Extracellular vesicle-localized miR-203 mediates neural crest-placode communication required for trigeminal ganglia formation

While interactions between neural crest and placode cells are critical for the proper formation of the trigeminal ganglion, the mechanisms underlying this process remain largely uncharacterized. Here, we show that the microRNA-(miR)203, whose epigenetic repression is required for neural crest migration, is reactivated in coalescing and condensing trigeminal ganglion cells. Overexpression of miR-203 induces ectopic coalescence of neural crest cells and increases ganglion size. Reciprocally, loss of miR-203 function in placode, but not neural crest, cells perturbs trigeminal ganglion condensation. Demonstrating intercellular communication, overexpression of miR-203 in the neural crest in vitro or in vivo represses a miR-responsive sensor in placode cells. Moreover, neural crest-secreted extracellular vesicles (EVs), visualized using pHluorin-CD63 vector, become incorporated into the cytoplasm of placode cells. Finally, RT-PCR analysis shows that small EVs isolated from condensing trigeminal ganglia are selectively loaded with miR-203. Together, our findings reveal a critical role in vivo for neural crest-placode communication mediated by sEVs and their selective microRNA cargo for proper trigeminal ganglion formation.

(v/v) cacodylate buffer, pH 7.2-7.4. Samples were gradually dehydrated with serial solutions 159 of 50%, 70%, 80%, 90%, 95%, 100% acetone and then embedded in epoxyPolybed 8120 160 resin. Next, ultra-thin sections (~70 nm thick) were harvested on 300 mesh copper grids, 161 stained with 5% uranyl acetate and 1% lead citrate, and observed with a FEI Tecnai G2 Spirit 162 transmission electron microscope, operating at 120 kV. The images were randomly acquired 163 with a CCD camera system (MegaView G2,Olympus,Germany). 164 For preparation of sEVs, trigeminal ganglia were treated with a solution of dispase and 165 trypsin to obtain a single cell suspension. Samples were centrifuged at 10000 x g for 10 min 166 and the supernatant was recovered to isolate sEVs. Then, the sample was filtered through 167 0.2 μm filter and then pelleted by centrifugation at 100,000 x g for 90 min to obtain an sEVs 168 enriched fraction. As the protocol of EVs isolation include a filtration step using a 0.2 µm filter, 169 the term sEVs (that include exosome and small size microvesicles) will be used throughout 170 the text. sEVs were resuspended in PBS-DEPC containing a protease inhibitor cocktail 171 (cOmplete™ ULTRA Tablets, Mini, EASYpack. Sigma). For characterization of particle 172 quality size and abundance of the isolated sEVs nanoparticle tracking analysis methodology 173 (NTA -Nanoparticle Tracking Analysis, Nanosight LM10 (Malvern™, U.K.)) was used, with 174 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 for the isolation of small RNA. 176 was treated with amplification-grade DNaseI (Invitrogen). The reverse transcription reaction 181 to obtain the cDNA was performed with the MystiCq® microRNA cDNA Synthesis Mix kit 182 (Merck) and amplified by PCR using the following primers (miR-34-5p Fw: 5'-GCC GCT GGC 183 AGT GTC TTA G-3'; miR-203 Fw: 5'-CCG GCG TGA AAT GTT TAG G-3'; and miR-UNI Rev: 184 5'-GAG GTA TTC GCA CCA GAG GA-3'). in chick embryos. miR-203 was previously shown to be present in premigratory NC cells but 201 down-regulated prior to their delamination from the neural tube (Sánchez-Vásquez et al., 202 2019). In agreement with this, we noted that mature miR-203, detected using locked nucleic 203 acid-digoxigenin-labeled probes, was absent at HH13 from migrating HNK1 immunoreactive 204 cranial NC cells (Fig. 1B). However, miR-203 expression was again noticeable at stage HH16 205 when cranial NC cells are coalescing at the site of trigeminal ganglion formation (Fig. 1C. 206 Black arrowheads). Later at stage HH20 when the ganglion is almost fully condensed, signal 207 was robust, particularly at the center of the lobe (Fig. 1D). Taken together, these data indicate 208 that miR-203 expression is reactivated at the time of NC coalescence and condensation into 209 ganglia, consistent with the intriguing possibility that miR-203 may be required for NC 210 aggregation during trigeminal ganglion formation. 211 212 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023    . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532527 doi: bioRxiv preprint  . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532527 doi: bioRxiv preprint condensation having ectopic condensation and/or denser TG) on Control and miR-203 OE. Numbers Given that miR-203 re-initiates during trigeminal condensation, we next asked whether 248 its loss of function would disrupt proper ganglion formation. To test this possibility, we utilized as a control. The neural tube was electroporated at stages HH9 to target one side of the 253 embryo, and embryos were then examined after the ganglia had condensed (HH17-18). 254 Surprisingly, we failed to detect defects in the morphology of the trigeminal ganglion after the 255 loss of miR-203 in NC cells, as visualized by ISH for Sox10 (Fig. 3A). 256 The trigeminal ganglion has a dual origin from both NC and ectodermal placodal cells. 257 Therefore, we next explored the possible functional role of miR-203 in the trigeminal placodes 258 by electroporating the right placodal ectoderm at HH9 with the miR-203 or scrambled sponge 259 plasmids. Intriguingly, the miR-203 sponge resulted in trigeminal ganglia that displayed a 260 more loosely organized and less aggregated morphology than those in the non-injected side 261 or observed in control embryos ( Fig. 3B. Black arrowhead). The effect was statistically 262 significant and more severe in the ophthalmic (OpV) than in the maxillo-mandibular (MmV) 263 lobe; this is highly reminiscent of the phenotype observed after trigeminal ectoderm ablation 264 (Shiau et al., 2008). Our findings raise the intriguing possibility that miR-203 is produced in 265 the neural crest (donor cell) but exerts its biological effect in the placode cells (recipient cell). 266 This is consistent with the putative role of placode cells as crucial mediators of NC 267 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is  be releasing exosomes-like structures into the extracellular space (Fig. 4B). Shedding 292 vesicles protruding from the plasma membrane are also observed (Fig. 4C). 293 As a further confirmation that EVs produced by the NC cells are able to reach and be 294 internalized by placode cells, we have adapted the pHluo_M153R-CD63-mScarlet vector 295 (Sung et al., 2020) to work in chick embryos. This plasmid allows dynamic subcellular 296 monitoring of exosome lifecycle, including MVB trafficking and exosome uptake. The plasmid 297 contains a modified pH-sensitive GFP sequence (pHluo) inserted into the first extracellular 298 loop of the tetraspanin CD63. Of note, pHLuo-CD63 does not fluoresce in the acidic 299 endosomal pH of MBV; however, once exocytosed into the neutral pH of the extracellular 300 environment, it emits a bright fluorescence. In addition, the plasmid is tagged with pH-301 insensitive red fluorescent protein (mScarlet) that allows the visualization of exosome 302 trafficking (see scheme in figure 4D). 303 pHluo-CD63-mScarlet was introduced into the NC by electroporating into the neural 304 tube of HH9chick embryos. After NC migration and TG condensation, we observed active 305 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532527 doi: bioRxiv preprint vesicle release in the TG condensing region (dashed white line bordering the ganglia) 306 compared with the midbrain region where only mScarlet fluorescence was detected (Fig. 4E,  307 white arrowheads). Transverses sections through these embryos reveal mesenchymal 308 migratory NC cells containing MVBs in acidic conditions (mScarlet only) and in neutral pH 309 during the secretion process (mScarlet and pHluo positive) (Fig. 4F). Similar as previously 310 described (Sung et al., 2020), we observed trails that may represent EVs deposition. Finally, 311 to demonstrate the ability of NC cells to release EVs that can reach placode cells, sections 312 of electroporated embryos with pHLuo-CD63-mScarlet were immunostained at HH16 for the 313 neuronal marker Tuj1, which will mark placodally-derived neurons at this stage (Fig. 4G).

336
To gain deeper resolution and reveal interactions in real-time, we next performed co-337 cultures with NC and placode explants from embryos electroporated with pHluo 338 (pseudocolored yellow) and pCIG-mRFP (pseudocolored magenta), respectively (Fig. 5A). 339 The explant pairs were cultured until migratory cells from both populations contacted one 340 another at which time we generated time-lapse movies of the co-cultures. Live imaging 341 revealed NC cells that were surrounded by numerous extracellular pHluo+ puncta (possibly 342 EVs deposits), as well as trails (possibly migrasomes or retraction fibers) and cytonemes 343 (Fig. 5B, Supplementary Movie 1). Interestingly, the cytoneme-like structures produced by 344 migratory NC cells were in dynamic contact with placode cells (Fig. 5B'). In this sense, it has 345 been shown that small EVs can travel along cytonemes and are released in close proximity 346 to recipient cells during development (Chen et al., 2017;González-Méndez et al., 2017), 347 raising the intriguing possibility that may be similar during neural crest-placode interactions. 348 It was previously shown that exosomes could be captured by filopodia or 349 macropinocytosis events and endocytosed by recipient cells (Heusermann et al., 2016). We 350 observed a similar event in which placode cells contained vesicular structures in their 351 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 Interestingly, we also noted that placode cells produced filopodia that move toward EVs 355 deposits and engulf them (Fig. 5E-F, white arrowhead. Time lapse and 3D rotation in 356 Supplementary Movie 5 and 6, respectively). These observations demonstrate that neural 357 crest-produced EVs are internalized by placode cells. 358 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532527 doi: bioRxiv preprint

miR-203 produced in NC cells regulates translation in recipient placode cells. 379
To determine whether miR-203 produced in NC cells can reach placode cells to exert 380 a biological effect, we designed an experiment in which we drove overexpression of miR-203 381 and EGFP in the NC cells by electroporation into the premigratory NC. In the same embryos 382 we electroporated the trigeminal placode in the ectoderm with a dual-colored sensor vector 383 which expresses both nuclear d4EGFPn, containing two mature miR-203 recognition sites 384 such that the miRNA can bind and affect protein translation (Sánchez-Vásquez et al., 2019), 385 and mRFPn (see scheme in Fig. 6A). Embryos were electroporated at HH9-with the two 386 vectors and allowed to grow until HH17. Transverse sections through the embryos were then 387 immunostained for Tuj1 to identify the trigeminal ganglion cells (Fig. 6B). Although some of 388 the placode cells reaching the condensing area were EGFP+/RFP+ (white arrowhead), some 389 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023 with controls when the EGFP/RFP fluorescence intensity ratios were quantified on placode 394 cells (Fig. 6D). and showed a significant decrease compared with controls (Fig. 6E). 405 Taken together, our in vivo and co-cultured explant results demonstrate that miR-203 406 produced in the NC reaches and suppresses translation in placode cells, thus supporting the 407 idea that miRNAs may act as intercellular signals mediating proper neural crest-placode 408

communication. 409
. CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532527 doi: bioRxiv preprint

In vivo
Ex vivo co-culture d4GFPn . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532527 doi: bioRxiv preprint  (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532527 doi: bioRxiv preprint caused in placode cells. Thus, our study is one of the first to use an in vivo system to study 519 the role of miRNA cargo in sEVs during intercellular communication in early development. 520 Finally, we demonstrate that neural crest-produced cytoneme-like structures contact 521 placode cells. This structure participates in contact-mediated cell communication during early 522 organogenesis, where cell protrusions are used to deliver signals (Mattes and Scholpp, 523 2018). Importantly, it was demonstrated that sEVs are transported along cytonemes (Gradilla 524 et al., 2014). Based on this, we speculate that cytonemes produced by NC cells may enable 525 a high local concentration of sEVs released near placode cells. Active engulfment of sEVs 526 by placode cells ensures that a sufficient load of miRNAs reach the cytoplasm to mediate 527 target inhibition. Altogether, the neural crest-placode interaction during sensory ganglion 528 development offers an excellent in vivo model system in which to examine the mechanism 529 by which selective miRNAs cargo is delivered into sEVs and transported to specific target 530 cells. 531 532 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532527 doi: bioRxiv preprint . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted March 15, 2023. ;https://doi.org/10.1101https://doi.org/10. /2023