1. Introduction
Extracellular vesicles (EVs) encompass a wide variety of small, membrane-bound particles that provide paracrine, autocrine and endocrine cell signaling
[5,6][1][2]. A number of comprehensive reviews offer an excellent introduction to EVs generally
[7,8,9][3][4][5].
Although EVs are recognized to contribute to developmental signaling
[81[6][7],
82], most EV research has been dedicated to understanding their role in cancer biology and characterizing EVs for use in therapeutics or diagnostics
[83,84][8][9]. Despite their diminutive size, EVs have been documented in a variety of developmental contexts, although it is often unclear which type of EV is being released. Similarly, while the embryonic significance of membrane protrusions such as cytonemes or tunneling nanotubes is apparent, the classifications and use of protrusion terminology in developmental systems are often vague or misconstrued. Though EVs and protrusions can function separately during signaling, the possibility of their cooperation is not always taken into consideration when researching one topic or the other. Growing evidence in developmental biology has begun to connect these two mechanisms, demonstrating the importance of protrusions and their characterization alongside EV research. Moreover, as the technology to study EVs expands, so too does the ability for developmental biologists to adapt EV protocols and apply them to developmental systems (reviewed in
[85][10]). The following section
s offer
s examples to illustrate the role of EVs and protrusions during development.
2. Pre-Gastrulation
EVs have been implicated in the earliest stages of development. Upon fertilization in
Xenopus, exosomes are released into the perivitelline space from the zona pellucida, spermatozoa, and cortical granule
[86][11]. Likewise, post-fertilization zygotes and early cleavage human embryos secrete CD9-positive EVs into the perivitelline space
[87,88][12][13]. In mice, the pre-implantation inner cell mass releases MVs carrying laminin and fibronectin to initiate migration of extraembryonic trophoblasts, increasing implantation efficiency
[89][14]. Maternally derived EVs in the oviduct also promote healthy development and implantation of the blastocyst, and their absence is thought to contribute to the low success rate of assisted reproductive technologies (reviewed in
[90][15]).
Within the early embryo, EVs and protrusions work together. The blastocoel fluid of human blastocysts contains small EVs
[91][16]. Accordingly,
Xenopus blastomeres extend long filopodia across the blastocoel that shed EVs from their tips which are taken up by nearby blastomeres. Meanwhile on the basolateral surface of the blastocoel, short, motile filopodia interact with and traffic EVs for transfer of morphogens or maternal ligands
[86,92][11][17]. Cells of the murine blastocoel also project dynamic short and long traversing filopodia, with the latter connecting trophectoderm with distant inner cell mass cells. These filopodia contain actin, exhibit signs of vesicle release, and express FGFR2 and/or ErbB3 receptors, suggesting these protrusions are involved in signal transduction
[93][18]. While these studies do not specify the type of EVs or fine cellular protrusions involved, it is clear that EVs and protrusions play a crucial role in post-fertilization events, mammalian implantation, and blastula/blastocyst signaling.
3. Gastrulation
EVs continue to be important during gastrulation.
Xenopus gastrulation movements force EVs from the perivitelline space and blastocoel into the archenteron, where they are taken up by post-involution epithelial cells, though the function of this is unknown
[86][11]. The zebrafish yolk syncytial layer produces exosomes that are apparent in the bloodstream as soon as circulation commences
[94][19]. While this exosome release is syntenin-A-dependent and syntenin-A is required for epiboly, the role of exosomes in zebrafish gastrulation is unclear
[94,95][19][20]. Zebrafish gastrula-stage mesoderm and endoderm also produce migrasomes, which are necessary for proper specification of mesodermal lineages, reflected in defective organ morphogenesis
[96][21].
Cellular protrusions also enable cell communication during the complex embryonic germ layer rearrangements of gastrulation. In sea urchin gastrulae, primary mesenchyme, secondary mesenchyme and ectodermal cells contact one another with thin filopodia, indicating directed cell–cell communication across germ layers
[97][22]. Analogously, gastrulating murine mesodermal cells communicate with fine cellular protrusions as they collectively migrate
[98][23]. In zebrafish, intercellular bridges that form between pre-gastrula daughter cells are maintained through gastrulation across distant cells
[99][24]. Meanwhile, the enveloping layer produces fine actin-based protrusions during epiboly
[100][25] and Vangl2-dependent filopodia enable dorsal convergent mesodermal cell migration
[101][26]. Overall, gastrula stage cellular protrusions appear to provide spatiotemporal information to coordinate cell positions and movements; however, it is unknown whether EVs accompany these protrusions.
4. Patterning
While it was originally thought that morphogen gradients were created by diffusion and/or direct cell–cell contact, those models have changed with the discovery of EVs and fine cellular protrusions
[102][27]. In particular,
Drosophila has revealed a great deal about delivery of morphogens via EVs and protrusions. This includes Hedgehog transport by cytonemes in germline stem cells
[103][28] and EGF signaling via protrusions in the mechanosensory organs of the legs
[104][29]. Cytonemes containing specific receptors respond to different signals depending on the cell type. Examples include cytonemes that detect Decapentaplegic in eye discs, Delta-Notch in the air sac primordium, FGF in both air sac primordium and tracheal cells, and Wingless in wing imaginal discs
[105,106,107][30][31][32]. Wing imaginal discs also release exosomes containing Hedgehog
[108,109][33][34] and Wingless
[110][35]. Although Wingless EVs do not diffuse or contribute to the wing imaginal disc Wingless gradient, they do affect local signaling
[111,112][36][37]. More in depth descriptions and examples of EVs involved in morphogen transport of Hedgehog, Wnt, Notch, and BMPs in
Drosophila are reviewed in
[82][7].
Cytonemes or cytoneme-like protrusions have been implicated in vertebrate patterning as well. In the mouse and chick limb bud, SHH-producing cells present SHH on the surface of fine cellular protrusions which extend several cell diameters to interact with responding cells’ filopodia that contain SHH co-receptors
[113][38]. Zebrafish epiblast cells transfer Wnt from Vangl2-dependent long cytonemes to pattern the neural plate, and while EVs were not directly studied, there is evidence that Wnt may be packaged within vesicle-like structures
[114,115,116,117][39][40][41][42].
In other cases, protrusions partner with EVs to enable patterning. During zebrafish pigment stripe pattern formation, airinemes extend from unpigmented xanthoblasts to pigmented embryonic melanophores. Notch ligands carried within vesicular structures found at the tips of these airinemes activate Notch signaling to promote migration, separation, and stripe pattern formation
[66][43]. In a related example of protrusion and EV coordination, EVs released from cilia in
C. elegans neurons have been implicated in neuron-glia communication to pattern sensory organ morphogenesis
[118][44]. Meanwhile, cytonemes from SHH-producing cells in
Drosophila wing discs transport SHH-containing exosomes and MVs toward receptor cells. Although it is still unclear if exosomes travel within cytonemes, small vesicles were observed moving inside these protrusions
[77][45]. These examples highlight the importance of considering EVs in the context of protrusions and vice versa, as well as the challenge to fully characterize their functions in complex developing embryos.
5. Migration
Migration is a crucial function of cells during development. Migrating mesenchymal cells dynamically extend filopodia to sense their environment and create new adhesive contacts
[119,120][46][47]. Some of these protrusions become extended and specialized
[121][48]. In addition to migration during gastrulation
(discussed in Section 4.2), lateral plate mesoderm expressing EphB3b and hepatoblasts expressing EphrinB1 use long cellular protrusions to repel one another and coordinate directional migration during zebrafish liver bud formation
[122][49]. Long filopodia are also used for migration-related processes throughout nervous system development (reviewed in
[123][50]). For example, chick cranial neural crest cells exhibit both short and long protrusions to enable cell–cell communication as they migrate
[124][51]. Strikingly, cytoplasmic transfer of photoconvertible GFP was apparent through these neural crest cell membranous passages
[125][52]. This demonstrates a mechanism to communicate positional information between migratory cells, reminiscent of what was observed in sea urchin gastrulation
[97][22].
Evidence for EVs in developmental cell migration is just beginning to emerge.
Dictyostelium in particular have become a model for understanding EV signaling during migration
[126][53]. Studies on
Dictyostelium have demonstrated that the chemoattractant cAMP can be found within MVBs and released in EVs to promote directed cell migration
[127][54]. Chick neural crest cells also release exosomes as they migrate, suggesting this may be a conserved form of communication between migratory cells; however, the contents of neural crest cell exosomes remain elusive
[23][55]. Although sample size constraints hinder the collection of EVs from embryonic migratory cells, metastatic cancer cells release exosomes that influence migration direction
[67,128][56][57] and invasiveness of less aggressive cells
[129,130,131,132][58][59][60][61]. Cancer recapitulates development
[133][62], supporting the idea that embryonic cells regulate migration via exosomes.
Since only migrating cells form migrasomes, it is not surprising that they are involved in developmentally relevant migration events. Zebrafish gastrulae release migrasomes carrying chemokines, which are required for Kupfer’s vesicle formation and left/right laterality
[96][21]. Chick neural crest cells also deposit migrasomes as they migrate
[23][55]. Interestingly, murine neutrophils release migrasomes to discard damaged mitochondria and maintain cellular homeostasis
[134][63]. Neural crest cells are subject to high rates of oxidative stress
[135][64], raising the possibility that this could be a function of neural crest cell migrasomes as well. While EVs were originally expected to disperse soluble gradients, migrasomes are deposited into the ECM and act as a bread crumb trail-like mechanism that cells pick up from the substrate itself. It will be interesting to consider whether substrate-bound and free floating EVs are used interchangeably or serve different functions.
6. Morphogenesis
Cells engaged in morphogenetic movements use protrusions to reach across long distances to contribute signals or even mechanical forces. During the complex cell movements of mouse neural tube closure, fine filamentous extensions similar to cytonemes emerge from cells on either neural fold to create bridges as the neural tube comes together
[136][65]. This behavior has also been observed in fly embryos during dorsal closure as epithelial cells project microtubule-rich, finger-like protrusions that ‘zip’ the tissues together
[137][66].
Drosophila also require filopodia as myoblasts and myotube muscle cells fuse, as their loss results in a failure of adhesion-based foci to form between cells
[138][67]. In addition, the FGF gradient necessary for
Drosophila tracheal branching morphogenesis is created by reciprocal protrusions, where wing imaginal disc cytonemes present membrane-tethered FGF that is received in a contact-dependent manner by FGF receptor-expressing air sac cytonemes
[139,140][68][69]. Later, EVs are required for tracheal tube fusion
[141][70]. It is, however, unclear whether any of these examples of morphogenetic protrusions also involve the release of EVs. In
C. elegans, neurons and glia exchange EVs produced by cilia during sensory organ morphogenesis
[118][44], supporting this possibility.
7. Differentiation
EVs regulate differentiation in many ways. Neurons produce EVs that stimulate neural differentiation of neural stem cells
[142][71] and promote neural circuit development and function
[143][72]. Meanwhile, osteoblasts generate MVs and apoptotic bodies that promote differentiation of zebrafish scale osteoclasts by activating Rankl signaling
[144,145][73][74]. On the other hand, bone mesenchymal stem cells release apoptotic bodies to transfer proteins and miRNA to distant mesenchymal stem cells to regulate osteogenic differentiation in mice
[146][75]. Another study showed that primary chick notochord cells release two distinct SHH-containing exosome populations with different miRNA and protein profiles; only SHH exosomes containing integrins activated ventral spinal cord differentiation
[147][76]. Meanwhile, cultured mouse retinal progenitor cells transfer EVs containing mRNA, miRNA, and proteins to initiate retinal differentiation
[148][77]. In the subventricular zone, neural stem cells release EVs that regulate microglia morphology, which feedback cytokines to affect neural differentiation
[149][78]. These examples highlight the importance of signaling by EVs during differentiation.
Protrusions can also regulate differentiation. This is illustrated by the zebrafish spinal cord, where early differentiating spinal neurons extend basal protrusions that express Delta protein, which inhibits neurogenesis. These protrusions project over several cell diameters to preferentially interact with neural progenitors of the same subtype, leading to the spatiotemporal spacing pattern of neurons in the zebrafish spinal cord
[150][79]. A latticework of protrusions also forms in pre-neurogenic chick and human embryonic spinal cord, but their function is unknown
[151][80]. It is unclear what types of protrusions are involved in spinal cord differentiation, or whether EVs are also released.
EVs also modulate developmental potential. Embryonic stem cells release EVs that promote stemness and prevent differentiation
[152][81]. As lineage is restricted, EVs continue to modulate potency versus differentiation. For example, EVs released by nervous system stem cells affect cell fate and morphogenesis, and may be leveraged therapeutically (reviewed in
[153][82]). EVs and protrusions also work together. During brain development, EVs containing Prominin-1 (CD133) are present in the neural tube lumen, and neuroepithelial cells extend CD133-positive protrusions. While it is unclear if the EVs originate from these protrusions, because CD133 is a stem cell marker, this process may enable down-regulation of stem/progenitor properties and contribute to differentiation
[43][83].
8. Homeostasis
Homeostasis is equally crucial in developing and adult organisms. As mentioned
previ
nously, Section 4.4, migrating mouse neutrophils discard damaged mitochondria in migrasomes to maintain cellular homeostasis in a process termed ‘mitocytosis’
[134][63]. Inversely, rather than expunging dysfunctional or dying materials, zebrafish basal epithelial cells uptake apoptotic bodies to stimulate Wnt signaling and maintain cell numbers
[154][84].
EVs are critical for vascular homeostasis. As the embryo grows, CD63-positive exosomes released from the zebrafish yolk syncytial layer travel through the bloodstream to trophically support growth and vasculogenesis of the caudal vein plexus
[94][19]. Later, neurons in larval zebrafish and rodent brains secrete miR-132-containing exosomes that are taken up by endothelial cells and required for brain vascular integrity
[155][85]. It has been proposed that migrasomes may also contribute to vascular homeostasis due to known functions of migrasomes and the amount of migrasome-associated Tspans expressed within the cardiovascular system
[156][86].
Cellular protrusions are critical for homeostatic maintenance of the stem cell niche (reviewed in
[157][87]). This is most apparent in
Drosophila, where loss of filopodial extensions disrupts the transfer of signaling molecules that sustain the hematopoietic stem cell niche
[158][88]. Similarly in the
Drosophila germline, microtubule-rich nanotubes provide Decapentaplegic that maintains germline stem cells of the testis stem cell niche
[159][89]. Meanwhile niche support cells deliver cytoneme-based Hedgehog to neighboring somatic cells to maintain the
Drosophila ovary stem cell niche
[103][28].
9. Regeneration
Model organisms that regenerate offer unique opportunities to study regenerative mechanisms, especially when body parts are reproduced with the morphology and pattern normally created during development (see
[160][90]). As in development, EVs play a role in regeneration. For example,
Hydra secrete EVs containing Wnt signaling regulatory factors that modulate head and foot regeneration
[161][91]. During zebrafish fin regeneration, blastema cells release EVs to communicate with other cells in the fin
[162][92]. Meanwhile, highly regenerative newt cells secrete EVs that offer protective effects to mammalian cardiomyocytes
[163][93].
Indeed, the ability of EVs to elicit behaviors in receiving cells holds great promise for regenerative biology (reviewed in
[164][94]). Mesenchymal stem cell (MSC)-derived EVs in particular have garnered intense interest as a therapeutic approach (reviewed in
[165][95]). MSC-derived MVs can promote kidney, cardiac, liver, and neural regeneration, demonstrating the multi-effectiveness of EVs in promoting repair
[166][96]. MVs may do this by delivering mRNA and proteins that effect epigenetic reprogramming in target cells, as EVs stimulate upregulation of early pluripotency markers
[167][97]. Exosomes and MVs also contribute to wound healing. During post-injury angiogenesis, mesenchymal stem/stromal cells, leukocytes, platelets, erythrocytes and endothelial cells all release EVs that either stimulate or inhibit angiogenesis, depending on the cell type
[168][98].