Extracellular Vesicles and Nanoparticles: Emerging Complexities

Abstract

The study of extracellular vesicles and nanoparticles is rapidly expanding as recent discoveries have revealed a much greater complexity and diversity than was appreciated just a few years ago. New types of extracellular vesicles and nanoparticles have recently been described. Proteins and nucleic acids previously thought to be packaged in exosomes appear to be more enriched in different types of extracellular vesicles and in two recently identified amembranous nanoparticles, exomeres and supermeres. Thus, our understanding of the cell biology and intercellular communication facilitated by release of extracellular vesicles and nanoparticles is in a state of flux. In this review, we will describe the different types of extracellular vesicles and nanoparticles, highlight recent advances and present major unanswered questions.

Extracellular Vesicles and Nanoparticles: A Field of Expanding Complexity

Cells communicate with each other by secreting signaling molecules including proteins, lipids and nucleic acids. Cells can package these signaling molecules in extracellular vesicles (EVs) (see Glossary) in an effort to avoid rapid degradation and escape immune surveillance, amongst other purposes, resulting in local and long-distance intercellular communication. EVs are lipid-bilayer membrane-enclosed vesicles that are released by all cell types studied thus far under both normal and pathological conditions and are detected in all tissues and bodily fluids [1–6]. Recent advances in isolation and analytical methods have led to the identification of an ever-increasing number of EV types [1,6,7] (Figure 1 and Table 1).

Figure 1. The world of extracellular vesicles and nanoparticles.

Most mammalian cells are between 10 to 100 μm in diameter and can release a diversity of heterogenous EVs and NVEPs ranging in size from about 5 nm up to more than 5 μm. The overlapping sizes of many EVs and NVEPs make it difficult to separate them completely. ACLY, ATP citrate lyase; ARMM, arrestin domain-containing protein 1-mediated microvesicle; ARRDC1, arrestin domain-containing protein 1; FASN, fatty acid synthase; HDL, high-density lipoprotein; HSPA13, heat shock protein family A (Hsp70) member 13; IDL, intermediate-density lipoprotein; LDL, low-density lipoprotein; MVP, major vault protein; SMAP, supramolecular attack particle; TGFBI, transforming growth factor beta-induced; TSP1, thrombospondin 1; VLDL, very-low-density lipoprotein.

Table 1.

Extracellular Vesicles (EVs) and Non-Vesicular Extracellular Nanoparticles (NVEPs)

Name	Category	Size	Membrane	Origin/release mechanism	Markers	Refs
Exosomes	Small EV	30–120 nm	Lipid bilayer	Multivesicular endosomes, amphisomes	CD63, Syntenin-1	[1,32,44]
Arrestin domain-containing protein 1-mediated microvesicles (ARMMs)	Small EV	30–150 nm	Lipid bilayer	Ectosome	ARRDC1, TSG101	[29,54,55]
Small Ectosomes	Small EV	30–150 nm	Lipid bilayer	Ectosome	CD147, CD9	[36,37]
Microvesicles	Small to large EV	150–1,000 nm	Lipid bilayer	Ectosome	Annexin A1, Annexin A2, α-Actinin-4	[1,2,45]
Apoptotic bodies/vesicles	Small to large EV	100–5,000 nm	Lipid bilayer	Apoptosis	Annexin V (phosphatidylserine)	[1,68,69]
Migrasomes	Large EV	500–3,000 nm	Lipid bilayer	Retraction fiber	TSPAN4, Integrin alpha 5, Integrin beta 1, damaged mitochondria	[79–81]
Large oncosomes	Large EV	1–10 μm	Lipid bilayer	Ectosome	ARF6, V-ATPase G1, CK18, Annexin A1	[1,62,64]
Exophers	Large EV	3.5–4 μm	Lipid bilayer	Unknown, autophagy-related?	Huntingtin, Tau protein, LC3, Annexin V (phosphatidyl serine), damaged mitochondria	[83,84]
Supermeres	NVEP	22–32 nm	No	Unknown	TGFBI, HSPA13, ENO2	[11,13]
Exomeres	NVEP	28–50 nm	No	Unknown	FASN, ACLY	[8,9,125]
Lipoproteins	NVEP	5–1,200 nm	Single lipid layer	Exocytosis (VLDL), Plasma membrane assembly (HDL)	ApoA1 (HDL), ApoB100 (LDL. IDL, VLDL), ApoE (IDL, chylomicron), ApoB48 (chylomicron)	[90,92,97]
Vault	NVEP	70 nm	No	Unknown, cell death and amphisomes?	MVP, vPARP, TEP1	[1,117,119]
Supramolecular	NVEP	120	No	Secretory	Thrombospondin-1,	[98,99]
attack particles (SMAPs)		nm		granules	Perforin-1, Granzyme B
Viral particles	NVEP (some EV-like, some can also be inside EV)	30–300 nm	Lipid bilayer (some)	Plasma membrane, exocytosis, cell lysis	Env, Gag (HIV-1), Spike protein S1 (SARS-CoV-2), Hexon (Adenovirus), VP1 (Polyomavirus)	[94,96,122]

It is increasingly appreciated that non-vesicular extracellular nanoparticles (NVEPs), lacking a lipid bilayer membrane, are present and often plentiful in the extracellular space and bodily fluids [1,8–11]. NVEPs include well known entities like lipoprotein particles, nucleosomes, and vaults but also recently discovered exomeres [8,9] and supermeres [11–13] (Figure 1 and Table 1). The expanding world of EVs and NVEPs are largely due to improvements in techniques and methodology but also to the growing awareness of their complexity and heterogeneity, which has spurred an active search for new particle subsets. Amongst NVEPs, we will focus on exomeres and supermeres in this review, and speculate on the biogenesis of supermeres.

Extracellular Vesicles: Biogenesis and Heterogeneity

EVs were initially suspected of being a means for normal cells to dispose of unwanted material for maintaining normal tissue homeostasis or for cancer cells promoting tumor progression and metastasis [14,15]. However, it is now well-established that EVs play numerous roles in intercellular communication, facilitated by transfer of EV cargo to recipient cells following uptake or by the interaction of EV surface proteins with cellular receptors [5,7,16–18]. Active cell-cell communication by EV secretion is important to maintain normal physiological function [18,19], and aberrant signaling through EVs has also been linked to numerous disease states, including cancer, cardiovascular disease, neurological and immunological disorders [17,20–22]. EVs and their attendant cargo of proteins, lipids and nucleic acids, including microRNAs (miRNA), are rich sources of potential biomarkers and therapeutic targets, with EVs being explored as vehicles for drug delivery [22–25].

Different types of EVs have been classified based on their origin and biogenesis (Figure 2, Key Figure). At least three major modes of biogenesis are known: apoptotic EVs (termed apoptotic bodies and vesicles) are generated by fragmentation of cells undergoing apoptosis, ectosomes are generated by direct outward budding of the plasma membrane, and exosomes are generated by inward budding of endosomal compartments that later fuse with the plasma membrane. In response to the plethora of EV types and the uncertainty in many instances of their biogenesis, MISEV has taken a pragmatic approach by classifying EVs as ‘large EVs’ (>200 nm) or ‘small EVs’ (<200 nm) [26]. However, it may be useful and timely to remind the research community of the different types of EVs and NVEPs, along with recent advances related to these extracellular particles.

Figure 2. Current understanding of the biogenesis pathways of EVs and select nanoparticles.

The ectosome class of EVs, ranging in size from 30 nm to 10 μm, are generated by direct outward budding of the cell plasma membrane into the extracellular space. Microvesicles, large oncosomes, small ectosomes and ARMMs are ectosomes but they differ in size, molecular composition, and mechanistic details of their release. Microvesicles, small ectosomes and ARMMs are secreted from both normal and cancer cells while large oncosomes are released from cells having undergone malignant transformation. Apoptotic bodies and vesicles range in size from 100 nm to 5 μm and are released when cells undergo apoptosis. Inward budding of the limiting membranes of late endosomes generates intraluminal vesicles (ILVs) causing formation of multivesicular endosomes (MVEs). The MVEs may traffic to lysosomes for degradation or to the cell surface for fusion with the plasma membrane whereupon the ILVs are released to the extracellular space as 30 – 120 nm exosomes. Alternatively, the MVE may fuse with an autophagosome to generate a hybrid organelle, termed the amphisome. Amphisomes may traffic to lysosomes for degradation or to the plasma membrane for release of its contents to the extracellular space. Migrasomes are 500 nm to 3 μm EVs that are released from the retraction fibers of migrating cells and are dependent on the formation of tetraspanin- and cholesterol-enriched microdomains. Exophers are large, 3.5 – 4 μm EVs that contain damaged mitochondria and protein aggregates that appear to be released under stress conditions to maintain tissue homeostasis. The exact release mechanism of exophers is currently unknown but may involve autophagic machinery. Exomeres and supermeres are amembranous 20 – 50 nm nanoparticles released from both normal and cancer cells, and they can be detected in the circulation. They are enriched for specific proteins and nucleic acids compared to EVs but the biogenesis of exomeres and supermeres is currently unknown.

Exosomes

The most studied type of small EV is the exosome (Figure 2). The formation of exosomes begins in endosomes with inward budding of the limiting late endosomal membrane to generate intraluminal vesicles (ILVs) within a multivesicular endosome (MVE). This process of inward budding results in reorientation of endocytosed transmembrane proteins such that they are signaling competent with the ectodomain facing outwards and the cytoplasmic tail into the interior of the ILV with its enclosed cargo. Fusion of MVEs with the plasma membrane results in ILVs being released to the extracellular space as 30 – 150 nm exosomes [1,6,7]. Endosomal sorting complex required for transport (ESCRT) machinery is important for sorting of proteins to ILVs in the MVE and thus exosome formation. However, ESCRT and ESCRT-accessory proteins like TSG101, ALIX and VPS4 are also active in outward budding of the plasma membrane, both for vesicle [27–29] and virus release [30,31]. Syntenin-1, another protein involved in exosome formation and protein loading [32,33], is highly abundant in small EV samples [1,34], but also may be found in large EV samples [1,35]. Likewise, while many of the tetraspanin (TSPAN) proteins like CD63, CD81 and CD9 are highly enriched in exosomes [1,6,7], and have long been used as marker proteins for exosomes, it is increasingly recognized that small TSPAN-containing EVs can bud directly from the plasma membrane and these would be classified as ectosomes/microvesicles based on their biogenesis [36,37].

The RAB family of small G proteins is well known for being involved in vesicle trafficking, including the release of exosomes. RABs are involved in many aspects of transport through the endocytic pathway. RAB7 [32,38], RAB11 [39,40] and RAB35 [41,42] have been found to regulate exosome secretion, and RAB27A and RAB27B are involved in the later steps of MVE-dependent trafficking of ILVs to the cell surface for release as exosomes [14,38,43,44]. Inhibition of the normal function of these RAB proteins has been shown to decrease secretion of exosomes [14,32,39,42,44]. In polarized epithelial cells, RAB27A and RAB37 regulate apical release of exosomes, whereas RAB39A and RAB39B mediate their basolateral release [45,46]. It is becoming increasinglhy clear that membrane contact sites between endoplasmic reticulum (ER) and endosomes may be involved in exosome biogenesis [47,48] and cargo loading [49]. ER-endosome contact site formation is dependent on RAB7 [47,48] and supports a small GTPase conversion from ARL8B to RAB27A in exosome secretion[48]. Much remains to be discovered about the roles of RABs not only in exosome biogenesis but also for other types of EVs as well. For example, RAB13 recently has been implicated in release of β1-integrin-containing small EVs, likely directly from the plasma membrane, that are distinct from CD63-positive exosomes [50]. As another example, RAB31 was found to control an ESCRT-independent exosome biogenesis pathway by promoting formation of ILVs and suppressing degradation of MVEs [51].

Ectosomes

Ectosomes, are by definition derived from outward budding from the plasma membrane followed by pinching off to release the vesicle (Figure 2). Ectosome shedding is thought to occur from most, if not all, healthy cells and is thus distinct from EV release that occurs when cells are dying through apoptosis or necrosis (see below). While initially assumed to be mostly large EVs, ranging in size from 200 – 1000 nm, it has become clear that outward budding can also generate ectosomes in the small EV size range, that is, smaller than 200 nm in diameter [1,29,36,37]. It is also clear there is more than one type of ectosome. There are ‘classical’ microvesicles ranging in size from 150 – 1000 nm in diameter; they are characterized by expression of Annexin A1 and A2 [1,45,46,52,53], possibly α-Actinin-4 [1,2], and exhibit lower flotation densities than small EVs [1]. There are also small (<150 nm) arrestin domain-containing protein 1 (ARRDC1)-mediated microvesicles (ARMMs) that characteristically express ARRDC1 and TSG101, and require VPS4 activity [1,29,54,55]. Additionally, T cells release approximately 70 nm synaptic ectosomes at the immunological synapse when they make contact with antigen-presenting cells in a process that requires TSG101 for sorting of T cell receptors and VPS4 for vesicle scission at the plasma membrane [28,56]. Perivascular dendritic cells release 500 – 1000 nm ectosomes, dependent again on VPS4, to elicit anaphylaxis by relaying allergens to mast cells [27]. Further complicating matters, cells can release an abundance of tetraspanin (CD9, CD63, CD81)-positive exosome-sized sEVs (‘small ectosomes’) by direct budding from the plasma membrane [36,37]. It is not clear at this point if, and how, small ectosomes might differ from ARMMs in their composition and biogenesis.

Generation of microvesicles requires rearrangement of the actin cytoskeleton to promote plasma membrane budding followed by scission and release of vesicles. The small GTP-binding protein ADP ribosylation factor 6 (ARF6) is a critical regulator of classical microvesicle biogenesis. ARF6 activates phospholipase D to recruit ERK to the plasma membrane where ERK phosphorylates and activates myosin light chain kinase (MLCK). Activated MLCK phosphorylates MLC to promote contraction of the actin cytoskeleton at the neck of the budding vesicle for scission and release of the microvesicle [57]. ARF6 is also involved in delivery of pre-miRNA and DNA cargo to microvesicles that are shed from tumor cells [58,59]. Large oncosomes are a class of atypically large 1 – 5 μm microvesicles released from tumor cells induced by overexpression or constitutive activity of oncoproteins [60–63]. They also feature enrichment of ARF6, which may serve for abscission and shedding [62,64] and, similar to classical microvesicles, express Annexin A1 [1]. It should be noted that ARF6 is enriched in both small and large EVs [1], and it has been shown to regulate inward budding of CD63-positive ILVs into MVEs with its depletion resulting in decreased secretion of CD63/Syntenin-1/ALIX-positive exosomes [65], indicating that ARF6 may be present in multiple EV types.

Apoptotic EVs

Apoptotic EVs are released as fragments of cells undergoing apoptosis (Figure 2) and are thought to function in immune regulation and inflammation within the tumor microenvironment [20,66]. They have been divided into larger apoptotic bodies (~1 – 5 μm) and smaller apoptotic vesicles (~100 – 1000 nm) [1,20,67–70]. Apoptotic bodies are significantly smaller than intact cells, which facilitate their engulfment and removal by phagocytes. The distinction between apoptotic bodies and apoptotic vesicles based on size is somewhat arbitrary with the release of apoptotic EVs likely representing a broad continuum; however, it is possible that the molecular cargo of apoptotic EVs differs based on size [68,71].

Phosphatidylserine (PS) is a phospholipid component of the plasma membrane that under normal conditions is confined to the inner leaflet of the plasma membrane. Upon induction of apoptosis, PS is flipped to the outer leaflet where it serves as an “eat me signal” for clearance by phagocytes. Annexin V binds outer leaflet PS and is commonly used to detect apoptotic cells and EVs [68,69,72]. Annexin V-positive small apoptotic EVs have very low or undetectable levels of CD63, CD81, CD9, ALIX and TSG101 [1,73]. However, small apoptotic EVs that express high levels of CD63, but low levels of CD9 and ALIX, are reportedly released from cells through MVEs after induction of apoptosis [74]. This would represent a distinctly different biogenesis pathway than the larger apoptotic vesicles that are produced by membrane blebbing, and once again highlights that CD63 may not be an exclusive marker of classical exosomes. Characteristic of apoptotic bodies and large apoptotic vesicles (~ 500 nm) is the presence of histones, DNA, nuclear, ER and mitochondrial proteins [68,70,72]. Non- apoptotic cell death, such as necrosis, necroptosis, and ferroptosis may also cause release of EVs with a range of sizes [75–77], and a feature of necroptotic EVs, like apoptotic EVs, is the externalization of PS on the outer leaflet of the membranes and Annexin V-positivity [77,78].

Other types of EVs

In addition to exosomes, microvesicles and apoptotic EVs, other types of EVs have been described (Figure 2). These include 500 – 3000 nm migrasomes, which are released from the retraction fibers of migrating cells [79,80]. Migrasome formation is mediated by assembly of large macrodomains enriched for TSPAN4 and cholesterol [81]. Migrasomes may function to remove damaged mitochondria from cells [82]. Another large type of EV are the 3.5 – 4 μm exophers that contain damaged mitochondria and protein aggregates and have been found to be released from C. elegans neurons and murine cardiomyocytes in response to neurotoxic and metabolic stress, impairment of autophagy and inflammation [83,84]. Autophagy-related EVs are currently poorly understood but contain autophagy-related proteins and span in size from small EVs (< 200 nm) to larger vesicles with diameters ranging from 350 – 500 nm [1,85–88].

Extracellular Nanoparticles Enter the Fray

Although the majority of research on secreted entities has focused on EVs, it has long been recognized that proteins, RNA and DNA can be released from cells in NVEPs that include lipoproteins and nucleosomes. In 2018, a new type of extracellular nanoparticle named exomere was reported [8]. Exomeres were isolated using asymmetric flow field-flow fractionation (AF4) of an initial 100,000 x g ultracentrifugation pellet and further separated based on size and hydrodynamic properties yielding this new amembranous nanoparticle, as well as sEVs. Smaller than exosomes, exomeres were found to be enriched in metabolic enzymes and contain RNA and DNA. A year later the first functional transfer of cargo by exomeres was demonstrated [9]. Rather than using AF4 to isolate exomeres, this study adopted a simpler ultracentrifugation-based approach while demonstrating that the isolated extracellular nanoparticles were similar to those purified by AF4, including the enrichment of metabolic enzymes and the presence of RNA and DNA.

It has become clear that preparations of EVs contain NVEPs like exomeres [89], vaults [1], lipoproteins [1,90–92], nucleosomes [1,93] and viral particles [94–96], due, at least in part, to their overlapping size with EVs. Unlike EVs, most NVEPs lack a lipid bilayer membrane although lipoproteins have an outer lipid shell. Other extracellular nanoparticles like exomeres and vaults are amembranous and have little lipid content. While EVs are derived from plasma or endosomal membranes, the biogenesis of extracellular nanoparticles is varied but not directly derived from membranes, or, in the case of exomeres and supermeres (see below), biogenesis mechanisms are currently unknown. Although many extracellular nanoparticles are of similar size to EVs and are therefore present in samples of crude EVs prepared with common methods such as ultracentrifugation and size-exclusion chromatography, their different molecular composition (levels of lipids, proteins and nucleic acids) provides them with dissimilar densities, which can be utilized to separate non-EV material from EVs by density gradient fractionation [1,91,93]. Alternatively, the expression of particular proteins on EVs and NVEPs has been used to specifically isolate the targets of interest [1,2,45,46,97]. Specialized cells may release particular types of extracellular nanoparticles as is the case for cytotoxic T lymphocytes that release supramolecular attack particles (SMAPs) to kill target cells [98]. SMAPs are 120 nm particles surrounded by a thrombospondin-1 shell and a core of cytotoxic proteins, including perforin-1 and granzyme B [98]. Recent work has found that SMAPs reside in multicore granules and represent a second wave of targeted cell killing following the immediate release of cytotoxic proteins by single core granules [99].

The modern era of EV research largely began in 2007 with the discovery that exosomes can protect and transfer mRNA and miRNA to recipient cells, thus enabling RNA-mediated intercellular communication [100]. In the ensuing years, EV research has focused largely on extracellular RNA (exRNA), its packaging inside EVs, release mechanisms, possible functional roles [101], and potential use of exRNA as biomarkers of disease in minimally invasive liquid biopsies [16,102–104]. Many different studies have found that exRNA released to the extracellular space does not match the relative RNA abundance of the releasing cell, suggesting the enrichment of exRNA is specific [1,14,16,105–107]. Considerable effort has been devoted to discovering exRNA sorting motifs or protein-mediated mechanisms that control selective loading to exosomes or other EVs [108–112]. However, it has become increasingly appreciated that exRNA is not only associated with exosomes and other EVs but also with many other extracellular carriers [1,11,113,114]. This may partly explain why it has been so difficult to identify robust sorting mechanisms of exRNA to presumed EVs – investigators may have been looking at the wrong exRNA carrier. Several types of extracellular nanoparticles contain exRNA. These include lipoproteins, ranging in size from 5 – 12 nm for high-density lipoprotein (HDL), 18 – 25 nm for low-density lipoprotein (LDL), all the way up to ~1200 nm for chylomicrons that are highly abundant in circulation and likely represent the predominant particle type in plasma EV preparations [1,90,92,102]. HDL and LDL lipoproteins can transport miRNA and may therefore be a major carrier of miRNA in circulation [97,115], as are non-vesicular AGO2-miRNA complexes [1,116]. Vaults are large (41 nm by 72.5 nm) cytoplasmic ribonucleoprotein particles that are highly conserved in eukaryotes [117]. Vaults contain major vault protein (MVP) and small non-coding vault RNA (vtRNA), which is a type of exRNA commonly identified in EV samples [107,118]. However, vaults and associated vtRNA, are detected in the non-vesicular fractions following density gradient purification of exosomes [1]. Interestingly, vaults have been observed in amphisomes [119], suggesting that they may be actively released by secretory amphisomes in addition to their passive release from dying cells [1]. Extracellular DNA can be found independently of EVs in the form of nucleosomes (DNA wound around histones), in exomeres and perhaps other as yet poorly defined extracellular nanoparticles [1,9,89,93,120]. DNA and histones have also been reported in EVs, predominantly in larger types of EVs [1,20,58,72,121].

Viral particles can be released independently of EVs, as enveloped viruses with an outer lipid bilayer membrane (akin to an EV) or non-enveloped viruses (akin to an NVEP), ranging in size from to 30 – 300 nm [94–96]. Additionally, viruses can be transmitted inside EVs, which may enhance replication and infectivity, as well as enable the virus to evade immune recognition [85,94,122]. EVs can interact with viruses in many ways and EVs from infected cells can contain viral proteins and RNA [95,96], with ACE2-containing EVs and exomeres being shown to directly bind with the S1 subunit of the SARS-CoV-2 S protein [10,123]. Because of their size and shared characteristics, some viral particles and EVs will co-purify with many of the commonly used isolation methods [94,96,124].

Supermeres: The New Kid on the Block

The newest member of the family of NVEPs are supermeres. The name ‘supermere’ derives from ‘supernatant of exomeres’ as they were discovered in the supernatant remaining after isolation of the exomere pellet [11,12]. Supermeres are smaller than exomeres as determined by atomic force microscopy, exhibit selective enrichment of proteins and RNA, and elicit different patterns of hepatic gene expression when injected intravenously into mice [11,13]. However, exomeres have been isolated from 100,000 g centrifugation pellets subjected to subsequent AF4 fractionation [8], or from direct 167,000 g centrifugation [9,10], whereas supermeres have been isolated by 367,000 g centrifugation [11], it cannot at this stage be excluded that exomeres and supermeres are derived from a larger overlapping population of extracellular nanoparticles ‘sampled’ by using different centrifugation speeds. Both exomeres and supermeres appear to be broadly released by many human and mouse cell types and have been detected in human plasma [8–11,125]. Supermeres contain many proteins previously ascribed to exosomes, other EVs, and exomeres (Table 2). This follows several recent reports that have argued that many of the proteins assumed to be contained in exosomes are rather associated with different EVs or are secreted in a non-vesicular manner, and that exomeres contain many previously presumed EV proteins [1,9,89,126].

Table 2.

Select proteins previously associated with EVs that are highly enriched in supermeres

Protein name	Gene name	Category
GAPDH	GAPDH	Metabolic enzyme
Enolase 1	ENO1	Metabolic enzyme
Enolase 2	ENO2	Metabolic enzyme
Pyruvate Kinase M 1	PKM1	Metabolic enzyme
Pyruvate Kinase M 2	PKM2	Metabolic enzyme
Aldolase A	ALDOA	Metabolic enzyme
Glucose-6-Phosphate Isomerase	GPI	Metabolic enzyme
Lactate Dehydrogenase A	LDHA	Metabolic enzyme
Lactate Dehydrogenase B	LDHB	Metabolic enzyme
Triosephosphate Isomerase 1	TPI1	Metabolic enzyme
Hexokinase 1	HK1	Metabolic enzyme
Hexokinase 2	HK2	Metabolic enzyme
Glypican-1	GPC1	Cell surface heparan sulfate proteoglycan
Amyloid Beta Precursor Protein	APP	Cell surface receptor, myloid plaques
MET Proto-Oncogene	MET	Receptor tyrosine kinase
Heterogeneous Nuclear Ribonucleoprotein A2/B1	HNRNPA2B1	miRNA binding
Argonaute 1	AGO1	miRNA binding
Argonaute 2	AGO2	miRNA binding

We are thus left with at least two possibilities: either that cells can secrete a specific protein via different types of EVs and nanoparticles simultaneously or that the isolation methodologies used until recently have not been able to effectively separate disparate types of EVs and nanoparticles. However, these two possibilities are not mutually exclusive. As an example, glycolytic enzymes have been detected in exosomes and sEVs [6,127], but, with improved methodology, glycolytic enzymes were not detected in exosomes [1], but were enriched in exomeres [8,9], and even more highly enriched in supermeres [11]. Other examples are amyloid precursor protein (APP) and glypican-1, previously identified as an exosomal cargo, [128,129], but far greater levels of both are found to be associated with exomeres and supermeres [9,11]. Supermeres from many cell types are highly enriched for the protein transforming growth factor beta-induced (TGFBI) while fatty acid synthase (FASN) is more associated with exomeres [11]. Although full-length transmembrane proteins were not detected in supermeres, a striking finding was the presence of ectodomain cleavage products of APP, glypican-1, MET, amphiregulin and EGFR [11]. Exomeres contain the ectodomain fragment of ACE2 [10], the entry receptor for SARS-CoV-2, and the ectodomain fragments of both ACE and ACE2 were present in supermeres [11]. As supermeres do not contain membranes, the absence of embedded full-length membrane proteins is not surprising. Interestingly, the full-length untethered version of some full-length GPI-anchored proteins, including glypican-1 and CEACAM5, are present in supermeres while other GPI-anchored proteins, including dipeptidase 1 (DPEP1) and the ectonucleotidase CD73, appear to be restricted to EVs in their tethered form [11].

Surprisingly, supermeres may be a more significant source of small exRNA than both EVs and exomeres [11]. Conversely, it is not yet known if supermeres are a source of large exRNA. Supermeres also contain an abundance of YRNA and small nuclear RNA (snRNA), and, like exomeres, the majority of small exRNAs are miRNAs [11]. Perhaps it is not surprising that supermeres are enriched in RNA-binding proteins that carry miRNA, including AGO1, AGO2, hnRNPA2B1 and exportin-5 [11]. These RNA-binding proteins have been shown to be in exomeres and non-vesicular entities [1,11,116], although they have previously been implicated in sorting of miRNAs to EVs [59,112,130]. In contrast, the levels of transfer RNA (tRNA), or tRNA fragments, in exomeres, supermeres, and the non-vesicular material that contaminates small EV preparations, are very low compared to highly purified small EVs [1,9,11], indicating the close association of extracellular tRNA with EVs as previously reported [131]. Many exRNA appear to be modified or fragmented versions of cellular RNA. As an example, extracellular miR-1246 in supermeres is a fragment of the U2 snRNA [11]. A later analysis revealed that the presence of extracellular miR-126 could be a footprint of abundant extracellular ribonuclease activity on U2 snRNA complexed with ribonucleoproteins [132]. Ribosomal RNA fragments are highly abundant not only in exomere and supermere samples but also in EVs, and ribosomes can co-isolate with exomeres [132], leading to the suggestion that extracellular ribosomes and ribonucleoprotein complexes are distinct nanoparticles present in EV, exomere and supermere samples [132].

Currently, the biogenesis of supermeres (or exomeres) is unknown. Are they preformed within the cell or do they self assemble upon release from the cell? It is unlikely they are formed due to cell death as supermeres are obtained from cells with greater than 95% viability when conditioned medium is harvested. If supermeres are derived from broken EVs they would be expected to contain full-length transmembrane proteins characteristic of EVs but they do not. If supermeres represented the luminal content from broken EVs the fragmented EV membrane parts containing full length should be present in isolated samples but they are not. Nor is it likely that supermeres form as aggregation artifacts of the ultracentrifugation-based purification. Size separation by fast protein liquid chromatography (FPLC) of cell-conditioned media not subjected to ultracentrifugation results in supermere peaks that are identifcal to those observed when supermeres obtained by ultracentrifugation are subsequently separated by FPLC (unpublished data). Constituents and characteristics of supermeres may offer clues as to their biogenesis. We present a number of cargo that might suggest research directions to pursue to shed light on their biogeneis. HSPA8 and HSP90 subunits are consistently present and are amongst the most abundant proteins found in supermeres [11]. Many supermere constituents, including metabolic proteins, have KFERQ motifs [133,134] that are targeted by HSP8A in chaperone-mediated autophagy and in endosomal microautophagy [135]. The cargo-selective complex (VPS35, VPS29 and VPS26) of the retromer [136] are found associated with supermeres [11]. Cleaved extracellular portions of transmembrane proteins are enriched in supermeres, including neurodegeneration-associated proteins like APP, and many phase separation-associated RNA-binding proteins, including hnRNPA2B1 and hnRNPA1 [137,138], and metabolic enzymes [139,140] are enriched in supermeres [11]. Secretory amphisomes might also be relevant to these processes, as secretion of cytosolic nucleosomes, and likely vaults, can occur by this mechanism [1,119]. It can be speculated that supermeres may result from a stress response that is adaptive and active in cancer cells and/or as a survival mechanism for normal cells. The composition of supermeres [11,13] and exomeres [9,11,89] examined so far have been quite consistent suggesting that whatever processes causes their formation is likely general and predictable.

NVEPs versus EVs: Why It Matters

While the complexity of heterogeneous mixtures of EVs and NVEPs that contribute to the cellular secretome represents a daunting technical challenge, it is becoming clear that various types of vesicles and particles are enriched in different proteins and exRNA. Over the last few years, there has been a growing recognition of the extent to which the presence of exRNA is a general feature of the extracellular space with many different types of carriers, ways of driving release, and potential functions in intercellular communication [1,116,126,141]. With the improved methodology of EV separation from non-vesicular material [1,89,111], and the discovery of exomeres and supermeres [8,11], it is also becoming clear that many of the proteins previously assumed to be present in the extracellular space in EVs are rather associated with various types of NVEPs. To understand the basic cell biology of secretion used by cells to release specific biomolecules, we first have to correctly identify the extracellular carrier(s) of a particular biomolecule. Optimized methods can then be developed to isolate or enrich for the specific EV or nanoparticle of interest whether for basic or translational research that aims to explore the potential for biomarker discovery. It is worth noting that the newly discovered supermeres are enriched in many clinically relevant proteins such as APP (Alzheimer’s disease), proprotein covertase subtilisin/kexin type 9 (PCSK9), ACE and ACE2 (cardiovascular disease) and MET, glypican-1, TGFBI and AGO2 (cancer) [11]. Next, the biogenesis of the carrier, whether a specific type of EV or nanoparticle, can be explored and possible ways to inhibit or enhance secretion can be identified with obvious implications for biomarker discovery or therapeutic intervention. It is assumed that different types of EVs could interact differentially with cells based on their physical characteristics, including their specific surface biomolecules, which, in turn, may determine the signaling capacity of the EV, as well as the tissue/cell type through which the EV may signal [21,142]. Internalization of EVs can occur rapidly through several different cellular mechanisms, including endocytosis, macropinocytosis and phagocytosis [143–145], while the uptake of exomeres and supermeres appears to be significantly slower [11]. Less is known about the mechanisms by which extracellular nanoparticles interact with cells. Both exomeres and supermeres are taken up by, and can signal to, recipient cells and tissues [8,9,11], in a process that is likely driven, at least in parts, by macropinocytosis [11]. Organ biodistribution studies indicate that liver, kidney, spleen, lung, colon and bone are common sites for uptake of EVs and exomeres and supermeres, while supermeres are preferentially taken up by the brain, indicating they are able to traverse the blood brain barrier [8,11]. Thus, for reasons both related to basic and applied research, the field should focus on improving separation of distinct EVs and nanoparticles rather than continuing to rely on bulk isolation methods that yield mixed populations. This is not to say that bulk isolation methods that yield complex mixtures of EVs and particles cannot provide important cell biological insights or identify potential clinically relevant biomarkers [125]. Two contrasting concerns must be considered when deciding upon the best approach for clinical biomarker discovery and monitoring – the improved signal-to-noise that may be obtained by specifically isolating the vesicle or nanoparticle containing the relevant biomarker rather than using bulk isolation versus the extra time, labor and expense that may be required for more sophisticated purification and analytical methods.

Concluding Remarks

Despite the explosion of publications on EVs in recent years, and the anticipated increase in those on extracellular nanoparticles, more questions arise, and definitive answers seem more elusive (see Outstanding Questions). However, it is becoming clear that advances over the last few years have greatly increased our awareness that cells utilize a multitude of EVs and NVEPs to facilitate communication with neighboring and distant cells. What is much less clear is how to effectively separate, analyze and characterize this abundance of novel agents of intercellular communication. The outcome of studies of EVs and NVEPs is dependent on the details of how the work was performed. Small differences in protocols can lead to major shifts in results and conclusions. However, there is a general consensus that EVs represents a diverse category of vesicles of which endosome-derived exosomes are just a small part. Many types of EVs, large and small, are generated by direct budding of the plasma membrane, and these EVs carry many of the proteins that were previously assumed to be mostly secreted in exosomes. The field now must conted with the recent realization that NVEPs, including supermeres, exomeres, vaults, lipoproteins and protein assemblies, including ribosomes and ribonucleprotein complexes, may be responsible for much of the extracellular transport of RNA. As noted above, the biogenesis of supermeres and exomeres is currently unknown and the mechanisms that determine how and why a given RNA or DNA is packaged for export in an EV or NVEP are also to a large extent unexplored. It is also not clear what further heterogeneity of NVEPs that might still be uncovered and whether some constituents of exomeres and supermeres might be attributable to other distinct nanoparticles such as ribosomes or ribonucleoprotein complexes. These are hotly debated topics that will no doubt be the focus of intense research for years to come.

Outstanding questions.

Do different classes of EVs and NVEPs have different functions in intercellular communication?

What is the degree of cargo selectivity to different classes of EVs and NVEPs?

Is sorting of proteins and RNA to EVs and NVEPs selective or random?

What are physiologically relevant concentrations of EVs and NVEPs to be added to recipient cells?

Are there different functional consequences of a cargo if it is delivered to recipient cells by different carriers?

To what extent are EVs and NVEPs agents of cellular secretion of extracellular RNA?

To what extent are EVs and NVEPs that contain RNA and DNA taken up by cells and functionally alter recipient cells?

Can definable structures of exomeres and supermeres be detected with cryo-election microscopy?

What is the biogenesis of exomeres and supermeres? Do exomeres and supermeres derive from a larger overlapping population of extracellular nanoparticles?

The revolution in our understanding, and appreciation, of the diversity of EVs and other extracellular carriers of proteins, lipids, RNA and DNA that has the potential to contribute to intercellular communication has been heavily dependent on development of methods and techniques used for isolation and characterization. This trend is likely to continue. Technical developments in the study of EVs and NVEPs continue to be a major research and commercial focus as expected for a relatively young but burgeoning field. While a considerable amount of research has been performed to illuminate the roles of EVs in cell-cell communication, the roles of NVEPs in intercellular communication is relatively unexplored. Another major challenge will be to translate our rapidly expanding understanding of the large repertoire of EVs and NVEPs that a given cell has at its disposal into an understanding of how communication is facilitated in the intact organism.

Highlights.

Research from the past few years has overturned our understanding of the field of EVs and nanoparticles by demonstrating a much greater diversity and complexity of extracellular carriers of proteins and nucleic acids.

Cells release not only lipid-bilayer EVs, such as exosomes, but also NVEPs, such as exomeres and supermeres, both of which contain distinguishing proteins and nucleic acids.

Supermeres are a newly discovered type of 15–25 nm amembranous extracellular nanoparticle that contain RNA and many proteins previously associated with exosomes and other EVs.

Technical improvements in the tools used for EV and NVEP research will continue to drive future insights in the cell biology of these distinct modes of intercellular communication.

Glossary

Amphisome

a hybrid organelle, generated by the fusion of an MVE with an autophagosome, that can traffick to the lysosome for degradation or the plasma membrane for release of its contents

Apoptotic extracellular vesicle (apoptotic EV)

a class of extracellular vesicles, ranging in size from 100 – 1000 nm apoptotic vesicles to 1–5 μm apoptotic bodies that are generated when cells undergo programmed cell death, apoptosis

Arrestin domain-containing protein 1 (ARRDC1)-mediated microvesicles (ARMMs)

a type of small 30 – 150 nm ectosome derived by direct outward budding of the plasma membrane followed by pinching off of the vesicle

Ectosome

a category of EVs that are derived by direct outward budding of the plasma membrane followed by pinching off of the vesicle

Exomere

a type of small amembranous non-vesicular 28 – 50 nm non-vesicular extracellular nanoparticle

Exopher

a type of large 3.5 – 4 μm EV that contains damaged mitochondria and protein aggregates

Exosome

a type of small 30 – 150 nm EV derived from the release of ILVs upon fusion of an MVE or amphisome with the plasma membrane

Extracellular vesicle (EV)

a membranous, lipid-bilayer enclosed, vesicle released by cells that may contain full-length transmembrane proteins, soluble proteins in the vesicular lumen, as well as RNA and DNA, depending on vesicle type

Intraluminal vesicle (ILV)

a small vesicle that bud from the limiting membrane into the endosomal lumen of developing MVEs

Large oncosome

a class of atypically large 1 – 5 μm ectosomes released by cancer cells induced by overexpressed or constitutively active oncoproteins

Lipoprotein

a class of extracellular nanoparticles containing a core of lipids and a phospholipid outer shell into which apolipoproteins may be embedded

microRNA (miRNA)

small non-coding RNA that exerts post-transcriptional regulation of gene expression through RNA silencing

Microvesicle

a type of small to large 150 – 1000 nm ectosome derived by direct outward budding of the plasma membrane followed by pinching off of the vesicle

Multivesicular endosome (MVE)

a type of late endosome containing multiple ILVs inside

Non-vesicular extracellular nanoparticle (NVEP)

a non-vesicular extracellular nanoparticle (distinct from an EV) that does not contain a lipid-bilayer membrane and does not contain full-length transmembrane proteins although it may contain cleaved fragments of such proteins

Nucleosome

the structural unit of DNA packaging, consisting of a segment of DNA wound around eight histone proteins

Supermere

a type of small amembranous non-vesicular 22 – 32 nm non-vesicular extracellular nanoparticle

Supramolecular attack particle (SMAP)

a 120 nm nanoparticle released by cytoxic T lymphocytes to kill target cells

Transfer RNA (tRNA)

an RNA adapter molecule that transfers an amino acid to the protein translational machinery

Vault

a 41 nm by 72.5 nm amembranous cytoplasmic ribonucleoprotein eukaryotic organelle that may also be found in the extracellular space as a NVEP

Viral particle

an enveloped or non-enveloped 30–300 nm extracellular nanoparticle carrying viral protein, viral RNA or DNA, and represents the form of a virus outside of an infected cell