SEARCH

SEARCH BY CITATION

Keywords:

  • Sef;
  • FGF2;
  • FGF4;
  • FGF8;
  • Shh;
  • chick embryos;
  • limb development;
  • progress zone;
  • AER

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The signaling pathways leading to growth and patterning of various organs are tightly controlled during the development of any organism. These control mechanisms usually involve the utilization of feedback- and pathway-specific antagonists where the pathway induces the expression of its own antagonist. Sef is a feedback antagonist of fibroblast growth factor (FGF) signaling, which has been identified recently in zebrafish and mammals. Here, we report the isolation of chicken Sef (cSef) and demonstrate the conserved nature of the regulatory relationship with FGF signaling. In chick embryos, Sef is expressed in a pattern that coincides with many known sites of FGF signaling. In the developing limb, cSef is expressed in the mesoderm underlying the apical ectodermal ridge (AER) in the region known as the progress zone. cSef message first appeared after limb budding and AER formation. Expression was intense at stages of rapid limb outgrowth, and gradually decreased to almost undetectable levels when differentiation was clearly apparent. Gain- and loss-of-function experiments showed that FGFs differentially regulate the expression of cSef in various tissues. Thus, removal of the AER down-regulated cSef expression, and FGF2 but not FGF4 or FGF8 beads substituted for the AER in maintaining cSef expression. At sites where cSef is not normally expressed, FGF4 and FGF2, but not FGF8 beads, induced cSef expression. Our results demonstrate the complexity of cSef regulation by FGFs and point to FGF2 as a prime candidate in regulating cSef expression during normal limb development. The spatiotemporal pattern of cSef expression during limb development suggests a role for cSef in regulating limb outgrowth but not limb initiation. Developmental Dynamics 233:301–312, 2005. © 2005 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Fibroblast growth factors (FGFs) are heparin-binding polypeptide mitogens that signal by means of binding and activating a family of cell surface tyrosine kinase receptors (FGFR1–FGFR4; Shaoul et al., 1995; Powers et al., 2000). FGFs regulate the growth, differentiation, survival, and migration of a wide variety of cell types, and are involved in essentially every process in animal development from invertebrates to humans (Powers et al., 2000; Ornitz and Itoh, 2001; Niehrs and Meinhardt, 2002). The developing limb is one of the best-studied organs with respect to the role of FGFs in morphogenesis and pattern formation, and has long been considered an excellent model for studying the signals that direct embryonic growth and pattern formation (Martin, 1998; Tickle, 2002; Niswander, 2003).

Limbs develop from small buds that appear at particular axial levels. Each bud is composed of a lateral plate mesoderm (LPM) and an overlying surface ectoderm. Skeletal elements, tendons and other connective tissues of the mature limb are generated from the mesoderm, whereas skin and cutaneous appendages are formed from the surface ectoderm (Martin, 1998). A few hours after limb budding, changes in cell shape and position within the surface ectoderm result in the appearance of the apical ectodermal ridge (AER) that develops along the anterior–posterior distal margin of each bud. Vertebrate limb development depends on signals from the AER. The AER acts to maintain the underlying mesenchyme, called the progress zone, in a highly proliferative and undifferentiated state (Summerbell et al., 1973; Martin, 1998; Wolpert, 2002). Removal of the AER results in the formation of a truncated limb (Summerbell, 1974; Johnson et al., 1994).

Developmental studies in chick and mice, as well as analysis of human limb genetic disorders, highlighted the importance of FGF-mediated signaling in multiple stages of limb development. Mutations in several FGFRs were shown to be associated with disorders of human limb patterning (Webster and Donoghue, 1997; Wilkie et al., 2002; Tickle, 2002). Gene knockout in mice, as well as studies of FGF gene expression in early stages of chick limb development, identified FGF10 and FGF8 as potential regulators of limb development (Martin, 1998; Tickle and Münsterberg, 2001). FGF10 is expressed in the mesenchyme of the prospective limb region and is essential for limb formation (Min et al., 1998; Sekine et al., 1999). FGF8 is expressed in the prospective limb bud surface ectoderm, which form the AER (Crossley and Martin, 1995; Mahmood et al., 1995; Crossley et al., 1996; Vogel et al., 1996). Conditional knockout of FGF8 resulted in skeletal defects, but showed no effect on the initiation of limb outgrowth (Moon and Capecchi, 2000; Lewandoski et al., 2000). Other FGFs such as FGF2, FGF4, FGF9 or FGF17, are expressed in the AER and are thought to play redundant functions in limb outgrowth and patterning (Savage et al., 1993; Niswander et al., 1993; Moon et al., 2000; Sun et al., 2000; Niswander, 2003; Tickle, 2002). Beads coated with members of the FGF family, such as FGF2, FGF4, or FGF8, can replace AER function in the outgrowth and patterning of the chick limb, and application of FGFs to the flank can induce the growth of an ectopic limb (Johnson et al., 1994; Cohn et al., 1995; Tickle and Munsterberg, 2001).

Recent studies revealed that FGF signaling pathways are negatively regulated by complex intracellular mechanisms that involve ligand-induced antagonists, such as the Sprouty proteins family, and MKP3 (Minowada et al., 1999; Pouyssegur et al., 2002; Dikic and Giordano, 2003; Christofori, 2003; Eblaghie et al., 2003; Kawakami et al., 2003; Tsang and Dawid, 2004). Whereas MKP3 is a MAP-kinase specific phosphatase, Sprouty proteins lack an enzymatic activity (Pouyssegur et al., 2002; Christofori, 2003). They act by means of an interaction with, and the inhibition of, signaling molecules in the Ras/MAPK pathway (Christofori, 2003). In addition, Sprouty members are expressed in limbs in a pattern that correlates with known sites of FGF signaling, and overexpression of certain Sprouty members in chick repress FGF-mediated limb development (Minowada et al., 1999; Chambers and Mason, 2000; Lin et al., 2002). An additional feedback inhibitor of FGF signaling has been identified recently in zebrafish (Furthauer et al., 2002; Tsang et al., 2002). This inhibitor, termed Sef (Similar Expression to Fgf genes), encodes a putative transmembrane protein that is conserved among vertebrates. The spatial pattern of expression of Sef in zebrafish embryos correlates with that of the Fgf3 and Fgf8 genes (Furthauer et al., 2002; Tsang et al., 2002). In zebrafish embryos, Sef expression is positively regulated by FGFs, and misexpression of Sef inhibits FGF signaling causing characteristic malformations in zebrafish embryos (Furthauer et al., 2002; Tsang et al., 2002). The mammalian homologue of zebrafish Sef (zfSef) was isolated by several groups, including ours, and studies in cultured cells revealed that Sef inhibits FGF-induced cell division (Kovalenko et al., 2003; Yang et al., 2003; Xiong et al., 2003; Preger et al., 2004). In mammals, additional Sef isoforms are generated by means of an alternative splicing mechanism. These Sef isoforms display differential pattern of expression and vary in biological and biochemical properties from the originally identified trans-membrane Sef protein (Preger et al., 2004; and D. Ron, unpublished results).

In the present work, we have investigated the relationship between Sef expression and FGF signaling pathways during chick embryogenesis, as well as studied the regulation of cSef expression during limb development. We found that cSef expression domains are in intimate association with known sites of FGF action. In the limb, the AER regulates cSef expression in the progress zone, and FGF2, but not FGF4 or FGF8 can effectively substitute for the AER in inducing cSef expression.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Cloning of Chicken Sef and Analysis of Its Expression Pattern During Development

To clone chicken Sef, a database search was applied using the human Sef (hSefa; Preger et al., 2004) sequence as a query for the avian homologue. Several expressed sequence tags (ESTs), were obtained, with the largest EST containing the entire open reading frame (ORF) of chicken Sef, extending into the 3′ untranslated region (UTR; GenBank accession no. AY278204). The deduced amino-acid sequence of chicken Sef (designated cSef) revealed an ORF of 733 amino acid residues. The predicted cSef product is a type I transmembrane protein containing eight potential N-linked glycosylation sites and a conserved immunoglobulin-like domain in the extracellular region, a putative tyrosine phosphorylation site and an interleukin 17 receptor-like domain in the intracellular region (Fig. 1a). Comparison of the amino-acid sequence of cSef with that of zfSef and mammalian Sefs (Furthauer et al., 2002; Tsang et al., 2002; Lin et al., 2002; Preger et al., 2004) showed that the Sef gene is highly conserved throughout vertebrate evolution. The cSef gene product shares 70% and 80% similarity with zfSef or mammalian Sef, respectively.

thumbnail image

Figure 1. Sequence comparison between chicken, zebrafish, and mammalian Sef and detection of cSef transcripts in chick embryonic tissues. a: The deduced amino acid sequence of cSef was aligned with amino acid sequences of the indicated Sef gene products. Dashes represent gaps inserted to maximize alignment. Amino acids conserved between all Sef species are marked by asterisks, and those that are conserved between two or more species are marked by dots. Conserved domains and consensus sequences are indicated: Potential N-linked glycosylation sites (arrowheads), signal sequence (dashed box), Ig-like domain (light gray box), transmembrane domain (closed box), conserved tyrosine (closed box and asterisk), and the IL17 receptor-like domain (dark gray box). b: Expression was determined by reverse transcriptase-polymerase chain reaction using total RNA from whole embryos or from separated body or brain from embryonic day (E) 3 to E17. Amplification was performed with cSef specific primers as described in the Experimental Procedures section. Templates for positive controls are plasmids containing cSef or glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Lanes 1–2, E3 and E4 embryos; lanes 3, 5, 7–9, brain from E6, E8, E10, E12, and E17 embryos; lanes 4, 6, bodies of E6 and E8 embryos.

Download figure to PowerPoint

To determine Sef expression during chick embryonic development, we initially used reverse transcriptase-polymerase chain reaction (PCR) using cSef-specific primers. RNA was extracted from either whole chick embryos (embryonic day [E] 3 and E4), or from separated body and brain (E6 and E8), or brain alone (E10, E12, and E17). All the samples that have been examined were positive for cSef transcripts. Expression levels of cSef were the lowest in E3 embryos (Fig. 1b).

To establish a close correlation between cSef and known sites of FGF signaling, we studied the expression of cSef in chick embryos from stages 10–25, using both sections and intact embryos. Sef-positive domains were present at all the stages that have been examined. At stage 10, cSef transcripts were detected in Hensen's node and the most posterior neural plate. At this stage, cSef transcripts were also observed in the forebrain, the presumptive midbrain–hindbrain boundary (MHB), and the otic placode (Fig. 2a). By stage 12, cSef expression was intensified in Hensen's node, in the posterior neural plate, and in the MHB. In addition, cSef expression was extended to the branchial arches and the intermediate mesoderm (Fig. 2b). At stage 15, Hensen's node, the posterior neural tube, and the intermediate mesoderm posterior to the 10th somite axial level still expressed cSef. In the head region, cSef was strongly expressed in the MHB, prosencephalon and the branchial arches (Fig. 2c). Furthermore, cross-section through the developing eye revealed cSef expression in the thickening overlying ectoderm that will give rise to the lens (Fig. 2k). At stage 17, faint expression of cSef was observed for the first time in the emerging forelimb. At this stage, no staining was observed in the prospective hind limb region (Fig. 2d, indicated by a black arrow). At stage 18, expression was intensified in the forelimb and faint cSef staining become apparent in the emerging hindlimb (Fig. 2e, indicated by black arrows). At this stage (Hamburger and Hamilton [HH] 18), the tail bud was strongly stained for cSef transcripts, and in the head region, cSef expression persisted in the MHB, the branchial arches and the otic vesicle. The expression in the forebrain at this stage was restricted to the telencephalon. In addition, strong expression was observed in the first maxillary and mandibular pharyngeal arches (Fig. 2e). By stage 20, cSef expression was significantly intensified in the hindlimb bud (Fig. 2f, black arrow). Cross-section through the head at this stage revealed weak expression of cSef in the telencephalon and the developing retina, but strong expression in the pharyngeal arches (Fig. 2i). At this stage cSef was weakly expressed in the somites. By stage 23, cSef expression in the MHB was decreased and was almost diminished by stage 25. Weak expression was still apparent in the branchial arches and the otic vesicle (Fig. 2g,h). The expression of cSef in the pharyngeal arches, at these two stages, was restricted only to the first maxillary arch. Cross-section through the trunk region (Fig. 2j) of stage 23 embryos revealed weak expression of cSef in the dermomyotome. At stage 25, strong expression of the gene was observed in the developing retina. From stage 20 onward, cSef staining in the limbs was intensified and was significantly higher as compared with other Sef-positive tissues (Fig. 2f,g,h, indicated by black arrows). cSef expression was also detected in the feather buds of E8 embryos (Fig. 2l).

thumbnail image

Figure 2. The expression pattern of cSef during chick development. ah: Whole-mount RNA in situ hybridization of chick embryos of Hamburger and Hamilton stages (HH) 10–25. i: Cross-section through the head region (marked by white line in f) of an HH20 embryo. j: Cross-section through the anterior somites (marked by black line in g) of an HH23 embryo. k: Cross-section through the eye placode of an HH15 embryo. l: cSef expression in the feather buds of the hindlimb of embryonic day 8 embryo. ba, branchial arches; dm, dermomyotome; Hn, Hensen's node; im, intermediate mesoderm; MHB, midbrain–hindbrain boundary; nc, notochord; nt, neural tube; op, otic placode; ov, otic vesicle; pa, pharyngeal arch; pc, prosencephalon; rt, retina; tb, tail bud; tc, telencephalon.

Download figure to PowerPoint

Kinetics of cSef Gene Expression in Limbs

The striking expression pattern of cSef in the developing limbs, led us to focus our study on cSef regulation in this organ. First, we performed a detailed study of cSef spatiotemporal expression pattern during limb development. Whole-mount RNA in situ hybridization and sections of limbs from HH23 embryos revealed that cSef is uniformly expressed along the anterior–posterior axis of the distal mesoderm. Expression was extended inward for approximately 150 μm from the limb tip in the mesenchyme underlying the AER, but not in the AER (Fig. 3a–c, indicated by arrows). The expression domain of cSef is located within the progress zone, a region containing rapidly proliferating and undifferentiated cells that are maintained at this state by the AER (Summerbell et al., 1973; Johnson et al., 1994). In addition, cSef displayed a graded expression pattern in both the proximo-distal (PD) and dorso-ventral (DV) axis (Fig. 3b). In the PD axis, staining was more intense in the distal mesenchyme immediately subjacent to the AER, and then decreased proximally. In the DV axis, staining was stronger on the dorsal side.

thumbnail image

Figure 3. The kinetics of cSef expression in the developing limb. Whole-mount RNA in situ hybridization of limb buds at different stages of development. a: Distal to proximal view of Hamburger and Hamilton stages (HH) 23 forelimb bud show the apical ectodermal ridge (AER, indicated by an arrow) as white structure on the background of cSef expression in the progress zone. b: Sagittal section through an HH24 forelimb bud reveals a dorsoventral (D and V, respectively) gradient of cSef expression in the progress zone. cf: Different stages of limb development exhibit cSef expression in the progress zone. Black arrows point to the AER; red arrowheads point to cSef expression. In all panels, anterior is up. Note that, in a, the symmetry in the distribution of cSef transcript at the anteroposterior axis, at this stage, is not evident owing to the angle at which photograph was taken (compare with Fig. 4d–g).

Download figure to PowerPoint

thumbnail image

Figure 4. Regulation of cSef expression by fibroblast growth factor (FGF) signaling. Heparin acrylic beads soaked with FGF8, FGF4, or FGF2 were implanted into the flank (a–f) or into the anterior (g,i,k) or posterior (h,j,l) proximal aspect of a forelimb bud. ac: Ectopic limbs (indicated by arrows) can be observed after implantation of beads soaked with FGF2 (a), FGF4 (b), and FGF8 (c). df: Only FGF4 (e) and FGF2 (f) were able to induce weak ectopic expression of cSef in flank regions, whereas no discernible expression of cSef was observed with FGF8 soaked beads (d). g,h: FGF8 beads did not induce ectopic expression of cSef in proximal regions of the limb bud. i,j: FGF4 induces faint expression of cSef in proximal regions of the limb bud. k,l: Note the distant circular pattern of cSef expression around the beads in i and j is different from the ectopic intense cSef expression achieved with FGF2 beads. For the induction of an ectopic limb, photographs were taken 3.5 days after FGF bead implantation; for induction of cSef expression in the flank and limbs, photographs were taken 19 and 24 hr after bead implantation, respectively. Arrows point to the site of implantation.

Download figure to PowerPoint

As demonstrated in Figure 2, cSef transcripts first appeared in the forelimb of stage 18 embryos. By stage 20, cSef was expressed in both the fore- and hindlimbs and this expression gradually increased in the next few stages. The highest levels of cSef transcripts were observed in limbs from embryonic stages 22 and 23, and its expression was evenly distributed throughout the anterior–posterior axis of the progress zone. However, in limbs from embryonic stage 25, an anterior–posterior gradient of cSef expression was clear with higher expression levels in the two-thirds anterior aspect of the distal limb region (Fig. 3d). On day 6 (stage 28), when digits start to form, cSef expression domain was dramatically decreased in size, and by day 8 (stage 33), it remain as a narrow strip at the tip of the posterior digits (Fig. 3e,f).

Regulation of cSef Gene Expression by FGF Signaling

Expression of feedback antagonists is regulated by the signaling pathway they inhibit (Niehrs and Meinhardt, 2002; Dikic and Giordano, 2003). In zebrafish embryos, FGFs control Sef expression (Furthauer et al., 2002; Tsang et al., 2002). To elucidate whether FGF signaling can regulate expression of cSef during chick embryogenesis, we investigated the effect of exogenous addition or removal of endogenous FGF signals on cSef expression. For ectopic cSef induction, we examined whether FGF signaling is sufficient to induce cSef expression in tissues that are competent to respond to FGF signals and where cSef normally is not expressed. Heparin beads soaked with FGF2, FGF4, FGF8 or control beads, soaked with PBS, were implanted in the flank and the anterior or posterior proximal aspects of the developing limb, in HH18–HH20 embryos. Expression of cSef was determined by whole-mount in situ hybridization between 6 and 28 hr after implantation. As a control for the biological activity of each growth factor, we assessed their ability to induce the growth of an additional limb when implanted in the flank of HH15 embryos as described (Cohn et al., 1995). Representative results are shown in Figure 4. Each growth factor readily induced the growth of an extra limb, whereas no outgrowth was detected when beads alone were implanted (Fig. 4a–c, and data not shown). cSef transcripts could not be detected at 6, 19, and 28 hr after FGF8 beads implantation, in neither the flank (0 of 9 cases, Fig. 4d), nor in the limb mesoderm (0 of 15 cases, Fig. 4g,h). In contrast, both FGF4 and FGF2 induced cSef expression at all implantation sites. Modest but reproducible cSef induction was observed by 19 hr after implantation of FGF4 or FGF2 beads in the flank (4 of 6 and 12 of 13 cases for FGF4 and FGF2, respectively, Fig. 4e,f). Similar results were obtained with beads implanted at stage 15 (data not shown). In the limbs, the expression pattern obtained with FGF2 (13 of 13 cases, Fig. 4k,l) was uniform and concentrated in a small circle immediately around the implanted bead. The signal observed with FGF4 was diffuse, occupied a relatively larger circle, and staining was weak or negative around the bead (13 of 13 cases, Fig. 4i,j). In addition, FGF2 induced consistently higher levels of cSef mRNA than FGF4 at all three implantation sites, although both growth factors have similar biological potency as determined by their efficient induction of an ectopic limb (Fig. 4a,b), and their mitogenic activity in NIH/3T3 cells (half maximal activity at 0.5 ng/ml, data not shown; Reich-Slotky et al., 1995; Sher et al., 1999).

Analysis of the Ability of FGFs to Modulate Endogenous cSef Expression in the Apical Mesoderm of the Developing Limb

To examine the effect of loss of FGF signals on cSef expression, we surgically removed the AER, which is the main source of FGFs in the developing limb (Johnson et al., 1994; Martin, 1998; Tickle and Münsterberg, 2001). The AER was removed from forelimbs of embryos at stages 18–20, and cSef gene expression was evaluated by whole-mount RNA in situ hybridization. Expression of cSef was down-regulated in the mesenchyme underlying the removed AER region already at the earliest time point examined (6 hr). Parts of the AER in the most anterior and posterior regions of the limb were left untouched to serve as an internal control. These regions exhibited normal cSef expression (Fig. 5a, compare with control intact limb in 5b). These results reveal that signaling emanating from the AER control the expression of cSef.

thumbnail image

Figure 5. Fibroblast growth factor-2 (FGF2) but not FGF4 or FGF8 can induce cSef expression after apical ectodermal ridge (AER) removal. a: The AER was partially removed (area marked between the arrowheads), and cSef expression was assessed by whole-mount RNA in situ hybridization. b: Control limb where the AER remains untouched shows the normal expression of cSef. c–g: The AER was partially or completely removed from limbs of Hamburger and Hamilton stage (HH) 18–HH20 embryos, and FGF (0.1 mg/ml or 1 mg/ml) beads were implanted at the apical mesoderm of the limb. Embryos were incubated at 37°C for different time intervals after bead implantation, and cSef expression was then detected by whole-mount RNA in situ hybridization. Shown here are photographs of embryos incubated for 24 hr with FGF beads. c: FGF4 beads (0.1 mg/ml). d,e: FGF2 beads (0.1 mg/ml). f,g: FGF8 beads (0.1 and 1 mg/ml, respectively). h,i: FGF8 at both low (0.1 mg/ml, h) and high (1 mg/ml, i) concentrations can up-regulate Shh in AER(−) limbs. j: Shh expression is completely down-regulated after AER removal. k: Control limb showing the normal expression of Shh in the zone of polarizing activity of the limb. Arrows indicate the implanted beads. Note that, in all panels except b and k, anterior is to the left.

Download figure to PowerPoint

To investigate whether FGFs can support cSef expression in the apical mesoderm, FGF-soaked beads (FGF2, FGF4, or FGF8) were implanted at the tip of limb buds after surgical removal of the AER [designated AER(−) limbs]. Embryos were analyzed for cSef expression between 6 and 40 hr after implantation. FGF4, at a concentration that induced readily detectable ectopic cSef expression (0.1 mg/ml; 13 cases), or at threefold higher concentration (6 cases), failed to maintain cSef expression by 20, 24, and 26 hr after bead implantation (Fig. 5c, and data not shown). In contrast to these results, beads soaked with FGF2 (0.1 mg/ml) maintained extremely strong cSef expression in the underlying mesoderm of AER(−) limbs 20, 24, 27, and 40 hr after bead implantation (22 of 25 cases; Fig. 5d,e). FGF8, at a similar concentration, failed to induce detectable levels of cSef transcripts 6, 17, 19, 22, and 24 hr after bead implantation (total of 33 cases; Fig. 5f, and data not shown). There was, however, a noticeable outgrowth in the vicinity of the FGF8 bead within 24 hr after implantation (see Fig. 5f). To assess the possibility of a synergistic effect between FGF4 and FGF8, we implanted beads soaked with a combination of both growth factors (0.1 mg/ml each), but obtained negative results (5 of 5 cases, data not shown). These results were unexpected, since FGF8 was shown to induce cSef expression in zebrafish (Furthauer et al., 2002; Tsang et al., 2002), these results were unexpected. To address this problem, we have tested whether increasing the concentration of FGF8 would result in a positive effect, and also tested the ability of FGF8 to maintain Shh expression in the posterior mesoderm of AER(−) limbs at both the low, and the high FGF8 concentrations. Within 24 hr after implantation of FGF8 beads (soaked with 1 mg/ml of FGF8) in the apical mesoderm, a significant outgrowth of the limb bud was observed but only very faint cSef expression could be detected in the underlying mesoderm (7 of 9 cases; Fig. 5g). In contrast, strong Shh expression was detected 24 hr after implantation of FGF8 beads in the posterior mesoderm at both low and high FGF8 concentrations (9 of 9 cases; Fig. 5h,i, respectively). Shh transcripts were not observed in limbs where the AER was removed, with or without control beads (compare Fig. 5j with the control limb in Fig. 5k, and data not shown). Taken together, these results suggest that cSef expression can be differentially regulated by FGFs, and point to a unique role for FGF2 in regulating cSef expression in the developing limb.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Sef is a newly identified feedback antagonist of FGFs. In zebrafish, the expression of Sef is induced by FGF signaling (Furthauer et al., 2002; Tsang et al., 2002). By isolating chick Sef and investigating its expression pattern during normal chick embryogenesis and under various experimental settings, we demonstrated the evolutionary conservation of this regulatory loop. Chick Sef is expressed in many sites of FGF action, and its expression is dependent on FGF signaling. FGFs induce ectopic cSef expression, and in the developing limb, cSef expression is not maintained in limbs where the AER is removed, unless FGF beads are applied. Moreover, we found that various members of the FGF family differentially regulate the expression of cSef.

The expression pattern of cSef was examined from stages 10–25 in whole embryos. cSef transcripts were detected in many sites, including the intermediate mesoderm, Hensen's node, posterior neural plate, MHB, forebrain, tail bud, branchial arches, limbs, otic vesicles, somites, eyes, and feather buds. These expression domains are intimately associated with known FGF signaling centers (Savage et al., 1993; Ohuchi et al., 1994; Shamim and Mason, 1999; Mayordomo and Alvarez, 2000; Walshe and Mason, 2000; Ford-Perriss et al., 2001; Tickle, 2002; Stolte et al., 2002; Trokovic et al., 2003; Liu et al., 2003) and are consistent with the role of Sef as an inhibitor of FGF signaling (Furthauer et al., 2002; Tsang et al., 2002; Niehrs and Meinhardt, 2002; Kovalenko et al., 2003; Xiong et al., 2003; Preger et al., 2004). The pattern of cSef expression during chick development closely resembles that reported for zebrafish and mouse (Furthauer et al., 2002; Lin et al., 2002), suggesting that the function of Sef is conserved throughout vertebrate evolution.

In the developing limb, cSef is expressed in the mesenchyme underling the AER. In this region, known as the progress zone, the AER maintains the cells in a rapidly proliferating, undifferentiated, and labile condition (Summerbell et al., 1973; Stark and Searls, 1973). In addition, this region is responsible for limb outgrowth and specification of skeletal elements (Summerbell et al., 1973). The most intense cSef signal was located in the distal portion of the progress zone, and then decreased proximally. Cells beyond the progress zone did not express detectable levels of cSef mRNA. The kinetics and expression pattern of cSef in the limbs can shed light on its role during normal limb development. Our expression studies localized the onset of cSef expression in the forelimb at stage 17. Because in the chick embryo the limb bud appears at stage 16 (Stark and Searls, 1973; Crossley et al., 1996), our findings indicate that cSef is not involved in regulating limb initiation. The changes in the size of cSef expression domain and its expression levels tightly correlate with the known proliferative state of the cells in the distal limb mesenchyme (Summerbell et al., 1973; Stark and Searls, 1973). Thus, the levels of cSef were the highest in areas with high growth rate, and its expression levels and overall size of expression domain decreased when differentiation was clearly apparent. These correlations strongly suggest that cSef is involved in regulating limb outgrowth at the proximal–distal axis. It can either fine tune or restrict the rate by which cells proliferate in the progress zone. Because cells differentiate only after they leave the progress zone (Summerbell et al., 1973; Johnson et al., 1994), another plausible function of cSef can be the maintenance of the undifferentiated state of cells in this area.

The expression of cSef in the underlying mesoderm is regulated by signaling emanating from the AER. We tested the ability of three members of the FGF family, known to be expressed in the ridge, to perform the AER function in maintaining cSef expression and found considerable differences among them. At a concentration as low as 0.1 mg/ml, beads soaked with FGF4 or FGF8 were unable to rescue cSef down-regulation. At higher concentrations (0.3 and 1 mg/ml for FGF4 and FGF8, respectively), FGF4 was still negative and FGF8 could maintain very faint cSef levels. FGF2 was by far more potent as it maintained normal or even higher cSef levels at the lowest concentration that was tested. Several lines of evidence exclude the possibility that the three growth factors differed in biological activity: (1) Beads soaked with each growth factor (0.1 mg/ml) induced the growth of an ectopic limb when implanted in the flank of HH15 embryos. (2) FGF4 readily induced ectopic cSef expression at the lowest concentration tested and was as potent as FGF2 in a mitogenic assay. (3) In the absence of AER, FGF8 induced limb outgrowth and up-regulated Shh expression in the posterior mesoderm at a concentration that was negative in inducing cSef expression. Taken together, these results suggest that FGF2 can fulfill the function of cSef inducer during normal limb development. Comparison of the spatiotemporal expression pattern of cSef and AER-FGFs during limb development further supports the candidacy of FGF2. Of the six FGFs known to be expressed in the AER, only FGF2 and FGF8 are expressed throughout the AER, a pattern that is complementary with that of cSef (Savage et al., 1993; Savage and Fallon, 1995; Crossley et al., 1996; Vogel et al., 1996). The remaining FGFs, FGF4, FGF9, FGF17, and FGF19, are localized to the posterior portion of the AER (Martin, 1998; Tickle and Münsterberg, 2001; Niswander, 2003; Kurose et al., 2004). In addition, their expression does not precede that of cSef as would be expected from an inducer (Laufer et al., 1994; Tickle and Münsterberg, 2001). The expression of FGF2 and FGF8 precede that of cSef, but FGF8 continues to be expressed in the ridge at stages where cSef expression is markedly reduced (Savage et al., 1993; Crossley et al., 1996; Vogel et al., 1996; Pizette and Niswander, 1999; Montero et al., 2001). The expression of FGF2, on the other hand, overlaps with that of cSef in the distal mesenchyme, and similar to cSef (present work), FGF2 expression decreases during differentiation (Savage et al., 1993; Dono and Zeller, 1994; Savage and Fallon, 1995). Interestingly, FGF2 mRNA and protein, similar to cSef, display dorsoventral and proximodistal distribution in the progress zone at stages of rapid proliferation (Savage et al., 1993; Savage and Fallon, 1995). Thus, FGF2 is expressed in the right time and locations to serve as cSef inducer and to maintain cSef expression in the apical mesoderm during normal limb development.

In addition to its expression in the AER, FGF2 is known to be expressed in the underlying mesoderm (Savage et al., 1993). Despite these, the mesodermally derived FGF2 could not maintain cSef expression in the absence of the AER. Since previous studies suggested that ridge derived-FGF2 regulates its own expression in the underlying mesoderm (Flott-Rahmel et al., 1992), it is reasonable to propose that following AER removal, the levels of FGF2 in the mesoderm drop below a threshold that is required for maintaining cSef expression. FGF2 is also expressed in the peripheral dorsal ectoderm and the underlining mesenchyme (Savage et al., 1993; Dono and Zeller, 1994) where cSef expression was not observed. FGF2 beads, however, readily induced ectopic expression of cSef when implanted in the dorsal margins of the limb. These results could suggest that the lack of endogenous cSef expression in these regions is due to the presence of an inhibitory factor whose effect can be alleviated by excess of FGF2. Alternatively, the endogenous levels of FGF2 are below the threshold required for normal cSef induction. In this respect, it is noteworthy that the levels of FGF2 in the ridge are known to be higher than those in the dorsal limb ectoderm (Savage et al., 1993).

Similar to their observed effect on endogenous cSef expression, FGFs differed in their ability to induce ectopic expression of cSef. Thus, only FGF2 and FGF4, but not FGF8, induced ectopic cSef expression. The intensity of expression in the flank was significantly weaker than in the limb sites, suggesting differences in tissue competence to FGF signals with respect to cSef expression. There was, however, an apparent difference between the expression pattern obtained with FGF2 and FGF4, which can be readily seen in the limbs (see Fig. 4). FGF2 induced an intense cSef signal immediately around the implanted bead, whereas FGF4 induced a weaker and a diffused signal at a distance from the bead. Since the response to FGFs is in general biphasic, a possible explanation for the different pattern of induction is that the concentrations used in the study were inhibitory in the case of FGF4 but not FGF2. Therefore, FGF4 could induce cSef expression only away from the bead where its concentrations are lower. Another possibility is that FGF2 directly induces cSef expression, whereas FGF4 acts indirectly. It is generally accepted that FGFs do not diffuse very far due to their tight binding to extracellular matrix and cell-associated heparan sulfate proteoglycans (Basilico and Moscatelli, 1992). FGF4 may act, therefore, on cells near the implantation site to trigger signals that were transmitted further away and resulted in cSef induction.

The expression domains of Sef and FGF8 closely correlate during both chick and mouse embryonic development (Furthauer et al., 2002; Lin et al., 2002; and present work). Therefore, the observation that FGF8 was not able to induce ectopic cSef expression or efficiently maintain endogenous cSef expression in the limb mesoderm was rather unexpected. The most likely explanation is that FGF8 is capable of inducing cSef expression but at times and locations that are different from those applied in the current study. For example, as in zebrafish (Tsang et al., 2002), FGF8 may induce cSef expression at earlier stages of development. Alternatively, a cofactor may be necessary in order for FGF8 to induce cSef expression, as with the induction of Shh expression in the anterior portion of the limb, which requires the cooperation of FGF8 with retinoic acid (Crossley et al., 1996). Further studies will be required to delineate the exact role of FGF8 in regulating cSef expression.

The observation that exogenous FGF4 and FGF8 could not maintain normal cSef expression in the apical mesoderm after AER removal is rather intriguing. Our current study, and studies by others (Niswander et al., 1993; Niswander and Martin, 1993; Johnson et al., 1994; Crossley et al., 1996; Vogel et al., 1996) demonstrated that the apical mesoderm is competent to respond to signals induced by exogenous FGF4 and FGF8. Thus, their failure to maintain normal levels of cSef after AER removal (as compared with FGF2) cannot be explained simply by different affinities for their signaling receptors. One possibility is that these FGFs require a cofactor to induce efficient cSef expression when added exogenously. Another possibility may be related to the mechanism by which cSef expression is regulated. FGFs can trigger several signal transduction cascades, including the Ras/MAPK, the PI3-kinase pathway, and the p38-MAPK (Maher, 1999; Powers et al., 2000; Boilly et al., 2000; Ong et al., 2001). FGFs can also differentially activate some of these pathways (Boilly et al., 2000). An attractive possibility, therefore, would be that FGF2, but not FGF8 or FGF4, can efficiently trigger the pathway leading to cSef induction in the limb. If correct, this hypothesis suggests that cSef expression is regulated downstream of the FGF-signaling receptor and that the pathway responsible for its expression can be differentially activated by the various FGFs. Therefore, an interesting avenue of research will be to identify the signaling pathway that is responsible for transcriptional activation of the cSef gene during normal limb development and how the different ridge derived FGFs affect this pathway.

In summary, we showed that cSef is expressed during chick embryogenesis from early to late stages of development and that its expression pattern overlaps with known sites of FGF activity. The kinetics of cSef expression in limbs suggests that cSef is involved in regulating limb outgrowth but not limb initiation. We showed that cSef expression in the developing limb is regulated by signaling emanating from the AER and that FGF2, but not FGF4 or FGF8, can efficiently replace AER function in maintaining cSef expression. To the best of our knowledge, this is the first demonstration for a unique effect of FGF2 during limb development.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Growth Factors, Reagents, and Chemicals

Restriction enzymes, enzymes used in cDNA cloning, and plasmid construction were from NEB, Pharmacia, and Roche. Recombinant human FGF2 was produced in bacteria and purified as previously described (Reich-Slotky et al., 1995). Recombinant human FGF4 and FGF8b were from R&D Systems. Bovine serum albumin was from ICN, and all other chemicals and reagents were from Sigma.

Cloning of cDNA and Detection of Chicken Sef Transcript

Database search using zfSef or hSef sequences as query revealed several ESTs of avian origin. Primers synthesized based on these sequences were used to clone the chicken Sef cDNA by polymerase chain reaction with reverse transcription from whole chick embryos total RNA. The GenBank accession numbers of cSef are AY278204 for the entire sequence of the Sef gene; and BF724076, 603816822F1, 603105-617F1, 603843254F1, 604146238F1, 603104329F1 for different portions of the gene. For the detection of Sef transcripts, total RNA was extracted from whole chick embryos or from embryonic brain as previously described (Eisemann et al., 1991). A total of 2 μg of total RNA were used for first-strand synthesis with primer from the 3′ UTR of cSef transcript (5′-ATGACATCTATAAACAGTATAACA). Amplification was carried out with primers 5′-GAAGAACCTCCTGTAATCACT and 5′-ATGACATCTATAAACAGTATAACA. Radiolabeled PCR products were obtained essentially as described (Munsterberg et al., 1995; Maroto et al., 1997).

Whole-Mount RNA In Situ Hybridization

The procedure was performed essentially as described by Wilkinson (1992). Embryos were treated with 2 mg/ml proteinase K for 5–8 min at room temperature before the hybridization. Digoxigenin-labeled sense and antisense probes were prepared according to the manufacturer's instruction (Roche). The following probes were used for the whole-mount in situ hybridization: a cDNA fragment corresponding to nucleotides 2122-2519 of cSef; Shh transcript, a 1.7-kb fragment of pHH2 (kindly provided by Clifford Tabin; Riddle et al., 1993). Specific activities of the probes were tested according to manufacturer instructions and were comparable. Incubation with alkaline phosphatase was between 3 and 4 hr. Embryos were staged according to Hamburger and Hamilton (1951). Frozen sections (18–20 μm) were made after the in situ hybridization, and photographs were taken using the differential interference contrast system of an inverted Leica microscope DMIRE2.

Implantation of FGF Beads Into Chick Embryos

Heparin acrylic beads (Sigma) were washed several times with PBS and soaked in PBS solution containing 0.1 mg/ml of FGF2, 0.1 or 1 mg/ml FGF8b, and 0.1 or 0.3 mg/ml FGF4. Control beads were placed in PBS solution. Beads were incubated for at least 1 hr at room temperature, placed on ice, and used within few hours for in ovo implantation. The bead was transferred to the egg using fine forceps and placed into a slit made in sites as indicated in the Result section.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Sharon Mink, Dr. Ifat Sher, and Dr. Orit Goldschmidt for critical review of the manuscript, and for useful discussions. D.R. was supported by grants from the Israel Cancer Research fund and the Technion-Neidersachsen Foundation, and R.R. was funded by the Malat Family Foundation.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES