Shedding of ciliary vesicles at a glance

Cilia are surface-exposed and evolutionarily conserved organelles with central roles in motility and signaling (Carvalho-Santos et al., 2011; Mitchell, 2017; Sung and Leroux, 2013). The prototypical cilium consists of an array of nine microtubule doublets (the axoneme) that projects from a modified centriole (the basal body) and is ensheathed by a specialized membrane with a unique protein and lipid composition (Mukhopadhyay et al., 2017; Nachury and Mick, 2019; Satir et al., 2010). The ciliary proteome is shaped by trafficking complexes that move proteins in and out of cilia and by the transition zone (TZ), a diffusion barrier at the base of the ciliary shaft. Intraflagellar transport (IFT) complex A mediates membrane protein entry into cilia, IFT complex B moves proteins inside cilia and the BBSome complex retrieves select proteins from cilia into the cell body (Nachury and Mick, 2019; Shinde et al., 2020; Ye et al., 2018). In vertebrates, primary cilia are integral to vision (Bachmann-Gagescu and Neuhauss, 2019; Insinna and Besharse, 2008), olfaction (Jenkins et al., 2009; Kleene, 2008), body weight homeostasis (Omori et al., 2015; Song et al., 2018; Vaisse et al., 2017) and signaling pathways such as Hedgehog (Bangs and Anderson, 2017; Hu and Song, 2019). Early hints of extracellular vesicles (EVs) being shed by cilia into the extracellular environment have now been generalized to a wide variety of organisms and biological contexts (Ikegami and Ijaz, 2021; Wang and Barr, 2018). Roles in regulation of cilium length, in removal of excess material from cilia and in cell–cell communication have all been proposed for cilia-derived EVs. However, a major barrier to testing the physiological consequences of ciliary EV shedding lies in our limited understanding of the molecular mechanisms that underlie this process. Here, we review the diverse instances of shedding by cilia before assessing the current mechanistic understanding of ciliary shedding and evaluating the proposed roles of EVs in disposal of excess material by cilia and in cell–cell communication.

Depending on the biological context and the size of the shed particle, ciliary shedding has been described as ectocytosis, decapitation or autotomy. Despite the divergence in names, all described instances of ciliary EV shedding might be part of an encompassing spectrum (see poster).

Ectocytosis describes the budding of small EVs, also called ectosomes or microvesicles (see Box 1), directly from the plasma membrane or ciliary membrane (Mathieu et al., 2019; van Niel et al., 2018). In Chlamydomonas, cilia are the only conduit of EV release, because the remainder of the cell is covered by an impermeable cell wall. It has been estimated that the constitutive rate of ciliary ectocytosis in Chlamydomonas can consume an entire cilium in 6 hours (Dentler, 2013). Regulation of the length of cilia may thus be accomplished via shedding, in addition to the regulation of cargo entry into cilia (Wren et al., 2013). Not all instances of ciliary ectocytosis are constitutive, and signal-dependent ectocytosis has been observed upon fertilization of Chlamydomonas gametes when the adhesion receptor SAG1 becomes activated (Cao et al., 2015), and upon stimulation of ciliary G-protein-coupled receptors (GPCRs) in mammalian cells when retrieval fails (Nager et al., 2017) or when endocytosis is blocked (Stilling et al., 2022). In both constitutive and signal-dependent contexts, ectocytosis can be viewed as an effective means to rid the cilium of unwanted excess material. Congruently, ciliary ectocytosis is greatly elevated when ciliary proteins are overexpressed in the sensory neurons of Caenorhabditis elegans (Razzauti and Laurent, 2021). These results emphasize the importance of studying ciliary shedding with cargoes expressed at near-endogenous levels.

Decapitation (also called excision) describes the shedding of large fragments of the cilium that accompanies the transition from the ciliated to non-ciliated state. Decapitation has been reported in cultured cell lines (RPE1-hTERT, IMCD3, NIH-3T3 and mouse embryonic fibroblasts) upon exit from quiescence and entry into the cell cycle, coincident with the switch from a ciliated state to the liberation of centrioles for mitotic progression (Phua et al., 2017; Wang et al., 2019). In developing chick and mouse neural tubes, instances of cilium decapitation precede neuronal migration and differentiation, indicating that decapitation may allow a switch in cell fate (Das and Storey, 2014; Toro-Tapia and Das, 2020).

Although most shedding events have been reported to originate from the tip of the cilium, electron microscopy and super-resolution imaging have found vesicles bulging from or juxtaposed along the ciliary shaft in the cultured kidney cell lines IMCD3 and MDCK, and in the retinal pigmented epithelial cell line RPE1-hTERT (Chacon-Heszele et al., 2014; Huang et al., 2016; Sun et al., 2019), suggesting that budding may occasionally take place along the cilium. Alternatively, the static nature of these observations leaves open the possibility that EVs associated with the cilium shaft represent intermediates in EV uptake by cilia.

Autotomy (or cilium amputation) describes the en bloc shedding of an entire cilium via a precise scission event at a location immediately distal to the TZ (Quarmby, 2004). Autotomy is well-documented in free-swimming organisms, such as Chlamydomonas and Tetrahymena cells or sea urchin larvae, and can be triggered by a variety of environmental stresses, such as high salt, low pH or anesthetics (Child, 1959; Lefebvre, 1995; Morris and Vacquier, 2019; Wheeler, 2017). The autotomy response may be construed as a protection mechanism from environmental insults. While it remains to be determined whether precise autotomy takes place past echinoderms in the evolutionary tree, the combination of high concentrations of extracellular Ca²⁺ and membrane disruption that triggers autotomy in free-swimming organisms has been successfully applied to purify motile cilia from mammalian olfactory cells, tracheal epithelial cells and ependymal cells for proteomic studies (Mayer et al., 2009; Narita et al., 2012; Ostrowski et al., 2002). Primary cilia may also be subject to autotomy, as IMCD3 cells have been reported to shed their entire cilium upon exit from quiescence (Mirvis et al., 2019), suggesting that decapitation and autotomy might both contribute to the cell cycle-dependent removal of cilia (see poster).

Our knowledge of the mechanisms that lead to shedding of ciliary material remains fragmented. Below, we review the role of known molecular players in the various types of shedding behavior, but we warn the reader that these mechanisms might not be universally used by all cilia in all ciliated organisms.

Although little attention had historically been given to actin in mammalian primary cilia, cryo-electron tomography and live imaging of actin biosensors have convincingly established the presence of actin filaments in cilia (Kiesel et al., 2020; Lee et al., 2018). Polymerization of branched actin networks plays a key role in both signal-dependent ectocytosis and decapitation (Nager et al., 2017; Phua et al., 2017; Stilling et al., 2022; Wang et al., 2019). Pharmacological inhibition of the actin regulators drebrin and myosin 6, as well as inhibition of actin polymerization by cytochalasin D, prevent signal-dependent ectocytosis in cultured IMCD3 cells (Nager et al., 2017), and targeted sequestration of actin monomers in the cilium blocks decapitation (Phua et al., 2017; Stilling et al., 2022). The precise mechanisms by which actin facilitates ectocytosis are not known and might entail phase separation of membrane lipids and subsequent scission via line tension (Lee et al., 2018) or formation of a contractile ring that initiates membrane closure.

Given their central roles in outward vesicle budding, it should come as no surprise that the endosomal sorting complex required for transport (ESCRT) complexes (see Box 2) have been implicated in ciliary EV shedding. First, ESCRT proteins have been detected at the TZ of Chlamydomonas cilia (Diener et al., 2015) and in cilia of mammalian cells, including RPE1-hTERT, IMCD3 and NIH-3T3 cell lines, and mouse photoreceptors (Datta et al., 2015; Jung et al., 2020; Mick et al., 2015; Ott et al., 2018; Skiba et al., 2021). Second, the ESCRT-III subunit VPS4 is required for shortening of cilia in Chlamydomonas, a process that might be mediated at least in part by ectocytosis (Long et al., 2016), and VPS4 is upregulated in Paramecium upon deflagellation (Gogendeau et al., 2020). However, EV release from male nematode cephalic cilia is unaffected by mutations in some ESCRT-0 or ESCRT-I subunits (Carter and Blacque, 2019; Wang et al., 2014). Nevertheless, considering the multiplicity of factors that funnel cargoes into the ESCRT-III machinery, ectocytosis in C. elegans might utilize other early ESCRT components (Box 2). Cargo sorting by the ESCRT complexes into multivesicular bodies is achieved by recognition of and binding to ubiquitylated proteins (Christ et al., 2017; Raiborg and Stenmark, 2009). Interestingly, polyubiquitylation of ciliary proteins has recently been shown to earmark them for their retrieval (Desai et al., 2020; Shinde et al., 2020). Given that ubiquitin itself has been found in Chlamydomonas EVs, ubiquitylation may also participate in sorting of ciliary proteins into EVs (Long et al., 2016). Future work will be needed to determine whether ubiquitylation regulates both ectocytosis and retrieval, and how a common mechanism can dictate such divergent trafficking routes.

Phosphoinositides are phospholipids with an inositol headgroup that can be phosphorylated at positions 3, 4 and 5. Each of the seven phosphoinositide species marks a compartment, with phosphatidylinositol 4-phosphate [PI(4)P] enriched in cilia and phosphatidylinositol (4,5)-bisphosphate [PI(4,5)P₂] enriched in the plasma membrane. PI(4,5)P₂ is a well-characterized activator of branched actin polymerization via Arp2/3 complex regulators of the WASP/Scar family (Pollard, 2016; Sechi and Wehland, 2000), and forced production of PI(4,5)P₂ in cilia is sufficient to trigger ciliary actin polymerization (Phua et al., 2017; Stilling et al., 2022). Congruently, removal of the ciliary phosphatidylinositol polyphosphate 5-phosphatase INPP5E leads to the elevation of PI(4,5)P₂ levels at the ciliary membrane, subsequent actin polymerization, decapitation or ectocytosis, and cilia shortening (Phua et al., 2017; Sharif et al., 2021). Furthermore, photoreceptors shed their entire cilia as EVs in Inpp5e mutant mice (Sharif et al., 2021). Similarly, targeting a phosphoinositide 5-kinase to cilia of NIH-3T3 fibroblasts increases the ciliary levels of PI(4,5)P₂ and promotes actin polymerization and ciliary shedding in a pathway that requires the well-characterized cilium disassembly factors Aurora kinase A and histone deacetylase 6 (HDAC6) (Stilling et al., 2022). Conversely, forced targeting of the yeast PI(4,5)P₂ phosphatase Inp54p to cilia of NIH-3T3 fibroblasts results in cilia elongation (Stilling et al., 2022). Together with the data from Chlamydomonas (Dentler, 2013), these data further support ciliary shedding as a major mechanism for ciliary length regulation.

Exposure to high extracellular Ca²⁺ concentrations combined with membrane tears deflagellates Chlamydomonas, Tetrahymena and sea urchin larvae (Quarmby and Hartzell, 1994; Rosenbaum and Carlson, 1969; Stephens, 1991), and Ca²⁺ influx is required for autotomy of Chlamydomonas cilia (Finst et al., 1998; Quarmby and Hartzell, 1994). Yet, how Ca²⁺ is transduced into a signal for autotomy remains unknown. The microtubule-severing enzyme katanin was considered a plausible candidate, as knockdown of katanin in Chlamydomonas attenuates Ca²⁺-mediated deflagellation (Lohret et al., 1998; Rasi et al., 2009). It is, however, unlikely that Ca²⁺ triggers autotomy via direct activation of katanin, because Ca²⁺ negatively regulates the microtubule-severing activity of katanin in vitro (Iwaya et al., 2012). Moreover, overexpression of katanin in IMCD3 cells triggers cilium shedding even when intracellular Ca²⁺ is artificially depleted (Mirvis et al., 2019). Furthermore, a multitude of TZ proteins are predicted to bind Ca²⁺ (Zhang and Aravind, 2012), and it is conceivable that Ca²⁺ alters the TZ structure to license axoneme severing at the distal end of the TZ. Highlighting an active role for the TZ in autotomy, Paramecia mutants for one TZ module are resistant to Ca²⁺-mediated autotomy, whereas mutants for a different TZ module continuously shed cilia (Gogendeau et al., 2020) (see poster). Finally, centrins are small Ca²⁺-binding proteins present in basal bodies that assemble Ca²⁺-sensitive contractile filaments with their partner Sfi1 (Gogendeau et al., 2007), and this contractility might underlie the decrease in centriole diameter observed when centrioles are exposed to Ca²⁺in vitro (Sanders and Salisbury, 1989, 1994). Here, it is conceivable that torsion propagating from the basal body engages a mechanical cascade at the distal end of the TZ that triggers localized severing of microtubules. Beyond autotomy, it remains to be determined whether Ca²⁺ plays a role in more focused instances of ciliary shedding.

In mammalian cells, loss of BBSome function leads to the ectocytosis of GPCRs that would normally be subject to retrieval from the cilium into the cell body (Nager et al., 2017) and to a global increase in EV shedding by kidney cells (Volz et al., 2021). In C. elegans, BBSome dysfunction leads to accumulation of vesicles in the vicinity of neuronal cilia (Akella et al., 2020; Wang et al., 2014). Ectocytosis thus represents a means to rid the cilium of unwanted proteins when the retrieval machinery is defective.

Small GTPases of the Rab family are key regulators of endomembrane trafficking. Ten Rabs have been associated with cilia and three with ciliary shedding (see poster). Interference with Rab7, a late endosomal GTPase, prevents actin polymerization and blocks serum-induced decapitation (Wang et al., 2019). Rab35, best known for its role in fast endosome recycling and actin remodeling during cytokinesis (Dambournet et al., 2011), limits the amount of ciliary INPP5E, possibly regulating ectocytosis (Kuhns et al., 2019). Expression profiling in C. elegans has suggested an association between Rab28 and cilia (Jensen et al., 2016), and loss of Rab28 leads to the accumulation of EVs around neuronal cilia of C. elegans (Akella et al., 2020). Meanwhile, knockout of Rab28 prevents disk shedding in mouse photoreceptors, suggesting context-dependent roles of Rab28 in the regulation of ectocytosis (Ying et al., 2018). Finally, depletion of the exocyst, a vesicle-tethering complex and known effector of Rab8 and Rab10, results in a reduced number of EVs in kidney cells (Zuo et al., 2019).

Photoreceptors detect light by concentrating the light-sensing GPCRs opsins and the downstream phototransduction machinery in a hypertrophied primary cilium (also known as the outer segment) filled with stacked membrane disks (Bachmann-Gagescu and Neuhauss, 2019). The outer segment is a remarkably dynamic organelle that turns over its contents every 10 days, most likely to replenish photodamaged rhodopsin. At the tip of the outer segments, discrete packages of old disks are shed at daily intervals and engulfed by the supporting retinal pigmented epithelium (RPE). At the base of the outer segment, new disks are continuously formed and push older disks towards the tip (Mazzoni et al., 2014; Young and Bok, 1969).

Interestingly, ciliary shedding might participate in both disk shedding and biogenesis (see poster). Actin-mediated extrusion of a protodisk initiates disk formation at the base of the outer segment. Protodisk extrusion entails branched actin polymerization via the Arp2/3 complex (Spencer et al., 2019) and its regulators WASF3 and PCARE (Corral-Serrano et al., 2020), and actin filaments subsequently retract from the protodisk to enable flattening of the disk. Interestingly, in the absence of the disk membrane protein peripherin, outer segments are reduced to cilia surrounded by EVs (Salinas et al., 2017). Peripherin has been shown to actively suppress ectocytosis of the protodisk via its cytoplasmic tail, although the downstream intermediates remain to be identified. Disk formation may therefore be likened to a frustrated ectocytosis event interrupted at the membrane extension stage (Molday and Goldberg, 2017; Salinas et al., 2017; Spencer et al., 2020).

Disk shedding is traditionally considered to be driven by the phagocytic activity of the RPE, as isolated photoreceptors do not shed disks (Williams and Fisher, 1987). However, emerging evidence suggests that photoreceptors actively participate in shedding. First, immediately prior to the shedding of disks, photoreceptors flip phosphatidylserine from inner to outer membrane leaflet at the prospective shed tip of the outer segment (Ruggiero et al., 2012). Outer-leaflet phosphatidylserine is a well-characterized ‘eat-me’ signal for apoptotic cells (Nagata et al., 2016; Segawa and Nagata, 2015) and performs a similar role in photoreceptor outer segments by engaging integrin α_vβ₅ and the Mer receptor tyrosine kinase (MerTK) at the RPE. Engagement of these signaling receptors in the RPE cell promotes activation of the Rac1 GTPase and the extension of filipodia that wrap around the photoreceptor tip and engulf the shed stack (Nandrot et al., 2004, 2007; Ruggiero et al., 2012). Furthermore, Rab28 is required for shedding of cone outer segments. Mutation of Rab28 in zebrafish or mice results in elongated outer segments and a reduced number of phagosomes in the RPE (Carter et al., 2020; Ying et al., 2018), and Rab28 mutation causes cone–rod dystrophy in human patients (Roosing et al., 2013). As Rab28 localizes to cone outer segments and participates in ciliary ectocytosis in nematodes (Akella et al., 2020; Jensen et al., 2016), it is tempting to propose that Rab28 promotes disk shedding in a photoreceptor-autonomous manner. Nonetheless, photoreceptor-specific knockouts of Rab28 will be required to test this hypothesis, because Rab28 is also expressed in the RPE. Interestingly, the primary cilium of photoreceptors is not the only source of rhodopsin-laden EVs. In Xenopus laevis, a mutant form of rhodopsin that targets inefficiently to the outer segment becomes secreted into EVs by the cell body (Ropelewski and Imanishi, 2020). This study highlights the role of EV release in disposal of unwanted material and the challenges in assessing the origin of EVs based on their contents (see also Box 1).

With their unique protein and RNA composition, EVs are often sufficient to elicit specific responses when added to recipient cells. However, whether EV secretion is necessary to mediate communication between cells in physiological contexts remains, for the most part, elusive. The role of EVs in cell–cell communication has been best documented in the field of cancer research, where tumor cells secrete EVs to downregulate ‘eat-me’ signals and promote nefarious paracrine signaling (Guo et al., 2017; Patton et al., 2015; Poggio et al., 2019). In the case of ciliary EVs, multiple reports suggest a role for EVs in communication. However, the relative paucity of mechanistic insights has made it difficult to interfere with the process of EV shedding.

First, Chlamydomonas cilia contain a lytic enzyme that is shed in ectosomes and is required for the breakage of the sporangium wall into which gametes are enclosed after gametogenesis (Kubo et al., 2009; Wood and Rosenbaum, 2015; Wood et al., 2013). Cilia-less Chlamydomonas cells cannot break the sporangium wall on their own, but addition of concentrated ciliary EVs enables sporangium rupture and gamete liberation. Second, the encounter of Chlamydomonas gametes of opposite mating types triggers an initial adhesion event mediated by their cilia, followed by an activation cascade that culminates in gamete fusion (see poster). The engagement of adhesion receptors triggers the release of ciliary EVs packed with the adhesion receptor, and EVs from activated gametes can trigger mating-like responses when added to gametes from the opposite mating type (Cao et al., 2015). Intriguingly, the machinery to amidate peptides is present in Chlamydomonas cilia and is released into EVs upon mating (Luxmi et al., 2019). Amidation converts the carboxy group at the C terminus into a primary amine to confer stability and bioactivity to a range of secreted peptides in vertebrates and invertebrates. In the above study, amidated peptides were detected on EVs, and chemical synthesis of one such peptide produced a chemotactic agent (Luxmi et al., 2019). Peptidergic signaling might thus represent an ancestral signaling modality that enables eukaryotes to extend the range of EV bioactivity. Thus, although the importance of EV-mediated communication awaits further study, ectocytosis appears to be relevant for the homeostasis of Chlamydomonas cilia.

The shedding of ciliary EVs has been documented in the parasites Leishmania spp. (Silverman et al., 2008) and Trypanosoma cruzi (Bayer-Santos et al., 2013), which are responsible for leishmaniasis and Chagas disease, respectively. Humans are innately immune to Trypanosoma brucei brucei owing to the expression of trypanosome lytic factors. However, Trypanosoma brucei rhodesiense, the parasite that causes sleeping sickness, can neutralize trypanosome lytic factors by expressing the serum resistance-associated protein SRA (see poster). Interestingly, nanotubes extending from the ciliary membrane of T. brucei rhodesiense cells fragment into EVs to release SRA-positive EVs into the host bloodstream. Uptake of SRA-containing EVs by T. brucei brucei helps them evade host immunity (Szempruch et al., 2016).

The environment surrounding nematodes is replete with EVs that contain a select subset of ciliary proteins, such as the polycystin ion channel (which comprises PC1 and PC2, also known in C. elegans as LOV-1 and PKD-2, respectively; Maguire et al., 2015; Wang et al., 2014). Surprisingly, PC2-containing EVs originate either from the tip of the cilium proper or from a region of the plasma membrane below the TZ named the periciliary membrane (Maguire et al., 2015; Wang et al., 2021). That EVs containing ciliary proteins can originate from a non-ciliary location should serve as a cautionary note for inferring the origin of EVs based on their content (see Box 1). In addition, given that the mechanisms that underlie EV production by the periciliary membrane are likely distinct from ciliary shedding, reserving the term ‘ciliary EVs’ to the population that buds from cilia will facilitate future mechanistic discussions.

Mechanistically, the presence of PC2-containing EVs in the environment requires factors that shape the architecture of cilia (IFT complexes and motors, tubulin post-translational machinery), as well as the myristoylated protein CIL-7 and the kinesin KLP-6 (Maguire et al., 2015; O'Hagan et al., 2017; Wang et al., 2014). Live imaging of the release of PC2-containing EVs from the tip of cilia paints a hierarchical model for these factors, with the IFT kinesin required for ciliary entry of PC2, KLP-6 required for CIL-7 enrichment just below the tip and CIL-7 required for PC2 enrichment at the very tip (Wang et al., 2021). These data place PC2 tip enrichment as the linchpin of PC2 packaging into EVs and pose the question of whether tip enrichment is sufficient to direct a given membrane protein into EVs.

Several lines of evidence suggest that PC2-containing EVs regulate mating behaviors in nematodes (see poster). First, the number of PC2-containing EVs released by male nematodes increases when they are exposed to mating partners (Wang et al., 2021). Second, male cil-7 mutants release fewer EVs and display mating defects (Maguire et al., 2015). Third, PC2-containing EVs released by males adhere to the hermaphrodite vulva after mating (Wang et al., 2020). Finally, EVs purified from the environment of wild-type strains can increase the frequency of tail-chasing behavior when added to males (Wang et al., 2014), whereas EVs isolated from mutants defective in PC2 shedding do not (Silva et al., 2017; Wang et al., 2014). Altogether, the research performed in C. elegans presents the most compelling case for a role of EVs in cell–cell communication.

Kidney epithelial cells shed ectosomes that contain PC1 and PC2 (also known as PKD1 and PKD2, respectively; Hogan et al., 2015). Proteomic studies on human urinary EVs have identified PC1 and PC2, the Hedgehog signaling protein Smoothened, as well as several of their first-degree interactors (Chacon-Heszele et al., 2014). In addition, proteomic studies of murine endothelial EVs, which appear to be shed in a cilium-dependent manner in response to shear flow (Mohieldin et al., 2021b), have found proteins potentially important for neuronal maintenance (Mohieldin et al., 2021a). In a recent study, small EVs purified from BBSome mutant kidney cells have been found to downregulate Wnt signaling activity when added to cells (Volz et al., 2021). Remarkably, primary cilia may act as EV-receiving sites, as urinary EVs added to primary biliary epithelial cells or IMCD3 cells attach to cilia within 1 minute and disappear shortly thereafter (Hogan et al., 2009). While the data suggest that EVs attach and fuse with cilia, an alternative interpretation is that EVs transiently attach to cilia and then detach. Primary cilia of glioblastoma cells shed EVs with mitogenic activity (Hoang-Minh et al., 2018). This activity is cilium-dependent both in donor and host cells, suggesting that cilium mediates both the shedding and the uptake. Suppression of ciliogenesis in these glioblastoma lines promotes cell death and sensitization against chemotherapeutic agents (Hoang-Minh et al., 2016). Finally, the addition of purified biliary EVs to cultured rat cholangiocytes decreases levels of phosphorylated ERK and mitotic activity in a cilium-dependent manner (Masyuk et al., 2010). In conclusion, although many of the observations reported in vertebrate cell systems remain phenomenological, the tantalizing links that have been uncovered warrant further testing of the role that cilia-derived EVs may play in cell–cell communication.

As discussed above, nearly all cilia are capable of shedding extracellular material of diverse sizes through conserved mechanisms. Cilia shedding modifies the behavior of the donor cell in a cell-autonomous manner, either desensitizing the cell to a particular signaling pathway or an environmental insult, inducing cell cycle re-entry, or switching between transcriptional programs. Circumstantial evidence suggests various roles for cilia shedding in cell–cell signaling, with the regulation of nematode mating being the strongest case established so far. Ultimately, the demonstration that ciliary EV shedding is required in select biological processes awaits the development of tools that specifically and efficiently interfere with the shedding of ciliary EVs.

We thank Swapnil Shinde for comments on the manuscript.