Summary
FGFs are key developmental regulators which typically direct cell proliferation, survival, and migration following the engagement of tyrosine kinase receptors. We find that coordinate loss of Fgfr1 and Fgfr2 in cranial neural crest cells results in facial clefting and mandibular defects with high levels of apoptosis, and that suppressing cell death alleviates the mutant phenotype. To identify critical downstream signaling pathways that regulate these processes, we generated allelic series of knock-in point mutations in each gene that disrupt binding of signaling effectors to the receptors, alone or in combination, as well as a kinase dead allele of Fgfr2 which broadly phenocopies the null mutant. While signaling mutations in either receptor, even when combined, failed to recapitulate the null mutant phenotypes, they revealed discrete roles for various pathways in regulating specific aspects of craniofacial development. We furthermore found that these signaling mutations together abrogate multiple established FGF-induced signal transduction pathways, while other FGF functions such as cell-matrix and cell-cell adhesion remain unaffected. Our studies establish combinatorial roles of both Fgfr1 and Fgfr2 in development and identify novel kinase-dependent cell adhesion properties for FGF receptors, beyond their well-established roles in intracellular signaling.
Introduction
The development of the face involves the coordination of multiple morphogenetic processes including the formation of the frontonasal, maxillary and mandibular processes, and their convergence at the midline. The pharyngeal arches (PA) appear on each side of the future head and neck and grow ventrally, and are composed largely of mesenchyme derived from cranial Neural Crest Cells (cNCCs), which surrounds a mesodermal core and is covered by surface ectoderm. The primitive mouth, or stomodeum, is flanked rostrally by the frontonasal prominence (FNP), laterally by the maxillary (Mx) processes, and caudally by the mandibular (Man) processes. Subsequently, the FNP widens, grows and bulges around the nasal placodes. By E10.5, rapid growth of the maxillary mesenchyme pushes the nasal pits medially. Closure of the midface initiates with the convergence of the medial (MNP) and lateral (LNP) nasal processes. Rapid growth of the Mx and LNP then pushes the MNP to converge.
NCCs differentiate into multiple cell types, including, cartilage, bone, neuronal, glial and smooth muscle cells (Bronner and LeDouarin, 2012). In the head, cNCCs collectively migrate through various mechanisms (Szabo and Mayor, 2018) and give rise to the majority of the bone and cartilage in the head and the face (Chai et.al. 2000; Couly et al., 1993). Paracrine signaling between the facial ectoderm and the underlying cNCC-derived mesenchyme is particularly important for craniofacial morphogenesis, implicating a number of signaling pathways. FGF8 has been shown to play an important role in regulating mandibular development (Shigetani et al., 2000; Trumpp et al., 1999) and midface integration (Griffin et al., 2013). In mammals, 22 FGFs have been identified by sequence homology, with 18 acting as secreted ligands for 4 FGF receptors (FGFR1-4) (Brewer et al., 2016; Ornitz and Itoh, 2015). All Fgfr genes have been disrupted in mice demonstrating both specific and redundant functions in vivo. Fgfr1 null mutants fail to gastrulate and exhibit a defect in epithelial to mesenchyme transition required for mesoderm formation (Ciruna and Rossant, 2001; Ciruna et al., 1997; Deng et al., 1994; Yamaguchi et al., 1994). Recent studies of Fgfr1 and Fgfr2 mutants however have documented an earlier genetic background-dependent role for Fgfr1 in primitive endoderm and trophectoderm development (Brewer et al., 2015; Hoch and Soriano, 2006; Kurowski et al., 2019), and a combined role for Fgfr1 and Fgfr2 in both of these lineages (Kang et al., 2017; Kurowski et al., 2019; Molotkov et al., 2017). Null mutants for Fgfr2 exhibit embryonic lethality at E10.5 associated with placenta deficiency and exhibit multiple additional defects including the absence of limb bud development (Molotkov et al., 2017; Xu et al., 1998b; Yu et al., 2003). Additional evidence further supports a role for Fgfr1 and Fgfr2 in craniofacial development, as conditional mutagenesis of Fgfr1 in cNCCs or of Fgfr2 in the epithelium leads to facial or palatal clefting (Brewer et al., 2015; Hosokawa et al., 2009; Rice et al., 2004; Wang et al., 2013) while deletion of both receptors in cNCCs prevents midface closure (Park et al., 2008). However, the signaling mechanisms by which FGFs regulate craniofacial development have not been elucidated.
Upon ligand binding, FGFRs dimerize which leads to transactivation of the kinase domain, subsequent phosphorylation of intracellular tyrosines, and binding of signaling effectors that in turn orchestrate activation of downstream signaling pathways (Brewer et al., 2016; Lemmon and Schlessinger, 2010). Different thresholds in dimer strength and stability also may come into play, as well as signaling dynamics engaged by each downstream signaling pathway (Li and Elowitz, 2019; Vasudevan et al., 2015; Zinkle and Mohammadi, 2018). The signaling pathways that operate downstream of FGFR1 have been particularly well studied (reviewed in (Brewer et al., 2016)). One of the most important signaling pathways induced by FGF signaling is ERK1/2, which is mainly thought to be activated by constitutive binding of adaptor proteins FRS2 and FRS3 to the juxtamembrane portion of FGFR1, leading to subsequent recruitment of GRB2 and SHP2 and activation of the MAPK pathway. However, GRB2 can also associate with the adaptor protein GAB1, which conversely promotes activation of PI3K/AKT (Kouhara et al., 1997; Ong et al., 2000; Ong et al., 2001; Xu et al., 1998a). In addition, CRKII or CRK-L binding to FGFR1 at Y463 also drives activation of ERK1/2, as well as JNK (Larsson et al., 1999; Moon et al., 2006; Seo et al., 2009). PLCγ binds FGFR1 at Y766, and regulates both receptor intracellular trafficking and PKC activation (Mohammadi et al., 1992; Mohammadi et al., 1991; Sorokin et al., 1994). Finally, GRB14 binding requires phosphorylation of FGFR1 Y766 and Y776 and negatively regulates PLCγ recruitment (Browaeys-Poly et al., 2010; Reilly et al., 2000). Signaling interactions with FGFR2 have been less well studied, but FRS2 and PLCγ have been confirmed to bind to FGFR2 at the juxtamembrane domain and Y769, respectively (Ceridono et al., 2005; Eswarakumar et al., 2006). CRK-L, but not CRK, binds to FGFR2 presumably at Y466 (Moon et al., 2006; Seo et al., 2009). Together, these studies indicate that FGF signaling converges on ERK1/2, as well as other downstream pathways. FGF signaling through these effectors regulates canonical RTK outputs, such as such as cell proliferation, survival, and migration (Brewer et al., 2016; Lemmon and Schlessinger, 2010). Beyond these canonical RTK signaling outputs however, FGF signaling is also known to regulate other cellular processes, through less well established mechanisms, such as cell-matrix (Meyer et al., 2012) or cell-cell adhesion (Kurowski et al., 2019; Rasouli et al., 2018; Sun and Stathopoulos, 2018). It remains unclear if all activities engaged by the FGF receptors are dependent on activation by FGFs, on signaling through the kinase domain, or if the receptors can engage cell adhesion receptors through interactions of their extracellular domains or by acting as scaffolds in specific cell surface compartments.
Although characterization of effector binding to RTKs provides critical insights on their signaling specificity, assessing their relative significance requires in vivo validation. Numerous lines of evidence point to ERK1/2 downstream of FRS proteins as a major FGF signaling output (Corson et al., 2003; Eswarakumar et al., 2006; Gotoh et al., 2004; Hadari et al., 2001; Lanner and Rossant, 2010; Vasudevan et al., 2015). An allele of Fgfr1 deficient for FRS2 and FRS3 binding by deletion of the juxtamembrane part of the receptor did not recapitulate the Fgfr1 null phenotype, however, suggesting that multiple developmental contexts do not require FRS dependent FGF signaling (Hoch and Soriano, 2006). PLCγ has been proposed to play a role as a negative regulator of FGF signaling (Partanen et al., 1998). To further study FGFR1 signaling pathways, we generated an allelic series of knock-in point mutation at the Fgfr1 locus that disrupt binding of multiple signaling effectors, alone or in combination (Brewer et al., 2015). Analysis of these signaling mutant alleles suggested that FGFR1 requires combinatorial signaling during development. Strikingly, the most severe Fgfr1 signaling mutants did not recapitulate the Fgfr1 null phenotype, despite almost eliminating ERK1/2 signaling, suggesting that additional signaling pathways remain to be identified. In vivo functions of FGFR2 signaling effectors are still largely unknown, however evidence to date indicates that signaling through FRS2 is not required by FGFR2 during development (Eswarakumar et al., 2006; Sims-Lucas et al., 2009).
To understand how FGF signaling regulates craniofacial development, we analyzed Fgfr1; Fgfr2 double conditional cNCC mutants. We found that loss of FGF signaling dramatically affects midface as well as mandibular development. To identify the downstream signaling pathways that are involved, we generated an allelic series of knock-in point mutations at the Fgfr2 locus that disrupt binding of FRS2, CRK-L, PLCγ and GRB14, alone or in combination. To our surprise, analysis of the Fgfr2 allelic series demonstrated that none of these signaling effectors are critically required downstream of FGFR2 during development. We furthermore intercrossed Fgfr1 and Fgfr2 mutants carrying the same signaling mutations to each other or to null mutants. These studies revealed that while various signaling pathways impair discrete aspects of craniofacial development, in the most severe combination they still did not recapitulate the double null phenotype. Analysis of intracellular pathways demonstrated that classical FGF-induced signal transduction pathways were abrogated in our mutants, while non-canonical cell-matrix and cell-cell adhesion functions of the FGF pathway remained unaffected even in the most severe signaling mutants and depended on kinase activation. Our results thus ultimately point to an unknown kinase-dependent output of FGFRs that regulates aspects of the cytoskeleton and cell adhesion and is important in embryonic development. Our studies further establish combinatorial roles of both Fgfr1 and Fgfr2 in craniofacial development and provide support for the notion that RTKs have adopted multiple functions in cells, beyond their well-established roles in intracellular cell signaling.
Results
Disruption of Fgfr1 and Fgfr2 in NCCs leads to defective craniofacial morphogenesis
We had previously generated Fgfr1GFP and Fgfr2mCherry reporter mice (Molotkov et al., 2017) that allowed us to identify spatial domains and overlap of Fgfr1 and Fgfr2 expression in E10.5 embryonic heads. Fgfr1 expression was observed primarily in the mesenchyme (Figure 1A, yellow arrow). In contrast, strong Fgfr2 expression was seen in the epithelia except for a small domain surrounding the nasal pit (Figure 1A, yellow asterisk). Weaker widespread Fgfr2 expression was also observed within the mesenchyme (Figure 1A, red arrow), indicating that both Fgfr1 and Fgfr2 are co-expressed in multiple regions (Figure 1A and Supplementary Figure S1A).
To interrogate FGFR1/2 functions, and to establish a baseline to study cell signaling mutations, we initially investigated how loss of both receptors in NCCs influence craniofacial development. To this end, conditional null alleles of Fgfr1 and Fgfr2 were combined with the Wnt1Cre driver that is active in NCCs as they delaminate from the neural tube (Danielian et al., 1998), and Wnt1Cre conditional mutants are referred to as cKO. Similar results were obtained with the Wnt1Cre2 driver (Lewis et al., 2013). Throughout this work, all Fgfr1, Fgfr2 and Cre driver alleles were analyzed on a 129S4 co-isogenic background, to avoid phenotypic variations that might be attributable to second-site modifiers. No overt defects were observed in Fgfr1cKO/+; Fgfr2cKO/+ double heterozygous or in Fgfr2cKO/cKO conditional mutants (Figure 1B and Supplementary Figure S1B). In contrast, conditional ablation of Fgfr1 (Fgfr1cKO/cKO) led to a fully penetrant facial cleft (Figure 1B) indicating a critical role for Fgfr1 rather than Fgfr2 in cNCC development.
Although Fgfr2cKO/cKO mutants did not display any overt phenotype, loss of Fgfr2 significantly enhanced the phenotype of Fgfr1cKO/cKO conditional mutants (Figure 1B and Supplementary Figure S1B). Both Fgfr1cKO/cKO and Fgfr1cKO/cKO; Fgfr2cKO/+ embryos exhibited defects in mandible development where proximal structures including angular and coronoid processes were hypomorphic (Supplementary Figure S1C). Conditional deletion of both Fgfr1 and Fgfr2 (Fgfr1cKO/cKO; Fgfr2cKO/cKO) led to severe agenesis of most NCC derived craniofacial structures including the frontal and nasal bones, nasal cartilage, maxilla, and mandible (Figure 1B and Supplementary Figure 1C). All Fgfr1cKO/cKO; Fgfr2cKO/cKO conditional mutants died perinatally. However, mesoderm-derived structures such as the parietal, interparietal and supraoccipital bones remained unaffected (Supplementary Figure S1B). Overall, we observed reduced or no ossification of all neural crest derived craniofacial skeletal structures in the double mutants at E18.5.
Next, we analyzed E9.5 and E10.5 Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG double mutants, where NCCs were GFP labeled, to investigate when noticeable defects emerge. Although GFP+ NCCs were distributed throughout their migration streams at E9.5, the PA1 and PA2 arches appeared hypoplastic in double mutants (Figure 1C, yellow arrows). By E10.5, Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG double mutants were morphologically identifiable with wider midline separation and hypoplastic MNP, LNP, maxillary and mandibular prominences (Figure 1C, yellow asterisk). A similar wide midline separation was also observed in Fgfr1cKO/cKO; Fgfr2cKO/+; ROSA26mT/mG mutants (Figure 1C, yellow asterisk). We analyzed molecular changes during morphogenesis at E10.5 using whole mount in situ hybridization. We examined expression of facial prominence markers, Alx3, Msx1, and Six3, along with midline morphogenesis markers, Shh and Nkx2.1. In Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants, we observed the absence of Alx3 and Six3 expression in the MNP (Figure 1D). Msx1 expression was significantly reduced (Figure 1F), suggesting a reduction in number of NCC cells in the double mutants. Along the midline, while Shh expression remained unaffected, Nkx2.1 expression was lost indicating a midline morphogenesis defect (Figure 1D). We also examined the expression of Fgf8 which is expressed in the ectoderm. Interestingly, Fgf8 was expressed in the Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants and was comparable to the Fgfr1cKO/+; Fgfr2cKO/+ controls (Figure 1D). These results indicate that the morphological defects observed upon loss of FGF receptors are not due to patterning defects, as the domains of Fgf8 and Shh expression at E10.5 remain unaffected, and instead suggest changes in the number of NCCs.
By E12.5, acute defects in midline integration were apparent, with a fully penetrant facial cleft (Figure 1E, yellow asterisk) and a severe mandibular defect (Figure 1E, yellow arrow). By E15.5, histological examination showed defective organogenesis in multiple organs, including the palate, tongue, and skeleton (Supplementary Figure S1D). Conditional double mutants also showed severe defects in trigeminal ganglion development. Whole mount staining with a neurofilament marker revealed that in Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants, the mandibular branch of trigeminal nerve failed to reach the anterior mandibular arch, while the maxillary branch showed fewer neural fibers compared to the Fgfr1cKO/+; Fgfr2cKO/+ control (Figure 1F). Interestingly, Fgfr1cKO/cKO; Fgfr2cKO/+ mutants, which exhibited a mild mandibular phenotype, did not show a trigeminal nerve defect (Figure 1F).
The anterior part of the craniofacial skeleton, including the maxilla and mandible, nasal cartilage, Meckel’s cartilage, frontal bone and anterior cranial base (ethmoid and sphenoid bones), are derived from NCCs. Analysis of alcian blue/ alizarin red stained skeletal preparations at E14.5 revealed that in contrast to the Fgfr1cKO/+; Fgfr2cKO/+ controls, Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants developed an anteriorly truncated skull due to loss of nasal cartilage (Figure 1G). Meckel’s cartilage, a transient cartilage template that directs the formation of bony mandible, was also severely affected and was completely missing in Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants (Figure 1G). The size of Meckel’s cartilage was significantly reduced in Fgfr1cKO/cKO; Fgfr2cKO/+ conditional mutants (Figure 1G, yellow arrow). We also observed absence of alizarin red staining in Fgfr1cKO/cKO; Fgfr2cKO/cKO mutant heads at this stage suggesting these defects are accompanied by reduced ossification. Defects in ossification were also observed in Fgfr1cKO/cKO; Fgfr2cKO/+ mutants which displayed an intermediate phenotype, specifically reduced alizarin red staining in the mandible and a complete absence of ossification in the maxilla (Figure 1G, red arrow). Taken together, these results suggest that FGF signaling plays an important role in skeletal differentiation of cNCCs.
Skeletal differentiation is a multi-step process starting with formation of cartilage progenitors which undergo maturation and eventually terminally differentiate into bone. To investigate the role of FGF receptors during skeletal differentiation of NCCs in the head we analyzed the expression of chondrogenic (Col2a1) and osteogenic markers (Col10a1 and RUNX2), first at an early stage at E14.5 (Supplementary Figure 1E) when cartilage progenitors are formed, and then at E17.5 (Supplementary Figure 1E) during which terminal differentiation is largely complete. Our analysis revealed that at E14.5 in Fgfr1cKO/+; Fgfr2cKO/+ control, anterior cranial base primordia express Col2a1, in a broader domain than Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants (Supplementary Figure 1E). This suggests that the skeletal differentiation program is initiated and cartilage progenitors are formed in Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants, but is probably delayed or attenuated. By E17.5, we observed similar levels of Col2a1 expression in both controls and Fgfr1cKO/cKO; Fgfr2cKO/cKO double mutants. However, expression of Col10a1, a terminal differentiation marker, was undetectable at this stage in the Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants while it was still expressed in broad domains in Fgfr1cKO/+; Fgfr2cKO/+ controls (Supplementary Figure 1E), suggesting that the terminal differentiation process during endochondral ossification is blocked. We also looked at expression of Runx2, a mature cartilage and bone differentiation marker. In our analysis, we found that Runx2 was expressed at similar levels in both controls and Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants at E17.5 (Supplementary Figure 1E). The extent of ossification was assessed by micro-CT at E18.5, which showed reduced ossification of NCC derived structures in the anterior skull as well as the mandible (Figure 1H). Overall, we conclude that FGF receptors play an important role during terminal differentiation of skeletal structures that primarily affects the anterior NCC-derived structures, including the anterior chondrocranium, the Meckel’s cartilage, the nasal cartilage and bone. The observation that FGF signaling is important for skeletal differentiation is consistent with previous studies of differentiation of long bones (Karuppaiah et al., 2016).
Increased apoptosis is observed in mutant facial primordia
We observed reduced fluorescence from GFP+ cells in the mid-face in Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG mutants at E9.5, which became more noticeable by E10.5 (Figure 2A), suggesting reduced numbers of NCCs in the midface. To further investigate this observation, we determined the percentage of GFP+ cNCCs at E10.5 using flow sorting of single cell suspensions from the five facial prominences for all genotypes. Compared to controls, Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG double mutants showed a 50% reduction in the number of GFP+ cells (Figure 2B). We did not observe a significant reduction in either Fgfr1cKO/+; Fgfr2cKO/cKO; ROSA26mT/mG or Fgfr1cKO/cKO; Fgfr2cKO/+; ROSA26mT/mG mutants, suggesting that the skeletal defects observed at E14.5 or E17.5 result from both reduction in NCC numbers as well as Fgfr1; Fgfr2 dependent differentiation of NCCs into skeletal lineages. A reduction in numbers of NC lineage cells in the midface might occur due to reduced proliferation, increased cell death, or both. Therefore, we first investigated the extent of cell proliferation by EdU incorporation at E10.5 in mutant facial primordia (Figure 2C-D). Compared to controls, we observed a similar extent of EdU incorporation in the Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG mutant NCC derived craniofacial mesenchyme (Figure 2C-D), suggesting that the extent of cell proliferation between control and Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG mutants remains unaffected. Next, we investigated cell death in Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG conditional homozygous double mutants. TUNEL staining revealed a striking increase in apoptosis in the craniofacial mesenchyme at E10.5 in Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants, most notably in the LNP relative to the MNP (Figure 2E-F). These observations suggested that cell-survival/ regulation of apoptosis downstream of FGF receptor signaling plays a crucial role during craniofacial development.
Although we found a clear increase in cell death in double mutants, we asked to what extent this observation could explain the overall morphological defects. The BH3-only protein BIM can bind to and repress the function of several pro-survival BCL-2 family members and therefore plays a critical role in initiating apoptotic pathway in multiple cell types (Chipuk and Green, 2008; Czabotar et al., 2014; Youle and Strasser, 2008). A recent report showed that mutations in the pro-survival genes Mcl-1 and Bcl-x leads to severe holoprosencephaly, which can be corrected by further loss of Bim, indicating that cell survival and cell death are finely balanced during development (Grabow et al., 2018). To investigate the role of cell survival and determine if a reduction of Bim levels can rescue craniofacial defects in Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants, we first compared Fgfr1cKO/cKO; Fgfr2cKO/+; Bim+/- mutants with control Fgfr1cKO/cKO; Fgfr2cKO/+ embryos at E17.5. Fgfr1cKO/cKO; Fgfr2cKO/+; Bim+/- mutants were obtained at expected ratios and exhibited a 55% reduction in midline separation in Fgfr1cKO/cKO; Fgfr2cKO/+; Bim+/- mutants compared to Fgfr1cKO/cKO; Fgfr2cKO/+ controls (Figure 2G), thus showing a partial rescue of craniofacial defects. We also observed a partial rescue of medial skeletal structures in Fgfr1cKO/cKO; Fgfr2cKO/+; Bim+/- embryos, where defects in anterior nasal cartilage and palatine process of the premaxilla were partially alleviated, as well as other medial structures including the primary and secondary palate, pterygoid process and basisphenoid bone. The reduction in Bim levels significantly rescued defects observed in Fgfr1cKO/cKO; Fgfr2cKO/cKO; Bim+/+ embryos, particularly at the level of the nasal mesenchyme and the midface (Figure 2H). Other morphometric parameters such as skull length and intercanthal distances also showed a partial rescue.
We next examined if reduced Bim levels can alleviate cell survival defect at E10.5. We used the TUNEL assay to assess apoptosis in embryonic heads at E10.5 on sections of Fgfr1cKO/cKO; Fgfr2cKO/cKO; Bim+/- and Fgfr1cKO/cKO; Fgfr2cKO/+; Bim+/- embryos. Compared to wild type levels of BIM, both Fgfr1cKO/cKO; Fgfr2cKO/cKO; Bim+/- and Fgfr1cKO/cKO; Fgfr2cKO/+; Bim+/- embryos showed reduced cell death (Supplementary Figure S1F). We observed an overall 25% reduction of cell death in Fgfr1cKO/cKO; Fgfr2cKO/+; Bim+/- embryos, as opposed to a 17% reduction in Fgfr1cKO/cKO; Fgfr2cKO/cKO; Bim+/- embryos. Interestingly, Bim heterozygosity resulted in a greater morphological rescue in Fgfr1cKO/cKO; Fgfr2cKO/cKO; Bim+/- mutants than in Fgfr1cKO/cKO; Fgfr2cKO/+; Bim+/- mutants, which already had less cell death. Taken together, these results underscore the importance of cell death in the etiology of the Fgfr1cKO/cKO; Fgfr2cKO/cKO conditional double mutant phenotype.
Generation of an allelic series of Fgfr2 signaling mutations
To interrogate FGFR2 signaling mechanisms in vivo, we generated an allelic series of knock-in point mutations at the Fgfr2 locus by gene targeting in ES cells (Figures 3). We introduced mutations in critical residues, preventing binding of signaling effectors to FGFR2 (Figure 3A-C) and verified targeting by PCR and Southern blots (Figure 3D). The Fgfr2F mutation (amino acid substitutions L424A and R426A), Fgfr2C mutation (amino acid substitution Y463F), and Fgfr2PG (amino acid substitutions Y769F and Y779F) mutation were designed to disrupt binding of FRS2, CRK-L and PLCγ/GRB14, respectively. We also generated compound Fgfr2CPG and Fgfr2FCPG signaling mutants by combining multiple signaling mutations. We provided biochemical validation of effector binding disruption for all signaling mutations. To validate the disruption of effector binding, 3T3 cells were transfected with triple Flag-tagged cDNAs of FGFR2c isoforms for each signaling mutants. We confirmed disruption of FRS2, CRK-L and PLCγ binding in Fgfr2PG, Fgfr2CPG, Fgfr2F and Fgfr2FCPG mutations, respectively, via co-immunoprecipitation and western blot analysis (Figure 3B).
Fgfr2 signaling mutant alleles were evaluated for their ability to partially or completely recapitulate the Fgfr2-/- E10.5 phenotype associated with placenta and limb deficiencies (Molotkov et al., 2017; Xu et al., 1998b; Yu et al., 2003). Surprisingly, all signaling allele mutants were at least partially viable and fertile as homozygotes (Table 1). We observed a decreased growth rate for Fgfr2F/F, Fgfr2PG/PG, and Fgfr2FCPG/FCPG mutant mice compared to controls (Supplementary Figure S2A). Interestingly, the growth retardation observed for Fgfr2PG/PG mutants was rescued by the concomitant disruption of CRK-L binding site in Fgfr2CPG/CPG mutants, suggesting opposite roles for these effectors in mediating FGFR2 signaling. Skeletal preparations at birth revealed a kinked tail phenotype for Fgfr2PG/PG (7/9) and Fgfr2CPG/CPG (6/12) neonates, the most extreme cases resulting in a curly tail phenotype (Supplementary Figure S2B).
To further identify in vivo phenotypes associated with our signaling mutations, we crossed Fgfr2 signaling mutant mice with the null allele. Fgfr2C/-, Fgfr2PG/- and Fgfr2CPG/- mice were viable, with growth retardation apparent in Fgfr2PG/- mutant mice (not shown). In contrast, hemizygous Fgfr2F/- and Fgfr2FCPG/- mutant mice were recovered in expected Mendelian ratios at E18.5, but died at birth (Table 2). None of the Fgfr2F/- and Fgfr2FCPG/- neonates were able to suckle, as evidenced by the absence of an abdominal milk spot. This may be a consequence of cranial nerve defects, since Fgfr2F/- and Fgfr2FCPG/- E10.5 mutant embryos exhibited decreased trigeminal nerve projections into facial prominences compared to control littermate (Supplementary Figure S2C).
Both Fgfr2F/F and Fgfr2FCPG/FCPG mice develop periocular lesions in the eye as early as postnatal day P15 that worsened over time. Similar phenotypes have been associated with defects in lacrimal glands, which provide protection to the non-keratinized epithelial surface of the eye (Inaba et al., 2018). Upon further investigation, we indeed observed a defect in lacrimal gland development in both Fgfr2F/F and Fgfr2FCPG/FCPG mice (Supplementary Figure S2D). In mice, lacrimal gland development starts at E13.5 by an epithelial invagination into the surrounding mesenchyme, and progresses by branching morphogenesis to become a fully functional organ by P7. FGF10-FGFR2 signaling plays a critical role during this process where it regulates proliferation in epithelial cells (Garg et al., 2017; Steinberg et al., 2005). Both Fgfr2F/F and Fgfr2FCPG/FCPG mutants showed loose clusters of acinar cells, which populate the distal end of the ducts, and occupied a much smaller area at P7. A significant reduction in the size of the lacrimal gland was also brought about by reduced number of branches and smaller lengths of the tubes in the Fgfr2F/F and Fgfr2FCPG/FCPG signaling mutants (Supplementary Figure S2E and S2F).
The observation that all Fgfr2 signaling mutants were viable and the fact that the Fgfr2FCPG/FCPG mutants do not recapitulate the Fgfr2-/- phenotype raised the possibility that a critical downstream adaptor might still interact with the FGFR2FCPG receptor. We data-mined a recent proteomic screen identifying FGFR2b dependent phosphorylation events (Francavilla et al., 2013) and identified IRS2 (insulin receptor substrate) as a possible new putative FGFR2 candidate binding partner. IRS2 belongs to the same superfamily of adaptor protein as FRS2 and shares a similar protein architecture, with membrane targeting and PTB (phosphotyrosine binding) domains, as well as a C-terminal tail containing multiple tyrosine phosphorylation sites (Supplementary Figure S2G). We found that IRS2 binds weakly to both WT and FGFR2FCPG receptors in primary MEFs, independent of FGF stimulation (Supplementary Figure S2G and S2H). However, we failed to observe a genetic interaction between Irs2-/- and Fgfr2FCPG/FCPG mutant mice at E10.5, E13.5 or at birth (data not shown), indicating that IRS2 is a not a critical missing effector of FGFR2 signaling in vivo.
Coordinate roles of signaling pathways in development
Our results indicate that signaling mutations in Fgfr2 do not lead to overt craniofacial defects. Likewise, a previous study focusing on Fgfr1 harboring a similar allelic series found that Fgfr1F/cKO and Fgfr1FCPG/cKO mutants showed only a low incidence of cleft palate (Brewer et al., 2015). Because Fgfr1 and Fgfr2 have a significant degree of co-expression in NCC-derived mesenchymal cells (Figure 1A), we reasoned that discrete functions of signaling pathways that operate downstream of one receptor could be masked by the presence of the other, wild-type receptor. To test this hypothesis, we examined compound conditional hemizygous Fgfr1F, Fgfr1FCPG and Fgfr2F mutations over the Fgfr1- and Fgfr2- null alleles in the context of craniofacial development. We were unable to perform a conditional mutation for the Fgfr2FCPG allele due to the serial retention of multiple lox sites during the generation of this allele (Figure 3C).
To characterize discrete phenotypes resulting from signaling mutations in Fgfr1 (Fgfr1F or Fgfr1FCPG alleles), we crossed Fgfr1F/cKO; Fgfr2cKO/+ or Fgfr1FCPG/cKO; Fgfr2cKO/+ males to Fgfr1cKO/cKO; Fgfr2cKO/cKO females. We examined Fgfr1F/cKO; Fgfr2cKO/cKO or Fgfr1FCPG/cKO; Fgfr2cKO/cKO (Figure 4A) embryos at E16.5 for skeletal defects. Compared to Fgfr1cKO/+; Fgfr2cKO/cKO controls, Fgfr1F/cKO; Fgfr2cKO/cKO and Fgfr1FCPG/cKO; Fgfr2cKO/cKO conditional mutants developed severe agenesis of the midface structures, but the phenotype was not as severe as in Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants. Notably, the nasal cartilage and the mandible were more severely affected in Fgfr1FCPG/cKO; Fgfr2cKO/cKO conditional mutants compared to Fgfr1F/cKO; Fgfr2cKO/cKO conditional mutants (Figure 4A, red arrow). A similar analysis was performed for Fgfr2F allele to examine its phenotype (Figure 4A). We examined Fgfr1cKO/cKO; Fgfr2F/cKO embryos and found that they developed more severe midline fusion and mandible defects than Fgfr1F/cKO; Fgfr2cKO/+ controls. Anterior skeletal structures such as the nasal cartilage, premaxilla and maxilla, and sphenoid bones were either severely reduced or absent. A striking reduction of the mandible was observed along with a complete loss of Meckel’s cartilage (Figure 4A). For a more detailed phenotypic analysis, we performed micro-CT analysis for Fgfr1F/cKO; Fgfr2cKO/cKO, Fgfr1FCPG/cKO; Fgfr2cKO/cKO and Fgfr1cKO/cKO; Fgfr2F/cKO conditional mutants (Supplementary Figure S3A). Several structures were found to be differentially affected in Fgfr1F/cKO; Fgfr2cKO/cKO versus Fgfr1FCPG/cKO; Fgfr2cKO/cKO (summarized in Table 3). In all these cases non-neural crest derived structures such as the parietal, interparietal and supraoccipital bones remained unaffected.
Similar to Fgfr1cKO/cKO; Fgfr2cKO/cKO mutants, morphological defects in Fgfr1F/cKO; Fgfr2cKO/cKO and Fgfr1FCPG/cKO; Fgfr2cKO/cKO embryos arose as early as E10.5. Fgfr1FCPG/cKO; Fgfr2cKO/cKO embryos and, to a lesser extent, Fgfr1F/cKO; Fgfr2cKO/cKO embryos developed hypoplastic nasal prominences (Supplementary Figure S3B and S3C, yellow arrow) and mandibular prominences (Supplementary Figure S3B and S3C, red arrow) along with a midfacial cleft (Supplementary Figure S3B and S3C, yellow asterisk) compared to the littermate controls. When we examined cell death, we observed LNP specific TUNEL positive foci in both Fgfr1FCPG/cKO; Fgfr2cKO/cKO and Fgfr1F/cKO; Fgfr2cKO/cKO (Supplementary Figure S3D and S3E), as we had observed in Fgfr1cKO/cKO; Fgfr2cKO/cKO double conditional mutants.
We next examined the role of specific signaling pathways that may be engaged coordinately by both receptors upon ligand activation. Fgfr1C/C; Fgfr2C/C and Fgfr1CPG/CPG; Fgfr2CPG/CPG double mutants were recovered in normal numbers, were fertile, and did not exhibit a craniofacial phenotype. We therefore focused our attention on the F and FCPG mutations. At E10.5, Fgfr1F/-; Fgfr2F/- compound mutants showed growth retardation compared to Fgfr1F/+; Fgfr2F/+ mutant embryos. The phenotype was more severe in Fgfr1F/-; Fgfr2-/- embryos (Supplementary Figure S3F). We observed phenotypic variability in these mutants even though the study was conducted on the same co-isogenic 129S4 genetic background. Axis truncation along with a defect in somite patterning was previously observed in Fgfr1FCPG/FCPG mutants (Brewer et al., 2015). Fgfr1FCPG/- mutant mice also showed growth retardation and axial patterning defects (data not shown). A similar, but more severe phenotype was observed for Fgfr1FCPG/-; Fgfr2-/- mutants compared to Fgfr1F/-; Fgfr2-/- (Supplementary Figure S3G). In both compound mutants, developmental delays were more acute and severe disruption of posterior structures was observed. Fgfr1FCPG/-; Fgfr2-/- mutants had more severe defects in somite formation and patterning compared to Fgfr1F/-; Fgfr2-/- mutants (data not shown). Both Fgfr1F/-; Fgfr2-/- and Fgfr1FCPG/-; Fgfr2FCPG/- were obtained at very low frequency and craniofacial defects were difficult to characterize as structures were either growth retarded or did not develop.
We did not recover Fgfr1F/F; Fgfr2F/F mutant embryos at E10.5. Interestingly, we did not observe craniofacial defects in Fgfr1F/F embryos at E10.5. However, compound Fgfr1F/F; Fgfr2+/F mutants displayed severe hypoplastic nasal prominence defects but mandibular development was not affected (Supplementary Figure S3F). A similar approach was taken to analyze Fgfr1FCPG and Fgfr2FCPG compound signaling mutants. Earlier we reported Fgfr1FCPG/FCPG mutant mice showed embryonic lethality at E10.5 with severe posterior truncations (Brewer et al., 2015). We observed similar defects in Fgfr1FCPG/FCPG; Fgfr2+/+ and Fgfr1FCPG/FCPG; Fgfr2+/FCPG compound mutations. Morphological analysis of embryonic heads at E10.5 showed that Fgfr1FCPG/FCPG; Fgfr2+/FCPG mutants developed hypoplastic nasal and mandibular prominences (Supplementary Figure S3G). Fgfr1FCPG/FCPG; Fgfr2FCPG/FCPG mutants survived until E7.5 but were retarded compared to control littermates. They nonetheless still formed mesoderm, as evidenced by Eomes staining (Figure 4C), in contrast to double Fgfr1-/-; Fgfr2-/- mutants which fail at implantation (Kurowski et al., 2019). These results indicate that signaling mutations in both Fgfr1 and Fgfr2 genetically interact during development, but that the combination of the most severe signaling mutations fail to recapitulate the double null mutant phenotype.
Signaling mutations disrupt multiple intracellular pathways
FGF signaling activates numerous signaling pathways upon ligand stimulation (Brewer et al., 2016). We performed a preliminary time course for pathway activation using primary Fgfr2+/+ cells at E10.5. In these cells, FGF1 led to robust ERK1/2 activation, however, the amplitude of the response was lowered by half with FGF-8b (Supplementary Figure S4A). We therefore used FGF1 for further analysis since this ligand gave the more robust response. To evaluate intracellular pathway activation downstream of wild-type FGFR2 and the FGFR2F, FGFR2CPG and FGFR2FCPG signaling mutant receptors, we generated immortalized E10.5 FNP cell lines from the respective mouse strains also carrying a mutation in Ink4A/Arf to facilitate rapid immortalization of these cells. Similar to a previous study with palatal mesenchymal cells (Fantauzzo and Soriano 2017), we found that expression of facial mesenchyme markers was similar between primary and immortalized FNP cells (iFNPs; Supplementary Figure S4B). We isolated iFNPs for respective genotypes (wt-iFNP, F-iFNP, CPG-iFNP and FCPG-iFNP). These cells express undetectable levels of Fgfr3 and Fgfr4, and express predominantly Fgfr1 and, to a lesser extent, Fgfr2 (data not shown). To interrogate signaling functions that operate specifically downstream of FGFR2, we then eliminated FGFR1 expression in each cell line by CRISPR/Cas9 mutagenesis, leaving FGFR2 as the only receptor expressed (Supplementary Figure S4C).
Using cells derived from each signaling mutant, we set forth to interrogate activation of six pivotal pathways, ERK1/2, PI3K/AKT, PLCγ, p38, STAT and JNK, known to be engaged by FGF signaling (Brewer et al., 2016). Upon stimulation with FGF1, we observed robust ERK1/2 activation in wt-iFNPs. ERK1/2 activation was diminished by around half for either Fgfr2F/F or Fgfr2CPG/CPG iFNPs and was only eliminated in Fgfr2FCPG/FCPG mutant cells compared to wt-iFNPs (Figure 5). Our data suggests therefore that both FRS2 and CRKL/PLCγ binding is necessary for ERK1/2 activation, similar to previous observations with Fgfr1 (Brewer et al., 2015). PI3K/AKT signaling peaked following 5 minutes of FGF1 treatment in wt-iFNP cells. A similar response was also observed in F-iFNP cells. In CPG-iFNPs, a slight reduction was observed compared to either wt-iFNP or F-iFNP cells. AKT activation was reduced to a background level only in iFNP cells derived from Fgfr2FCPG/FCPG mutants (Figure 5). Therefore, CRKL and PLCγ binding appear to have more of an effect on AKT activation downstream of FGFR2 than interaction with FRS2. However, again loss of FRS2, CRKL and PLCγ binding together was necessary to abrogate PI3K/AKT activation. We observed a robust PLCγ activation upon FGF1 treatment in wt-iFNP cells (Figure 5). We observed that PLCγ activation was only slightly reduced in F-iFNP cells. However, in FCPG and CPG mutants FGFR2 dependent PLCγ activation was reduced to background levels (Figure 5). Therefore, our results indicate that a subset of this pathway activation is FGFR2-FRS2 dependent, while for the most part interaction with Tyr769 plays a critical role. P38 activation downstream of FGF signaling is known to play an important role in vasculature development and STAT3 is involved in chondrocyte proliferation. Interestingly, a longer and sustained activation of p38 and pSTAT was observed in FGF1 treated wt-iFNP cells over a period of 60 mins. However, F-iFNPs, CPG-iFNPs and FCPG-iFNPs all failed to show activation of both P38 and pSTAT suggesting P38 and pSTAT activation require FRS2, CRKL and PLCγ for an appropriate response (Figure 5). Last, FGF receptors can signal through CRK or CRK-L to recruit Cas and activate the C3G-RAP1 pathway and JNK serine threonine kinases, independently of Ras. We observed robust JNK activation in wt-iFNP cells suggesting FGFR2 can also activate JNK upon FGF1 stimulation. We also saw dampened activation of the pathway in both F-iFNP and CPG-iFNP cells while JNK activation was lost in FCPG-iFNP cells (Figure 5). Our analysis suggests that CRK-L binding play important roles in JNK activation, but that engagement of FRS2 is also required in this pathway.
In a previous report (Brewer et al., 2015), we analyzed FGFR1 signaling output upon FGF1 stimulation. We found that FNP cells from Fgfr1FCPG/FCPG mutants resulted in a near total reduction of ERK1/2 response. However, the signaling output of FGFR1FCPG was not investigated in a Fgfr2 null background. We used CRISPR/Cas9 to create Fgfr2 null iFNP cells from Fgfr1FCPG/FCPG; Ink-/- embryos. Upon FGF1 stimulation, we observed low pERK activation in Fgfr1FCPG/FCPG iFNP cells which peaks at 2 mins. In Fgfr2CRISPR-KO; Fgfr1FCPG iFNP cells, the overall amplitude of ERK1/2 activation was lower than in Fgfr1FCPG/FCPG iFNP, approximating the background level of activation (Supplementary Figure S4D). pAKT activation was undetectable in both Fgfr2CRISPR-KO; Fgfr1FCPG and Fgfr2+/+; Fgfr1FCPG iFNP cells (Supplementary Figure S4D and E). These results indicate that the most severe signaling mutations combinations in Fgfr1 and Fgfr2 effectively abrogate classic signal transduction pathways for both receptors.
A kinase dead mutation in Fgfr2 exhibits Fgfr2 null and additional phenotypes
The lack of a more severe phenotype in Fgfr2 signaling mutants prompted the question of whether FGFR2 predominantly exerts its functions in a kinase dependent fashion. To address this question, we generated a kinase dead (KD) mutation at the Fgfr2 locus in which Lys517 in the ATP binding site was converted to an alanine (Bellot et al., 1991; Hanks et al., 1988) (Figure 3A, 3C and 3D). A correctly sized FGFR2KD protein of ∼130Kd was found to be expressed in Fgfr2KD/KD protein lysates (Figure 3E). Although no morphological defects besides a partially penetrant kinked tail phenotype were observed at birth, fewer than expected (24/78) Fgfr2KD/+ heterozygotes were recovered when Fgfr2KD/+ heterozygous males were crossed to wild type females. Fgfr2KD/+ heterozygous embryos were recovered in normal Mendelian ratios at E10.5 (27/60) and showed no obvious morphological defects in LNP and MNP, maxillary or mandibular prominences. During later developmental stages up to P0, Fgfr2KD/+ heterozygotes also appeared normal with no limb or craniofacial abnormalities. Interestingly, Fgfr2KD/+ heterozygotes exhibited a semi-dominant phenotype by postnatal day 15 (P15), when characteristic peri-ocular defects started to appear. This defect was significantly more severe than that observed in Fgfr2F/F and Fgfr2FCPG/FCPG homozygous signaling mutants. By P21, this phenotype was aggravated (Supplementary Figure S2D) and was characterized by reduced branching in the lacrimal glands along with fewer loosely held acinar cell clusters (Supplementary Figure S2E and S2F).
Upon intercrossing of Fgfr2KD/+ heterozygotes, we did not recover Fgfr2KD/KD homozygotes at birth (Table 1). Since Fgfr2-/- homozygous mutants are embryonic lethal at E10.5, we undertook analysis at this stage. 6/36 (17%) Fgfr2KD/KD embryos resulting from Fgfr2KD/+ heterozygote intercrosses were recovered at E10.5. Morphological examination of Fgfr2KD/KD embryos showed a characteristic absence of limb buds, a defect in the allantois which was loosely or incompletely held by the chorion to the ectoplacental cone, and a dilated pericardium (Figure 6A). These defects phenocopy those observed in Fgfr2-/- homozygous mutants at this stage. Upon further examination of Fgfr2KD/KD embryos however, we observed additional phenotypes including severe posterior truncations and a defect in forebrain development, that resulted in a rounder head compared to Fgfr2KD/+ heterozygote littermates (Figure 6A). Craniofacial defects affected the FNP, as well as maxillary and mandibular prominences (Figure 6A).
To further evaluate differences in severity between the Fgfr2KD and Fgfr2- alleles, we performed an allelic complementation study by intercrossing Fgfr2KD/+ and Fgfr2+/- heterozygotes. At E10.5, 18/79 (23%) Fgfr2KD/- embryos were recovered, but six of these did not show a heartbeat. In parallel, we intercrossed Fgfr2+/- heterozygotes, resulting in a similar recovery (8/36; 22%) of Fgfr2-/- mutant embryos at E10.5. Taken together, these results demonstrate that both Fgfr2KD and Fgfr2- alleles exhibit lethality at similar stages. Morphologically, Fgfr2KD/- embryos showed an absence of limb buds and defects in the chorio-allantoic junction along with a dilated pericardium, similar to Fgfr2-/- embryos (Figure 6A). We also observed posterior truncations defects in Fgfr2KD/- mutants, similar to Fgfr2KD/KD mutant embryos. In contrast, defects in the forebrain, medial and lateral nasal prominences, and maxillary and mandibular prominences of Fgfr2KD/- mutants appeared less severe than in Fgfr2KD/KD mutants.
We used mRNA in situ hybridization to characterize the expression of Msx1 (Figure 6B) and Nkx2.1 (Figure 6C) in the facial prominences of Fgfr2-/-, Fgfr2KD/-, and Fgfr2KD/KD mutants at E10.5. We observed robust expression of Msx1 mRNA in Fgfr2-/- and Fgfr2KD/- mutants, but it was reduced significantly in Fgfr2KD/KD homozygous mutants. Nkx2.1 mRNA was expressed in the floor plate along facial midline in Fgfr2-/- mutants, however, it was severely affected in both Fgfr2KD/KD and Fgfr2KD/- mutants, most probably due to a semi-dominant effect of Fgfr2KD allele. We investigated Meox1 mRNA expression (Figure 6D) to assess axial truncation defects in Fgfr2KD/KD and Fgfr2KD/- mutants. Compared to Fgfr2-/- mutants, both Fgfr2KD/KD and Fgfr2KD/- mutants showed reduction in Meox1 expression and reduced somite numbers. Axial defects were more severe in Fgfr2KD/KD homozygous mutants (Figure 6A). Taken together, these results indicate that the Fgfr2KD mutation affects the same processes as the Fgfr2- allele, indicating that overall FGFR2 exerts its action in a kinase dependent way. The Fgfr2KD mutation also exhibited additional semi-dominant effects during embryonic and postnatal development.
Cell matrix and cell adhesion properties are retained in FCPG relative to null mutants
Since signaling mutant cells showed near complete inactivation of classic RTK signaling activities, but the corresponding mutant mice failed to recapitulate the null mutant phenotype, we reasoned that some function engaged by FGF signaling must still be retained in the most severe FCPG mutants. Previous lines of evidence have implicated FGF signaling in the control of cell migration and adhesion. To further investigate the role of FGF signaling in directing cell movement, we subjected starved primary FNP cells to a transwell migration assay (Figure 7A). We observed that both PDGF (a known chemoattractant) as well as FGF treatment stimulated migration of control cells to an extent comparable to serum. Fgfr1FCPG/cKO; Fgfr2cKO/cKO cells migrated upon PDGF stimulation similar to serum, however these cells showed severe migration defects upon FGF1 stimulation indicating that this process depends on active FGFR1-induced signal transduction. Fgfr1cKO/cKO; Fgfr2cKO/cKO double-null mutant FNP cells also did not migrate efficiently upon FGF1 stimulation, as expected due to lack of FGF receptors. However, these cells migrated in serum and PDGF stimulated conditions, albeit to an extent lesser than control cells.
We next examined Fgfr1; Fgfr2 dependent cell migration on extracellular matrix, using a scratch/ wound assay. Primary FNP cells harvested from E11.5 facial mesenchyme of either control (Fgfr1cKO/+; Fgfr2cKO/+), Fgfr1FCPG/+; Fgfr2cKO/+, or Fgfr1cKO/cKO; Fgfr2cKO/cKO embryos were observed upon FGF and PDGF stimulation. Spreading of control cells over the wound during a twelve-hour period was comparable upon FGF, PDGF, and serum stimulated conditions. Interestingly, Fgfr1FCPG/cKO; Fgfr2cKO/cKO cells also showed comparable spreading capacities to control cells in all three stimulation conditions. However, Fgfr1cKO/cKO; Fgfr2cKO/cKO double mutant cells failed to spread within this time frame into the wound area in response to FGF, while responses to PDGF and serum were normal (Figure 7B).
Defects in migration arise from impaired focal adhesion formation during cell spreading and cell-matrix interaction. However, it is unclear what role FGF signaling play during this process. A forward scatter plot of freshly harvested FNP cells remained unchanged across all genotypes (Supplementary Figure S5A), indicating that cell spreading defects could not be attributed to differences in cell size. We then looked at Paxillin localization by immunofluorescence as cells spread in Fgfr1cKO/+; Fgfr2cKO/+; ROSA26mT/mG, Fgfr1FCPG/cKO; Fgfr2cKO/cKO; ROSA26mT/mG, and Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG double null mutant FNP cells upon treatment with FGF, PDGF, or serum over a three-hour period. We observed that both GFP+ Fgfr1cKO/+; Fgfr2cKO/+; ROSA26mT/mG control cells and Fgfr1FCPG/cKO; Fgfr2cKO/cKO; ROSA26mT/mG mutant cells formed numerous Paxillin enriched focal adhesions during PDGF and FGF stimulated cell-spreading, that resembled serum enriched conditions. Double null mutant cells however failed to form any Paxillin+ foci upon FGF treatment and subsequent cell spreading, but still responded normally to PDGF and serum (Figure 7C and Supplementary Figure S5B). Western blot analysis showed that FAK, a key regulator involved in maturation of focal contacts, was expressed at similar levels across all genotypes; however, we observed reduced pFAK levels in freshly harvested Fgfr1cKO/cKO; Fgfr2cKO/cKO double mutant cells compared to Fgfr1cKO/+; Fgfr2cKO/+ control cells (Supplementary Figure S5C and S5D). Interestingly, double null mutant cells also showed reduced levels of Paxillin (Supplementary Figure S5C and S5D). In contrast, Fgfr1FCPG/cKO; Fgfr2cKO/cKO mutant FNP cells expressed pFAK at levels comparable to Fgfr1cKO/+; Fgfr2cKO/+ control cells (Supplementary Figure S5C and S5D). We next investigated if FGF was also necessary for activation of pFAK in Fgfr1FCPG/cKO; Fgfr2cKO/cKO mutant cells. To this end, we treated Fgfr1cKO/+; Fgfr2cKO/+ control, Fgfr1FCPG/cKO; Fgfr2cKO/cKO, and Fgfr1cKO/cKO; Fgfr2cKO/cKO double mutant FNP cells with serum or FGF1 before analyzing FAK activation. pFAK levels were increased to similar levels in Fgfr1FCPG/cKO; Fgfr2cKO/cKO cells as in Fgfr1cKO/+; Fgfr2cKO/+ control cells, upon serum or FGF stimulation (Supplementary Figure S5E). Fgfr1cKO/cKO; Fgfr2cKO/cKO double null mutant cells failed to show a significant response upon FGF1 stimulation, likely due to absence of FGF receptors (Supplementary Figure S5E). These results parallel the immunofluorescence observations and indicate that cell spreading and stabilization of cell-matrix interaction are actively governed by FGF signaling to its receptors. Interestingly, FGFR2FCPG still retained this residual function and responded to FGF ligand stimulation by forming focal adhesions.
Next, we analyzed cell spreading properties in Fgfr2FCPG/FCPG-iFNP cells (Fgfr1CRISPR-KO Fgfr2FCPG/FCPG) in which Fgfr1 was inactivated. Addition of either serum, PDGF or FGF resulted in robust cell spreading and formation of Paxillin+ focal adhesion in Fgfr1CRISPR-KO Fgfr2FCPG/FCPG and Fgfr1CRISPR-KO Fgfr2+/+-iFNP cells (Figure 7C and Supplementary Figure S5F). Fgfr1CRISPR-KO Fgfr2-/--iFNP cells failed to show a similar response (Figure 7C and Supplementary Figure S5F), similar to what we observed in Fgfr1FCPG/cKO; Fgfr2cKO/cKO mutant primary FNP cells upon FGF1 treatment. These observations indicate that although FGFR1FCPG and FGFR2FCPG lose most FGF1 dependent intracellular kinase signaling outputs, they still retain functions pertaining to cell-matrix interactions. Taken together, our results suggest that FGFR1/2 regulate several cell biological processes in NCCs pertaining to cell-matrix adhesion and migration.
We were also curious to know if Fgfr1FCPG cells could form stable cell-cell contacts comparable to control cells. We cultured freshly harvested primary GFP+ FNP cells from Fgfr1cKO/+; Fgfr2cKO/+; ROSA26mT/mG control embryos and Fgfr1FCPG/cKO; Fgfr2cKO/cKO; ROSA26mT/mG embryos and compared their behavior to Fgfr1cKO/cKO; Fgfr2cKO/cKO; ROSA26mT/mG double null mutant cells. Both Fgfr1cKO/+; Fgfr2cKO/+; ROSA26mT/mG control cells and Fgfr1FCPG/cKO; Fgfr2cKO/cKO; ROSA26mT/mG cells formed extensive cell-cell contacts in culture (Figure 7D and Supplementary Figure 5G). We observed that both control and Fgfr1FCPG/cKO; Fgfr2cKO/cKO; ROSA26mT/mG mutant cells formed extensive adherens junctions, marked by localized β-catenin along cell boundaries (Figure 7E). In contrast, double null mutant cells formed far fewer cell-cell contacts with no localized β-catenin accumulation, suggesting that the contacts are either unstable or that they do not mature (Figure 7E, red arrows). Last, we examined NCC cell-cell contacts in vivo, in the LNP. GFP+ cells in the mesenchyme maintained extensive cell-cell contacts within the LNP in both control and Fgfr1FCPG/cKO; Fgfr2cKO/cKO; ROSA26mT/mG at E11.5. Strikingly, GFP+ cNCCs in the double null mutant LNP were mostly isolated and interspersed (Figure 7F). Across all genotypes, cell contacts in the MNP remained unaffected (Figure 7F). Taken together, these results indicate that the most severe signaling mutations in Fgfr1 and Fgfr2 still retain cell-matrix and cell-cell interactions, while abrogating classic signal transduction pathways.
Discussion
Previous work has implicated FGF signaling as a critical regulator of craniofacial development. FGF8 conditional null or hypomorphic mutants exhibit defects in midface integration and mandible development (Griffin et al., 2013; Trumpp et al., 1999). Fgfr1 conditional mutagenesis in cNCCs leads to facial clefting (Brewer et al., 2015; Wang et al., 2013), and constitutive Fgfr2b or conditional Fgfr2 mutants in the palate epithelium display a cleft palate (Hosokawa et al., 2009; Rice et al., 2004). Previous work from our laboratory and others has investigated the signaling pathways by which FGFs operate in vivo (Brewer et al., 2015; Eswarakumar et al., 2006; Hoch and Soriano, 2006). Although signaling through ERK1/2 has been widely thought to be a predominant pathway through which FGFs operate (Brewer et al., 2016; Lanner and Rossant, 2010), our analysis of an allelic series of signaling mutations in Fgfr1 failed to recapitulate the null mutant phenotype, despite abrogating ERK1/2 signaling (Brewer et al., 2015). In this work, we sought to address questions raised by these findings, focusing on the relative roles for FGFR1 and FGFR2 in craniofacial development and the signaling mechanisms by which these receptors operate, and to investigate if there are non-canonical functions of FGF signaling beyond known engaged pathways that may reconcile the gap in our phenotypic analyses.
The fact that Fgfr1c mutants recapitulate many aspects of the Fgfr1-/- phenotype (Partanen et al., 1998), and that Fgfr2b mutants are reminiscent of Fgfr2-/- embryos (De Moerlooze et al., 2000), has led to the view that FGFR1 and FGFR2 predominantly function in mesenchymal or epithelial contexts, respectively. In this work, we documented extensive co-expression of the two receptors in the cNCC derived mesenchyme and in the overlying epithelia, using fluorescent Fgfr1 and Fgfr2 reporter alleles. We showed that both receptors function coordinately within the neural-crest derived mesenchyme as combined loss of Fgfr1 and Fgfr2 together in the neural crest leads to significantly more severe midface and mandibular defects than loss of each receptor alone. Midline fusion defects, increased cell death and defects in migration of the trigeminal ganglion to the mandibular branch arose at E10.5 upon loss of both receptors in the neural crest lineage. By E15.5 severe defects in the craniofacial skeleton were observed. The phenotype that we described is considerably more severe than has been previously noticed (Park et al., 2008), possibly the result of the 129S4 co-isogenic background used throughout this study, as we found a near total lack of mandible development.
Despite the known activity of FGFs as mitogens in many cell types, we did not detect a significant change in cell proliferation in Fgfr1cKO/cKO; Fgfr2cKO/cKO double null mutant embryos. We observed however high levels of apoptosis in conditional double null mutants, suggesting that this process might be involved in establishing the overall mutant phenotype. Increased cell death has also been observed previously in the branchial arches of hypomorphic or conditional Fgf8 mutants (Griffin et al., 2013; Trumpp et al., 1999). Cell death was highest in the LNP, which normally together with the maxillary prominence expand considerably and push cells towards the midline in development. To functionally test the role of cell death, we crossed a null mutant allele for Bim, which antagonizes anti-apoptotic members of the BCL2 family, into the double conditional null background. This resulted in decreased cell death accompanied by a partial rescue of frontal structures, highlighting a critical role for FGF-mediated cell survival during craniofacial development. How FGF signaling might regulate cell survival remains to be determined, but BIM is a known target of phosphorylation by several MAP kinases, particularly ERK1/2 which phosphorylates BIM and targets it for ubiquitination and proteasomal degradation (Clybouw et al., 2012). JNK and PI3K/AKT activation are also known to stabilize BIM levels in cells (Lei and Davis, 2003). Cell survival through BIM may therefore be regulated by FGF since several of these signaling pathways are engaged by FGFR1 and FGFR2.
FGF signaling has been known to be important for craniofacial development, but the mechanisms by which it operates in this context had not been identified for both receptors. To determine the functionally relevant pathways that operate in vivo, we generated an allelic series of signaling mutations at the Fgfr2 locus, identical to an allelic series previously designed for Fgfr1 (Brewer et al., 2015), that prevent the binding of individual effectors. We thus disrupted binding of FRS2, a prominent pathway leading to the engagement of ERK1/2, as well as binding of CRK/CRKL, and PLCγ/GRB14, alone or in combination. Surprisingly, despite effectively preventing the binding of these signaling effectors to the receptor, we found that each of these Fgfr2 signaling mutations gave rise to viable mice. Furthermore, we explored the effect of signaling mutations in the context of craniofacial development. We had previously observed mild craniofacial defects leading to a partially penetrant cleft palate for Fgfr1F/cKO and Fgfr1FCPG/cKO conditional NCC mutants (Brewer et al., 2015). Because Fgfr1 and Fgfr2 synergize in craniofacial development, we analyzed phenotypes of the signaling mutation of one receptor over the null mutation of the other receptor. Indeed, both Fgfr1F and Fgfr2F mutations placed over the null background of the other receptor showed defects in mandibular development, as well as midface closure. For Fgfr1, we were further able to show that specific structures, namely the frontal cartilage, as well the mandible, maxilla and squamosal bones, were more severely impacted in Fgfr1FCPG than in Fgfr1F mutants. Interestingly, Fgfr1F/F; Fgfr2+/F embryos exhibited broad defects in midline closure, but not in mandible development, also underscoring the coordinate role of both receptors in craniofacial development. These results identify specific roles or thresholds for individual signaling pathways in discrete developmental contexts.
FGF signaling is widely thought to proceed through ERK1/2 as a downstream effector, but it is also known to engage numerous other pathways (Brewer et al., 2016). In the context of early development, the connections between FGF and ERK1/2 signaling have been particularly well noted (Brewer et al., 2016; Lanner and Rossant, 2010). Supporting the notion that not all FGF responses are dependent on the ERK1/2 pathway, however, an RNAseq study in primary mouse embryonic palatal mesenchyme cells showed that only half of FGF regulated transcripts were sensitive to exposure to a MEK inhibitor (Vasudevan et al., 2015). ERK1/2 is thought to be primarily engaged through FRS2, but Fgfr2F/F mice were viable and our analysis of Fgfr2F/F cell signaling in a Fgfr1 null background revealed that this mutation only resulted in about 50% reduction of ERK1/2 signaling. For Fgfr2, the ERK1/2 signaling pathway was only abrogated in Fgfr2FCPG/FCPG signaling alleles, similarly to Fgfr1 (Brewer et al., 2015). Taken together, these results indicate that for both receptors, ERK1/2 engagement relies on the coordinate engagement of FRS2, CRK, and PLCγ signaling. We furthermore showed that in the absence of any other FGFR, the Fgfr2FCPG mutation abrogates not just FGF-induced ERK1/2 signaling, but also FGF-induced PI3K/AKT, PLCγ, p38, JNK, and STAT signaling. Last, the fact that Fgfr1FCPG/FCPG; Fgfr2FCPG/FCPG mutants fail to recapitulate the peri-implantation lethal double null mutant phenotype, despite abrogating expected FGF signaling activity, suggests that functions beyond those classically expected from a receptor tyrosine kinase are still active in these mutants. While our studies have been restricted to cells of neural crest origin, it will be of interest to investigate if these signaling requirements are also conserved in other contexts, for instance within epithelia.
The phenotypic gap between Fgfr2FCPG/FCPG and Fgfr2-/- mutant phenotype could either be due to heretofore unrecognized kinase-dependent signaling activity or to a kinase-independent function. To distinguish between these possibilities, we generated an inactivating Fgfr2KD allele by introducing a mutation in the ATP binding site of the kinase domain (Bellot et al., 1991; Hanks et al., 1988). Although such kinase dead mutations have been introduced in cells or organisms ectopically, to our knowledge this is the first such knock-in allele generated in an RTK gene. This Fgfr2 mutation resulted in lethality at E10.5, with defects in limb outgrowth and chorio-allantoic junction, highly reminiscent of the Fgfr2-/- mutant phenotype. The fact that this mutation recapitulates hallmark Fgfr2-/- mutant phenotypes supports the model that FGFR2 broadly operates in a kinase dependent fashion. Moreover, since the Fgfr2b constitutive mutation leads to similar limb phenotypes but no placental insufficiency (De Moerlooze et al., 2000), and phenotypes in both tissues are observed in Fgfr2-/- or Fgfr2KD/KD mutants, FGFR2 activity must be kinase-dependent in both mesenchymal and epithelial contexts. Efforts to generate a similar mutation at the Fgfr1 locus were unsuccessful, perhaps because FGFR1 is thought to be essential for exit from pluripotency (Molotkov et al., 2017).
Interestingly, we observed semi-dominant effects in Fgfr2KD/+ heterozygous mutants, primarily during postnatal development in the lacrimal gland, and more severe phenotypes in Fgfr2KD/KD mutants relative to the null, affecting craniofacial and mesoderm development. FGFRs have been shown to be able to form heterodimers in culture (Bellot et al., 1991; Ueno et al., 1992). Although these have never been demonstrated in vivo in the absence of any over-expression, it is possible that the Fgfr2KD allele not only inactivates Fgfr2 but also suppresses Fgfr1 activity through FGFR2KD: FGFR1 heterodimers, wherever they are co-expressed. This would explain why Fgfr2KD mutants still generally resemble Fgfr2-/- mutants which are lethal at a similar stage, rather than Fgfr1-/- mutants which fail at implantation on the 129S4 genetic background (Brewer et al., 2015; Kurowski et al., 2019; Molotkov et al., 2017). Previous studies have shown that other RTKs such as PDGF receptors can also function as heterodimers, giving rise to robust signaling both in terms of amplitude and duration during craniofacial development (Fantauzzo and Soriano, 2016). In cNCCs however, FGFR1: FGFR2 heterodimers would consist of FGFR1c and FGFR2c isoforms which would bind the same ligands, unlike the case of the PDGFRs which bind different ligands. Also, we would still expect homodimers of both receptors to be important because loss of one receptor enhances the phenotype in a conditional null background for the other. Last, while cell signaling might occur through heterodimers, this mechanism cannot fully account for the discrepancy between FCPG and null allele phenotypes for either receptor as Fgfr1FCPG/FCPG; Fgfr2FCPG/FCPG double mutants develop until E7.5 with a significant degree of mesoderm formation whereas Fgfr1; Fgfr2 double null mutants on the same genetic background die at implantation (Kurowski et al., 2019; Molotkov et al., 2017). These considerations notwithstanding, the endogenous formation and functional relevance of FGFR heterodimers remains to be tested but is a tantalizing possibility. Alternatively, it is possible that the FGFR2KD receptor accumulates at the plasma membrane, soaking up ligand and thus acting as a dominant negative by titrating away ligands and/or effectors. This possibility is less probable as both alleles were knocked-in at the Fgfr2 locus and are thus expressed at the same levels, unless the FGFR2KD receptor accumulates to a much higher level than the FGFR2FCPG receptor.
The inability of Fgfr1FCPG or Fgfr2FCPG mutations to recapitulate the Fgfr1-/- or Fgfr2-/- phenotypes, respectively, while broadly eliminating classic RTK signaling outputs for ERK1/2, PI3K, PLCγ, and additionally for FGFR2, p38, JNK and STAT1, is at first puzzling and raises multiple questions. For FGFR2, the similarity between the Fgfr2KD/KD and Fgfr2-/- mutant phenotypes indicates that an unknown signaling output which is engaged upon kinase activation has not been tested. The near complete abrogation of multiple signal transduction outputs in the most severe Fgfr1 and Fgfr2 signaling alleles indicates that we have interrogated relevant cell signaling pathways and raises the possibility that FGFRs possess non-canonical functions that are not impacted by our signaling mutations. Importantly, while the FCPG mutations disrupt the ability of intracellular effectors to engage classical RTK activity, they do not abrogate the kinase activity of the receptors suggesting that both the FGFR1FCPG and FGFR2FCPG receptors may be able to phosphorylate targets regulating non-canonical activities.
A wide body of literature has correlated FGFRs with various aspects of cell adhesion through interactions of the extracellular domain, which remains untouched in any of our signaling mutations, with cell adhesion receptors. FGFR1 and FGFR2 interaction with FGF ligands is known to involve a third player, heparan sulfate proteoglycans (HSPGs)(Rapraeger et al., 1991; Yayon et al., 1991). Here, heparan sulfate is linked covalently to cell surface transmembrane type proteins such as Syndecans or GPI-anchored type proteins such as Glypicans (Ornitz and Itoh, 2015). Several of these HSPGs like ANOSMIN have been shown to play a role in neural crest development (Endo et al., 2012). In turn these proteins can interact with integrins, regulating cell-matrix adhesion (Geiger and Yamada, 2011; McQuade et al., 2006; Moser et al., 2009). Consistent with a role for FGFRs in regulating integrin signaling, we observed that Fgfr1FCPG/cKO; Fgfr2cKO/cKO FNPs and Fgfr1CRISPR-KO; Fgfr2FCPG/FCPG iFNPs were still able to form focal adhesions during cell spreading on fibronectin, in contrast to Fgfr1; Fgfr2 double mutant FNPs. Likewise, Fgfr1FCPG/cKO; Fgfr2cKO/cKO FNPs were able to migrate in scratch assays in response to FGF, unlike Fgfr1; Fgfr2 double mutant FNPs. We furthermore showed that these phenotypes correlated with the ability for FGFs to induce FAK and Paxillin phosphorylation through wild type, FGFR1FCPG or FGFR2FCPG receptors. The fact that cell-matrix adhesion was FGF dependent, as seen by changes in focal adhesion formation as well as FAK or Paxillin phosphorylation, suggests that FGF binding to the receptor induces a signaling cascade that may further facilitate the activity of cell adhesion receptors. This may be furthermore enhanced by the cell adhesion complexes being brought into proximity through interaction with FGFR extracellular domains. This kinase dependent activation could occur either by direct phosphorylation of a focal adhesion component by the FGFRs, or through an intermediary such as SRC family kinases which have known roles in integrin signaling (Chen et al., 2018; Klinghoffer et al., 1999). Engagement of SRC family kinases could occur indirectly, or directly as they can bind to the FGFRs (Brewer et al., 2016; Dudka et al., 2010; Schuller et al., 2008) at a site that we have not disrupted in any of our alleles. Additionally, it is possible that defects in FGF-dependent adhesion could result in the induction of anoikis (Frisch and Francis, 1994), a process that has been previously shown to be regulated by Bim (Mailleux et al., 2007), potentially linking our observed adhesion defects and increase in cell death in the LNP. Last, the fact that focal adhesion assembly and phosphorylation defects have also been observed in Fgfr1-/-; Fgfr2-/- keratinocytes (Meyer et al., 2012), although signaling pathways were not investigated in that study, suggests that engagement of FGF signaling has a broad function in regulating cell adhesion in both mesenchymal and epithelial contexts.
In addition, both FGFR1 and FGFR2 are known to interact through the acid box in their extracellular domain with various cell adhesion molecules such as cadherins (Kon et al., 2019), which regulate critical events such as polarity, cell-cycle, EMT, cell-cell contacts and migration and differentiation of cNCCs (Scarpa et al., 2015). The intracellular domain of cadherins is tightly associated with the cytoskeleton through catenin adaptors, and localized RhoA and Rac1 activity at the incipient contacts stabilize these interactions and act as signaling nodes (Perez et al., 2008; Vasioukhin et al., 2000; Wheelock and Johnson, 2003). Since several cadherins might be involved during maturation of cell-cell contacts, accumulation of β-catenin has been used to investigate stable cell-cell contacts during neural crest migration (Nakagawa and Takeichi, 1995). Consistent with a role for FGFRs in mediating cell-cell adhesion through cadherins, we observed that both Fgfr1FCPG/cKO Fgfr2cKO/cKO FNPs and Fgfr1CRISPR-KO Fgfr2FCPG/FCPG iFNPs were able to make strong cell contacts between themselves, as shown by β-catenin accumulation at the junctions, in contrast to double null mutant cells. We observed poor cell-cell adhesion not only among primary FNP cells in culture, but also in vivo, where Fgfr1-/-; Fgfr2-/- mutant cells showed very limited aggregation in the LNP. FGFR1/2 interaction with cadherins may thus have a dual role, one in promoting cell motility through FGF-induced ERK1/2 signaling (Kon et al., 2019), and an opposite one in cell adhesion, as ERK1/2 signaling is abrogated in Fgfr1FCPG and Fgfr2FCPG mutants. During epithelial to mesenchymal transition, including neural crest cell delamination, FGF signaling regulates cadherin switching (Ciruna and Rossant, 2001; Nieto et al., 2016; Sun et al., 1999). FGF signaling also regulates E-cadherin localization, and in the absence of Fgfr1, E-cadherin polarization is affected in the mural trophectoderm in mouse (Kurowski et al., 2019) and in zebrafish cardiomyocytes (Rasouli et al., 2018). Taken together, our results indicate that FGFRs regulate processes such as cell adhesion, and possibly more, beyond their classic signaling cascades (Figure 7G). Additional genetic, biochemical and cell biological studies may identify further non-canonical roles for these receptors beyond their traditional activities in signal transduction.
STAR Methods
Detailed methods are provided in the online version of this paper and include the following:
Key resource table
Contact for reagent and resource sharing
Method details
Generation of knock-in mice
Mouse strains
Generation of Fgfr2-Flag3x expression vector and stable 3T3 expression lines
Coimmunoprecipitation and Western blotting
Cell derivation and culture conditions
Skeletal preparations
Acetocarmine and hematoxylin and eosin staining
Scratch assays
Transwell assays
Immunofluorescence and Antibodies
In situ hybridization
Micro-CT imaging
Cell proliferation assay
TUNEL assay
RT-qPCR
Quantification and statistical analysis
Key resource table
Contact for reagent and resource sharing
Further information and requests for resources and reagents should be directed to the corresponding author, Philippe Soriano (philippe.soriano{at}mssm.edu).
Method details
Generation of knock-in mice
Four distinct targeting vectors carrying the Fgfr2F, Fgfr2C, Fgfr2PG, and Fgfr2KD mutations were generated. The Fgfr2F targeting vector was generated by cloning a short homology arm (1.7kb region between exon 9-10) and a long homology arm (5.1kb, spanning exon10) into PGKneolox2DTA.2 (Hoch and Soriano, 2006). To allow recombineering into SW105 bacteria, the neo cassette was subsequently replaced by PGKEm7neo flanked by FRT sites, which contains both a eukaryotic and a prokaryotic promoter. Similarly, for the Fgfr2C targeting vector, we cloned a long homology arm (5.3kb region spanning exon11) and a short homology arm (1.7kb region between exon 11-12) into PGKneolox2DTA.2 and used a PGKEm7neo flanked by both FRT and LoxP sites for recombineering. For the Fgfr2PG targeting vector, we cloned a short homology arm (1.9kb region 5’ of exon19) and a long homology arm (5.4kb spanning exon19 and 3’-UTR) into PGKneolox2DTA.2 and used a PGKEm7neo flanked by FRT sites for recombineering. For Fgfr2KD targeting vector, we cloned short homology arm (1.9kb SmaI to MfeI, spanning exon 12) and a 3.7kb long homology arm (MfeI to BclI, spanning exon 13) into PGKneolox2DTA.2 (Hoch and Soriano, 2006). Details of regions corresponding to homology arms are provided in Table S1.
For all four alleles, site-directed mutagenesis (SDM) was performed using Phusion polymerase. Nucleotide substitutions introduced by SDM in exon10 for Fgfr2F allele (introduces an XmaI site), exon11 for Fgfr2C allele (introduces a SacI site), exon 19 for Fgfr2PG allele (introduces an EcoRI site) and in exon12 for Fgfr2KD allele (introduces an AluI site) are provided below. All introduced mutations were verified by sequencing. Details of nucleotide substitutions in Fgfr2 signaling mutations are provided in Table S2.
The targeting vectors for Fgfr2F (linearized with NotI), Fgfr2C (linearized with XhoI), Fgfr2PG (linearized with NotI) and Fgfr2KD (linearized with NotI) were electroporated into 129S4 AK7 ES cells. For generating the allelic series of signaling mutations, ES cells were targeted first with the C targeting vector generating Fgfr2+/C mutant cells. After verifying for correct targeting events, the neo cassette was removed by transient transfection with PGKCrebpA, leaving a single LoxP site behind (Figure 3C). Fgfr2+/C ES cells were then targeted using the PG targeting vector generating either Fgfr2+/PG or Fgfr2+/CPG mutant cells, as determined by breeding of the chimeras to ROSA26Flpo mice (Raymond and Soriano, 2010). After verifying for correct targeting events, the neo cassette was removed by transient transfection with PGKFlpobpA (Raymond and Soriano, 2007), leaving both an FRT site and a LoxP site behind (Figure 3C). Fgfr2+/CPG neo- ES cells were finally targeted with the F targeting vector, resulting in Fgfr2+/F or Fgfr2+/FCPG mutant ES cells, as determined by breeding. After verifying for correct targeting events, the neo cassette was removed by transient transfection with Flpe, which is less efficient than Flpo in ES cells (Raymond and Soriano, 2007), in order to not recombine sequences between exons 10-18 due to the retention of the FRT site during the generation of the Fgfr2PG allele. We screened targeting events initially by PCR coupled with restriction digestion to identify incorporation of nucleotide substitutions. Proper targeting was confirmed by Southern blotting using 5′ external and 3′ external probes amplified using the following primer pairs and then an internal probe against Neo. Primers used to generate probes for confirming targeted clones using Southern blots are described in Table S3.
ES cell chimeras were bred to Meox2-Cre or ROSA26Flpo deleter mice (Raymond and Soriano, 2010; Tallquist and Soriano, 2000) maintained on a 129S4 genetic background to remove the neomycin selection cassette and the deleter alleles were subsequently crossed out. Two independent mouse lines were generated from independent ES cell clones for each allele, and phenotypes were confirmed in both lines. The Fgfr2C, Fgfr2PG, Fgfr2F, Fgfr2FCPG and Fgfr2KD alleles were maintained on the 129S4 genetic background. Fgfr2C, Fgfr2PG, Fgfr2F and Fgfr2KD mice were genotyped using oligonucleotides listed in Table S3, with the F, C or PG primers all being able to genotype Fgfr2FCPG mice.
Mouse strains
All animal experimentation was conducted according to protocols approved by the Institutional Animal Care and Use Committee of the Icahn School of Medicine at Mount Sinai. Fgfr1cKO/cKO, Fgfr2cKO/cKO, Fgfr1-GFP and Fgfr2-mCherry were previously described (Hoch and Soriano, 2006; Molotkov et al., 2017). Fgfr1 signaling mutations (Brewer et al., 2015) are referred to as Fgfr1C, Fgfr1F, Fgfr1CPG and Fgfr1FCPG. Tg(Wnt1-cre)11Rth, Tg(Wnt1-cre)2Sor, Cdkn2a+/tm1Rdp, Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo, and Bcl2l11tm1.1Ast are referred to in the text as Wnt1-Cre, Wnt1-Cre2, Ink, ROSA26mT/mG, and Bim respectively (Bouillet et al., 1999; Danielian et al., 1998; Lewis et al., 2013; Muzumdar et al., 2007; Serrano et al., 1996). All lines were maintained on a 129S4 co-isogenic background, except for Bim which was crossed into the Fgfr1/2 deficient backgrounds after only six generations of backcrossing to 129S4.
Generation of Fgfr2-Flag3x expression vector and stable 3T3 expression lines
An Fgfr2 isoform “c” cDNA isoform was PCR amplified from primary MEFs derived from Fgfr2+/+, Fgfr2PG/PG, Fgfr2CPG/CPG, Fgfr2F/F and Fgfr2FCPGFCPG and subsequently digested with HindIII and XhoI. The fragments were cloned in the pcDNA expression vector and sequence verified. Linearized pcDNA-FGFR2 plasmids were transfected in 3T3 cells cultured in DMEM supplemented with 10% calf serum (HyClone Laboratories) with 50 U/mL each penicillin and streptomycin. Stable clones were selected in 500 μg/mL G418. 10 clones from each construct (FGFR2wt-, FGFR2PG-, FGFR2CPG-, FGFR2F-, or FGFR2FCPG-Flag3x) were expanded and assessed for FLAG expression by western blot. Clones expressing high FGFR2-FLAG levels were selected for further analysis.
Coimmunoprecipitation and Western blotting
Stable 3T3 cells expressing FGFR2WT, FGFR2PG, FGFR2CPG, FGFR2F, or FGFR2FCPG-Flag3x were serum-starved (0.1% calf serum supplemented DMEM) overnight, stimulated for 15 mins with 50 ng/mL FGF1 (PeproTech, 450-33A) or FGF8b (PeproTech 100-25B) and 5 μg/mL heparin, and lysed in ice-cold NP-40 lysis buffer (20 mM Tris HCL at pH 8, 137 mM NaCl, 10% glycerol, 1% Nonidet [NP-40], 2 mM EDTA, 25 mM β glycerol phosphate, 1 mM Na3VO4, 10 mM NaF, 1× cOmplete, EDTA-free Protease Inhibitor Cocktail. 800 μg cell lysates were subsequently used for immunoprecipitation with Anti-FLAG M2 magnetic beads using the manufacturer’s protocol. We incubated lysates with anti-FLAG M2 magnetic beads overnight at 4°C followed by five washes with lysis buffer, and precipitated proteins were eluted in Laemmli buffer (10% glycerol, 2% SDS, 0.002% bromophenol blue, 0.062M Tris-HCl, pH 6.8) containing 10% β-mercaptoethanol, heated for 5 min at 95°C, separated by SDS-PAGE and analyzed by western blots.
Western blot analysis was performed according to standard protocols using horseradish peroxidase-conjugated secondary antibodies (1:10,000 dilution) developed by chemiluminescent HRP substrate. Primary antibodies were used at the following dilutions for Western blotting: FGFR2 (1:500 dilution), CRKL (1:500 dilution), FRS2 (1:500 dilution), Flag2 M2 (1:500 dilution), phospho-p44/42 MAPK (1:1,000 dilution), p44/42 MAPK (1:1,000 dilution), GAPDH (1:1000 dilution), phospho-AKT (1:1,000 dilution), AKT (1:1,000 dilution), phospho-p38 (1:500 dilution), p38 (1:500 dilution), phospho-PLCγ1 (Y783) (1:200 dilution), PLCγ1 (1:1000 dilution), pJNK (1:500 dilution), STAT3α (1:1000 dilution), phospho-FAK (1:1000 dilution), FAK (1:1000 dilution), β-catenin (1:1000 dilution), Paxillin (1:1000 dilution), and IRS2 (Cell Signaling Technology, 3089).
Cell derivation and culture conditions
Primary iFNPs were generated by dissecting the maxillary and nasal prominences of E11.5 Fgfr2+/+; Ink-/-, Fgfr2F/F; Ink-/-, Fgfr2CPG/CPG; Ink-/-, and E9.5 or E11.5 Fgfr2FCPG/FCPG; Ink-/- embryos in PBS. The tissue was disassociated with 0.125% Trypsin-EDTA and cultured in DMEM supplemented with 20% FBS, 50 U/mL each penicillin and streptomycin on fibronectin coated plates (0.5μg/cm2). Cells were subsequently split 1:5 through for at least 5 passages before immortalized cell lines were obtained. Cells were allowed to grow until sub-confluent. All experiments were performed between passage 15 and 25. We used PX459 V2.0 vector (Addgene plasmid # 62988) to CRISPR out either Fgfr1 or Fgfr2 and create Fgfr1 null, Fgfr2 null, or Fgfr1: Fgfr2 double null cells. gRNA sequences for Fgfr1 and Fgfr2 were selected using CHOPCHOP gRNA design web tool and were cloned using the oligonucleotides (Table S3), as previously described (Ran et al., 2013). Plasmids were transfected in respective iFNP cells cultured in DMEM supplemented with 10% calf serum with 50 U/mL each penicillin and streptomycin. Stable clones were selected in 5 μg/mL Puromycin. Clones were verified (homozygous deletion of exon6 for Fgfr1 and deletion exon5 for Fgfr2 which also introduces a frameshift mutation) using PCR (Table S3). Primary MEFs were derived from E12.5 wild type mice embryos. Embryos were eviscerated and after removing the head, remaining tissue was chopped into 1 cm pieces and incubated in 1 mL trypsin-EDTA (0.25%) for 30 mins with intermittent shaking. 10 mL DMEM 10% calf serum was added and the mixture was allowed to pass through a cell-strainer. Cells collected from each embryo was plated in 0.2% gelatin coated 15cm plate.
Skeletal preparations
Embryos at E14.5, E16.5 or E18.5 embryos were skinned, eviscerated, fixed in 95% ethanol overnight, and stained (0.015% Alcian blue, 0.005% Alizarin red, 5% glacial acetic acid, in 70% ethanol) overnight at 37°C. Skeletons were then cleared in 1% KOH and transferred to decreasing concentrations of KOH in increasing concentrations of glycerol until clear.
Acetocarmine and hematoxylin and eosin staining
Freshly harvested tissue was fixed in 4% PFA at 4°C overnight followed by dehydration in 70% ethanol. For acetocarmine staining, tissues were incubated in 0.5% aceto-carmine (0.5 g carmine stain dissolved in 100 ml boiling 45% acetic acid for 15 minutes), followed by de-staining in 70% ethanol for 1 minute and 1% acid alcohol (1% HCl in 70% ethanol) for 2 minutes and 5% acid alcohol (5% HCl in 70% ethanol) for 1 minute. For hematoxylin and eosin staining, freshly harvested tissues were dissected in PBS, and fixed in 4% PFA followed by dehydration through a graded ethanol series, and embedded in paraffin. 5mm sections were cut. After deparaffinization and rehydration, sections were stained with Harris modified hematoxylin followed by a 10 second wash in acid-alcohol (1% v/v HCl in 70% EtOH), followed by counterstaining with 1% eosinY. Tissues were washed and mounted with Permount (Thermo Fisher Scientific).
Scratch assays
Cells were seeded onto glass coverslips coated with 5 μg/mL human plasma fibronectin purified protein. At ∼90–100% confluency, cells were scratched with a P1000 pipet tip, washed with PBS and incubated in fresh medium containing either 0.1% FBS, 10% FBS, 50 ng/mL FGF1 and 5 μg/mL heparin or 10 ng/mL PDGF-AA supplemented DMEM for 12 hrs.
Transwell assays
All cells were serum-starved for 24 hrs in 0.1% FBS supplemented DMEM prior to migration. Cell culture inserts compatible for 24-well plates containing polyethylene terephthalate membranes with 8 μm pores (Corning Inc., Corning, NY, USA) were coated with 5 μg/mL human plasma fibronectin purified protein. 100,000 cells were loaded in each insert in 250 μL medium containing 0.1% FBS and inserts were immersed in 500 μL medium containing either 10% FBS, 50 ng/mL FGF1 and 5 μg/mL heparin or 10 ng/mL PDGF-AA for 10 hr. Migrated cells were subsequently fixed in 4% PFA in PBS for 10 min and stained in 0.1% crystal violet in 10% ethanol for 10 min. Dried inserts were photographed using an Axiocam 105 color camera fitted onto a Stemi 508 stereo microscope (Carl Zeiss Microscopy, LLC). Five fields of cells from each of three independent trials were photographed and quantified.
Immunofluorescence and Antibodies
For immunostaining whole mount embryos were fixed in 4% paraformaldehyde solution (PFA) in PBS overnight and washed with PBS five times, permeabilized with 0.5% Triton X100 in PBS for 30 min and blocked in 2% donkey serum for 2h at room temperature. Primary anti-neurofilament antibody (Clone 2H3, DSHB) was used at a 1:20 dilution in 1% donkey serum in PBST; embryos were incubated overnight at 4°C. The next day, embryos were washed 4 times in PBST and incubated with anti-mouse HRP-conjugated secondary antibodies at a 1:1000 dilution for 4hrs at room temperature followed by washing in PBST 4 times and signal was developed using ImmPACTDAB kit. For whole-mount immunofluorescence at E7.5, embryos were fixed overnight in 4:1 methanol:DMSO. Primary antibodies for Eomes (1:100 dilution) and Cdx2 (1:100 dilution) were used. For immunostaining cells, cells were fixed for 10 mins in 4% PFA in PBS at room temperature. Cells/ tissues were subsequently processed for immunofluorescence analysis as detailed above using anti-paxillin primary antibody (1:250 dilution) with Alexa647 conjugated phalloidin (1:40 dilution). For immunofluorescence on sections, antibodies for GFP (1:100 dilution), mCherry (1:100 dilution) and SMA (1:100 dilution) was used. Embryos were stained with DAPI following fixation as previously described (Sandell et al., 2012). Cells / tissues were photographed using an Olympus DP71 digital camera fitted onto an Olympus BX51 fluorescence microscope, Leica SP5 confocal or a Hamamatsu C11440 camera fitted to a Zeiss Observer Z1 microscope. Epifluorescence was imaged in Zeiss Axioplan fitted to ProgRes CT3 camera.
In situ hybridization
Labeled antisense-RNA probes were synthesized for Alx3, Msx1, Six3, Nkx2.1, Fgf8, Shh, Col2a1, Col10a1 and Meox1. Digoxigenin-labeled anti-sense probes were generated as described, and mRNA in situ hybridization on paraffin sections for chromogenic detection was performed using standard protocols.
Micro-CT imaging
Micro-CT imaging of the skulls were performed using a SkyScan 1172 scanner (Bruker, Kontich, Belgium). The mouse heads were dissected and fixed in 10% neutral buffered formalin and washed and stored in PBS at 4 °C. The skull bones were scanned with settings of 50 kV, 500 μA, 10 μm pixel resolution, 0.3° rotation steps, and 4 frames average imaging with a 0.5-mm Al filter at Micro-CT Core, School of Dentistry, NYU, New York. The acquired X-ray projections were reconstructed using the Imirus software (Oxford Instruments).
Cell proliferation assay
For EdU labeling in mice, pregnant females were injected intraperitoneally with 100 mg/kg body weight of EdU. EdU detection was carried out as per manufacturer’s instruction for Click-iT EdU Cell Proliferation Kit.
TUNEL assay
Sections were deparaffinized and were rehydrated in PBS, followed by post-fixation in 4% PFA. In Situ Cell Death Detection Kit, TMR red user protocol was used to detect cell death.
RT-qPCR
Cells were lysed, and mRNA was extracted according to Qiagen RNeasy kit standard protocol. cDNA was synthesized using a 2:1 ratio of random primers to Oligo(dT) with SuperScript IV RT (Invitrogen). qPCR was performed with PerfeCTa SYBR Green FastMix for iQ (Quanta Biosciences) with Bio-Rad iQ5 multicolor real-time PCR detection system and analyzed with Bio-Rad iQ5 optical system software (version 2.0). Cycling conditions were as follows: step 1, 3 min at 95°C; step 2, 10 sec at 95°C; step 3, 30 sec at 60°C; and repeat steps 2 and 3 for 40 cycles. Proper amplification was confirmed using a melting curve and by running samples on a gel to ensure that the correct size band was obtained. Graphs were made using Microsoft Excel and Prism. Primer sequence for respective genes used for RT-qPCR analysis is listed below.
Quantification and statistical analysis
Statistical analysis was performed using GraphPad Prism6.0 and Microsoft Excel. Values are presented as mean ± s.e.m..
Acknowledgments
We thank Jia Li and Chantel Dixon for technical assistance; Elaine Fuchs for helpful insights into cell adhesion mechanisms; Jerry Chipuk for conversations about cell death; Colin Dinsmore for extensive discussions; and our laboratory colleagues, Stu Aaronson, Rob Krauss, and Sergei Sokol for critical comments on the manuscript. We thank the NYU School of Dentistry Micro-CT Core and the Mt. Sinai Flow Cytometry and Microscopy facilities for assistance and advice. This work was supported in part by the Tisch Cancer Institute at Mount Sinai (P30 CA196521 Cancer Center Support Grant for access to Mt. Sinai cores) and by grant RO1 DE022778 from NIH/NIDCR to P.S.
Footnotes
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵