Trophoblastic Extracellular Vesicles and Viruses: Friends or Foes?

Abstract

Cells produce cytoplasmic vesicles to facilitate the processing and transport of RNAs, proteins, and other signaling molecules among intracellular organelles. Moreover, most cells release a range of extracellular vesicles (EVs) that mediate intercellular communication in both physiological and pathological settings. In addition to a better understanding of their biological functions, the diagnostic and therapeutic prospects of EVs, particularly the nano-sized small EVs (sEVs, exosomes), are currently being rigorously pursued. While EVs and viruses such as retroviruses might have evolved independently, they share a number of similar characteristics, including biogenesis pathways, size distribution, cargo, and cell-targeting mechanisms. The interplay of EVs with viruses has profound effects on viral replication and infectivity. Our research indicates that sEVs, produced by primary human trophoblasts, can endow other non-placental cell types with antiviral response. Better insights into the interaction of EVs with viruses may illuminate new ways to attenuate viral infections during pregnancy, and perhaps develop new antiviral therapeutics to protect the feto-placental unit during critical times of human development.

Introduction

Extracellular vesicles (EVs), existing in the plasma, urine, and likely in all biofluids^1,2, influence a variety of physiological and pathological processes, such as aging³, neutralization of bacterial toxins⁴, viral infection^5–7, immune response^8,9, and tumorigenesis.^10,11 EV subtypes and function are defined by their size, surface molecules, release mechanisms, and cargo¹², including proteins; RNAs such as microRNAs (miRNAs)^13,14, long non-coding RNAs¹⁵, and transfer RNAs;¹⁶ and other bioactive molecules. The spectrum of EV types includes, in order of decreasing size, tumor cell–derived oncosomes^17,18 (1,000~10,000 nm), apoptotic blebs (ABs, 100~5,000 nm), microvesicles¹⁹ (MVs, 100~1,000 nm), and exosomes^20,21 (50~150 nm). Other, less-defined EVs include amphisoms, arrestin-domain-containing protein 1 (ARRDC1)-mediated microvesicles (ARMM), non-classical MVs and exosomes, and the large macrolets.^21–24 Other than obvious size differences among EV subpopulations, exosomes are notably distinct from other EVs as they are formed by cellular endocytic pathways, whereas oncosomes, ABs, and MVs are directly shed from the plasma membrane. Recently, by using asymmetric flow field-flow fractionation, a new type of non-vesicular nanoparticle, termed exomeres (30~50 nm), was identified within exosome preparations.²² The cargo of exomeres is strikingly distinct from exosomes, adding to the heterogeneity of EVs and the nanoparticle repertoire produced by host cells.^22,25 Following the consensus recommendations by the International Society for Extracellular Vesicles (ISEV)^21,26,23, and reflecting the difficulty in defining EV subtypes based on size and biophysical properties, we herein designate exosomes as small extracellular vesicles (sEVs). Extensive reviews on the biogenesis and properties of EVs, which are not a focus of this text, have been published elsewhere.^19,26,23,27 Here we focus on the intriguing interplay between EVs and viruses, particularly as it pertains to human placental trophoblasts. These cells use sophisticated defense mechanisms to confront viruses that may reach the placenta, and thus protect against transmission of viruses to the developing fetus.

Human placental EVs have been shown to convey intercellular messages at the fetal-maternal interface²⁸ and play a pivotal role in cellular adaptation processes, including remodeling of maternal endothelial cells^29–31, immuno-modulation and tolerance^32–36, and defense against viral infection.^37,5 Based on their potential role in maternal-fetal interaction and disorders of pregnancy, the impact of placental EVs and their cargo has been studied in the context of gestational diseases. These include studies of biomarkers of environmental pollutant-induced injury or pregnancy disorders, including intrauterine growth restriction, preterm birth, and preeclampsia.^38–44 For more details, the readers are referred to other reviews in this special issue on placental EVs and disorders of pregnancy. Here we focus on current knowledge of trophoblastic EVs and relevant viral infections during pregnancy.

EVs and viruses: similarities and dissimilarities

Before reviewing the complex interaction of EVs with pregnancy-related pathogenic viruses and the role of EV-based defense against viruses, we compare several key properties that define EVs and viral pathogens, emphasizing similar characteristics and differences.

Size and Shape

The size of most viruses is in the range of 20~500 nm in diameter, which overlaps with the size distribution of EVs. Unlike the roughly spherical structure of EVs, with an outer phospholipid bilayer membrane, many viruses have less spherical nucleocapsid shapes, such as rods or polygonal spheres, attributed to viral nucleoproteins (or capsid proteins) and their interaction with viral nucleic acids. Additionally, certain viruses, such as human immunodeficiency virus (HIV-1) and vesicular stomatitis virus (VSV), have an extra lipid membrane bilayer intercalated with viral enveloped proteins, leading to spherical or rod-like shapes.

Vesiculated or Non-vesiculated forms

Depending on the encasement of the nucleocapsid core in a lipid layer derived from host cell membrane, viruses are categorized as non-enveloped or lipid-enveloped. As such, vesiculated viruses resemble EVs and may use similar routes to enter target cells. Notably, differences in vesiculation among non-enveloped viruses, enveloped viruses, and vesiculated EVs may be blurred in some circumstances. For instance, hepatitis A viruses, presumably non-enveloped, cloak their nucleocapsids with host cell–derived lipid membrane when circulating in the plasma of infected humans.⁴⁵ Upon egress from host cells, some enveloped viruses, such as hepatitis C virus, disguise progeny virions with an additional lipid membrane bilayer and associated host cellular proteins.⁴⁶ In a more compound scenario, EVs derived from virus-infected cells might enclose viral miRNAs⁴⁷, further obscuring the boundary between cellular EVs and EVs derived from progeny virions.

Formation and Propagation

Key proteins that control the intricate steps of exocytosis are utilized in biogenesis of both EVs and viruses.^48,49 Processing within the cellular endosomal network, including both early and late endosomes, and the subsequent fusion of multivesicular bodies (MVBs) with the plasma membrane or the alternative direct budding of plasma membrane, are obligatory for EV generation and egress from parental cells.⁷ Likewise, viruses also leverage the endosomal network and MVBs to facilitate their replication, assembly, and release from host cells. Mechanistically, Rab27a/b⁵⁰, SNAP protein and its cognate SNAP REceptor (SNARE) protein complexes⁵¹, and the Endosomal Sorting Complexes Required for Transport (ESCRT) multiple subunit family members, including ALIX1 and TSG101, promote fusion of MVBs with the plasma membrane and facilitate the release of both viruses and intraluminal vesicles (which will become EVs) from host cells⁵². As expected, certain virus families may preferentially deploy discrete subsets of host cell proteins to support their replication, assembly, and release. Similarly, EVs derived from various cell types also evolve unique mechanisms to modulate their biogenesis pathways, such as ceramide-dependent, but ESCRT-independent, signaling.^53,54 Nevertheless, viruses and EVs are fundamentally distinct in the ability to replicate: while some viruses may not propagate or egress from host cells, most can synthesize progeny virions in the infected recipient cells and resort to new-round infection events. In contrast, currently there is no evidence that EVs bear the capability to propagate in recipient cells.

Cargo content

The DNA or RNA genomes enclosed within viruses encode viral proteins that engage host cells in support of viral replication. Moreover, DNA viruses such as human cytomegalovirus (hCMV) and retrovirus HIV-1 leverage host machinery to transcribe viral miRNAs, which functionally repress host cell mRNA transcripts and therefore facilitate viral replication.^55–57 Interestingly, EV’s RNA cargo, such as miRNAs, has been demonstrated to attenuate the expression of host target genes and therefore redirect recipient cells into pathophysiological pathways, including aging, lipodystrophy, and tumor metastasis.^14,13,58,3 Viruses package viral genomes as nucleocapsids and transcribe viral miRNAs following delivery of their genomes to the host. EVs encapsulate miRNAs upon egress from donor cells, yet pathways for processing EV miRNA cargo and its delivery to Argonaute proteins in RNA-induced silencing complex (RISC)⁵⁹ remain unclear. The possibility that Argonaute proteins might be a part of the sEVs’ cargo was recently ruled out by Coffey’s group, who demonstrated that Argonaute proteins are present in non-vesicular forms rather than in sEVs.²¹ These data are consistent with our previous findings based on human trophoblastic sEVs.³⁷ Compared to viruses bearing DNA genomes, sEVs have no double-strand DNA (dsDNA) and associated histone proteins.

Given that both EVs and viruses exploit similar biogenesis pathways, it is not surprising that they may share similar lipid composition.^48,27 Notwithstanding the resemblance, there are unambiguous differences in the protein landscape between EVs and viruses. Unlike EV proteins that were derived from donor cells, viruses contain non-cellular, virus-specific proteins, which assist them in entry into the host and multiply progeny virions.^60,61 Collectively, while cargo delivery by both viruses and EVs can reprogram some functions of the recipient cells, the consequences are unequivocally distinctive in that viruses may lead to detrimental effects on the host, whereas EVs, depending on affected signaling cascades and the cellular context, might elicit both beneficial and deleterious effects on recipient cells.

Target cells: Entry and endocytosis

EVs enter target cells to execute their function, whereas viruses must enter target cells for replication. Based on this notion, and the structural similarity between viruses and EVs, it is not surprising that EVs and viruses may utilize similar cell recognition, attachment, and endocytic mechanisms to target intracellular domains. When compared to EV processing, more is currently known about virus uptake and processing in target cells.

The entry of viruses into cells is a stepwise process, with initial interaction taking place upon virus attachment to surface factors on the cellular plasma membrane, followed by receptor clustering into macrodomains, signaling activation, formation of endocytic vesicles with delivery of viral cargo into endosomes, endosomal sorting and trafficking, and a final escape from the endosomes. Similarly to viruses, EVs utilize multiple nonexclusive entry mechanisms to deliver their cargo into recipient cells and execute their functions.^19,62 The initial interaction that guides EV attachment to a specific cell may take place using “barcode and reader” recognition proteins that are also involved in endocytosis pathways, including clathrin-mediated and caveolae-mediated endocytosis.^63,64 Cell entry by enveloped viruses, such as herpes simplex virus (HSV), also entails virus and plasma membrane fusion, primed by viral fusion proteins to overcome energy barriers imposed by the two merging lipid bilayer structures.⁶⁵ Intriguingly, the fusogenic proteins syncytin1/2 (ERVW-1 and ERVFRD-1), derived from envelope proteins of human endogenous retroviruses (ERV) and predominantly expressed in the placenta, are present in trophoblastic sEVs as well^66,67,37 and are postulated to regulate sEV fusion with target cells.^37,66 Cellular surface protein integrins that guide viral entry are also packaged in sEVs, including human trophoblastic sEVs³⁷, leading to tumor sEV-mediated organotropic metastasis.^68,62 In addition to protein barcode and reader-conferred viral entry, viral envelope phospholipid phosphatidylserine (PS), alongside PS-adaptor protein Gas6, mediate viral entry of cells by binding to Axl, a TAM receptor tyrosine kinase on target host cells.⁶⁹ Notably, we found that phosphatidylserine is also expressed on human trophoblastic EVs, including ABs, MVs, and sEVs.³⁷

Upon internalization, viruses usurp the cell endocytic pathways for further processing. The assortment of viral receptor proteins and pathways used for cell entry guide the interaction of viruses or EVs with the cellular endocytosis pathways, including clathrin-mediated endocytosis, caveolae-mediated endocytosis, and macropinocytosis.^68,19,70 Fascinatingly, the cells deploy similar endocytic pathways to block viral replication, thus staging the cell-virus war at the endocytosis stage. Clathrin-mediated endocytosis is used by many viruses, such as VSV⁶⁸, HIV-1⁷¹, and Zika.^72,73 This process includes pre-existing clathrin-coated pits or newly assembled clathrin sites, which mediate virus or EV delivery to early endosomes in the cell periphery.⁶³ Viruses (such as simian virus 40, SV-40) that utilize caveolae-mediated endocytosis, colocalize with caveolin, with subsequent tyrosine phosphorylation–dependent virus uptake. Clathrin/caveolin-independent pathways commonly begin with a virus-induced membrane curvature and wrapping of the plasma membrane around the viral vesicles, followed by fission and delivery of the viral vesicle to early endosomes.⁶⁸ Macropinocytosis, triggered by many viruses⁷⁴ and likely used by EVs¹⁹, is sparked by integrins, receptor tyrosine kinase, or phosphatidylserine, followed by actin-stimulated membrane ruffling that forms filopodia, lamellipodia, or blebs, which engulf viruses or EVs, with subsequent macropinosome maturation into late endosomes and fusion with lysosomes. Consistently, filopodia are critical for sEV uptake in primary human fibroblasts.⁷⁵ Using primary human placental fibroblasts and human uterine microvascular endothelial cells as likely relevant recipient cells for trophoblastic sEVs, we uncovered that trophoblastic sEVs enter these two target cells mainly through macropinocytosis and clathrin-mediated endocytosis, but not caveolin-dependent endocytosis.⁷⁶

Once taken up by cells, viruses or EVs enter the cell endosomal network, sharing the space with intraluminal vesicles. If viruses or EVs do not exit the endosome into the cytosol, they presumably shuttle onward within early and late endosomes and subsequently undergo degradation in lysosomes. Notably, many viruses hijack the endosome machinery and leverage the acidic pH for fusion with the endosomal membrane, culminating in release of the viral genome. The fusogenic protein hemagglutinin (HA), expressed on the surface of influenza virus, is a canonical mediator of fusion with endosomes.⁷⁷ Likewise, the enveloped retrovirus HIV-1 expresses the trimeric fusion glycoprotein gp160⁷⁸, and membrane fusion largely occurs in the endosome following internalization by clathrin-mediated endocytosis, implying that HIV-1 can evade immune surveillance by reducing exposure of the viral epitope on the outer layer of host cells.⁷¹ Some viruses can escape early endosomes, late endosomes, lysosomes, endoplasmic reticulum, Golgi or even recycling endosomes and form their own independent, clustered macrodomains.⁷⁰ Likewise, following entry of target cells such as primary human placental fibroblasts and human uterine microvascular endothelial cells, human trophoblastic sEVs are transported to APPL1- and EEA1-positive early endosomes, and then translocated to late endosomes and lysosomes.⁷⁶ Intriguingly, the majority of endocytosed sEVs deploy diverse membrane interaction modes to relocate to regions near the endoplasmic reticulum (ER). Such a unique sEV-and-ER contact is believed to modulate cargo release.⁷⁵ Our recent findings also revealed that miR-517a-3p, one of trophoblast-specific sEV cargo molecules, localizes to intracellular P-bodies and binds GW182 and Argonaute 2 (AGO2) in target primary human placental fibroblasts.⁷⁶ The relevance of entry and endocytic pathways of placental EV may be germane to placental biology and the use of EVs in perinatal translational medicine.⁷⁹

Certain sEVs suppress viral infection via delivery of vesicular cargo

Among the intricate strategies used by cells to combat viral infections, donor cells may endow neighboring and/or distant recipient cells with EV-mediated RNAs and proteins that participate in antiviral signaling cascades, thus inhibiting viral replication and/or abrogating progeny virion production and release. Relevant to modulation of trophoblast differentiation by interferon-induced transmembrane (IFITM) protein family members⁸⁰, IFITM3 hinder cell entry of a broad spectrum of viruses, including the Flaviviridae family Dengue and Zika viruses, HIV-1, Influenza, and SARS coronavirus^81–83, and also reroute the internalized viruses to lysosomes for degradation, thus limiting viral replication.⁸⁴ Consistent with these observations, sEV-mediated transfer of vesicular IFITM3 proteins suppresses Dengue virus entry and subsequent progeny virion release.⁸⁵ Cytidine deaminase APOBEC3G restrains viral replication and infectivity of a panel of retroviruses, including HIV-1, by deamination of the viral genome. Notwithstanding, the HIV-1 Vif protein resists the antiviral activity of APOBEC3G.^86–89 APOBEC3G-laden sEVs also suppress viral replication of wild-type HIV-1, despite the fact that the viral genome maintains the least frequency of deamination⁹⁰, suggesting that APOBEC3G is unlikely to be the sole antiviral effector of HIV-1 inhibition.

At an organ scale, sEVs might reconstruct antiviral defense mechanisms that engage multiple distinct cell types. As such, induced packaging of functional antiviral molecules into sEVs derived from one cell type may augment an antiviral response by other cell types and, in turn, impede viral infection. Type I interferon (IFN), an innate immune defense molecule, confines viral infection at the host defense frontline until an adaptive immune response is mounted.⁹¹ In addition to the direct action of IFN on host cells, IFN can stimulate the packaging of antiviral factors into sEVs and therefore propagate its antiviral activity to other cell types, usurping sEVs to execute IFN action⁹². Beyond augmented delivery of proteins, miRNAs such as miR-423-5p may be induced and sorted into sEVs of human diploid MRC-5 cells infected by rabies viruses and exert a suppressive effect on viral replication by repressing the suppressor of cytokine signaling 3, an inhibitor of innate interferon signaling.⁹³

An antiviral response to sEVs takes place not only within an individual but also among individuals and across species. For instance, sEVs from healthy human semen attenuate HIV-1 infection by repressing HIV-1 proviral transcription.⁹⁴ Furthermore, bacteria-released sEVs may protect women from HIV-1 infection. For example, predominant vagina-resident lactobacilli can curtail HIV-1 infection, and human ex vivo tissues or cell lines treated with lactobacillus-derived sEVs from heathy women can reduce HIV-1 attachment and entry and, consequently, HIV-1 virus load.⁹⁵

Certain sEVs promote viral infection by cloaking viruses or changing vesicular cargo

The notion that certain viruses, such as retroviruses, engage inherent EV biogenesis machinery for the assembly of progeny virions⁹⁶ supports the concept that EVs might augment viral infectivity. As such, progeny virions are cloaked with host cell–derived EVs and egress from infected cells via the conventional EV exocytosis pathway, rendering them Trojan horses to evade immunosurveillance.⁹⁷ Supporting this notion, both non-enveloped hepatitis A viruses and enveloped hepatitis C viruses, disguised with host cell–derived lipid membrane, have been shown to circulate in the plasma of infected humans, where they act as pseudo-enveloped viruses that may acquire increased viral entry tropism or avoid neutralization by antibodies.^46,45 Similarly, by hijacking EVs to harbor multiple virion copies, enteroviruses, noroviruses, and rotaviruses can overburden the host viral defense mechanisms and thus promote viral replication.^98,99 Adding another dimension to virus dissemination prior to the lytic phase of viral infection, picornaviruses are released via a non-lytic route by encasing their progeny with membranes derived from multiple distinct EVs, including MVs and sEVs, which account for the discrete viral infectivity.¹⁰⁰

In some pathological settings, despite the fact that sEVs derived from infected cells do not contain intact virions and had therefore been presumed to be noninfectious¹⁰¹, sEV content, including vesicular protein and miRNA cargoes, may be altered by incorporating significant amounts of viral proteins or miRNAs following viral infection.^102,47 Such virus-primed sEVs can reprogram recipient uninfected cells to facilitate virus spread via increased cell-to-cell contact¹⁰¹ or by potentiating their susceptibility to viral infection.¹⁰³ In addition, a virus-favorable niche reconstructed by virus-educated sEVs is associated with tumor virus–initiated cell transformation.¹⁰⁴ For instance, Epstein-Barr Virus exploits sEVs to transport viral miRNAs into uninfected recipient cells and repress host cell transcription of immunoregulatory genes such as CXCL11.⁴⁷

Virus-derived factors packaged within sEVs may exert detrimental collateral damage on non-infected “bystander” cells and, therefore, in conjunction with the propagation of virions in permissive cells, ultimately result in tissue or organ dysfunction. Contributing to in vivo HIV-1-sparked clinical comorbidity, such as abnormal hematopoiesis and atherosclerosis, HIV-1 protein Nef-containing sEVs alter lipid raft composition through replacement of TLR4 and TREM-1 to lipid rafts of “bystander” cells and lead to boosted secretion of proinflammatory cytokines.¹⁰⁵ In some pathophysiological situations, tumorigenesis may be functionally associated with augmented viral infection due to suppression of innate immunity. Consistent with this notion, tumor-derived EGFR-positive sEVs enhance viral infection in host macrophages by dampening in vivo innate antiviral response.¹⁰⁶

Human pregnancy, viral infection, and trophoblastic sEV-mediated defense mechanisms

Antenatal vertical transmission of DNA viruses, including hCMVs and parvovirus B19 (PVB19), and RNA viruses, such as rubella virus (RV) and Zika virus (ZIKV), has been shown to cause human congenital syndromes.^107–109 For instance, maternal infection by RV during the first trimester of human pregnancy is associated with fetal congenital rubella syndrome, including cardiovascular anomalies, microcephaly, deafness and related conditions. While the exact mechanisms by which RV infects the human placenta and subsequently results in congenital rubella syndrome remain unclear, the problems associated with this syndrome has been largely eliminated by the prevalent deployment of an effective RV vaccine.

Placental proteins and glycosphingolipids are likely utilized as receptors to facilitate virus entry into the placenta, with subsequent intrauterine transmission.¹⁰⁷ For instance, ZIKV infection of pregnant women is associated with fetal neurological deficits, including microcephaly and retinal damage, and causes major fetal morbidity and mortality. Rather than replicating in syncytiotrophoblasts (STBs)¹¹⁰, ZIKV robustly replicates in placental cells expressing viral entry cofactors Axl and TIM1, including extravillous trophoblasts and placenta-specific Hofbauer macrophages.^111,112 PVB19 affects fetal erythropoiesis and may lead to hydrops fetalis due to fetal anemia.¹⁰⁷ Glycosphingolipid globoside¹¹³, a receptor for PVB19, in conjunction with viral coreceptors such as α5β1 integrin¹¹⁴ and Ku80¹¹⁵, are deployed for virion attachment and internalization through endocytosis pathways. Intriguingly, globoside is strongly expressed in the villous trophoblasts of first and second trimester placentas rather than in those of third trimester, which may, in part, contribute to the observation that maternal infection of PVB19 in early gestation may lead to worse outcomes when compared to near-term infection.¹¹³ Maternal hCMV infection poses a risk of fetal transmission, with subsequent birth defects, including cognitive impairment, hearing loss, and intrauterine growth restriction. hCMV replicates in villous cytotrophoblasts (CTBs) by resorting to neonatal Fc receptor–mediated transcytosis from STBs to the underlying CTBs, where most of its replication takes place.^116,117 In addition to transcytosis entry, hCMVs also infect chorionic trophoblast progenitors¹¹⁸, extravillous trophoblasts¹¹⁹, and amniotic epithelial cells.¹²⁰

Most recently, a novel bat-borne, RNA betacoronavirus, SARS-CoV-2¹²¹, which resembles the coronaviruses that caused severe acute respiratory syndrome (SARS) and middle eastern respiratory syndrome (MERS), caused viral pneumonia and the coronavirus disease 2019 (COVID-19) pandemic. While the precise mechanisms underlying SARS-CoV-2 pathogenesis are being uncovered, recent data indicate that the transmembrane TMPRSS2 serine protease is required for priming the viral surface Spike proteins that subsequently utilize angiotensin converting enzyme 2 (ACE2) act as receptors for virus entry.^121,122 Although these proteins are expressed in human placental trophoblasts^123,124, current data indicate that, even during maternal viremia, the feto-placental compartment is relatively spared from SARS-CoV-2 infection at least in the third trimester.^125–127 This suggests that the human placenta may mount effective viral defense pathways even when this typically airborne virus is present in the blood stream and can infect other internal organs.¹²⁸

The morphology of the human placenta constitutes a sophisticated physical barrier of syncytialized cells that may attenuate infection by viruses that circulate in the maternal blood.^129,130 Supporting this notion, primary human trophoblasts, isolated from normal full-term placentas, are relatively resistant in vitro to a diverse range of viruses, including DNA and RNA viruses such as HSV, hCMV, HIV-1, and ZIKV, when compared to other cell types.^131,132,110 Autophagy is engaged in host antiviral defense by activating pattern recognition receptor signaling due to pathogen-associated molecular patterns of the invading viruses, and subsequently diverting them to lysosomes for degradation.¹³³ We found that the relative viral resistance of primary human trophoblasts is not due to defects of viral entry, and instead, primary human trophoblasts exhibit high levels of basal autophagy, which in part contributes to the relative resistance of these cells to viruses.¹³¹ Conditioned primary human trophoblasts medium containing trophoblastic sEVs can also induce autophagy in non-placental cells and render them relatively resistant to viral infection including VSV, Coxsackievirus B, and Hepatitis C virus. Mechanistically, viral resistance is attributed, at least in part, to released trophoblastic sEVs and cargo miRNAs of the chromosome 19 microRNA cluster (C19MC) family.^5,131 C19MC miRNAs are primate specific and almost exclusively expressed in the placenta, where they represent the majority of the miRNAs expressed in trophoblasts.¹³⁴

Both sEVs and cargo C19MC miRNAs such as miR-517a-3p promote autophagy and viral degradation within autolysosomes, thus restraining a subsequent virus uncoating process and cytoplasmic viral replication.¹³¹ Of note, we showed that the expression of C19MC miRNAs, detectable in maternal plasma as early as 2 weeks after implantation, is not affected by maternal infection with hCMV, and that C19MC miRNA augment infection of other cell types by hCMV^135,131, implicating a more complex interaction of hCMV with trophoblastic antiviral pathways.

The C19MC-mediated effect on viral infection is largely executed by trophoblastic sEVs, not ABs or MVs.³⁷ Furthermore, C19MC-laden sEVs execute their antiviral activity without affecting interferons and interferon-stimulated genes such as IFI44L.¹³⁶ This antiviral phenotype is recapitulated from plasma sEVs of pregnant women compared to those from non-pregnant women.³⁷ Additional data also indicate that a C19MC-independent defense mechanism, type III interferon, IFN-λ, participates in protection of primary human trophoblasts against viral infection.^110,136 Collectively, both C19MC miRNAs and IFN-λ are expressed in primary human trophoblasts and orchestrate two independent antiviral defense mechanisms. Whereas IFN-λ or trophoblastic C19MC-containing sEVs attenuate VSV infection, concomitant exposure of target cells to IFN-λ and sEVs leads to an additive antiviral effect.¹³⁶ Moreover, C19MC-containing sEVs can communicate an antiviral activity to a broad range of non-trophoblastic cells, including other local placental cells, maternal cells, or possibly, fetal cells.¹³⁷ In line with this notion, we recently uncovered that two primary human cells that do not endogenously transcribe C19MC miRNAs, including primary human placental fibroblasts and human uterine microvascular endothelial cells, exhibit an antiviral effect after uptake of C19MC-laden sEVs.⁷⁶ While the mechanisms underlying these observations are currently being pursued, these findings illuminate unprecedented mechanisms employed by human trophoblasts at a systemic scale during pregnancy.

Future perspectives

While the early data on the function of placental sEVs in pregnancy disorders are intriguing, the underpinnings of placental sEV biogenesis and their interaction with maternal tissues and organs remain enigmatic.⁷⁹ EVs enter cells via various endocytosis routes, dictated by the nature of their interaction with recipient cells.^19,138 Deep insight into the mechanisms underlying the action of barcode-like molecules at the two apposing sEV and cell lipid membranes, followed by entry into recipient cells, is lacking. Advanced imaging technologies, such as super resolution microscopy, and novel genetic tracking systems are needed in order to tackle such fundamental questions. An additional, critical question is how processing of placental sEV cargo is intimately connected to the control of cell and tissue homeostasis and/or resilience. In this regard, fusion of sEVs with the endosome membrane might be essential for vesicular cargo release. Decoding these molecular mechanisms, using in vitro cell culture studies with in vivo animal tracking approaches, will shed light on the functional relevance of placental EVs to pregnancy adaptation and response to viral pathogens.

Depending on the specific cellular contexts, the interplay of EVs and viruses may have a profound impact on viral replication and infectivity. Deepening our knowledge of key factors that mediate such bidirectional effects may suggest new pathways for protection against viral infection. In light of our discovery that trophoblastic sEVs relay an antiviral program at the maternal-fetal interface, further research into effector molecules that mediate C19MC-conferred antiviral response may shed light on the interaction of sEVs and viruses and become instrumental in initiating therapeutic intervention against viral infection. Such research may also advance our understanding of the on-going conflict between viral infection and placental responses to such infection, and the antiviral and proviral effect of certain proteins and RNAs, as shown in other contexts.^139–142 It may also explain why some placentas and fetuses are infected by certain viruses during maternal viremia, and others are not. Within this context, a critical challenge lies in the unambiguous separation of EVs-derived from infected cells from released virions that co-exist in the extracellular space and may rely on understanding the contributions of virus-primed EVs and viruses alone to viral infectivity. Leveraging the unique biochemical and biophysical properties of EVs or encapsulated viruses, combined with enhanced flow cytometry-based single-vesicle sorting, may provide solutions to this challenge.^7,143–147

From the translational medicine viewpoint, EV-based therapeutics is an emerging area that explores the implementation of personalized and precision medicine approaches to diseases.¹⁴⁸ Research using engineered EVs, consisting of appropriate antiviral agents that facilitate their specific targeting to eradicate virus latency, is being vigorously pursued. One promising strategy is deploying an HIV-neutralizing antibody and apoptosis inducers into EVs in order to suppress viral replication.¹⁴⁹ Such breakthroughs can markedly advance our knowledge of the complex interplay between EVs and viruses.

Figure 1. A general schematic illustrating how EVs and viruses enter cells via endocytosis and fusion-triggered endosomal escape for cargo release and subsequent processing.

EVs and viruses, including lipid enveloped or non-enveloped forms, enter their target cell by binding to specific receptors on the plasma membrane (1). They are subsequently internalized into endosomal vesicles through multiple endocytosis pathways, mainly including clathrin-mediated endocytosis, caveolin-mediated endocytosis, and macropinocytosis (2). Direct virus and plasma membrane fusion-mediated cell entry, utilized by certain viruses, is not shown, but described in the text. This is followed by trafficking within the endosomal network, including early endosome, late endosome or endolysosome (3), EVs and viruses reach the low-pH endosomal compartment (4) where fusogenic proteins on the surface membrane of EVs or viruses trigger membrane fusion (5), with subsequent cargo release to the cytosol (6). EV cargo, including diverse coding or non-coding RNA types, are released and likely execute their function in target recipient cells. In contrast, viral DNA or RNA genomes are released and processed for transcription and replication. The processing of EV or viral proteins is not shown. Unlike the arrival of EVs to their “final destination”, viruses can usurp host cell machinery to synthesize viral nonstructural proteins such as viral polymerase and/or integrase to support viral replication. Consequently, assembled viral genome and proteins may begin replication (7). Lastly, progeny virions utilize exocytosis pathways, cell budding or cell lysis for exit and virus dissemination (8). Additional similarities and dissimilarities in EV and virus pathways are detailed in the text. This figure was created with BioRender.com.