Extracellular Vesicles and Immune Response During Pregnancy: A Balancing Act

SUMMARY

The mechanisms underlying maternal tolerance of the semi- or fully-allogeneic fetus are intensely investigated. Across gestation, feto-placental antigens interact with the maternal immune system locally within the trophoblast-decidual interface and distantly through shed cells and soluble molecules that interact with maternal secondary lymphoid tissues. The discovery of extracellular vesicles (EVs) as local or systemic carriers of antigens and immune-regulatory molecules has added a new dimension to our understanding of immune modulation prior to implantation, during trophoblast invasion, and throughout the course of pregnancy. New data on immune regulatory molecules, located on EVs or within their cargo suggest a role for EVs in negotiating immune tolerance during gestation. Lessons from the field of transplant immunology also shed light on possible interactions between feto-placentally derived EVs and maternal lymphoid tissues. These insights illuminate a potential role for EVs in major obstetrical disorders. This review provides updated information on intensely studied, pregnancy-related EVs, their cargo molecules, and patterns of fetal-placental-maternal trafficking, highlighting potential immune pathways that might underlie immune suppression or activation in gestational health and disease. Our summary also underscores the likely need to broaden the definition of the maternal-fetal interface to systemic maternal immune tissues that might interact with circulating EVs.

1 |. INTRODUCTION

A healthy pregnancy represents an immunological paradox, first described by Peter Medawar, in which the fetus is not rejected by the maternal immune system.¹ Indeed, the feto-placental unit constitutes a semi-allogeneic graft within the maternal host, where typically maternal T and B lymphocytes have not been previously exposed to and tolerized by the expression of paternal antigens (Ags). The paradox is heightened when pregnancy results from a sperm and a donor oocyte, or in a surrogate pregnancy, where the fetus is a full allograft. The intricate immune-regulatory mechanisms that operate at the fetal-maternal interface and in the mother’s lymphoid organs and avert the maternal immunologic attack on feto-placental tissues while maintaining immune homeostasis and maternal competent defense against pathogens remain largely unknown.²

Complex developmental processes underlie the development of the embryonic immune system.³ As the immune interactions evolve, multiple tolerance-promoting mechanisms become active at the fetal-maternal interface and negotiate between the fetal-derived trophoblast and the maternal leukocytes and decidual cells. These mechanisms include downregulation of uterine natural killer (NK) cell cytolytic function, M2-polarization of decidual macrophages, sequestration of decidual dendritic cells (DCs), enrichment of regulatory T cells (Tregs), epigenetic silencing of effector T cell chemoattractants in decidual cells, release of immunosuppressive cytokines and galectins, and downregulation of trophoblastic classic major histocompatibility complex (MHC) molecule expression.^4–9 Uterine NK (uNK) cells are the most abundant maternal leukocytes in first trimester decidua basalis, followed by decidual macrophages and T cells. The function of the uNK cells is regulated by interactions with decidual stromal cells and leukocytes and the semi-allogeneic trophoblasts, which express a unique repertoire of MHC class-I molecules that activate or inhibit uNK function.^10–12 Indeed, binding of inhibitory receptors on uNK cells is associated with insufficient spiral artery remodeling and shallow trophoblast invasion, which are themselves associated with placental dysfunction and clinical diseases such as preeclampsia and fetal growth restriction.¹³ Decidual uNK cells also generate mediators that may contribute to T cell tolerance against fetal Ags at the feto-maternal interface.¹³

At a systemic level, and akin to the immune response against transplants, T and B lymphocytes in the maternal lymphoid tissues become aware of the presence of embryonic non-self Ags during early gestation. In mouse models, T cells in the maternal secondary lymphoid tissues (SLTs) recognize fetus-derived peptides, presented by maternal Ag-presenting cells (APCs) in a maternal MHC-restricted fashion.¹⁴ This recognition is associated with deficient T cell priming, followed by T cell clonal deletion or anergy and generation of Tregs.^14–20 Interestingly, recent evidence based on mouse experiments has shown that the induction or maintenance of peripheral T cell tolerance against fetal Ags is associated with the presence of myeloid cells that express the transcription factor Autoimmune regulator (Aire). Aire promotes the expression of tissue-restricted self Ags, including fetal Ags, which induce peripheral and central T cell tolerance through T cell deletion and conversion to Tregs.²¹

Our understanding of the mechanism(s) by which the feto-placental unit delivers Ags and immunoregulatory signals locally and systemically to maternal immune cells in the decidua and lymphoid organs is inadequate. During the past two decades, it has become evident that horizontal, cell-to-cell transfer of extracellular vesicles (EVs) is a component in fetal-placental-maternal crosstalk. These EVs may mediate the feto-maternal exchange of immune regulatory proteins, Ags, mRNAs, non-coding RNAs, and lipids. Accumulating evidence reveals that EVs traffic between the feto-placental unit and the maternal regional and distal immune cells, implicating the presence of a unique, EV-based communication system that contributes to maternal immune modulation during pregnancy (Fig. 1). In this review we center on EV cargo and function in the context of gestational immunology. The science of EVs is young and rapidly developing, and thus, many of the research trajectories detailed here should be considered pieces within an expanding mosaic. Most cited publications are from the past 10–15 years, with information centering on pregnancy-relevant EV types, their cargo, trafficking patterns, EV-based immune suppression and activation functions, and the potential roles of EV-mediated immune responses in major obstetrical disorders. While we focus on EVs in the context of the biology of pregnancy, we avoid detailed analyses of the use of EVs as biomarkers of health and disease.²²

Figure 1: Fetal-placental-maternal trafficking of EVs with potential immune functions in vivo.

The red arrows depict suggested feto-placental trafficking to the maternal decidua and uterine tissues, blood, secondary lymphoid tissues (uterus-draining lymph nodes, spleen), systemic endothelial cells, and distant organs. The blue arrows depict suggested trafficking routes of maternal EVs to feto-placental tissues. The inset depicts the routes of potential bi-directional EV passage at the maternal-placental-fetal interface.

2 |. PREGNANCY-RELATED EVs: KEY PLAYERS

The family of EVs encompasses diverse phospholipid membrane-enclosed vesicles of different biogenesis pathways, size, membrane composition and intraluminal cargo. They are released by most, if not all, prokaryotic and eukaryotic cells, including plant cells. The most studied and best characterized EVs are bona fide exosomes, microvesicles (MVs) and apoptosis-related EVs. Recent technological advances enabled a greater resolution of EV and extracellular micro/nano particles (e.g., ectosomes, exomeres, supermeres).^23–26 These are not included in this text as the data about their presence and function during pregnancy are scant.

Exosomes are small EVs (sEVs) of 40–150 nm in size that originate from the cellular endocytic compartment by reverse budding of the limiting membrane of early or recycling endosomes.^27,28 The resulting membrane invaginations are severed and released as intraluminal vesicles (ILVs) within the endosome, which becomes a multivesicular body (MVB). The MVBs can fuse their limiting membrane with the cell surface membrane, leading to the release of the ILVs as exosomes to the extracellular milieu in vitro or to bodily extracellular fluids in vivo.^27,28 Alternatively, MVBs can fuse with lysosomes, where the ILVs are degraded.^27,28 The ILVs can also escape from the MVB lumen by undergoing back-fusion (retrofusion) with the MVB limiting membrane.²⁹ Besides the bona fide exosomes generated within MVBs, the heterogeneous family of sEVs includes other subpopulations of small vesicles that remain poorly characterized, with some formed directly at the plasma membrane.²⁶ This latter subtype was recently termed ectosomes, distinct from MVB exosomes.²⁵ To avoid confusion and reflecting the state-of-the-art in the field of gestational EVs, we collectively refer to these nanovesicles as sEVs.

MVs consist of small and large EVs ranging between 0.1–1 μm in size. They are shed from cells by budding directly from the plasma membrane. Older, lower-resolution isolation protocols tended to cluster these MVs as microparticles (MPs). Therefore, unless clearly defined, we commonly refer to these vesicles as MVs/MPs. Lastly, cells undergoing apoptosis generate (a) apoptotic blebs (0.1–1 μm in size) through evagination and pinching-off from the cell membrane, (b) apoptotic bodies (1–5 μm in diameter), resulting from the cell breakdown process, (c) EVs from fragmentation of beaded apoptopodia (< 1 μm in size), and (d) apoptotic cell bodies that do not undergo further disintegration and contain cell nucleus remnants.³⁰ Interestingly, during the initial stages of apoptosis, early apoptotic cells generate MVBs and release EVs with size and protein composition that resemble sEVs, hence termed apoptotic exosome-like vesicles.³¹ For the history of the EV terminology, see Couch, et al.³²

Pregnancy-related EVs, including placental EVs or EVs from other uterine sources that circulate in the maternal blood during pregnancy, have been shown to carry diverse types of immune regulatory molecules during normal and pathological pregnancies. They interact with a broad range of immune cell types and have been implicated in immunosuppressive or immune-activating phenotypes (reviewed by Bai and others^33–35). sEVs are present in the maternal circulation as early as 6 weeks of pregnancy, with concentrations that increase across normal gestation to reach a level of 50-fold higher when compared to non-pregnant women. Their size distribution is relatively stable across pregnancy.^36–41 Mice also exhibit a similar increase in the concentration of plasma sEVs across pregnancy.^42,43

Placental trophoblasts release sEVs, MVs, and apoptosis-related EVs^44,45 Profiling circulating placenta-specific sEVs, which are positive for CD63 and for a placenta-specific marker (placental alkaline phosphatase, PLAP), Salomon, et al, reported a gradual increase in the concertation of circulating placental sEVs in the second and third trimester of pregnancy when compared to the first trimester.^36,46 Interestingly, while the release of placental sEVs increased across pregnancy, the fraction of placental sEVs relative to the total circulating sEVs in human pregnancy remained stable through the first and second trimester of pregnancy (10–20%), and declined in the third trimester.^36,47

As we review the immunological aspects of EV cargo and function, it is essential to highlight principal sources of variability among studies, representing not only the evolving definition of EVs, as highlighted above, but also the source of biological material used for EV isolation (e.g., maternal blood, placental perfusate, explants, or cultured cells); EV purification method and purity; and the technology used to assess cargo and function.^48,49 The commonly interrogated protein cargo has been intensely analyzed using antibody-based detection methods (e.g., western blotting, immunocapture and immunodetection, immuno-electron microscopy, and mass spectrometry-based proteomics).⁵⁰ New biosensor-based high-throughput methodologies may minimize sample processing and improve sensitivity.^51–53

3 |. EV IMMUNOMODULATORY CARGO

The first proteomic analyses of sEVs by trypsin digestion and mass-spectrometry were carried out using a mouse DC line,^54,55 and established that, while some sEV cargo molecules are constitutively present and shared among sEV subtypes, other molecules are more restricted, and their repertoire is defined by their cell origin, differentiation state, and other phenotypic parameters. These and subsequent studies (reviewed by Choi⁵⁶) revealed that cargo proteins can be integral, glycosylphosphatidylinositol (GPI)-anchored or membrane-bound proteins or be carried within the vesicle lumen. EV membrane proteins are commonly found in the endocytic compartment membranes or in the parent cell’s plasma membrane. Relevant examples of EV membrane proteins are Ag-presenting molecules (MHC class-I and -II), tetraspanins (CD9, CD63, CD81), membrane transport proteins (flotillin-1), membrane receptors (transferrin receptor, epidermal growth factor receptor), T cell stimulatory or inhibitory molecules (CD86, PD-L1), apoptosis-inducing molecules (membrane-bound Fas-ligand (Fas-L) and tumor necrosis factor α (TNF-α)).⁵⁵ EV lumen proteins include cytosolic proteins or proteins associated with the cytosolic leaflet of cell membranes of the parent cells. Examples include intracellular transport proteins, the Rab GTPase family of proteins (Rab 7, Rab 11, Rab 27a), interleukin-1β (IL-1β) ribonucleoproteins, and proteins involved in signal transduction and exosome biogenesis (Alix, Tsg101, syntenin-1).^55–57 Meticulous attention to EV separation and isolation techniques is critical for distinguishing among EV subtypes.^25,57–59

Notwithstanding the significant data variability among studies, analysis of the sEV signature of healthy human pregnancy commonly reveals an immunosuppressive phenotype.^60,61 For example, early (8–16 weeks) or term placental sEVs carry the NKG2D ligands MHC class-I chain (MIC)-related proteins and UL16 binding protein (ULBP) 1–5 on their surface. Other immunosuppressive proteins, such as B7H1 (CD274), B7H3 (CD276), B7H6, and HLA-G5, are released in sEVs from first trimester placental explants, term placentas, and cultured cytotrophoblasts.⁶² Placental sEV B7H6 was shown to promotes the activity of NK cells.⁶³

Other examples of how sEV cargo proteins may influence target NK cells, macrophages, or monocytes have been described elsewhere.^64–68 Conversely, macrophage sEVs stimulate cytokine (IL-1α, IL-1β, IL-6, IL-8, IL-10, TNF-α) production by placental explants^69,70 (and reviewed in Nair and Salomon^35,36).

HLA-G molecules released by extravillous trophoblastic sEVs execute their immunosuppressive function by binding to leukocyte immunoglobulin-like receptor subfamily B members 1 and 2 (LILRB1 and LILRB2), where LILRB1 is expressed on subpopulations of T cells, B cells, and NK cells, and both LILRB receptors are expressed in monocytes/macrophages and DCs,^71–73 (and reviewed in Rebmann, et al⁷⁴). Trophoblastic glycoprotein 5T4 (TPBG), an immunoregulatory protein of unclear function, is expressed in trophoblastic sEVs across pregnancy.⁷⁵ High-throughput proteomic screens also identified a broad range of immune modulatory proteins within sEV cargo.^46,49,76 See section 5 for a description of the function of EV immunoregulatory proteins during pregnancy.

Since the original finding of mRNA and miRNA in sEVs and the trafficking of these cargo molecules to target cells,⁷⁷ diverse forms of sEV RNAs have been described under normal and pathological conditions.^78–82 The RNA molecules encapsulated within EVs are protected from degradation by RNases.^77,83 While the sEV RNA cargo commonly reflects the concentration of RNA within donor cells, enrichment of RNA and miRNA within sEVs has been described, with selective sorting and enrichment mediated by specific sEV RNA binding proteins.^80,84–88 Different types of RNA molecules are packaged within pregnancy-related EVs. Murine trophoblast stem cells and differentiated trophoblasts exhibit a set of pregnancy-specific miRNAs and many other types of small RNAs implicated in physiological and immune maternal-placental-fetal interactions.⁸⁹ Among human placental sEVs, these include numerous types of miRNAs and other short RNA species.^90–92 We showed that trophoblastic sEVs are enriched for miRNA species expressed from the placenta-specific chromosome 19 miRNA cluster (C19MC), and that the concentration of these C19MC miRNA in EVs is proportional to their concentration in trophoblasts.^93,94

Immunoreactive phospholipids comprise EV membranes and may also regulate immune responses in endothelial cells and target tissues.^44,95 Comprehensive information on the protein, nucleic acid, and lipid composition of different types of EVs released by a variety of mammalian and non-mammalian cell types is available at the free Web-based integrated databases Vesiclepedia (http://microvesicles.org)⁹⁶ and ExoCarta (http://www.exocarta.org).⁹⁷

Research into the MVs in pregnancy has been hampered by the less standardized technologies used to isolate, purify, and therefore, define EVs and, particularly, by the unclear distinction between MVs and MPs, as noted earlier.^35,44,76,98 Nonetheless, MVs/MPs harbor numerous proteins, many of which are capable of immunomodulatory function.⁹⁹ Placental MVs/MPs from explant cultures activate human monocytes in vitro by increasing expression of the adhesion molecule CD54 (ICAM-1) and release of the pro-inflammatory cytokines IL-1β, IL-6 and IL-8.^100,101 MVs also carry RNA molecules, including miRNA, which, as in sEVs, exhibit a correlation between their levels in MVs and in the trophoblasts that release them. Placental MVs/MPs have also been reported to carry DNA molecules,¹⁰² some of which may be of mitochondrial origin and may have an immunostimulatory effect.^95,103 Lastly, immunoregulatory phospholipids are also present in the membranes of MVs/MPs.^44,104

4 |. PLACENTAL-MATERNAL-FETAL EV TRAFFICKING PATTERNS DURING PREGNANCY

Horizontal transfer of sEVs and their cargo among cells, and particularly the passage of mRNAs, and miRNAs and the subsequent post-transcriptional regulation of gene expression in the acceptor cells, has been previously documented.^77,80,81,105 This biological effect requires the transfer of the transported RNAs to proper cytosolic sites in the acceptor cells. Membrane fusion assays suggest that sEVs are first internalized and then likely fuse with the membrane of the phagocytic vacuole.^81,106 Passage of the sEV content to the cytoplasm of the acceptor cells was originally demonstrated with content-mixing assays in which target cells expressing luciferase in the cytosol were incubated with sEVs containing luciferin in their lumen.⁸¹ In HeLa cells, 20–30% of the internalized sEVs release their cargo into the acceptor cell cytosol, a phenomenon triggered by endosomal acidification.¹⁰⁷ Recent studies demonstrated the transport of trophoblastic sEV cargo to RNA processing P-bodies in recipient cells.¹⁰⁸

Intercellular communication has long been recognized as central to the establishment of mammalian pregnancy. Recent data suggest that EVs play a key role in these complex processes.^109–111 EVs that contribute to bidirectional maternal-embryo communication during preimplantation/implantation may emanate from the blastocyst’s inner cell mass and trophectoderm and from the maternal oviduct, endometrial lining, endometrial stroma, and more distant tissues.^{109,112–121} Indeed, local bidirectional EV-mediated communication between blastocyst cells and recipient endometrium is believed to modulate maternal immune response and to prepare the trophectoderm cells for interaction with the endometrial epithelium.^122,123

Once pregnancy is established, EV traffic constitutes a bidirectional mechanism of feto-placental-maternal communication (Fig. 1). Propelled by blood and interstitial fluid movement, placental EVs can access the decidua and myometrium, where the EVs may interact locally with maternal leukocytes and decidual cells. Alternatively, placental EVs are released to the maternal circulation via the placental intervillous space or via the spiral arteries, which are lined by extravillous trophoblast. Once in the maternal blood, uteroplacental-fetal EVs that are not cleared from circulation can passively traffic to regional or distant maternal lymphoid tissues and systemic organs (Fig. 1).

Although the data are sparse, initial observations on the traffic of sEVs among the maternal, placental, and fetal compartments have recently emerged. Using lipophilic, fluorescently labeled dye, Tong, et al, injected sEVs from first trimester human trophoblasts into the tail veins of E12.5 mice and, within 24 h, found them to be distributed mainly to the lungs, liver, and kidneys.¹²⁴ Targeting of placental sEVs to the mother’s lung interstitial macrophages and Kupffer cells in mice is mediated through α₃/β₁ integrins, present on the vesicle surface.¹²⁵ MVs, obtained from comparable conditions, accumulated in the maternal lungs.¹²⁶ Using genetically labeled, fluorescent sEVs, feto-placental sEVs were detected in the maternal uterus and cervix.⁴³ This phenomenon was associated with increased NF-kB activation, elevated tissue content of TNF-α and IL-6, and cervix infiltration by macrophages.

Passage of maternal EVs to the feto-placental compartment has also been reported.^127,128 Fetus-to-placenta traffic was also inferred from studies showing the traffic of labor-associated umbilical artery sEVs and their miRNA cargo to placental cells in a manner that stimulated cytokine production and inflammation.¹²⁹ Some of these bidirectional EV traffic data were also provided using miRNA as a surrogate for EV biodistribution.¹³⁰ Together, although it is plausible that the communication of EVs in vivo mediates biological effects on the immune systems, maternal endothelium, and blood vessels, additional data are needed to establish the in vivo EV traffic patterns to the fetal, placental, or maternal immune cell types or tissues during pregnancy. Specifically, the pathways that mediate EV passage through the villous placental barrier (trophoblast, basement membrane, and endothelial cells) remain to be identified.

5 |. EVs AS IMMUNE FUNCTION MODULATORS DURING PREGNANCY

In certain physiological and pathological conditions, EVs released by leukocytes, non-hematopoietic cells, allografts, tumor cells, and microbes exert immune-stimulatory or suppressive effects on cells of the innate and adaptive immune systems.^131–134 These effects are mediated through horizontal transfer of pro-inflammatory or anti-inflammatory mediators to acceptor leukocytes or non-immune cells. These mediators include Ags, enzymes, membrane receptors, cytolytic molecules, T cell co-stimulatory or inhibitory ligands, mRNAs and non-coding RNAs.^28,135 EVs also interact with the complement and coagulation systems, which may affect immune responses.^136,137 EVs released by leukocytes or cells of non-hematopoietic lineages have been shown to affect activation of immune cells, as wel as their polarization, viability, tissue recruitment, cytotoxic function, cytokine and chemokine secretion, enzymatic activity, and Ag-presentation capability.^28,138,139

The impact of EVs on the innate and adaptive immune responses have been reviewed elsewhere^28,136–141 and will not be addressed in this text. Here, we focus exclusively on the immunoregulatory role of EVs during pregnancy. Notably, most knowledge on EV effects on the immune system during pregnancy emanates from in vitro or ex vivo experiments or following administration of exogenous EVs in mouse models. The development of animal models to track the fate of endogenous EVs and to selectively deplete subsets of EVs in vivo will be essential in order to conclusively establish the biological relevance of immune functions attributed to EVs.

5.1. Gestational EVs and immune suppression

Placenta-derived EVs have a multi-pronged leukocyte-suppressive effect in vitro on the innate and adaptive immune system.^{33,35,61,95,142,143} Ligands of the activating human NK cell receptor NKG2D, including the MIC proteins A and B and ULBPs are released in sEVs by human syncytiotrophoblasts.^144,145 The EVs bearing NKG2D ligands induce immune downregulation by internalization of the NKG2D receptor on NK cells, CD8 T cells, and γδ T cells, decreasing their cytotoxic function in vitro without altering their content of perforin.^144,145 This effect of trophoblastic EVs may explain the mechanisms underlying placental protection from attack by maternal cytotoxic leukocytes.¹⁴⁵ Interestingly, some of these EV-mediated immune escape pathways have been described in cancer cells.¹⁴⁶ Considering the proximity of NK cells to the extravillous trophoblasts and spiral arteries, and with their key role in implantation, trophoblast invasion, spiral artery remodeling and protection against microbial infections, it is not surprising that these cells were identified as pivotal targets of EVs.^5,147,148

Beginning in the first trimester of pregnancy, the human placenta constitutively releases sEVs that contain Fas-L (CD178) and TNF-α-related apoptosis-inducing ligand (TRAIL or CD253) and that were shown to efficiently induce apoptosis of human leukocytes, PBMCs, and T cell lines in vitro^37,149 (and reviewed in Mincheva-Nillson⁶¹). Both proteins are anchored to the sEV surface as highly bioactive membrane-bound trimers, which may protect them from degradation by matrix metalloproteinases. Ultrastructural analysis revealed that Fas-L and TRAIL are not expressed on the trophoblast plasma membrane, and that both molecules are sorted into ILVs in MVBs or secretory lysosomes located near the apical surface of syncytiotrophoblasts.^149–151 Interestingly, human trophoblast MVs also contain Fas-L, shown to mediate lymphocyte apoptosis.¹⁵¹ Circulating human placental sEVs or those isolated from human placental explants also carry the immunoregulatory molecule PD-L1 (CD274, B7-H1) on their surface. When incubated with human T cells in vitro, these sEVs reduce the expression of the signaling-transducing molecule CD3-ζ, a key component of the CD3-TCR complex.³⁷ sEVs isolated from human placental explants also express CD276 (B7H3), a B7-homologue that restrains T cell activation and proliferation.⁶² Another trophoblastic sEV molecule, Heat Shock Protein Family E (Hsp10) Member 1 (HSPE1) has been implicated in regulation of T cell expansion and differentiation of Treg cells, and this effect was abolished with sEVs from HSPE1 KO cells.^152,153

The fusogenic proteins Syncytin-1 and −2 (endogenous retrovirus (ERV) -W1 and ERV-FRD1, respectively) are endogenous retroviral proteins that are expressed on the membrane of trophoblastic sEVs and contain immunosuppressive domains.¹⁵⁴ The ectodomain of syncytin-1 reduces the production of the Th1 cytokines IFN-γ and TNF-α and of CXCL10 by whole human blood cells.¹⁵⁵ In vitro, trophoblastic sEV-associated syncytin-2 decreased the release of IFN-γ, IL-2, and TNF-α in activated PBMCs, and this effect was mediated by the Syn2 immunosuppressive domain.¹⁵⁴

These suppressive effects of placental EVs on T cell immunity could be one of the mechanisms by which women with multiple sclerosis develop fewer relapses during pregnancy.¹⁵⁶ In a mouse model of experimental autoimmune encephalomyelitis, pregnancy serum sEVs suppressed T cell activation, promoted the maturation of oligodendrocyte precursor cells and their migration to active CNS lesions, thus contributing to attenuation of the encephalomyelitis.¹⁵⁶

Several placental sEV RNA molecules were found to target NK cells.¹⁵⁷ The miRNA miR-141 was shown to suppress the proliferation of Jurkat T cells.¹⁵⁸ The C19MC member miRNA-519c, known to be carried within trophoblastic sEVs, silences phosphodiesterase 3B, a known stimulant of TNF-α.¹⁵⁹ Repeat exposure of the placenta to lipopolysaccharide (LPS) promotes endotoxin tolerance, which is at least partially mediated by transfer of sEV-packaged miRNA-519c.¹⁵⁹

Less well-defined cell debris, released from trophoblasts, may also contribute to immune modulation. For example, internalization of trophoblast debris (resembling syncytial knots) shed from normal placental villous explants decreased the expression of HLA-DR, T cell co-stimulatory molecules, and pro-inflammatory Th1-biasing cytokines in human macrophages and increased the release or expression of immune-suppressive molecules, including IL-10, IL-1Rα, and indoleamine 2,3-dioxygenase.¹⁶⁰

Other uterine cell types release sEVs that modulate immunity. Pathway analysis by proteomics and RNAseq of sEVs from amnion epithelial cells reveal numerous molecules that participate in immunomodulation.¹⁶¹ For example, amniotic fluid sEVs, obtained from mid-trimester pregnancy, contain the immune modulators HSP72 and HSC73.¹⁶² Amniotic fluid stem cell sEVs can suppress PBMC function.¹⁶³

5.2. Gestational EVs and immune activation

Normal pregnancy is considered a mild pro-inflammatory state that co-exists with a systemic shift towards a type 2–biased and Treg-adaptive immune function that results in a state of fetal immune tolerance via multiple mechanisms.^4–7,164 Indeed, decidualization, embryo implantation, trophoblast invasion, and the parturition-associated uterine changes take place within a sterile pro-inflammatory local environment. These observations led to the hypothesis that, during certain stages of normal pregnancy or under pathological circumstances, fetal-derived EVs may promote an immune disbalance, with enhanced immune response.

Trophoblast cell line–derived sEVs are chemoattractant for human monocytes in vitro, mimicking a phenomenon that occurs during embryo implantation and trophoblast invasion.⁶⁷ Human placental explant MVs/MPs activate peripheral neutrophils and trigger generation of neutrophil extracellular traps in an IL-8–independent manner, which is more pronounced at the intervillous space of the placentas of preeclamptic women,¹⁶⁵ and sEVs from second trimester human trophoblasts increase the transcription and release of NFkB targets (IL-8, IL-6 and other cytokines) from decidualized endometrial stromal fibroblasts.¹⁶⁶ Lastly, MVs isolated from human term trophoblast cultures bind to PBMCs and peripheral B cells in vitro and induce the release of TNF-α, IL-1α, IL-1β, IL-6, IL-8, and the chemokine CCL3, mostly by monocytes of non-pregnant and pregnant women.¹⁶⁷

The syncytiotrophoblast is also a target of EVs of non-placental origin that control placental immune function. sEVs from fetal cord arterial blood from women with spontaneous labor at term have a relative high content of miR-15b-5p that targets Apelin mRNA, which encodes for a peptide that inhibits inflammation and uterine contractility.¹²⁹ Thus, when internalized by primary human trophoblasts in vitro, miR-15b-5b augments the release of IL-1, IL-6, IL-8, and TNF-α, implicating this phenomenon in the fetal pro-inflammatory signals that may trigger labor.¹²⁹

Unlike the immune-inhibitory functions attributed to syncytin-1 in sEVs, the packaging of syncytin-1 within MVs from early and late pregnancy trophoblasts was shown to activate PBMCs,¹⁶⁸ which may be related to differential processing of sEV- vs MV-carried molecules. Syncytiotrophoblast MVs/MPs also cause the release of TNF-α, IL-18, IL-12, and IFN-γ from leukocytes.³⁸

While the mechanism of action and targets of C19MC miRNA in vivo remain to be identified, the addition of C19MC miRNA–containing sEVs attenuated viral replication in diverse target cells in vitro, a process mediated, at least in part, by autophagy,¹⁶⁹ (and reviewed by Sadovsky, et al¹⁷⁰). Further, discrete C19MC miRNA members that are harbored in trophoblastic EVs, such as miR-517a-3p, target Protein Kinase cGMP-Dependent 1 (PRKG1) in Jurkat T cells and NK cells.¹⁷¹ As PRKG1 suppresses T cell activation and proliferation, the net effect of miR-517a-3p would lead to activation of these immune cells. Finally, amnion cells also participate in immunostimulation.¹⁶¹ For example, amnion epithelial cell-derived sEVs enhance the release of IL-6 and IL-8 from myometrial and decidual cells.¹⁷²

5.3. EVs as Ag-presenting and Ag-transporting vesicles: the distinctive case of placenta-derived EVs

The initial observation that sEVs released by Epstein-Barr virus–transformed human B cells stimulate CD4 T cell clones suggested that sEVs released by APCs may function as Ag-presenting vesicles.¹⁷³ Indeed, APC-derived sEVs carry, on their surfaces, MHC class-I and -II molecules, along with T cell co-stimulatory and adhesion molecules. These immune molecules are topologically orientated such that their binding domains face outward, as they are positioned in the cells of origin. In vitro, as free-floating vesicles, APC-derived EVs stimulate T cell clones, lines, hybrids, and primed T cells, but weakly trigger the activation and proliferation of naïve T cells, which require a higher level of T cell receptor cross-linking and co-stimulation than Ag-experienced T cells.^28,139,174 The poor stimulatory ability of free-floating sEVs for naïve T cells in vitro is likely due to the small size of sEVs, low content of MHC molecules per vesicle, and dispersion of sEVs by Brownian motion. Indeed, when APC-derived sEVs are immobilized at high concentrations on latex beads, are retained in clusters on circumscribed areas on the surface of APCs, or alternatively, when the number of MHC-peptide complexes per sEV is higher through direct peptide loading, the APC-derived sEVs markedly increase their T cell stimulatory potential in vitro.²⁸ The ability of sEVs to stimulate T cells also depends on the stage of activation of the parent APCs. sEVs released by LPS-stimulated DCs carry a higher content of MHC class-II, CD86, and CD54, and exhibit a higher T cell stimulatory capacity than sEVs released by immature DCs.^81,175 They can even transfer their ability to activate naïve T cells to non-professional APCs.¹³⁵ Indeed, different types of EVs released by immature human DCs differ in their ability to bias CD4 T cell polarization in vitro.¹⁷⁴

sEVs can also function as a source of tumor, microbial, or allogeneic Ags for APCs. The Ag can be delivered in its native form, carried in the sEV membrane or lumen, or as Ag-derived peptides, loaded in the sEV MHC class-I or -II molecules.^28,138,139 After the sEVs are internalized by the acceptor APCs, the EV-borne native Ags are processed into peptides and the Ag-peptides carried in the sEV MHC molecules are transferred to new MHC molecules synthesized by the acceptor APCs for subsequent presentation to T cells.^28,139 When sEVs interact with APCs, the vesicles can be internalized or remain attached on the acceptor cell surface. Immature DCs are more efficient at internalizing sEVs than LPS-mature DCs, whereas the latter retain more sEVs on the cell surface.⁸¹ Thus, the transfer of MHC molecules through sEVs that remain bound to the acceptor APC surface but without reprocessing by the acceptor APCs, a process known as MHC cross-dressing, serves as another mechanism of intact MHC Ag presentation by the recipient’s APCs to directly alloreactive T cells.^139,140 This mechanism is known to play a key role in transplantation.¹⁴⁰ Through MHC cross-dressing via EVs, relatively few cells can transfer sEV-bearing MHC-peptide complexes to a much higher number of acceptor APCs. There is indirect evidence that this phenomenon may also amplify the T cell response against pathogens and cancer cells, during allergies or following immunization with cell-based vaccines.¹⁴⁰

Unlike EVs released by APCs and most cells in humans, EVs released by the human syncytiotrophoblasts are devoid of the classical, highly polymorphic MHC molecules, reflecting the notion that these Ag-presenting molecules are not expressed by the parent cells. Similarly, the extravillous trophoblasts that invade the decidua only express the classical MHC class-I molecule HLA-C and the non-classical non- (or low) polymorphic MHC class-I molecules HLA-G, HLA-E, and HLA-F. Interestingly, first trimester human placental explants release EVs expressing HLA-G, and HLA-G–positive EVs have been isolated from peripheral blood of pregnant women.^62,102,176 HLA-G is a ligand for the NK cell Ig-like Receptor (KIR) 2DL4 (KIR2DL4 or CD158d) and leukocyte Ig-like receptors ILT2 (CD85j) and ILT4 (CD85d), which are immune-suppressive receptors expressed by NK cells and macrophages.¹⁷⁷ KIR2DL4, the main HLA-G receptor expressed by decidual NK cells, is located intracellularly in the endosome membrane and signals after binding to internalized HLA-G.¹⁷⁷ Confocal microscopy analysis revealed that NK cells acquire HLA-G from extravillous trophoblasts via trogocytosis, followed by HLA-G endocytosis.¹⁷⁸ This suppresses the NK cytotoxic function by preventing the polarization of cytolytic granules, an inhibitory effect that can be reversed by NK cell activation by cytokines or viral products.¹⁷⁸ Thus, endocytosis of trophoblast-derived, HLA-G–bearing EVs by decidual NK cells could provide an alternative source of soluble HLA-G for signaling to KIR2DL4 in endosomes. Interestingly, EVs released by human amnion epithelial cells also express HLA-G and mediate suppression of T cell proliferation in vitro.¹⁷⁹

Despite the absence of classic polymorphic MHC molecules on the EV surface, placental EVs carry paternal minor histocompatibility Ags that can be processed and presented by maternal APCs, in the context of maternal MHC molecules, to T cells in the uterus-draining secondary lymphoid tissues (SLTs).¹⁸⁰ Placental EVs may also carry trophoblast-specific Ags, against which maternal lymphocytes may have not been tolerized.

5.4. Feto-placental EVs as information carriers to maternal SLTs: a lesson from transplantation

One of the central questions in the immunology of pregnancy is how feto-placental Ags are delivered to maternal lymphoid organs in a manner that allows their identification by the maternal leukocytes, and whether or not this mechanism promotes recognition of the feto-placental Ags in a pro-tolerogenic or immunogenic fashion. The prevailing hypothesis assumes that feto-placental cells circulate through the maternal blood to the maternal lymphoid organs, which is analogous to the way that donor passenger leukocytes (in particular graft-resident DCs), transplanted with organ or tissue allografts, traffic via blood or lymph to the graft-draining SLTs.

During pregnancy, antigenic feto-placental cells are relatively rare and commonly undetectable in the maternal blood or lymphoid organs.^181,182 Moreover, fetal DCs, the counterparts of donor passenger leukocytes in allografts, do not develop in the fetus until late in pregnancy,¹⁸³ and maternal decidual DCs do not contribute to presentation of fetal Ags to maternal SLTs because they are confined to the pregnant uterus.¹⁸⁴ Alternatively, feto-placental Ags could be shed as cell-free Ags in the form of EVs released by the fetal trophoblasts into the maternal circulation (Fig. 2). Consistent with this notion, T cell recognition of paternal allopeptides in mice begins when trophoblasts start invading the maternal decidual arterioles.¹⁴ The human trophoblasts shed various types of EVs and nuclear aggregates into the maternal blood during normal and pathological pregnancies. While the larger particles, such as syncytial knots, are mostly trapped in the pulmonary capillary bed,⁴⁵ the smaller EVs, including sEVs and MVs, circulate systemically and have access to the maternal SLTs (Fig. 1). Interestingly, trophoblast EVs shed from the placenta express paternally inherited minor histocompatibility Ags,¹⁸⁰ and placental EVs have been shown to regulate the function of immune cells in vitro.^{33,35,61,95,142,143}

Figure 2: Potential interactions between feto-placental EVs and maternal decidual leukocytes secondary lymphoid tissues.

Fetal tissues are noted in red shades, and maternal tissues in blue shades. The figure bottom compares mechanisms of allorecognition in the field of transplantation (gray shaded area) vs pregnancy, highlighting current knowledge. In mouse gestation, DCs develop relatively late, and a fairly low number of maternal leukocytes traffic to the fetal tissues. Thus, maternal T cells are unable to recognize allo-MHC molecules on fetal DCs mobilized to the mother’s SLTs (via the direct allorecognition pathway). Furthermore, because trophoblastic EVs do not express classic (highly polymorphic) MHC molecules on their surface, they are not recognized by directly alloreactive T cells on the surface of maternal APCs, when cross-dressed with the EVs (via the semi-direct allorecognition pathway, gray shaded area). Instead, feto-placental EVs may interact with maternal decidual uNK cells, macrophages, T cells, blood vessel cells, and decidual stromal cells, or traffic through peripheral blood or lymph to the maternal SLTs (spleen, uterus-draining lymph nodes), where the EVs are internalized and processed by maternal APCs (bottom blue shaded area.) In mice, fetus-derived antigenic peptides are presented within maternal MHC molecules by maternal APCs (likely conventional DCs) to T cells via the canonical pathway (also known, in transplantation, as the indirect pathway.) Feto-placental EVs could also be presented to maternal B cells by subcapsular sinus (SCS) macrophages in lymph nodes (or analogue macrophages in the spleen) of by follicular DCs located deep in B cell follicles.

Until recently, the prevailing hypothesis in transplantation was that recognition of donor Ags by T and B cells in the recipient’s SLTs required migration of donor passenger leukocytes from the graft to the draining lymph nodes or to the spleen. However, recent studies in mice revealed that, after transplantation of skin, heart, or pancreatic islet allografts, donor passenger leukocytes are undetectable or found in very low numbers in graft-draining SLTs, as occurs with fetal chimeric leukocytes in maternal SLTs during pregnancy.^185–187 However, we and others have shown that mouse allografts shed donor Ags in a cell-free format by releasing graft-derived sEVs that reach the recipient SLTs via lymphatic or blood circulation.^186–188 Once in the recipient SLTs, the graft sEVs carrying donor MHC molecules bind to recipient APCs.^186–188 The donor MHC molecules, expressed by the graft EVs and attached to the recipient APCs, are then recognized by directly alloreactive T cells (Fig. 2). Alternatively, the graft EVs are internalized and processed by the recipient APCs and the resulting allopeptides presented within the recipient MHC molecules to T cells via the canonical pathway, also known in transplantation as indirect allorecognition (Fig. 2).^185,186 Donor-reactive B cells also recognize donor intact MHC molecules on the surface of graft-derived EVs mobilized to the draining lymph nodes (Fig. 2).¹⁸⁸ Thus, in an analogous fashion, EV release by feto-placental cells could be a cell-free mechanism by which paternally inherited Ags and immunoregulatory mediators are dispersed in the maternal circulation, where the EVs are captured by maternal APCs in lymphoid organs for presentation to T and B cells (Fig. 2). Yet, unlike graft-derived EVs after transplantation, during a normal pregnancy the feto-placental EVs are somehow presented in a predominantly suppressive fashion to induce and maintain the systemic fetal pro-tolerogenic state.

5.5. EVs and maternal microchimerism

In placental mammals, feto-maternal cell microchimerism can be bidirectional. Feto-placental cells can migrate into the maternal blood circulation, particularly during labor and in complicated pregnancies, and transiently colonize maternal tissues. Conversely, a relatively low number of maternal cells can traffic to the fetal tissues and remain in the offspring after birth, a phenomenon known as maternal microchimerism. These maternal cells express non-inherited maternal Ags (NIMAs), including those maternally derived allogeneic intact MHC molecules not genetically inherited by the offspring. It was recently shown that rare maternal leukocytes release sEVs bearing non-inherited maternal MHC class-I and -II Ags together with CD86.¹⁸⁹ These “immunogenic” EVs cross-dress the offspring DCs and stimulate direct CD4 T cell alloreactivity against the non-genetically inherited maternal MHC molecules, whereas CD4 T cells, which recognize allopeptides from the same MHC Ags in an offspring MHC–restricted manner, are silenced.¹⁸⁹ Thus, transfer of maternal chimeric cell–derived EVs to the offspring APCs provides an explanation of T cell split tolerance against NIMAs and its potential consequences for reproductive fitness.¹⁹⁰

6 |. EVs’ MODULATED IMMUNITY IN GESTATIONAL DISEASES

EVs have been implicated in physiological processes during pregnancy, including preimplantation and implantation biology, decidualization, the establishment of the placenta and placental invasion, and fetal-placental-maternal interaction as pregnancy advances. EVs also play a role in common disorders of pregnancy, either as a part of the pathogenesis or as defense against disease (reviewed by Hashimoto¹⁹¹).

6.1. The establishment of pregnancy and pregnancy loss

The presence of EVs in the blastocyst culture medium has been documented in several species.^122,192,193 Blastocyst EV cargo was shown to include HLA-G, which, as previously noted, has a role as an inhibitory molecule for NK cells and possibly macrophages, T cells, and other endometrial immune cells.^177,192 Annexin V-binding blastocyst EVs also harbor progesterone-induced blocking factor, which may stimulate the production of IL-10 and modulate the function of CD4 and CD8 T lymphocytes in mice.¹⁹⁴ A broad range of miRNAs also define the blastocyst EV cargo. The blastocyst EV miRNA landscape is likely influenced by the donor’s age,¹⁹⁵ with some cargo miRNA members, such as miR-155i, implicated in immune regulation in other systems.^196,197

On the maternal side, uterine endometrial EVs contain different cargo proteins^198,199 and a repertoire of miRNAs. The uptake, by the trophectoderm, of one sEV miRNA species, miR-30d, enhances blastocyst adhesion to the endometrium.²⁰⁰ The uptake of endometrial sEVs by a trophoblast cell line in vitro enhances focal adhesion kinase signaling.¹¹⁷ In the ewe, endometrial sEVs are released to the uterine lumen during the estrous cycle and early pregnancy. They contain proteins and mRNAs, including enJSRV-ENV, HSC70, and IFN-regulatory factors that stimulate the trophectoderm to proliferate and secrete IFN-τ, which then control Toll-like receptor (TLR)-mediated signaling.²⁰¹

The presence of EVs in uterine lumen fluids during the preimplantation and implantation periods has been documented, and the impact of EV cargo components on the establishment of pregnancy has been interrogated in several species, even though the source of the EVs is commonly unknown. In general, during implantation, MVs/MPs are believed to have pro-inflammatory, pro-coagulant and pro-immunogenic properties and contain tissue factor. An increase in the number of circulating MVs/MPs was found to be associated with increased risk of pregnancy loss.^202–204 A decline in EV CD9 levels has been suggested to signal impending pregnancy loss.²⁰⁵ Bovine sEV miR-499 attenuates local inflammation during the implantation process, as it reduces LPS-stimulated production of TNF-α and IL-6 by bovine endometrial epithelial cells, suggesting that disruption of bta-miR-499 might increase local inflammation and the risk of pregnancy failure.²⁰⁶ Bovine uterine lumen sEVs derived from the embryo contain IFN-τ, and the application of these sEVs to uterine epithelial cells increases the production of IFN-stimulated genes.²⁰⁷ Other uterine flushing EVs are implicated in modulating immune-related gene expression in endometrial cells during implantation²⁰⁸

6.2. Preeclampsia and fetal growth restriction

Preeclampsia affects 3–8% of all pregnancies and is a leading cause of pregnancy-related maternal mortality and morbidity worldwide.^209,210 Although the etiology of preeclampsia remains to be identified, immune-mediated mechanisms have been implicated in the pathogenesis of this disease, with an emerging role for EVs in disease development and progression.^{142,211–215}

Several groups have identified an increase in the concentration of sEVs in the plasma of women with preeclampsia.^91,216–218 Moreover, the repertoire of cargo molecules in circulating sEVs of women with preeclampsia also changes, with augmented Flt-1, IL-2, and TNF-α and reduced IL-10, suggesting an acquired pro-inflammatory, Th1-biased phenotype.^216,218 This is also accompanied by increased TRAIL and Fas-L in sEVs released by placenta explants from early-onset, but not late-onset preeclampsia.²¹⁹ Preeclampsia-related sEVs change the cytokine repertoire of the BeWo trophoblast line, with increased IL-8 and decreased IL-10.²²⁰. Similarly, the miRNA cargo repertoire of these sEVs is different.^217,221 Functionally, these changes have been associated with defective angiogenesis, altered vascular smooth muscle migration, increased vascular tone, and increased platelet activation.^124,222

MVs/MPs have been also implicated in the pathogenesis of preeclampsia. Their plasma levels are higher, particularly in cases of early-onset preeclampsia.^{218,223–226} While their cargo and mechanism of action have been less thoroughly investigated than those of sEVs, preeclampsia-related MVs/MPs contain more Flt-1 and activate endothelial cells and slow down their proliferation.^101,218,226 Preeclamptic MVs are pro-inflammatory / -immunogenic and increase secretion of TNF-α, IL-12, and IL-18 in PBMCs.^38,168,227 Pro-inflammatory changes and preeclampsia-like phenotype were recapitulated in female mice by injecting a mix of human endothelial cell- and platelet-derived mixed EVs during mid-pregnancy, which activated the inflammasome in trophoblasts, a process implicated in the pathogenesis of preeclampsia.²²⁸

Although the pathogenesis of fetal growth restriction is different from that of preeclampsia, many women with severe preeclampsia who were included in studies cited here also presented with fetal growth restriction. Data regarding immune-mediated mechanisms and the role of EVs in pregnancies complicated by fetal growth restriction without preeclampsia are scant. Isolated sEVs from the plasma of women with fetal growth restriction exhibit reduced levels of sEVs that carry Fas-L, leading to less T cell suppression, as observed in normal pregnancy.²²⁹ In women with fetal growth restriction, we found no change in the number or size of maternal plasma sEVs or their miRNA cargo.²¹⁷

6.3. Preterm birth

Whereas sEVs from pregnant women’s plasma harbor pro-inflammatory mediators even without labor, there was no difference in those mediators between women in labor or not in labor. The number of placental EVs was higher preterm labor than in term without labor, with some changes in the repertoire of EV proteins.²³⁰ Similarly, altered protein landscape but without discrete immune changes were found in amniotic fluid sEVs, and most changes were not specific to the labor paradigms.²³¹

In mice, maternal plasma sEVs revealed a pro-inflammatory proteomic phenotype before birth (E18), and labeled E18 mouse sEVs injected intraperitoneally trafficked to intrauterine tissue and activated an inflammatory signal (COX-2, NFκβ, TNF-α, and IL-6) in the cervix, with similar changes in the fetal membranes.^43,232 Similar changes were also recorded using human samples, with changes in the protein and miRNA landscape in plasma sEVs from preterm birth vs controls. These changes, particularly those based on analysis of the protein cargo, suggested an involvement of immune pathways^233–235 (and reviewed in Menon, et al²³⁶). Finally, circulating MV/MP proteins were also different between preterm birth and controls. These proteins have been implicated in immune functions, inflammation, and oxidative stress.^237,238

6.4. Gestational diabetes

Gestational diabetes mellitus (GDM), a disease of glucose intolerance that starts during pregnancy, affects up to 15% of pregnant women worldwide.²³⁹ EVs are believed to contribute to endothelial cell dysfunction in GDM.²⁴⁰ sEVs extracted from the plasma of pregnant women were more abundant than those of healthy controls. When injected into pregnant mice, they also induced insulin release and peripheral insulin resistance²⁴¹ and stimulated a pro-inflammatory phenotype (GM-CSF, IL-4, IL-6, IL-8, IFN-γ, and TNF-α) in recipient endothelial cells.²⁴² Some of these changes were recapitulated using sEVs harvested from trophoblast media exposed to a high glucose concentration, and applied to human umbilical vein endothelial cells (HUVECs). Consistent with this observation, glucose was found to increase the release of sEVs and stimulate cytokine release.²⁴³ GDM plasma sEVs also exhibit an altered miRNA repertoire^244,245 (and reviewed by James-Allan²⁴⁶).

6.5. Infection and inflammation

Pregnancy-related EVs have been implicated in modulation of diverse types of infections and inflammatory disorders. Obese pregnant women have more circulating sEVs in the plasma; when these sEVs were added to endothelial cells (HUVEC) cultures, they increased the release of IL-6, IL-8, and TNF-α, with no change in IL-10.²⁴⁷ The effect of pregnancy-derived mouse serum sEVs on T cell activation and oligodendrocyte precursor cell proliferation and related parameters was described earlier.¹⁵⁶ We showed that human trophoblast–specific C19MC miRNAs are packaged within trophoblast sEVs, and when applied to cultured non-placental cells, these miRNAs or the sEVs that harbor them enhance resistance to a range of DNA and RNA viruses.^{93,169,170,248,249} This effect was recently recapitulated using expressed C19MC member miR-517–3p against CMV infection in fibroblasts and trophoblast lines.²⁵⁰ Another C19MC member, miR-519c, attenuated LPS-induced inflammation in placental explants or cultured trophoblasts.¹⁵⁹ Infection of pregnant women by malaria increased miR-517C levels in their plasma MVs/MPs. Lastly, HIV infection increased placental MV/MP production, and circulating MVs/MPs from HIV-positive women attenuated the production of monocyte chemoattractant protein-1 (MCP-1) in DCs.²⁵¹

7 |. FINAL PERSPECTIVES, QUESTIONS, AND FUTURE DIRECTIONS

Despite growing interest, our understanding of the role of EVs in the immune modulation of pregnancy is rudimentary. Listed below are some of the central questions that remain to be addressed.

7.1. What type(s) of EVs regulate the maternal immune response against feto-placental Ags?

To date, most studies on the immunobiology of EVs during pregnancy have centered on the assessment of cargo, traffic, and possible functions of EVs isolated from diverse sources by different EV purification methods, which in most cases were not standardized among laboratories. As discussed in this review, multiple EV subtypes may co-exist in the same biological samples, and isolation technologies of different EV subpopulations from small sample volumes have been evolving. There is a need for more meticulous EV purification and characterization technologies in order to unravel unknown functions of EV subtypes in the biology of pregnancy and pathogenesis of pregnancy-related disorders. Clarifying this issue may also explain some of the inconsistencies among studies regarding the biological effects of EVs during pregnancy, likely reflecting the diverse EV purification techniques.

7.2. Is EV-mediated signaling relevant to the immunology of pregnant organisms?

Most of the accumulating data in the field of EVs and pregnancy were supported by in vitro or ex vivo experiments using cell culture, explants, organoids, or an organ perfusion system, with subsequent transfer of purified EVs to another in vitro system for testing the biological effect. Very few studies were performed in vivo. Therefore, it is still unknown how feto-placental EVs traffic to maternal tissues, and whether or not feto-placental EVs are captured by specific maternal cells, tissues, or organs. It is likely that trophoblasts, which are directly bathed in maternal blood in the hemochorial (human, mouse) placenta, can shed EVs directly into the maternal circulation. An opposite route from the maternal plasma to trophoblasts is equally direct, but trafficking through the trophoblast and basement membrane into the villous core or the fetal vessels might be more involved and remains to be rigorously validated.

7.3. How does the EV cargo regulate the maternal immune response during pregnancy?

Once feto-placental EVs reach the maternal immune cells during pregnancy, does modulation of the maternal immune response require (i) interaction of EV ligands with receptors on the acceptor cells, (ii) endocytic processing of the EVs, (iii) cross-dressing of the surface of the acceptor cells with EVs, and/or (iv) release of the EV content into the cytosol of the acceptor cells? Do proteins, RNAs, and lipids delivered by the feto-placental EVs directly or indirectly instigate the maternal immune response?

7.4. Does EV-mediated regulation of maternal immunity impact gestational diseases?

Numerous studies have shown an association between the cargo or concentration of EVs in maternal peripheral blood and gestational diseases. These important findings strongly suggest that EVs may regulate the maternal immune response and that EVs could be used as biomarkers in pregnancy-related disorders. However, data are needed to mechanistically link feto-placental or maternal EVs with gestational immunology. The trafficking of EVs to the mother’s primary and secondary lymphoid tissues and a putative role in immune tolerance during pregnancy remains to be elucidated.

7.5. Could feto-placental EVs be used for therapeutic purposes?

The findings indicating that feto-placental EVs may transfer molecules that downregulate innate and adaptive immunity and promote tissue regeneration, have led to the idea that these EVs might have therapeutic potential. One example is human amnion-derived epithelial cells, which have long been considered for therapeutic use in the field of regenerative medicine due to their proliferative and differentiation capacities and their accessibility.²⁵² It is possible that EVs produced by the amnion may be useful for similar applications, as recently suggested.²⁵³ Amnion epithelial cell sEVs were deployed to attenuate bleomycin-induced lung inflammatory injury, with action that involves increasing macrophage phagocytosis, reducing neutrophil myeloperoxidases, and suppressing T cell proliferation.¹⁶¹ Interestingly, in utero mouse exposure to allogeneic MHC exosomes reduced the reactivity of recipient lymphocytes to the allo-Ags, but did not confer allograft tolerance.²⁵⁴ Placental EVs harboring antiviral proteins have also been suggested for inhibition of viral infection.²⁵⁵ These exciting leads underscore the need for additional research before pregnancy-related EVs are used therapeutically.