Entry and exit of extracellular vesicles to and from the blood circulation

Abstract

Extracellular vesicles (EVs) are biological nanoparticles that promote intercellular communication by delivery of bioactive cargo over short and long distances. Short-distance communication takes place in the interstitium, while long-distance communication is thought to require transport through the blood circulation to reach distal sites. EV therapeutics are frequently injected systemically and diagnostic approaches often rely on the detection of organ-derived EVs in the blood. However, the mechanisms by which EVs enter and exit the circulation are poorly understood. Here, the lymphatic system and transport across the endothelial barrier through paracellular and transcellular routes are discussed as potential pathways for EV entry and exit to and from the blood circulatory system.

MAIN

Extracellular vesicles (EVs) are cell-released nanoparticles that mediate intercellular communication over short and long distances. Numerous studies report the presence of organ-specific EVs in the blood,¹ however, the mechanisms by which such EVs exit the interstitium remain largely unknown. Similarly, it has been demonstrated that systemically administered EVs can reach specific organs,^2,3 although transport across the endothelial barrier remains elusive. The exponential growth of the EV field and the accelerated pace at which emerging therapeutic and diagnostic EV-based products are being developed,^4–8 highlights a pressing need to understand EV transport in the body. Accordingly, it could be argued that EV transport phenomena are the most critical component of a desired function, as an incorrect spatial context is futile and potentially detrimental. This Perspective explores potential nanoscale processes that EVs use to enter and exit the blood circulation (Fig. 1).

Figure 1 -. Potential pathways of extracellular vesicle (EV) entry to and exit from the blood circulation.

a) General representation of endothelium types. It is worth noting that several types of endothelium can be found in some organs, such as the kidneys. b) EVs may cross the endothelial barrier through transcellular or paracellular transport routes. In the case of paracellular transport, EVs may pass through gaps between endothelial cells, which is likely to be a prominent transport route in organs with discontinuous vasculature, such as the liver (1). EVs may also modify interendothelial junctions to enable paracellular transport (2). In the case of transcellular transport, EVs may pass through intra-endothelial fenestrations (3) or use various transcytosis pathways (4). Binding to lipoproteins to hijack caveolae-mediated transcytosis may also be exploited (5). Exocytosis-independent mechanisms of transcellular transport are also plausible, resulting in multilayered EVs released through membrane budding (6). Finally, cell protrusions could potentially be exploited as an alternative route for EV transport across the endothelial barrier (7).

Most organs have capillaries with a non-fenestrated continuous endothelial lining, which only enables water, small solutes, and gasses to pass through. By contrast, organs such as the liver and spleen have a discontinuous endothelial lining, which allows macromolecules and nanosized particles to traverse between the interstitial space and the blood circulation (Fig. 1a). It is also common for various subtypes of endothelia to be present in the same organ. For example, the kidneys contain several endothelial cell types that possess distinct structures and roles.^9,10 The kidney endothelial cells found in medium and large blood vessels form a non-fenestrated continuous layer, interconnected through intercellular junctions, and elongated in the direction of the blood flow. By contrast, glomerular endothelial cells possess numerous fenestrations and have a thick glycocalyx layer, which contributes to the filtration properties of the glomerular filtration barrier. The endothelial cells of peritubular capillaries are fenestrated and covered by a thin diaphragm made of glycoproteins, which facilitates the reabsorption and secretion of fluids and substances by neighboring tubular epithelial cells.^9,10 Endothelial cell heterogeneity within and among organs is likely to have a substantial impact on the mechanisms by which EVs traverse the endothelial barrier, of which plausible routes will be discussed in this article.

PARACELLULAR TRANSPORT OF EXTRACELLULAR VESICLES

Transport across discontinuous endothelium

With pathological conditions, such as cancer and inflammation, interendothelial junctions in organs with continuous vasculature can become compromised, enabling paracellular transport of macromolecules and nanosized particles.¹¹ This phenomenon has been exploited in the field of nanomedicine, where one of the main targeting mechanism for clinically approved nanoparticles relies on paracellular endothelial transport due to pathological vascular aberrations.¹¹ Therefore, paracellular transport could represent a major mechanism by which EVs enter and exit the blood circulation during pathological conditions (Fig. 1b). Accordingly, the numerous observations that cancer patients have increased levels of circulating EVs,⁶ may be attributed to leaky tumor vasculature, in addition to the common explanation of elevated EV production by cancer cells. It is important to note that increased vascular permeability is consistently observed in preclinical tumor models, although substantial interpatient, intertumor, and intratumor heterogeneity in terms of leaky vasculature exists in the clinical oncology setting.¹² This discrepancy is likely due to substantial differences in tumour growth and tumour-to-body weight ratios in animal models and patients, which affects vasculature morphology and function, consequently impacting macromolecular transport across the endothelial barrier. It is also important to note that the central paradigm of paracellular transport being responsible for leaky tumour vasculature has been questioned with new evidence suggesting a critical role of transcellular pathways.¹³

Transport across continuous endothelium

It is also possible that EVs have mechanisms of altering interendothelial junctions to enable paracellular transport across non-fenestrated continuous capillaries (Fig. 1b). In the case of immune cells, transendothelial migration through paracellular transport is an extensively studied phenomenon, involving the relocalization of structural proteins associated with endothelial tight and adherent junctions.¹⁴ Rolling adhesion between immune cells and endothelial cells results in the activation of multiple downstream signaling pathways within the latter, including production of reactive oxygen species and tyrosine phosphorylation of junctional components.¹⁴ These pathways induce cytoskeletal actin rearrangements that disturb binding interactions, enabling efficient leukocyte transendothelial migration across the vasculature.¹⁴ Several metastatic cancer cell lines appear to exploit similar mechanisms to enable EVs to cross the endothelial barrier. Such EVs have been shown to contain microRNAs (miRNAs)¹⁵ and proteins¹⁶ that destabilize the endothelial actin cytoskeleton or alter the expression and/or localization of junctional proteins, increasing vascular permeability.

Studies have linked miRNA-mediated endothelial dysregulation to EV-promoted metastasis. For example, EV-associated miR-181c was shown to destabilize the continuous blood-brain barrier, promoting metastasis of brain-tropic cancer cells.¹⁷ Administration of brain-metastatic EVs in endothelial cell culture models resulted in the cytoplasmic internalization of tight junction proteins in conjunction with the remodeling of associated actin filaments (Fig. 2a).¹⁷ These effects were attributed to miR-181c-induced downregulation of phosphoinositide-dependent kinase-1 (PDPK1), a protein involved in the phosphorylation of cofilin,¹⁷ which disassembles actin filaments when dephosphorylated.¹⁸ Taken together, the impact of miRNA cargo on cytoskeletal actin dynamics extends beyond the entry and exit of EVs from the circulation to more wide-reaching effects on cross-endothelial transport of 100 times larger structures, that is, metastatic cancer cells.

Figure 2 – Mechanisms for cancer cell-derived EVs to cross the blood-brain barrier endothelium.

a) Cancer EV-associated miR181c modulates the integrity of the tight junction by downregulating the expression of PDPK1, a protein involved in the phosphorylation of cofilin.¹⁷ Increased cofilin dephosphorylation as a result of this process leads to eventual actin filament disassembly, displacing attached tight junction proteins from their transmembrane localities into the cytoplasm, causing increased vascular permeability for paracellular movement of EVs. b) Breast cancer-derived EVs have been found to suppress the RAB7-associated degradation pathway following endocytosis, redirecting themselves to the RAB11-associated recycling pathway.³⁶ This enables subsequent transcytosis of the EVs through the basolateral membrane of the endothelial cells, into the interstitium. mRNA; messenger RNA; RISC, RNA-induced silencing complex; P, phosphate group; RAB, Ras-associated binding (protein).

Non-cancer-derived EVs have also been found to modulate endothelial cell-to-cell junctions. For example, during inflammation, the surface of endothelial-derived EVs becomes enriched with adhesion molecules that mediate endothelial barrier disruption.¹⁹ However, it is unclear to what extent EV-mediated modifications of interendothelial junctions take place under pathological and physiological conditions. It is also worth noting that the glycocode is an essential component of leukocyte extravasation, as membrane-bound surface glycans mediate interactions with the endothelium and associated junctional complexes.²⁰ However, glycans are a major overlooked component of EVs, and few studies have assessed the EV glycocode.^21–25 An increased focus on EV glycomics could aid in understanding interactions that take place at the EV-endothelium interface.

LYMPHATIC TRANSPORT OF EXTRACELLULAR VESICLES

In organs with discontinuous vasculature, such as the liver, paracellular transport of nanosized cargo occurs through inter-endothelial gaps in the vasculature wall. In addition to having discontinuous vasculature, the liver is the largest internal organ. Therefore, it could be reasoned that the liver contributes to the majority of non-hematopoietic EVs in the blood. It is challenging to determine the organ-specific origin of EVs, as many markers overlap. A recent study based on long RNA sequencing, predicted that EVs of liver origin were the fourth most abundant population in the blood, excluding hematopoietic-derived EVs.²⁶ Adipose tissue, muscle, and lung-derived EVs outnumbered those of liver origin,²⁶ indicating that if these results are accurate, other mechanisms of entering the blood circulation may be more efficient than paracellular transport across the vascular endothelium. In fact, lymphatic transport represents an alternative pathway for eventual entry into the blood circulation.

The lymphatic endothelium has junctions that are opened in response to interstitial fluid pressure, making it easier for macromolecules to exit the interstitium through the lymphatic system compared to the non-fenestrated continuous vascular endothelial barrier. Similar to macromolecules, it is likely that EVs enter the lymphatic system through fluid pressure-regulated flap valves, which are known to allow the passage of nano and micrometre-sized particles (Fig. 3).^27,28 Lymphatic vessels also express a myriad of receptors, such as integrins, that promote trafficking of cells throughout the lymphatic vessels. EVs, as a result of possessing cell-derived ligands on their surfaces, could potentially exploit receptor binding for lymphatic entry and transport. Given that lymph vessels are omnipresent, it is likely that lymphatic entry of EVs from the interstitium occurs across all organs, and could serve as the primary mechanism through which EVs enter the circulation in adipose tissue, muscle, lungs, and other organs with a non-fenestrated continuous endothelium.

Figure 3 -. Potential pathways of EV entry into the blood circulation through the lymphatic system.

EVs may exit the interstitium through fluid pressure-sensitive junctions (flap valves) in the lymphatic endothelium that create micrometre-sized gaps^27,28 (1), subsequently entering the blood circulation, either by exiting through high endothelial venules (HEVs) in the lymph nodes (2) or through the lymphovenous junction that drains into the great veins (3).

The lymphatic system is connected with blood circulation through the lymphovenous junction and lymph nodes.²⁹ On a daily basis, an average of 8–12 litres of lymph enters the blood; half through lymph nodes and half through the lymphovenous junction that empties into the great veins.²⁹ In the case of the lymph nodes, specialized blood vessels termed high endothelial venules, form a vascular transport barrier for macromolecules. It is possible that EVs can cross this barrier through either paracellular or transcellular pathways, although evidence of either is currently lacking. However, in the case of the lymphovenous junction, the lymph enters directly into the blood circulation without passing a vascular endothelial barrier. Therefore, lymphatic transport from the tissue interstitium to the great veins could represent a major mechanism by which EVs enter the blood circulation (Fig. 3).

Cancer cell-derived EVs have been found in the lymphatic system, where they promote metastatic spread.³⁰ In particular, cancer EVs can contain ‘don’t eat me’ surface proteins,³¹ which are likely to enable immunoevasion in the lymph. Transport in the lymphatic system may also occur in the opposite direction, as is the case for platelet-derived EVs, which have been found in the lymph in healthy and pathological settings.³² The prevention of retrograde flow of blood components into the lymph at the lymphovenous junction is thought to involve valves and platelet plugs.³³ Beyond the context of EVs, many outstanding questions remain regarding the mechanisms governing lymph-blood separation at the lymphovenous junction.³³

TRANSCELLULAR TRANSPORT OF EXTRACELLULAR VESICLES

In addition to the lymphatic system and paracellular transport through the vessel wall, other mechanisms exist for macromolecules to enter the blood circulation from the tissue interstitium. In tissues with fenestrated continuous vasculature, intra-endothelial fenestrations enable macromolecules to traverse the endothelium through channels within individual endothelial cells. In organs with a non-fenestrated continuous endothelial lining, proteins and other macromolecular are restricted from crossing the vascular endothelium through intra or inter-endothelial gaps. Consequently, a common mechanism by which macromolecules, such as lipoproteins and albumin, are transported across the endothelium relies on vesicle-facilitated cellular internalization and transcellular transport mediated by caveolae.^34,35 Either receptor-bound or fluid-phase macromolecules within caveolae pits are internalized, and vesicles are transported across the cytoplasm to the opposite side of the endothelial cell where they are released extracellularly through fusion with the cell membrane.³⁴ Therefore, it is possible that EVs use similar pathways of endothelial internalization and transcytosis.

Extracellular vesicle uptake by endothelial cells

Studies have shown that endothelial cells can internalize EVs through clathrin-dependent endocytosis, caveolin-mediated uptake, macropinocytosis, and lipid raft-mediated internalization.^36–39 Phagocytosis has also been observed to facilitate EV uptake in endothelial cells.⁴⁰ The presence of phosphatidylserine in the outer leaflet of EV membranes appears to facilitate phagocytic and macropinocytic uptake by immune cells,^41,42 and increases endothelial cell internalization of EVs.⁴³ Similar to synthetic nanoparticles,^44,45 it is likely that the internalization mechanisms are also dependent on the physical properties of EVs. EV biogenesis has resulted in three distinct classifications: exosomes (~30–150 nm), microvesicles (~100–1000 nm), and apoptotic bodies (~500–2000 nm). The significant overlap in size and absence of subtype-specific surface markers amongst these categories make it challenging to distinguish between vesicle types.^46,47 This intrinsic heterogeneity of EVs could potentially play a role in the cellular internalization profiles, as has been observed in the case of synthetic nanoparticles. For example, larger nanoparticles tend to be internalized through phagocytosis or macropinocytosis, whereas smaller ones are endocytosed by other mechanisms.⁴⁸

Multiple internalization mechanisms can be engaged by endothelial cells when taking up EVs. Protein and glycan interactions between the two interfaces have been shown to be cardinal in EV uptake, with certain ligands on EV surfaces complementing membrane receptors on target cells and triggering internalization.^38,39 For example, tetraspanins play a pivotal role in endothelial cell uptake. EVs overexpressing the tetraspanin 8-cluster of differentiation (CD)49d were more readily internalized, and CD106/vascular adhesion molecule 1 (VCAM-1) on endothelial cells was speculated to mediate enhanced uptake.⁴⁹ In the context of brain endothelial cells, EVs expressing lymphocyte function-associated antigen 1 (LFA-1) displayed enhanced internalization through binding to intercellular adhesion molecule 1 (ICAM-1).⁵⁰ Membrane-bound C-type lectin receptors on endothelial cells have been shown to mediate the internalization of macrophage-derived EVs.⁵⁰ Heparan sulfate proteoglycans on endothelial surfaces also play a role in EV internalization.⁵¹ Heparan sulfate proteoglycans are ubiquitous on the luminal face of the endothelium, being the most common endothelial cell-surface glycosaminoglycans, comprising between 50–90% of this glycan population.⁵² Several ligands commonly found on EV surfaces, such as fibronectin,^53–55 are known to interact with heparan sulfate proteoglycans.⁵⁶ Additionally, treatment of EVs with integrin blocking peptides, reduced uptake in endothelial cells.⁴⁰

Extracellular vesicle transcytosis in endothelial cells

Following uptake by cells, vesicles can be subject to various fates: fusion with endosomes and subsequent recycling of lipid and lipid-associated biomolecules to the plasma membrane or other organelles, transport to the lysosomal compartment, or transcytosis into the extracellular space. Cancer-derived EVs can engage in transcytosis to cross the endothelium of the blood-brain barrier (Fig. 1b).³⁶ In such cases, EVs downregulate the expression of genes responsible for endocytic degradation, diverting to a recycling pathway. Specifically, a study showed that brain-targeting EVs from metastatic breast cancer cells engage in caveolin-independent uptake mechanisms and increase innately low levels of transcytosis (Fig. 2b).³⁶ Studies assessing the intracellular fate of the endocytosed EVs revealed co-localization with Ras-associated binding protein (RAB)11 and vesicle associated membrane protein 3 (VAMP3), markers for recycling endosomes and/or exocytosis.³⁶ The transcytosis pathway was further confirmed by fusion of the endosomes containing the EVs with receptors on the basolateral membrane. Taken together, EVs from metastatic cancer cells can be transported from the luminal (blood-facing) to abluminal (brain interstitium-facing) side of the endothelium by circumventing intracellular degradation pathways and increasing the physiologically low levels of EV transcytosis at the blood-brain barrier. It is possible that similar mechanisms are employed by EVs when crossing non-fenestrated continuous endothelium encountered in other organs, such as the medium and large blood vessels in the kidneys. Adsorptive-mediated transcytosis is another mechanism proposed for blood-brain barrier crossing of EVs.^57,58 This mechanism relies on cationic moieties, which augment vesicular trafficking resulting in transcytotic release.⁵⁹ It is possible that cationic components are incorporated during biogenesis or after EV release as part of a biomolecular corona, enabling endothelial crossing through electrostatic interactions. Adsorptive-mediated transcytosis occurs with other macromolecular agents such as viruses, which are thought to co-opt EV trafficking mechanisms to cross endothelial barriers.^58,60–62

An ability to block transcytosis in specific endothelial cells would be a valuable tool to assess EV entry and exit into the circulation. In terms of the blood-brain barrier, major facilitator superfamily domain containing protein-2a (MFSD2A) was shown to inhibit transcytosis.⁶³ MFSD2A specifically hinders caveolae-mediated transcytosis⁶⁴ by facilitating the movement of unsaturated phospholipids from the outer to the inner layer of the endothelial cell membrane.⁶⁵ Consequently, the altered composition of the membrane lipids prevents the formation of caveolae.⁶⁴ It is likely that multiple other proteins specific to the central nervous system regulate transcytosis, and future studies are necessary to understand general and EV-specific mechanisms of transcytosis across the blood-brain barrier and other endothelial barriers.

Similar to EVs, lipoproteins are biological nanoparticles that enter and exit the circulation, serving as long-distance delivery systems for endogenous biomolecules and exogenous drugs.⁶⁶ Notably, lipoproteins are six orders of magnitude more abundant in the blood than EVs,⁶⁷ suggesting that EVs are likely to encounter numerous lipoproteins in the interstitial fluid and circulation. Proteomic studies have shown that the protein components of lipoproteins, that is, apolipoproteins, are bound to the EV surface.⁶⁸ Other studies have demonstrated that EVs can interact and bind to lipoproteins when mixed together.^69,70 Evidence that EVs bind to lipoproteins under physiological conditions has been reported,⁷¹ and a preprint has supported these findings.⁷² Notably, a recent study indicated that EVs and lipoproteins can form complexes that remain intact upon cellular internalization and release.⁷³ Therefore, it is possible that EVs form complexes with lipoproteins to hijack caveolae-mediated endothelial transcytosis pathways to traverse between interstitial spaces and the blood (Fig. 1b). Notably, EVs from a brain metastatic variant displayed enhanced interactions with low-density lipoprotein compared to a regular breast cancer cell line.⁷⁰ Further research is necessary to assess whether pro-metastatic EVs exploit lipoprotein interactions to promote transport across the endothelial barrier.

Exocytosis-independent transcellular transport

It is also possible that EVs cross the endothelial barrier through exocytosis-independent transcellular mechanisms involving endocytosis followed by membrane budding (Fig. 1b). The close proximity of membrane-bound organelles to the cell membrane can result in multilayered EVs upon membrane budding,⁷⁴ such as those containing mitochondria.⁷⁵ Similarly, it was speculated that endocytosed EVs in close proximity to the membrane could be released through membrane budding, resulting in the formation of multilayered EVs.⁷⁴ However, it is unknown whether endocytosis, structural preservation, and release through membrane budding, could serve as a mechanism of transcellular transport in general, and specifically, across the endothelial barrier.

CELL EXTENSION-MEDIATED TRANSPORT OF EXTRACELLULAR VESICLES

Cells produce various extensions, such as microvilli, filopodia, cilia, retraction fibres, and apoptopodia, all of which can release EVs.⁷⁶ Therefore, an additional speculative mechanism for entry of EVs into the blood circulation is through cellular extensions that interface with the vasculature. For example, cell extensions from osteocytes (dendritic processes)⁷⁷ and astrocytes (endfeet processes)⁷⁸ form tight connections with the vasculature (Fig. 1b).

OUTLOOK

Taken together, there are various possible mechanisms by which EVs traffic between the interstitium, the lymphatic system, and the blood circulation. Although individual examples of paracellular and transcellular transport have been described, a comprehensive understanding of EV entry and exit routes to and from the blood circulation in physiological and pathological contexts are lacking. This article highlights the pressing need to study EV transport across the vascular barrier, which is critical for understanding intercellular communication beyond the local interstitium, as well as for leveraging EV biology for mechanistic, diagnostic, and therapeutic uses.

In the case of therapeutics, an understanding of EV transport could aid in selecting the most efficient administration route for a given application. Subcutaneous, intraperitoneal, and intramuscular administration routes are likely to enhance lymphatic transport, as demonstrated for synthetic nanoparticles.^79–81 However, a poor understanding of EV fate following lymphatic drainage limits the exploitation of administration routes for improved transport and therapeutic effects. In addition, the impact of physical EV properties (such as their size) on interactions with the vascular barrier remains largely unknown, preventing the rational design of ideal therapeutics. Small EVs (approximately <100 nm) and large EVs (approximately >200 nm) have demonstrated varying distribution and accumulation profiles when administered in animal models. For example, a systematic review of biodistribution studies demonstrated that an elevated deposition of small EVs was observed in the liver and kidneys (first hour) and lungs and spleen (2–12 hours), whereas large EVs demonstrated increased accumulation within the lungs (first hour) and liver (2–12 hours).⁸² However, EV transport phenomena and vascular interactions that mediate distinct biodistribution profiles remain elusive.⁸³ An understanding of EV transport is also important for diagnostic applications. For example, depending on the importance of lymphatic transport, diagnostic capabilities could be improved by sampling EVs from lymph nodes and screening for disease-associated proteins,⁸⁴ nucleic acids,⁸⁵ and lipids.^86,87 Taken together, the lymph presents a promising avenue for diagnostic and therapeutic purposes, due to the likelihood of it being a primary route for EV trafficking.