when primary human umbilical vein endothelial cells stably expressing VP30 (HUVEC VP30 cells; Fig 5D) and human monocytic THP-1 cells stably expressing VP30 (THP-1 VP30 cells, pretreated with PMA to induce differentiation into macrophage-like cells; Fig 5E) were co-cultured with purified PLPs from exposed MEG-01 VP30 cells.

An increase in the amount of infectious virus in the supernatant of recipient cells co-cultured with PLPs isolated from exposed wild-type MEG-01 cells was also observed (Fig 5C-5E). However, the amount of infectious virus in each tested cell line was 1,000–10,000-fold lower at all time points compared to the recipient cells inoculated with PLPs isolated from exposed MEG-01 VP30 cells. These data suggest that not all the unbound residual viruses were removed from the exposed MEG-01 cells, or that virus remaining on the cell surface was released along with the PLPs, leading to carryover of infectious virus during the isolation of the PLPs from the MEG-01 cells.

Taken together, our data demonstrate that platelet-like particles released from EBOVΔVP30-infected megakaryocytes have the potential to transfer viral components by behaving like virus-like particles.

Discussion

By using our infectious but biologically contained EBOVΔVP30 system, here we showed the potential of platelets to behave as EBOV-like particles by receiving viral proteins and gRNA from their precursor EBOVΔVP30-infected megakaryocytes.

Megakaryocytic cells reside mainly in bone marrow, and although EBOV antigen-positive megakaryocytes have been reported, whether this represents direct infection or megakaryocyte emperipolesis has not been determined [32,33]. Moreover, an even more important question is whether platelets released from infected megakaryocytes retain viral components and possess infectious potential. Accordingly, here we examined and revealed the susceptibility of megakaryocytes to EBOVΔVP30 infection. In addition to residing in bone marrow, large numbers of megakaryocytes are trapped in the lungs and release platelets into the bloodstream [11]. While it is presumed that EBOV circulating in the bloodstream can travel to the lungs, it has been proposed that EBOV could directly enter the lungs through aerosols, resulting in spread to various tissues including the liver, spleen, heart, mesenteric lymph nodes, and bone marrow [79–82]. Given reports that megakaryocytes found in the liver and spleen can contribute to extramedullary platelet production under certain disease conditions [12,13], further investigation into how platelets and megakaryocytes contribute to this virus dissemination is warranted.

Previous studies have suggested [21] and shown [27,28,77] that platelets harboring virus particles can transfer the virus to other types of cells such as macrophages, dendritic cells, and CD4 T-cells, consistent with our finding that EBOVΔVP30 particles attached to PLPs can be transferred to recipient cells. Our study showed that platelets or platelet-like particles produced from EBOVΔVP30-infected megakaryocytes possess viral proteins, which may form viral inclusion bodies that can synthesize mRNA and gRNA. These PLPs also bear GP on their surface like virions, which could facilitate entry and progeny virus production from recipient cells such as endothelial cells, monocytes, macrophages, and neutrophils – cells capable of phagocytosing platelets [83,84]. Active mRNA and gRNA production in the PLPs containing viral proteins (e.g., NP, VP35) could be one potential explanation for facilitating efficient progeny virus production in recipient cells. However, the PLPs themselves seem unable to produce progeny viruses, perhaps due to limited cellular resources (e.g., transport machinery, lipid availability for viral membrane formation) or insufficient or imbalanced viral transcripts within PLPs. This observation is consistent with previous studies on DENV [25,85]. Unlike the viruses found within platelets (e.g., DENV, FLUAV, and HIV-1), EBOV particles were rarely internalized into PLPs, which may be due to their large, filamentous structure (80 nm in diameter and 800–14,000 nm in length). The average size of platelets is 1,000–2,000 nm in diameter. Similarly, vaccinia virus with its large virion size (roughly 360 × 270 × 250 nm) has been reported to bind to platelets at up to five virions per platelet, including internalized virions [86].

Although our EBOVΔVP30 system generates virus particles that closely resemble authentic virus in terms of replication and morphology, making it an ideal surrogate and model for in vitro studies using VP30-expressing cells [72], future studies using authentic virus in vivo and primary megakaryocytes are ultimately required to confirm our findings. Future studies should also include in vivo experiments using experimental animal models infected with authentic EBOV, incorporating a detailed time course with a focus on platelet-associated virus. Such studies could also evaluate whether platelet transfusion from infected animals to naïve recipients can transfer infectious virus similar to in vitro transfer studies. EBOV-like platelet particles are expected to be produced at very low frequency from infected megakaryocytes in the bone marrow due to the difficulty in accessing this tissue. However in future studies, it may be worth investigating whether EBOV-like platelet particles are more frequently released from other sites, such as lung, where infection can be observed despite the lung not being the primary site of infection [31,79,81,82,87]. In addition, EBOV-like platelet particles, through changes in their activation state or associated viral and cellular components, may modulate the activation or function of neighboring immune cells, such as neutrophiles, macrophages, and dendritic cells, thereby warranting further investigation. Given that platelets are increasingly recognized as key immune effectors that communicate with innate immune cells and endothelial cells [88,89], and that, when vascular integrity is compromised, it may allow platelet-virus complexes to access immune-privileged sites [90,91], it remains plausible, although still speculative, that platelets carrying viral components could contribute to characteristic hyper-inflammatory manifestations or facilitate localized persistence of EBOV infection, including potential relapse from abortively infected cells in survivors. In addition, since the biological function of the PLPs containing viral material is thought to be altered, further investigation is needed to fully understand EBOV-like platelet particles and discover their contribution to EBOV pathogenicity.

Overall, our findings shed light on the potential role of platelets and megakaryocytes in EBOV infection.

Methods

Ethics statement

The use of EBOVΔVP30 in biosafety level 2 (BSL-2) containment at the University of Wisconsin is approved by the Institutional Biosafety Committee, the National Institutes of Health (NIH), and the Centers for Disease Control and Prevention.

Cells

Wild-type MEG-01 cells (ATCC, CRL-2021), MEG-01 VP30 cells (MEG-01 cell line stably expressing EBOV VP30), MEG-01 CD41-mCherry cells (MEG-01 cell line stably expressing CD41 [GenBank accession no. NM_000419.5] fused with mCherry at the C-terminus [CD41-mCherry]), and MEG-01 CD41-mCherry/GP cells (MEG-01 cell line stably expressing CD41-mCherry and EBOV GP) were cultured in RPMI-1640 medium containing 10% fetal bovine serum (FBS) and antibiotics.

MEG-01 stable cell lines were generated by transducing cells with a pMXs-IRES-Neo (pMXs-IN; Cell Biolabs) or pMXs-Puro (pMXs-IP; Cell Biolabs) retroviral vector encoding EBOV VP30 (pMXs-IN VP30), CD41-mCherry (pMXs-IP CD41-mCherry), or EBOV GP (pMXs-IN GP) as previously described [73]. Stable cells were selected with 500 μg/ml G418 (InvivoGen) or 2 μg/ml puromycin (InvivoGen).

HUVEC VP30 cells (human umbilical vein endothelial cells stably expressing EBOV VP30) [73] were cultured in Endothelial Cell Growth Medium (Cell Applications, Inc; 211–500) with 500 μg/ml G418 (InvivoGen).

THP-1 VP30 cells (human monocytic THP-1 cells stably expressing EBOV VP30) were generated by transducing cells with a pMXs-IN VP30 retroviral vector and were cultured in RPMI-1640 medium containing 10% FBS and antibiotics plus 500 μg/ml G418 (InvivoGen). Prior to experiments, cells were treated with 50 ng/ml PMA for 2 days to induce differentiation into macrophage-like cells.

Vero VP30 cells (African green monkey kidney-derived Vero cells stably expressing EBOV VP30) [72] and Huh7 VP30 cells (Human hepatocarcinoma Huh7.0 cell line stably expressing EBOV VP30) [73] were cultured in DMEM supplemented with 10% FBS and antibiotics plus 3 μg/ml and 4 μg/ml puromycin, respectively. All cells were maintained at 37°C and 5% CO_2_ and were routinely tested for mycoplasma.

Viruses

EBOVΔVP30-GFP expressing green fluorescent protein instead of the viral VP30 gene was propagated in Vero VP30 cells [72].

Preparation of platelet-like particles

MEG-01 WT or VP30 cells were treated with 5 ng/ml phorbol 12-myristate 13-acetate (PMA) to induce differentiation into megakaryocytic cells [70,71] and cultured for four days. Supernatant containing platelet-like particles (PLPs) was collected and centrifuged at 100 x g for 20 min to remove cells. After filtration through a 5-μm filter (Millipore), the collected supernatant was centrifuged at 1,200 x g for 20 min to spin down the PLPs. The number of PLPs was counted on a hemocytometer (Millicell Disposable Hemocytometer, Millipore) and used for the assays.

EBOVΔVP30-GFP infection of MEG-01 VP30 cells

MEG-01 WT or VP30 cells were seeded on amine-coated dishes or plates (Corning) at 6 x 10^5^ cells/ml in the presence or absence of 5 ng/ml PMA 24 h prior to infection. Cells were exposed to EBOVΔVP30-GFP at an MOI of 5 for 1.5 h at 37°C and then unbound virus particles were removed by washing the cells with PBS. For assessment of GFP expression in infected cells, cells were collected on day 2 post-exposure and analyzed by flow cytometry. For assessment of virus titer, supernatants were harvested and centrifuged at 2,000 x g for 10 min to remove PLPs, cells, and cell debris. Viral titers were determined by use of a focus-forming assay on Vero VP30 cells as previously described [72].

Western blot analysis

Cells were harvested on day 2 post-exposure. PLPs were collected on day 4 post-exposure as described above. Virus particles remaining on the surface were removed with trypsin treatment and rinsed four times with PBS. The cells and PLPs were treated with lysis buffer [1% NP-40, 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.25% sodium deoxycholate], mixed with an equal volume of 2 × Laemmli sample buffer (Bio-Rad) containing 5% β-mercaptoethanol, and then incubated at 95°C for 10 min. Samples were analyzed by immunoblotting as previously described [73]. Primary antibodies used were against: EBOV GP (1:1000, cl. 254/3.12, in-house), EBOV NP (1:1000, R5071, in-house), EBOV VP40 (1:1000, cl. 6, in-house), EBOV VP35 (1:1000, cl. 5.82-6.3, in-house), EBOV VP30 (1:1000, PA5–112031, Invitrogen), EBOV VP24 (1:1000, cl. 21-5.2.5, in-house), CD41 (1:1000, ab134131, Abcam), Beta actin (1:1000, ab8224, Abcam), mCherry (1:1000, 43590, Cell Signaling Technology), and GFP (1:1000, 2956, Cell Signaling Technology).

Quantitative Reverse Transcription PCR (qRT-PCR)

Total RNA was extracted by using an RNeasy Mini Kit (Qiagen). Approximately 400–600 ng of total RNA was reverse-transcribed into cDNA by using a QuantiTect Reverse Transcription Kit (Qiagen) and an oligo dT primer or a forward primer targeting the EBOV leader sequence (5’-TTGTGTGCGAATAACTA-3’). The cDNA was amplified and analyzed by using PowerUp SYBR Green Master Mix (Life Technologies) on QuantStudio 6 Flex (Applied Biosystems) following the manufacturer’s protocol. The primer sequences were as follows: 5’-TTGACAGCAGGTCTGTCCGTTCAA-3’ (forward primer targeting EBOV NP gene), 5’-AACAACTGCTTCAAAGGCCTGTA-3’ (reverse primer targeting EBOV NP gene), 5’-CCTGATACTTGCAAAGGTTGGT-3’ (forward primer targeting EBOV Trailer sequence), 5’-ACGCAGGGAGAGAGGCTAAA-3’ (reverse primer targeting EBOV Trailer sequence). The primer targeting the human beta-actin (ACTB) gene, which was used as an internal control gene, was purchased from Qiagen. The results are expressed as fold-changes normalized to ACTB gene expression by using the comparative cycle threshold (ΔΔCt) method.

Detection of viral proteins in PLPs by use of the proximity ligation assay (PLA) and immunofluorescence assay

PLPs were collected and prepared from EBOVΔVP30-exposed MEG-01 VP30 cells as described above. Virus particles remaining on the surface were removed with trypsin treatment and rinsed with PBS. For detection of EBOV NP, VP35, and VP40 in PLPs, after being fixed with 4% paraformaldehyde (PFA) for 15 min and permeabilized with 0.005% Triton X-100 for 10 min, PLPs were subjected to the in-situ PLA using the Duolink Detection kit (Sigma) according to the manufacturer’s protocol with slight modifications [73]. Briefly, after blocking, cells were incubated overnight at 4°C with the following primary antibody pairs: Rabbit anti-EBOV NP (1:100, R5071) and mouse anti-EBOV VP35 (1:100, cl. 5.82-6.3) or mouse anti-EBOV VP40 (1:100, cl. 6). Incubation with PLA plus and minus probes (a pair of oligonucleotide-labeled secondary antibodies) and ligation (formation of a closed, circle DNA template) were performed as described in the manufacturer’s protocol. Amplification of the circled DNA template (rolling circle amplification) was performed for 180 min. Protein complexes were visualized by using an LSM 510 META confocal microscope (Carl Zeiss) with ZEN 2009 software (Carl Zeiss).

For detection of GP and CD41 on PLPs, PLPs were fixed with 4% PFA for 15 min. After being blocked with 1% BSA in PBS, PLPs were incubated overnight with the following primary antibodies: mouse anti-EBOV GP (1 μg/ml, 746/16.4) and rabbit anti-CD41 (1:100, ab134131, Abcam). After washes with PBS, PLPs were incubated with the following antibodies: an anti-mouse IgG–Alexa Fluor 488 (1:200, A11029, Life Technologies) and an anti-rabbit IgG–Alexa Fluor 546 (1:200, A11035, Life Technologies). Stained samples were analyzed by using an LSM 510 META confocal microscope (Carl Zeiss) with ZEN 2009 software (Carl Zeiss).

Flow cytometry

For assessment of cells that internalized PLPs possessing CD41-mCherry, Huh7 VP30 cells (2 x 10^5^) were co-incubated with PLPs (2 x 10^6^) collected and prepared from MEG-01 WT, CD41-mCherry, or CD41-mCherry/GP cells as described above. After co-culturing for 1, 3, or 5 h, cells were treated with 0.25% trypsin for 3 min to remove uninternalized PLPs and harvested. After being fixed with 4% PFA and washed several times with PBS, the cells were analyzed with a FACS Aria III flow cytometer (BD Biosciences) and FlowJo software (Tree Star).

For assessment of cells that internalized PLPs produced from EBOVΔVP30-exposed MEG-01 VP30 cells, Huh7 VP30 cells (2 x 10^5^) were co-incubated with PLPs (2 x 10^6^) that were prepared as described above. Virus particles remaining on the PLP surface were removed by 0.1% trypsin treatment for 1 min and rinsed four times with PBS. Supernatants of final washes were kept for a control experiment. After co-incubation for 2 days, cells were collected and fixed with 4% PFA. Then, the cells were analyzed with a FACS Aria III flow cytometer (BD Biosciences) and FlowJo software (Tree Star).

Assessment of virus titers from cells co-cultured with PLPs

Huh7 VP30 cells (2 x 10^5^), HUVEC VP30 cells (2 x 10^5^), and PMA-differentiated THP-1 VP30 cells (2 x 10^5^) were co-cultured with PLPs (2 x 10^6^) that were prepared as described above. PLPs were washed with 0.1% trypsin treatment for 1 min to remove virus particles remaining on the surface and rinsed four times with PBS before use in co-cultures. Supernatants were harvested at 2, 4, and 6 days post-co-culture and centrifuged at 2,000 x g for 10 min to remove PLPs and cell debris. Viral titers were determined by use of focus-forming assays as previously described [72].

Supporting information

S1 FigViral protein and gRNA expression levels in EBOVΔVP30-exposed MEG-01 VP30 cells.A. Expression of the indicated proteins in PMA-treated MEG-01 WT or VP30 cells, and Vero VP30 cells exposed to EBOVΔVP30-GFP at an MOI of 1 for MEG-01 cells, or at an MOI of 0.05 for Vero VP30 cells. Protein levels at 2 days post-exposure were analyzed by immunoblotting. B. Relative levels of EBOV gRNA in PMA-treated MEG-01 WT or VP30 cells exposed to EBOVΔVP30 at an MOI of 5. EBOV gRNA was quantified by RT-qPCR using the indicated genome-specific primer pairs and normalized to the level at 1 h post-exposure. Data are presented as means ± SD of two independent experiments performed in triplicate.(TIF)

S2 FigVirus titers in the supernatant of PLPs.EBOVΔVP30 titers in the supernatant of PLPs released from EBOVΔVP30-exposed MEG-01 VP30 cells. PMA-treated MEG-01 VP30 cells were exposed to EBOVΔVP30-GFP at an MOI of 5. PLPs were collected on day 4 post-exposure and subsequently cultured for 3 days. Supernatants were collected daily (Day 0 sample was collected at 1 h post-exposure). Viral titers were determined using Vero VP30 cells. Data are presented as means ± SD of three independent experiments.(TIF)

S3 FigCo-culture of Huh7 VP30 cells with EBOVΔVP30.GFP-positive cells following co-culture with EBOVΔVP30-bound PLPs. PLPs prepared from MEG-01 WT or VP30 were co-incubated with EBOVΔVP30-GFP at an MOI of 10. After washing with or without trypsin, PLPs (2 x 10^6^) were co-cultured with Huh7 VP30 cells (2 x 10^5^ cells) for 2 days. GFP-positive cells were quantified by flow cytometry. Data are representative of two independent experiments.(TIF)

S4 FigInternalization of PLPs by recipient cells.Percentage of mCherry-positive cells that internalized PLPs containing CD41-mCherry or CD41-mCherry/GP. Huh7 VP30 cells were co-incubated with the indicated PLPs for 1, 3, or 5 h. mCherry-positive cells were quantified by flow cytometry. Data are representative of three independent experiments.(TIF)

S1 DataSource data behind Figs 1–5 and S1–S2 Figs.(XLSX)

S1 MethodAdditional description of methods for flow cytometric analysis and assessment of virus titers from PLPs.(DOCX)