Restricted expression of Fgf16 within the developing chick inner ear

Summary

FGF signaling is required for otic placode induction and patterning of the developing inner ear. We have cloned the chick ortholog of Fgf16 and analyzed its expression pattern in the early chick embryo. Expression is restricted to the otic placode and developing inner ear through all the stages examined. By the closed otocyst stage, expression has resolved to anterior and posterior domains that partially overlap with those of Bmp4, a marker of the developing sensory patches, the cristae of the anterior and posterior semicircular canals. PDGFA, another growth factor with restricted otic expression, also overlaps with Fgf16 expression. The restricted expression pattern of Fgf16 suggests a role for FGF signaling in the patterning of the sensory cristae, together with BMP signaling.

Introduction

Inner ear patterning involves a complex, interrelated series of signaling events that include members of the FGF family of secreted signaling factors (Noramly and Grainger, 2002; Wright and Mansour, 2003). FGFs have important roles in proliferation, migration and differentiation (Ornitz and Itoh, 2001), especially in directional signaling across epithelial-mesenchymal boundaries (Hogan, 1999), reminiscent of the situation within the developing chick inner ear, where mesodermal FGF19 in synergy with WNT8c signals from the hindbrain are required for otic placode induction (Ladher et al., 2000; Ladher et al., 2005). We have cloned chick Fgf16 and analyzed its expression pattern by in situ hybridization within the early embryo, finding specific expression only within the otic placode and developing inner ear. The inner ear originates from the otic placode, an ectodermal thickening adjacent to rhombomeres 5 and 6, which becomes visible at HH stage 9 (HH-Hamburger and Hamilton) (Hamburger and Hamilton, 1951). By stage 11 the placode begins to invaginate, forming the otic pit/cup. The rims of the otic cup make contact, forming the closed otocyst at around stage 16, with the endolymphatic duct and sac arising from a dorsomedial projection of the otocyst, visible by stage 21. Further complex developmental steps lead to emergence of the vestibular system and auditory cochlea, including six sensory patches (organ of Corti, sacculus, utriculus, and three semicircular canal sensory cristae) (Barald and Kelley, 2004).

Fgf16 orthologues have been identified in human and rat (Miyake et al., 1998), mouse (Sontag and Cattini, 2003), and zebrafish and chimpanzee (Katoh and Katoh, 2005), with Ensembl reporting further orthologues in Canis familiaris, Monodelphis domestica and Xenopus tropicalis, amongst others (Birney et al., 2006). Fgf16 is a member of the FGF subfamily consisting of Fgf9/16/20, none of which possess the standard cleavable signal sequence. However, all are secreted due to a combination of a unique N-terminal region and a central hydrophobic region, containing the EFISIA motif, encoding for a non-classical mode of secretion (Miyakawa and Imamura, 2003; Popovici et al., 2004). Although a number of FGF receptors have been identified in the developing inner ear (Wright et al., 2003), FGF receptor affinity for FGF16 has not yet been determined. Thus far, only FGFR4 has been shown to bind to FGF16 in vitro (Konishi et al., 2000).

Expression of Fgf16 has been previously reported in late embryonic and adult cardiac tissues and embryonic brown adipocytes of mice and rats (Miyake et al., 1998; Konishi et al., 2000; Sontag and Cattini, 2003; Lavine et al., 2005). In an evaluation of 18 mouse Fgf gene expression patterns, Fgf16 was identified as a marker within the murine otic placode, developing otic vesicle, pharyngeal endoderm, branchial arches and olfactory placode (Wright et al., 2003). Otic expression was restricted to a single region, the dorsolateral region of the posterior otocyst. Our extensive analysis of the chick Fgf16 expression pattern demonstrates that in early developmental stages, Fgf16 is detected solely in the otic region, with no other expression in the developing embryo. Dynamic expression within the otic placode was refined, until at the otocyst stage, two defined spots of expression are detected in the patches of cells fated to become the anterior and posterior sensory cristae. Chick hindbrain rotation experiments suggest that signals from the hindbrain are responsible for dorso-ventral patterning of the inner ear, but not anterior-posterior posterior patterning (Bok et al., 2005). Thus, as yet unidentified intrinsic signals and/or extrinsic signals originating from tissues other than the hindbrain are required for anterior-posterior patterning. Based on the restricted expression pattern of Fgf16, we hypothesize that FGF16 is an intrinsic signal involved in patterning the anterior and posterior sensory cristae. This hypothesis does not exclude an extrinsic signal upstream of FGF16 arising from tissue surrounding the developing inner ear, such as head mesoderm, neural crest, surface ectoderm or branchial endoderm.

Results and Discussion

Cloning of chick Fgf16

Using the known mouse Fgf16 sequence we identified a chick EST ortholog that contained a partial chick Fgf16 fragment lacking the 5′ end. We cloned the full-length 624 bp coding sequence of chick Fgf16 using 5′ RACE GeneRacer Kit (Invitrogen). The predicted chick 207 amino acid sequence is 90.3% and 89.4% identical to that of human and mouse FGF16 (Miyake et al., 1998), respectively (Fig. 1). A fragment of the full-length sequence including a portion of the 3′ UTR was used to make an in situ hybridization probe, eliminating the possibility of cross hybridization to other Fgf family members. Analysis of the sequence within the respective genomes reveals that both the mouse and the human FGF16 genes are located on the X chromosome, whereas FGF16 in the Gallus genome is located on chromosome 4. The significance of this difference is not clear.

Figure 1.

Predicted amino acid sequence comparison of chick FGF16 with human FGF16 and mouse FGF16. The colon indicates identical amino acid residues of the compared sequences. 187/207 residues are identical to human (90.3%) and 185/207 are identical to mouse (89.3%). Mouse and human are 99.0% identical, with a two amino acid difference.

Restricted expression of Fgf16 within the developing inner ear

Analysis of the gene expression pattern of Fgf16 was undertaken by whole mount in situ hybridization (ISH), using embryos from pre-streak stages to stage 20. The earliest stage with detectable expression was Hamburger and Hamilton (HH) stage 8/9 (Hamburger and Hamilton, 1951), in an anterior-to-posterior stripe of dorsally expressing non-neural ectoderm adjacent to the hindbrain, corresponding to the future otic placode (Fig. 2A, B). Expression was detected exclusively in otic tissue at all stages examined. By HH12 expression was restricted to the dorso-medial lip of the invaginating otic placode (Fig. 2C, D). This expression expanded ventrally at HH13, with a weaker domain extending ventrally from the original dorso-medial expression domain (Fig. 2E, F). By HH16, the lips of the otic cup have fused to form the otocyst and two distinct, equally strong expression zones were observed in dorsal view: an anterior-medial and separate postero-lateral domains (Fig. 2G). Transverse sections of the embryo confirm strong staining in both the anterior (Fig. 2H) and posterior zones (not shown). Dynamic changes within the expression domains occur between stages HH16-20. From the two initially strong domains, the anterior-medial domain weakens at HH17 (compare Fig. 2G, J dorsal views). The dorsal view confirms down regulation of the anterior-medial domain, whereas the dorso-posterior domain in the otocyst remains unchanged (Fig. 2J). Dorsal expression of Fgf16 begins to separate into two separate domains when viewed from the lateral side, with the anterior spot being weaker than the posterior spot (Fig. 2I). The anterior transverse section of the otocyst reveals that the more anterior spot is composed of a weak medial and stronger lateral domain (Fig. 2 J, K). The more posterior section recapitulates this pattern (Fig. 2J, L). Within a few hours Fgf16 expression has resolved into two distinct domains, with separate anterior and posterior spots (Fig. 2 M, N) that bear a striking resemblance to Bmp4 and PDGFA expression within the putative cristae domains (see below). A dorsal view of the expression domain confirms that the spots are now fully separated, with a defined posterior domain now visible (Fig. 2N). Transverse sections from the anterior (Fig. 2O) and posterior otocyst reveal the posterior spot to be stronger laterally than medially (Fig. 2P). From HH18 the anterior domain is progressively restricted to the front edge of the otocyst (Fig. 2M, Q, U lateral, N, R, V dorsal), with stronger expression on the lateral side in transverse section (Fig. 2O, S, W). The posterior domain shifts from strong postero-lateral expression to more postero-medial expression (Fig. 2P, T, X). The anterior and posterior spots are strongly reminiscent of Bmp4 expression from HH17 onward.

Figure 2.

Fgf16 mRNA expression analysis. A, C, E, G are dorsal views of whole mount embryos from HH9-16 with anterior to the top of the page. Similarly J, N, R, V depict stages HH 17-20. I, M, Q, U are lateral views of the developing otocyst from HH 17-20, with anterior to the right. B, D, F, H are transverse gelatin sections (50 μm) through the more anterior part of the expression domain. K, O, S, W are transverse sections (50 μm) of the left otocyst (lateral to left) through the more anterior expression domain within the otocyst. L, P, T, X are sections from the more posterior domain (50 μm) of the same embryos. Scale bars: A–J, M, N, Q, R, U, V = 200 μm; K, L, O, P, S, T, W, X = 100 μm.

Double in situ hybridization analysis: Fgf16 and Bmp4

To determine the relationship between Fgf16 and Bmp4 expression patterns we used double ISH. The anterior and posterior Bmp4 expression domains mark the areas that give rise to the respective anterior and posterior crista ampullaris of the semicircular canals (Wu and Oh, 1996). In an example of remarkably similar developmental timing to Fgf16, Bmp4 expression becomes confined to two spots, an anterior and posterior domain, between HH16 and HH20. We show here that at HH19, both in the anterior and posterior expression domain that Fgf16 (Fig. 3G) and Bmp4 (Fig. 3H) partially overlap (Fig. 3A–F), with Bmp4 expression lying within the ventral domain of Fgf16 expression (Fig. 3C, F).

Figure 3.

Double in situ hybridization with Fgf16/Bmp4 and Fgf16/PDGFA at HH19. A–F are transverse sections (50 μm) from the same embryo double labeled with Fgf16/Bmp4. A–C are anterior sections through the otocyst; D–F are posterior sections through the otocyst. Insert G is a close up of a lateral view of a whole mount showing two spots of Fgf16 expression in the otocyst at HH19; anterior is to the right. Similarly, insert H shows two spots of Bmp4 expression in an HH19 lateral view of a whole mount. Note in the merged images that Bmp4 expression overlaps the ventral domain of expression of Fgf16 in both the anterior (C) and posterior (F) otocyst. I–N are transverse sections (50 μm) double labeled with Fgf16/PDGFA. I–K are anterior sections through the otocyst; L–N are posterior sections through the otocyst. Insert O is a close up of a lateral view of a whole showing PDGFA in the otocyst at HH19; anterior is to the right. Note in the merged images that PDGFA overlaps the expression domain of Fgf16 in both the anterior (K) and posterior (N) otocyst. Scale bar for all sections = 200 μm; lateral views G, H and O = 300 μm.

Double in situ hybridization analysis: Fgf16 and PDGFA

Interestingly, PDGFA (platelet-derived growth factor alpha) is also expressed at HH19 in the anterior and posterior domains (Fig. 3P). PDGFs are involved in proliferation, migration, differentiation and cell survival in vertebrate development. Previously PDGFA ligand was identified in later stages as an important mitogen in the growth and proliferation of hair cells in the developing rat cochlea (Lee et al., 2004). Overlap in the anterior domain between Fgf16 and PDGFA is incomplete (Fig. 3I–K), whereas the posterior domain appears to have exactly matching overlap (Fig. 3L–N). Further investigation is underway to determine the functional interactions between Fgf16/Bmp4 and PDGFA.

Conclusions

FGF signals arising from neural ectoderm, mesoderm and endoderm adjacent to the otic placode are important in the induction and early development of the inner ear (Ladher et al., 2005). We show that Fgf16 is expressed in a restricted domain of expression in the otic ectoderm from HH9 and is therefore unlikely to be directly involved in placode induction, although this remains to be tested. The restricted expression pattern of Fgf16 within the otic placode suggests a role in patterning a subset of cells within the developing otic cup and otocyst. Partial overlap in expression with the known sensory cristae marker Bmp4 and overlap with the growth factor PDGFA indicates a potential role for FGF16 signaling in patterning of the anterior and posterior cristae. Interestingly, experiments that disrupt or enhance FGF signaling provide evidence that FGFs (Fgf2, 3, 10) are upstream of BMPs (Bmp2, but not Bmp7) in the induction of cristae and associated semicircular canals (Chang et al., 2004). Functional experiments are underway to determine the role of Fgf16 in inner ear patterning.

Experimental Procedures

Incubation, harvesting, staging, in situ hybridization and vibratome sectioning were performed according to our standard protocols as described previously (Chapman et al., 2002). Blasting the Genebank EST Database with the mouse Fgf16 nucleotide sequence revealed a 3′ fragment of the chicken Fgf16 ortholog sequence (BU464707). 5′ RACE was performed using the GeneRacer Kit (Invitrogen). The following primers were used: 5′ RACE Reverse Primer: CTT GAG GTG CGC GAA GTC GGT G and Reverse Nest Primer: TTG AGG AAG CCG GGC GAA TC. Whole chick embryos, stages 10–19 were harvested and immediately placed in RNAlater RNA Stabilization Reagent (Qiagen). Total RNA was extracted using RNeasy Mini Kit (Qiagen). Utilizing the manufacturer’s GeneRacer protocol yielded a PCR fragment of 160bp that was ligated into pCRII-TOPO Vector (Invitrogen) and sequenced. Based on this 5′ sequence a Forward Primer: ATG GCC GAG GTG GGC GGC TT and Reverse Primer: TCA CCT GTA GTG GAA GAG GTC were designed to clone out the full length gene, which was ligated into pCRII-TOPO Vector (Invitrogen) and confirmed by sequencing. (NCBI full-length sequence accession number XM_420304). The primers used to make the in situ hybridization probe were as follows: Forward primer: GTA CGC CTC AAC GCT CTA C; reverse primer: AAC AGC AGT TGT CTG GAT TC.