The lipid raft-linker gene Raftlin-2 is expressed in migrating neural crest cells

Abstract

Transient plasma membrane domains called lipid rafts have emerged as important regulators of signal transduction. These territories are formed by lipid-lipid and lipid-protein interactions, and these local interactions can be scaffolded by resident lipid raft organizing protein family members. While roles for lipid rafts have been described for multiple signaling pathways in many contexts, their in vivo prevalence and roles during embryonic development remain incompletely understood. Here we examined gene expression for the Raftlin family of lipid raft organizing proteins, Raftlin-1 (RFTN1) and Raftlin-2 (RFTN2), over the course of early vertebrate development, with a focus on neural crest cell dynamics. By analyzing transcriptomic data across vertebrate species, we identified conserved patterns of RFTN1 and RFTN2 expression across species, where RFTN1 is broadly expressed at low levels, while RFTN2 is distinctly enriched in neural crest cells. We used fluorescent in situ hybridization to spatially define Raftlin gene family expression patterns in the early avian embryo. Our results show that RFTN1 is broadly expressed with periods of enrichment in the developing paraxial mesoderm. In contrast, RFTN2 expression is strongly enriched in neural crest cells, beginning during specification and persisting through migration, with additional expression in both the cranial and intermediate mesoderm. Together, these patterns suggest that Raftlin proteins may play important roles in regulating signaling during development with potential roles in somitogenesis and in neural crest and mesodermal cell migrations.

Introduction

Lipid rafts are nanoscale, transient membrane domains enriched in sphingolipids and cholesterol that often facilitate the enrichment or exclusion of critical signaling components, thereby acting as adaptable signaling platforms (Isik and Cizmecioglu, 2023; Roy and Patra, 2023; Sezgin et al., 2017; Simons and Ikonen, 1997). Accordingly, lipid rafts within the plasma membrane have been implicated in regulating numerous signaling pathways that play key roles in development (Mancini et al., 2026; Roy and Patra, 2023); however, the scope and prevalence of lipid raft function during embryogenesis remains unclear. While lipid rafts form due to transient lipid-lipid and lipid-protein interactions, their stability is regulated by families of lipid raft scaffolding proteins. Thus, insight into their involvement in developmental processes may come from carefully interrogating the spatiotemporal genes expression profiles of such lipid raft scaffolding proteins.

To this end, here we report the expression of two such lipid raft organizers over the course of vertebrate embryogenesis with a particular focus on neural crest cell development. Neural crest cells are a multipotent population of embryonic cells in the vertebrate embryo that contribute to many different adult tissues including the craniofacial skeleton, neurons and glia of the peripheral nervous system, septation in the heart, and melanocytes in the skin (Stundl et al., 2026; Vega-Lopez et al., 2018). During their development neural crest cells make complex specification and cell fate decisions, undergo coordinated proliferation, epithelial-to-mesenchymal transition (EMT), and extensive migration (Mancini et al., 2026; Piacentino et al., 2020; Shellard and Mayor, 2019; Stundl et al., 2026). These behaviors require carefully regulated levels of developmental signals including those of the Wnt, Fibroblast Growth Factor (FGF), Bone morphogenetic protein (BMP), Transforming growth factor β (TGF-β), and Notch pathways (Prasad et al., 2019; Stundl et al., 2026). Since lipid rafts have been implicated in regulating each of these signaling pathways (Mancini et al., 2026; Roy and Patra, 2023), neural crest cells may serve as a tractable in vivo model for studying the role of lipid rafts in a wide range of cellular processes.

Previous studies have shown the expression of several lipid raft organizing proteins to be enriched in neural crest cells, suggesting that lipid rafts play essential roles at specific times or in specific subpopulations of developing neural crest. Notably, a group of raft organizing proteins called Flotillins (also known as Reggies) are expressed in neural crest cells and neural crest cell derivatives in zebrafish and Xenopus embryos (Langhorst et al., 2005; Pandur et al., 2004; von Philipsborn et al., 2005). More recent work analyzing differential gene expression across neural crest EMT has shown upregulated expression of the another lipid raft organizing protein, RFTN2 (Raftlin-2), during avian cranial neural crest migration (Piacentino et al., 2022). These expression profiles raise the possibility that lipid raft function, regulated by these organizing proteins, serves an important role in regulating essential cellular processes during neural crest development.

Raftlin proteins, named for lipid raft-linking functions, are thought to act as scaffolds that localize to and regulate cell signaling on the plasma membrane (Lin et al., 2025; Matsumoto and Tatematsu, 2016). Raftlin-1 (gene symbol RFTN1; the protein product is also published under the name “Raftlin”, but we use “Raftlin-1” herein for specificity), and its paralog Raftlin-2 (gene symbol RFTN2), were first identified as necessary for T- and B-cell receptor signaling by concentrating receptor and downstream transmission networks in lipid raft domains (Matsumoto and Tatematsu, 2016; Saeki et al., 2009, 2003). Raftlin proteins localize to the plasma membrane by reversible acylation on N-terminal glycine and cysteine residues, and this membrane targeting strategy is required to mediate the interactions between B- and T-cell signaling components (Saeki et al., 2009, 2003). Despite lack of overt developmental phenotypes in Raftlin-1 and Raftlin-2 single and double knockout mice (Saeki et al., 2009; Tatematsu et al., 2016), their protein products have since been shown to function beyond T- and B-cell receptor signaling with essential functions in cell proliferation, survival, and even migration in other contexts. For example, Raftlin-1 promotes proliferation through the AKT/p38 Mitogen-Activated Protein Kinase (MAPK) signaling pathway in gastric cancer (Deng et al., 2022), and RFTN2 expression correlates with increased metastasis in B16 melanoma lines (Kienzler et al., 2025). On the other hand, Raftlin-1 negatively regulates endothelial cell migration by stabilizing vascular endothelial growth factor receptor 2 (VEGFR2) with its co-receptor neuropilin-1 on the plasma membrane and hinders their endocytosis, thereby modulating the downstream effects of VEGF signaling and diminishing focal adhesion engagement (Bayliss et al., 2020). This finding is particularly relevant to neural crest biology, since neural crest migration requires chemoattraction toward chemokines such as VEGF (McLennan et al., 2010; Wiszniak et al., 2015). Together, these observations suggest that Raftlin proteins may play important roles in neural crest cell development by tuning essential signaling events.

To better understand the possible roles of lipid raft-organizing proteins during neural crest cell development, here we carefully characterized the dynamic spatiotemporal expression pattern of the Raftlin genes RFTN1 and RFTN2 during early vertebrate development. Drawing upon previous transcriptomics studies, we show evidence that RFTN2 expression, but not RFTN1, is conserved in neural crest cells across amniotes. We then use fluorescent in situ hybridization in early avian embryos to define their spatiotemporal expression pattern at multiple stages of vertebrate development that encompass neural crest cell formation, migration, and fate decision. We show that while RFTN2 is highly expressed in specified and migratory neural crest cells along multiple axial levels, strong RFTN1 expression is instead restricted to the developing paraxial mesoderm. These findings implicate RFTN2 as a potential regulator of neural crest cell development, while RFTN1 may instead regulate cell signaling in other lineages.

Results and Discussion

Raftlin gene expression is dynamic during neural crest development across multiple species

To determine if Raftlin proteins may play a role in important developmental transitions, we sought to characterize their expression patterns in neural crest cells, a highly tractable in vivo model of several key developmental processes. To this end, we examined Raftlin gene expression in published transcriptome datasets from multiple amniote species including chick and mouse embryos, and embryonic stem cell (ESC)-derived human neural crest-like cells. First, we examined bulk RNA sequencing results collected from whole chicken embryos, as well as cranial neural crest cells sorted by TFAP2AE1::GFP expression, across multiple stages of development from Hamburger Hamilton stages HH6 through HH16 (Hovland et al., 2022). We plotted normalized expression levels for RFTN1 and RFTN2, together with SOX10 to display phases of neural crest specification and migration (Fig. 1A). The results show relatively stable expression of RFTN1 across development, both in the whole embryo and in the neural crest samples. Conversely, RFTN2 expression becomes prominently upregulated in neural crest cell samples following the onset of SOX10 expression, which is maintained from HH8 through HH16, with a slight downregulation by HH14 (Fig. 1A). This phase of peak expression coincides with the timeframe during which cranial neural crest cells are migrating most extensively through the head, suggesting that RFTN2 expression is regulated alongside the neural crest specification and migration gene regulatory networks (Stundl et al., 2026).

Fig 1. Raftlin genes show conserved expression patterns across amniotes.

(A) Expression levels of SOX10, RFTN1, and RFTN2 in regularized log-transformed counts per million (rlog(CPM)) results of bulk RNA sequencing from sorted chick neural crest cells (top panel) and the whole chick embryo (bottom panel); data from (Hovland et al., 2022). (B) Heatmap displaying variance-stabilizing transformation (VST)-normalized expression levels of indicated genes in human embryonic stem cells (ESCs), on Day 3 following Wnt activation to induce neural crest cell (NCC) fate (pre-NCC), and Day 5 of treatment representing NCC specification; data from (Leung et al., 2019). (C) UMAP cluster names (C) and indicated feature plots (C’, C”) displaying single-cell transcriptomes from murine cranial neural crest sorted from Wnt1-Cre/R26Tomato transgenic mouse embryos at embryonic days E8.5, E9.5, and E10.5; data from (Poverennaya et al., 2026).

Next we examined bulk RNA sequencing from human neural crest cells differentiated from ESCs (Leung et al., 2019). Over a five-day induction period, ESCs show decreasing expression of pluripotency factors like NANOG, and increasing expression of markers of neural crest specification including PAX3/7, TFAP2B, SNAI2, and SOX10 (Fig. 1B). Across this time course, RFTN1 expression was detected at moderate levels across differentiation, with a mild reduction as ESCs start to differentiate. Conversely, RFTN2 expression was detected within ESCs, and its expression strongly increased alongside bona fide markers of neural crest formation (Fig. 1B). Finally, we examined single-cell RNA sequencing results collected from sorted cranial neural crest cells from Wnt1-Cre/R26Tomato transgenic mouse embryos at embryonic days E8.5, E9.5, and E10.5 (Poverennaya et al., 2026). UMAP representation of these data capture neural tube (NT)-localized premigratory neural crest cells through their migration and fate restriction down either neuroglial or mesenchymal lineages (Fig. 1C). Feature plots displaying Rftn1 expression show scattered expression throughout all cell clusters, with a slight enrichment in the NT/premigratory neural crest cluster (Fig. 1C’). Conversely, Rftn2 shows sparse expression in the NT and premigratory neural crest, with clear expression biased to migratory and differentiating neural crest cell clusters (Fig. 1C”). Together, these results suggest that expression of Raftlin-2, but not Raftlin-1, undergoes conserved upregulation during neural crest development across amniotes.

Raftlin-1 is broadly expressed throughout the embryo, with strong enrichment in the developing somites

To spatially resolve Raftlin gene family expression in wild type chick embryos, we designed hybridization chain reaction fluorescence in situ hybridization (HCR-FISH) probes specific to RFTN1 and RFTN2 and performed time course expression analyses alongside markers of early development primarily focused on neural crest and related tissues. We used third-generation HCR-FISH due to the application of split probes for increased sequence specificity and reduction in off-target signal (Choi et al., 2018), and initial experiments produced a distinct spatial expression pattern for each probe set (Fig. S1). To further test specificity, we also performed negative control experiments utilizing fluorescent hairpins without the corresponding Raftlin gene-specific probes. The results indicate that the observed labeling patterns are specific to Raftlin probe hybridization and not non-specific binding of labeling reagents (Fig. S1).

Upon whole mount analysis, the RFTN1 expression pattern at HH9 shows broad expression, with apparent enrichment at the anterior neuropore, in the neural folds at the level of the first somites, and in the lateral plate mesoderm (Fig. 2A, white arrows indicating anterior neuropore (aNP), neural folds (NFs), and lateral plate mesoderm (LPM)). Interestingly, expression appeared weakest in the cranial region at the levels of the midbrain and hindbrain (Fig. 2A, open arrowhead). By HH10, in addition to broad diffuse expression, we observed qualitatively enriched expression of RFTN1 in the most posterior paraxial mesoderm-derived somites; this somite-specific enrichment persisted through HH16 (Fig 2B–C, white arrows indicating somites (som)) and is consistent with in situ hybridization experiments labeling Rftn1 expression in the mouse tail bud (Chen et al., 2025). Next, we performed transverse sectioning to more precisely identify which tissues selectively upregulate RFTN1 expression. In the head of HH10 embryos we observed broad, yet punctate staining across all major tissue populations including the midbrain/hindbrain neural tube (PAX2-positive), cranial mesoderm and neural crest-derived (TFAP2B-positive) mesenchyme, overlying non-neural ectoderm, pharyngeal endoderm, and the otic placode (PAX2-positive) (Fig. 2B’–B”, white arrows indicating neural tube (NT), cranial mesoderm (meso), neural crest derived mesenchyme (NC), non-neural ectoderm (NNE), pharyngeal endoderm (PE), and the otic placode (otic)). However, in more posterior sections at this stage, we observed similarly broad expression of RFTN1 in all tissues, as well as more concentrated puncta in the lateral plate mesoderm and throughout the somite; this expression becomes more restricted to the ventromedial region of the somite in the more anterior, further developed somites (Fig. 2B”’–B””, white arrows indicating lateral plate mesoderm (LPM) and somites (som)). This pattern is consistent with later stages of development at HH16, where we observed RFTN1 expression broadly in the most posterior somites (Fig. 2C””), restricted more ventromedially in the more anterior developing somites (Fig. 2C”’), further restricted to the medial most sclerotome (Fig. 2C”), and finally more diffuse, expression in the most mature somites (Fig. 2C’). Together these results are consistent with RNA-seq studies suggesting more ubiquitous and diffuse expression of RFTN1 across development with limited neural crest cell enrichment (Fig. 1).

Fig 2. RFTN1 is broadly expressed, with dynamic enrichment in developing somites.

Wild type chick embryos were probed using HCR-FISH for expression of RFTN1 alongside markers of neural crest cells (TFAP2B), and midbrain/hindbrain neural tube, otic placode, and intermediate mesoderm (PAX2) at the indicated stages. Representative wholemount images are shown of observed expression patterns at stages HH9, HH10, and HH16 (n=3/3 embryos examined per stage). Transverse sections of wholemount embryos at axial positions indicated by dashed lines (A’-A”’, B’-B””). Representative images are shown of observed expression patterns at stages HH10 and HH16 in section (n=2/2 embryos sectioned per stage). aNP, anterior neuropore; cranial, cranial region; otic, otic placode in B and otic vesicle in C; LPM, lateral plate mesoderm; NFs, neural folds; IM, intermediate mesoderm; NC, neural crest; mes, cranial mesenchyme; NNE, non-neural ectoderm; PE, pharyngeal endoderm; som, somite; ND, nephric duct; NT, neural tube; dm, dermamyotome; scl, sclerotome.

The expression pattern in the most posterior somites of HH10 and HH16 embryos raises the possibility that RFTN1 plays an important role during early somitogenesis. Our results showed RFTN1 expression to initially be highest in the newly epithelialized somites; however, after the ventromedial sclerotome undergoes epithelial-to-mesenchymal transition (EMT) and separates from the dermomyotome, this polarized enrichment appears lost (Fig. 2C’–C””). Interestingly, RFTN1 expression in the ventromedial quadrant of the somite is reminiscent of the expression pattern of PAX1 (Cauthen et al., 2001; Draga and Scaal, 2024). The transcription factor PAX1 is essential for sclerotome development, and its expression is induced by signaling molecules like Sonic hedgehog from neighboring tissues (Draga and Scaal, 2024; Fan and Tessier-Lavigne, 1994; Wallin et al., 1994). Given these observations, and the role of Raftlin-1 as a regulator of cell signaling in other contexts, we speculate that this restricted expression pattern may reflect a similar role for Raftlin-1 in the context of sclerotome development. Future experiments will be required to test this hypothesis and assess the relationship between Raftlin-1 and somite development.

Raftlin-2 expression is enriched in neural crest cells during specification and early migration

We next sought to define the spatiotemporal pattern of RFTN2 expression. Given evidence from RNA-sequencing studies that suggest enriched RFTN2 expression in neural crest cell populations (Fig. 1)(Piacentino et al., 2022), we probed for RFTN2 alongside PAX7 to mark the induced neural crest cells within the neural plate border (Basch et al., 2006; Stundl et al., 2026). We did not detect appreciable RFTN2 expression prior to neurulation within the well-defined PAX7-positive neural plate border at HH7 (Fig 3A). However as neurulation proceeds, specified neural crest cells at HH8 express TFAP2B (Rothstein and Simoes-Costa, 2020), at which point we observed the onset of RFTN2 expression in the dorsal neural folds, and expression the cranial mesoderm (Fig. 3B, white arrows indicating neural crest (NC) in dorsal neural folds and mesoderm (meso)). RFTN2 expression was expanded by HH9, detectable in nearly all TFAP2B-positive neural crest cells (Fig 3C–C’, white arrow indicating neural crest (NC)). Transverse sectioning confirms this pattern, revealing RFTN2 signal to be most enriched in neural crest cells, and present at qualitatively more sparse and lower levels in the adjacent cranial mesoderm (Fig. 3C”, white arrows indicating neural crest (NC) and cranial mesoderm (meso)).

Fig 3. RFTN2 is expressed in neural crest cells during specification and early migration.

Wild type chick embryos were probed using HCR-FISH for expression of RFTN2 and markers of neural crest cell development (PAX7 and TFAP2B) at the indicated stages. Representative wholemount images are shown of observed expression patterns at stages HH7 (n=4/4), HH8 (n=3/4), HH9–9+ (n=8/8), HH10–10+ (n=4/4). Transverse sections of wholemount embryos at the axial levels indicated by dashed lines (C’, E’, E”). Representative images are shown of observed expression patterns in section (n=2/2 embryos sectioned per stage). Regions of higher magnification are designated by dashed boxes. NC, neural crest; meso, cranial mesoderm; som, somite; IM, intermediate mesoderm.

Since RFTN2 appeared to be upregulated in the neural crest after onset of markers of specification, it is unlikely that Raftlin-2 is required for neural crest cell induction or specification. However, given the upregulation of RFTN2 expression after that of TFAP2B, it is tempting to speculate that RFTN2 may be a target of the specification gene regulatory network and could participate in maintaining specification and/or for neural crest cell survival before EMT and migration. Importantly, at this stage of development Wnt signaling remains an important facilitator of specification by activating expression of transcription factors like FOXD3 through the effector protein Axud1 (Simões-Costa et al., 2015). In zebrafish embryos, Wnt signaling relies on the localization of the GPI anchored protein Lypd6 to lipid raft domains to facilitate the appropriate axis patterning (Özhan et al., 2013). If future experiments show that Raftlin-2 functions in scaffolding lipid rafts in neural crest cells as well, these results could prompt testing the hypothesis that Raftlin-2 advances neural crest development by promoting Wnt signaling through similar mechanisms as described in zebrafish.

We next probed RFTN2 expression during early neural crest cell migration. RFTN2 expression is maintained in migratory neural crest cells at stages HH9+ through HH10+ as these cells undergo and complete EMT (Fig. 3D–E). By transverse sectioning we confirmed robust RFTN2 and TFAP2B expression in delaminating and migrating neural crest cells and found that RFTN2 is expressed throughout the cranial mesoderm, albeit at a qualitatively lower level (Fig. 3E’, white arrows indicating neural crest (NC) and cranial mesoderm (meso)). Transverse sections through the most posterior somites also reveal expression of RFTN2 in the intermediate mesoderm (Fig. 3E”, white arrow indicating intermediate mesoderm (IM)), but not in the somites as was seen for RFTN1 (Fig. 2). Since the intermediate mesoderm is the origin of nephric tissues and embryonic kidney (James and Schultheiss, 2003), this pattern suggests a possible role for Raftlin-2 in early kidney development.

As neural crest cells undergo EMT and delaminate, their migration is partially directed in response to chemotactic gradients, migrating toward chemoattractants such as vascular endothelial growth factor (VEGF) and stromal cell-derived factor 1 (SDF1) (Escot et al., 2013; McLennan et al., 2010; Olesnicky Killian et al., 2009; Shellard and Mayor, 2019; Wiszniak et al., 2015). In endothelial cells, Raftlin-1 regulates VEGFR dependent migration through regulating receptor trafficking to the plasma membrane (Bayliss et al., 2020). This, together with RFTN2 expression in migrating neural crest cells (Fig. 3D–E), raises the possible hypothesis that a similar mechanism may facilitate neural crest cell migration in response to VEGF, instead mediated by Raftlin-2. Additionally the SDF-1 receptor, CXCR4, co-localizes with the lipid raft marker GM1 ganglioside in hematopoietic stem cells, and to detergent-resistant lipid raft fractions upon activation (Wysoczynski et al., 2005). In B cells, Raftlin-1 also co-fractionates with GM1 in detergent-resistant membrane fractions (Saeki et al., 2003). By disrupting lipid rafts through cholesterol removal, hematopoietic stem cell chemotaxis in response to SDF-1 was significantly diminished (Wysoczynski et al., 2005). These data are consistent with the idea that some chemotactic signaling pathways may act through lipid rafts in a variety of cell types, and thus lipid raft organizing proteins may provide a means to regulate such signaling in cells such as the neural crest. Further, RFTN2 expression present in the cranial mesoderm may reflect regulation of signaling activities between neural crest cells and the paraxial mesoderm in the head. For example, cranial neural crest cells secrete inhibitors of Wnt and BMP signaling that are important for cranial myogenesis (Tzahor et al., 2003) and mesoderm cells will briefly migrate with invading neural crest cells (McKinney et al., 2020). Shared signaling pathways and migration could indicate a common role for RFTN2 in these two tissues.

Raftlin-2 expression is downregulated during late neural crest cell migration into the pharyngeal arches

After observing robust expression at early stages of neural crest cell migration, we then probed RFTN2 at later stages of migration and development. By HH13, TFAP2B expression in the more rostral mesencephalic neural crest was qualitatively reduced (Fig. 4A’, dashed lines enclosing cranial neural crest (cNC)) compared to the neural crest cells from the more caudal hindbrain streams emanating from rhombomeres r4 and r6 as they migrated around the otic vesicle (Fig. 4A’, white arrow indicating otic vesicle (otic)). We observed that RFTN2 displayed similar expression dynamics, with apparently stronger expression in the migrating r4 and r6 neural crest streams compared to the more rostral migrating cranial neural crest (Fig 4A’). This is consistent with our previous observations where the more rostral mesencephalic neural crest cells at HH10 show qualitatively higher RFTN2 enrichment during the earlier phases of migration (Fig. 3D–E). At HH13, RFTN2 expression also persisted in the most posterior intermediate mesoderm, whereas abundant PAX2 expression was observed over a broader territory stretching anterior past the domain of RFTN2 expression (Fig. 4A”, white arrow indicating intermediate mesoderm (IM)), suggesting that potential roles for Raftlin-2 in pronephros development may be transient. By HH17, RFTN2 expression appears more diminished across the embryo in whole mount (Fig. 4B). Transverse sections through the head reveal clear expression of TFAP2B and PAX2 in the pharyngeal arch ectoderm (Fig. 4B’), in tissues producing the cranial epidermis and epibranchial placodes, respectively (Baker et al., 2008; Van Otterloo et al., 2022; Washausen and Knabe, 2019). We observed a slight and qualitative enrichment in RFTN2 transcript detection in the pharyngeal arch 1 and 2 mesoderm of neural crest origin, which retained some sparse TFAP2B staining (Fig. 4B’, dashed outlines around neural crest-derived mesenchyme (NCM) in pharyngeal arch 1 and 2 (PA1 and PA2)); this enrichment was most clearly noted in comparison with the more medial mesoderm-derived mesenchyme (Fig. 4B’, mesoderm-derived mesenchyme (meso)). This neural crest population will contribute critical craniofacial skeletal elements including the mandible and hyoid bone (Frisdal and Trainor, 2014), thus, the scattered and diffuse expression pattern of RFTN2 in the neural crest derived mesenchyme may suggest that its downregulation is required to permit ectomesenchymal differentiation.

Fig 4. RFTN2 expression is downregulated in late neural crest cell migration and maintained at a lower level in cranial neural crest-derived mesenchyme.

Wild type embryos were probed using HCR-FISH for RFTN2, TFAP2B, and PAX2 at the indicated stages. Representative wholemount images are shown of observed expression patterns at stages HH13 and HH17 (n=3/3 per stage). Transverse sections are indicated by dashed lines (B’-B”). Representative images are shown of observed expression patterns in section (n=2/2 embryos sectioned). Regions of higher magnification are indicated by dashed boxes. (A’-A”) Dashed line highlights the midbrain neural crest region (A’). Dashed lines highlight neural crest derived mesenchyme in pharyngeal arches 1 and 2 (B’). cNC, cranial neural crest; r4, rhombomere stream 4; r6, rhombomere stream 6; NC, neural crest; IM, intermediate mesoderm; NT, neural tube; PA, pharyngeal arch; meso, mesoderm-derived mesenchyme; NCM, neural crest-derived mesenchyme; otic, otic vesicle.

Consistent with RFTN2 patterns of expression seen at earlier stages, trunk neural crest cells at HH13 and HH17 also showed some enrichment in RFTN2 labeling in the dorsal neural tube in the presumptive neural crest (Fig. 4B”, white arrow indicating neural crest (NC)). Interestingly, given its expression pattern in the migratory neural crest, the intermediate and cranial mesoderm, a role for RFTN2 in cell migration could be utilized by multiple tissues. The intermediate mesoderm contributes to the nephric duct, and undergoes extension by collective migration starting at HH10 where leader cells migrate towards an FGF8 gradient produced in the tail bud, depositing cells to epithelialize along their migratory path (Atsuta and Takahashi, 2015; Soueid-Baumgarten et al., 2014). RFTN2 expression is enriched in the most posterior portion of the intermediate mesoderm domain marked by PAX2 at HH13 (Fig. 4A”); this pattern reflects the location of the nephric duct leader cells as they actively migrate (Atsuta and Takahashi, 2015). Thus, it will be interesting to test if RFTN2 may be acting in the neural crest, cranial mesoderm, and intermediate mesoderm populations to modulate lipid raft activity to mediate migratory signaling events in each of these tissues.

Conclusions

Our analysis has revealed that RFTN1 and RFTN2 have dynamic and specific expression patterns throughout early embryonic development. Using transcriptomic data sets, we demonstrated that these expression patterns are similar across vertebrate species: RFTN1 is expressed lowly and broadly, while RFTN2 is upregulated during neural crest cell development. By probing RFTN1 and RFTN2 expression in the avian embryo, we showed that this pattern holds true in vivo. We observed sparse and broad RFTN1 across the embryo with enrichment in somites, suggesting that RFTN1 may play a role in somitic development. However, we found that RFTN2 expression in neural crest tissues is dynamic with an increase in expression after the onset of specification that is maintained during early migration before being downregulated as these cells complete their migration. Our investigation also revealed RFTN2 expression in the intermediate and cranial mesoderm which like neural crest cells have migratory abilities, suggesting that RFTN2 may have specific roles in cell migration. These tissues must interpret signaling cues from their environments, implicating the Raftlin proteins as facilitators of cell signaling during the development of numerous embryonic tissues. While classically described for a mechanistic role in scaffolding lipid rafts, future experiments will need to test if this mode of action holds true in the developing contexts discussed herein. These results call for more extensive in vivo functional studies to determine the specific and redundant roles of these proteins during development. Importantly, this area of inquiry presents an opportunity to inform our understanding of lipid raft functions during in vivo development.

Materials and Methods

Sequence analysis

Bulk and single cell RNA sequencing analysis was performed using previously published datasets. Briefly, regularized log-transformed counts per million (rlog(CPM)) were plotted from chick embryo sequencing data using a publicly available R Shiny app (http://ash274.shinyapps.io/RNA-Seq_App/)(Hovland et al., 2022). Human ESC-derived neural crest cell sequencing data was downloaded from the Gene Expression Omnibus (GSE125145, (Leung et al., 2019)). Reads were trimmed using fastp to remove low quality reads and overrepresented sequences (Chen et al., 2018). Reads were then mapped to a human reference genome (GRCh38 version 41) using STAR (Dobin et al., 2013) and count matrices were made using featureCounts (Liao et al., 2014). Variance-stabilized transformation (VST) was performed using DESeq2 (Love et al., 2014), and VST normalized expression values were plotted as heatmaps using Pheatmap in R (Kolde, 2015). For single cell RNA sequencing analysis, a processed An Anndata object describing gene expression from sorted Wnt1-Cre/R26Tomato transgenic mouse embryos was downloaded (Poverennaya et al., 2026), converted into a Seurat object using SeuratDisk, and custom feature plots were then generated using Seurat in R (Satija et al., 2015). Sequence analysis code was produced using Claude Sonnet v4.5 (Anthropic) and all code was manually reviewed, tested, revised, and verified by the authors. R code used for plotting human and mouse data are available (https://github.com/PiacentinoLab/2026_Jenne_RFTN_Expression).

Chicken embryo collection

Fertilized white leghorn chicken eggs were purchased from commercial sources (University of Connecticut) and cultured at 38°C in humidified incubators until embryos reached the desired Hamburger Hamilton stage. At this point, embryos were dissected on filter paper into 1x Ringer’s solution, fixed for one hour at room temperature in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS,) rinsed in PBS with 0.1% Tween 20 (PBST), and dehydrated in a methanol/PBST series before storage in 100% methanol at −80°C. Representative expression patterns shown reflect consistent patterns observed in at least three embryos imaged per stage in whole mount, and sections are representatives from at least two sectioned embryos.

Hybridization chain reaction fluorescent in situ hybridization

Generation 3 hybridization chain reaction fluorescent in situ hybridization (HCR-FISH) labeling kits were purchased from Molecular Instruments (Choi et al., 2018). Probe sets were purchased directly from Molecular Instruments or custom designed using published pipelines and ordered as oPools (Integrated DNA Technologies)(Kuehn et al., 2022). Probe sets used in this study include RFTN2 (B3 initiator), RFTN1 (B5), PAX7 (B2), TFAP2B (B4), PAX2 (B2) and were detected using appropriate hairpin labeled with AlexaFluor488, AlexaFluor546, AlexaFluor647. Experiments were performed following manufacturer’s instructions. Briefly, dehydrated embryos were rehydrated in a methanol/PBST series, permeabilized in a Proteinase K (New England Biolabs, #P8107S) 0.4U/mL PBST solution for 1–2 minutes, optimized to stage/size of embryos, and post fixed for 20 minutes in 4% PFA in PBS. Embryos were washed twice in PBST, twice in 50% PBST/50% 5x saline sodium citrate 0.1% Tween 20 (SSCT), then twice for in 5x SSCT, each for five minutes at room temperature. Embryos were incubated in warmed probe hybridization buffer for 30 minutes at 37°C with rocking, then incubated overnight in probe hybridization buffer at 1:250 probe dilution at 37°C with rocking. Embryos were washed four times for 15 minutes in prewarmed (42 °C) probe wash buffer at 37°C, then washed twice for five minutes in 5x SSCT, then incubated in room temperature hybridization buffer for 30 minutes. Hairpins were snap-cooled separately by heating for 90 seconds at 95°C, then cooling at room temperature for 30 minutes. Embryos were incubated in a solution of amplification buffer at a 1:50 hairpin dilution overnight in the dark. Finally, embryos were washed 2 times for 5 minutes in 5x SSCT, post-fixed in 4% PFA in 5xSSC 0.1% Tween for 20 minutes at room temperature, washed 2 times for 5 minutes in 5x SSCT, washed twice for 30 minutes in 5x SSCT with 1:1000 dilution of 1 mg/mL DAPI on the first wash, 5 minutes in 5x SSCT.

Sectioning

To generate transverse sections after whole mount HCR-FISH, labeled embryos were washed for five minutes in 50% PBST/50% 5x SSCT, washed five minutes in PBST, incubated in 5% sucrose in PBS for 15 minutes at room temperature, 15% sucrose in PBS overnight at 4°C, then 7.5% gelatin at 38°C overnight. Embryos were embedded in molds and flash frozen in liquid nitrogen before −80°C storage. Embryos were subsequently cryosectioned at a thickness of 18–20 μm.

Imaging

Wholemount and transverse section imaging was performed using a Zeiss LSM900 Confocal with Airyscan2 Detector. Fluorescent wholemount and transverse section images were collected as Z stacks. Images were prepared for display using Fiji and are presented as median-filtered maximum intensity projections.

Supplementary Material

1

Fig S1. Labeling patterns for RFTN1 and RFTN2 HCR-FISH are not due to non-specific hairpin binding. Wild type embryos were processed for HCR-FISH to probe for expression of RFTN1 and PAX2 (A) or RFTN2, TFAP2B, and PAX2 (C). Raftlin gene probes were detected using AlexaFluor 647-labeled hairpin amplification. Negative control experiments were performed using stage-matched wild type embryos, processed for HCR-FISH as above including AlexaFluor 647-labeled hairpins, but without the corresponding Raftlin gene probes (B, D). Representative images from RFTN1-labeled (n=3/3), RFTN2-labeled (n=8/8), and 647 hairpin-only (n=6/6 across multiple stages) embryos. som, somite; NC, neural crest.

Highlights.

• Raftlin family of lipid raft scaffolds are dynamically expressed during development

• RFTN1 is expressed most strongly in the developing paraxial mesoderm

• RFTN2 is upregulated in neural crest cells during their migration