FGF10 and Pax6 in Lacrimal Gland Development

Development 127, 2563-2572 (2000) 2563 Printed in Great Britain © The Company of Biologists Limited 2000 DEV4343

FGF10 is an inducer and Pax6 a competence factor for lacrimal gland development

Helen P. Makarenkova1,*, Masataka Ito1,*, Venkatesh Govindarajan2, Sonya C. Faber, Li Sun3, Gerald McMahon3, Paul A. Overbeek2 and Richard A. Lang1,‡ 1Skirball Institute for Biomolecular Medicine, Developmental Genetics Program, Cell Biology and Pathology Departments, New York University Medical Center, 540 First Avenue, New York, NY 10016, USA 2Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX 77030, USA 3SUGEN, Inc., 230 East Grand Ave., South San Francisco, CA 94080-4811, USA *These authors have contributed equally to this publication ‡Author for correspondence (e-mail: [email protected])

Accepted 11 April; published on WWW 23 May 2000

SUMMARY

We investigated the mechanism of tissue induction and FGF10-null mice strongly suggest that it is an endogenous specification using the lacrimal gland as a model system. inducer. This was supported by the observation that This structure begins its morphogenesis as a bud-like inhibition of signaling by a receptor for FGF10 (receptor 2 outgrowth of the conjunctival epithelium and ultimately IIIb) suppressed development of the endogenous lacrimal forms a branched structure with secretory function. Using bud. In explants of mesenchyme-free gland epithelium, a reporter transgene as a specific marker for gland FGF10 stimulated growth but not branching epithelium, we show that the transcription factor Pax6 is morphogenesis. This suggested that its role in induction is required for normal development of the gland and is to stimulate proliferation and, in turn, that FGF10 probably an important competence factor. In investigating combines with other factors to provide the instructive the cell-cell signaling required, we show that fibroblast signals required for lacrimal gland development. growth factor (FGF) 10 is sufficient to stimulate ectopic lacrimal bud formation in ocular explants. Expression of FGF10 in the mesenchyme adjacent to the presumptive Key words: FGF, FGF10, Budding morphogenesis, Paracrine lacrimal bud and absence of lacrimal gland development in induction, Lacrimal gland, Pax6

INTRODUCTION (FGF) families (Mason et al., 1994). In particular, FGF10, also known as keratinocyte growth factor-2 (KGF2), has been Development of the lacrimal gland is an example of an implicated in budding outgrowth of epithelia (Bellusci et al., epithelial-mesenchymal interaction. In the mouse, a single bud- 1997b; Ohuchi et al., 1997; Sekine et al., 1999). like invagination of the conjunctival epithelium at the temporal Both FGF7 and FGF10 can bind to and signal through extremity of the eye is the initial sign of lacrimal gland fibroblast growth factor receptor-2 (FGFR2 (Ornitz et al., formation (Kammandel et al., 1999). The mesenchymal cells 1996)). The mouse FGFR2 gene has alternative splicing for that surround the point of epithelial budding are the periocular exons 8 and 9 and this results in an alternative amino-acid cells, of neural crest origin (Johnston et al., 1979). The tubular sequence in the C-terminal half of the third immunoglobulin invagination of the lacrimal gland extends and branches fold of the extracellular domain. The IIIb (exon 8) isoform is multiple times to give the lobular structure of the mature gland known as the KGF receptor and the IIIc isoform as bek (Kammandel et al., 1999). (Johnson and Williams, 1993; Xu et al., 1998). bek binds with The type of morphogenesis (Hogan, 1999) that accompanies high affinity to FGF1, FGF2, FGF4 and FGF5 but not FGF7 development of the lacrimal gland has been studied in detail in (Ornitz et al., 1996). In contrast, the KGFR binds FGF7 and several other organ systems, including the limb (Martin, 1998), FGF10 as well as FGF1; the binding of FGF2 to the KGFR is the lung (Peters et al., 1994; Hogan and Yingling, 1998; at very low affinity (Miki et al., 1992; Lu et al., 1999). The Weaver et al., 1999) and teeth (Peters and Balling, 1999). As observation that expression of the KGFR is restricted to a result of these analyses, a selection of soluble signaling epithelial lineages and that bek is found only in mesenchymal molecules has been implicated in generating morphology of cells (Orr-Urtreger et al., 1993b; Iseki et al., 1997; Xu et al., this type. These include sonic hedgehog (Bellusci et al., 1997a; 1998) indicates that tissue-specific alternative splicing is one Pepicelli et al., 1998) and members of the bone morphogenetic way to provide FGF signaling specificity. protein (BMP) (Graff, 1997) and fibroblast growth factor The Pax6 gene is expressed in many tissues of the eye 2564 H. P. Makarenkova and others

(Grindley et al., 1995). Recent analysis has identified a 10 µM in the presence of Lipofectamine-2000 (Gibco BRL) in a 30% conserved transcriptional enhancer that, in the context of F 127 pluronic gel (Makarenkova and Patel, 1999). An 8 µl drop of reporter transgenes, is necessary and sufficient for directing oligonucleotide/Lipofectamine/pluronic gel was placed on the surface expression to the lens placode, the lens and corneal epithelium of the ocular region explants and at the temporal edge of the eye. In and, subsequently, to the epithelium of the lacrimal gland each experiment, one side of embryo was treated with sense and the (Williams et al., 1998; Kammandel et al., 1999). Pax6 is a other side with antisense. The statistical significance of the difference in control and experimental explant responses was assessed using member of the group of transcription factors that has both the nonparametric Wilcoxon signed-rank test (Ostle and Mensing, paired and homeodomain DNA binding motifs (Mansouri et 1975). P<0.01 was considered to reflect a statistically significant al., 1994). It has been broadly implicated in eye development difference. with the demonstration that Drosophila, mouse and human eye defects are a consequence of mutations in Pax6 (Quiring et al., Inhibitors of FGFR signaling 1994). Furthermore, in both invertebrates (Halder et al., 1995) SU5402 and SU9597 are compounds from a new family of inhibitors and vertebrates (Chow et al., 1999), Pax6 can direct the for the FGF receptor tyrosine kinases (Mohammadi et al., 1997; Sun formation of ectopic eyes. et al., 1999) and were employed to block FGF signaling in vitro. The In this study, we have taken advantage of a Pax6-based chemical name for SU5402 is 3-[4-methyl-2-(2-oxo-1,2-dihydro- reporter transgene (Williams et al., 1998) to study the indol-3-ylidenemethyl)-1H-pyrrol-3-yl]-propionic acid and its structure is given in Fig. 1. SU9597 is 3-{2-[6-(3-methoxy-phenyl)- mechanism of lacrimal gland development. We have shown, 2-oxo-1,2-dihydro-indol-3-ylidenemethyl]-4-methyl-1H-pyrrol-3- using the Small eye mouse, that normal lacrimal gland yl}-propionic acid (Fig. 1). For explant cultures, inhibitors were development requires a wild-type level of Pax6 expression in dissolved in dimethylsulfoxide (DMSO) as a 100 mM stock solution, conjunctival epithelium. In addition, we demonstrate that divided into portions and stored at –20oC. The compounds were then signaling through FGFR2IIIb is necessary for lacrimal gland diluted in culture medium to a final concentration of between 2.5 and induction and in the context of ventral periocular mesenchyme, 100 µM. is sufficient. This is consistent with the localized expression of FGF10 in the periocular mesenchyme adjacent to Pax6- Whole-mount in situ hybridization expressing presumptive lacrimal gland epithelium, and Whole-mount in situ hybridization was performed as described (Nieto suggests a model in which FGF10 provides a paracrine et al., 1996). Probes were generated from plasmids containing the full- length rat FGF10 (kindly provided by N. Itoh), mouse FGF7 (Mason stimulus for outgrowth of the lacrimal gland bud. et al., 1994) and mouse FGFR2 (Deng et al., 1996) cDNAs.

MATERIALS AND METHODS RESULTS Histological analysis The morphology of normal lacrimal gland Paraffin sections were prepared and stained either with Hematoxylin development and Eosin or only Hematoxylin, using conventional methods. We had previously shown that a highly conserved enhancer 3.5 Explant cultures kb upstream of the Pax6 gene P0 promoter could direct reporter Explant cultures were prepared from embryonic day (E)12.5-14.5 construct expression to the lens and corneal epithelium in fetuses of the P6 5.0 lacZ reporter line. The whole eye, together with transgenic mice (Williams et al., 1998). It has also been shown ectoderm and mesenchyme from one side of the head, was excised that this same region was sufficient for reporter expression in and placed on a filter (0.8 µm pore, Millipore) supported by a stainless the epithelial component of the lacrimal gland (Kammandel et steel grid. Explants were cultured in CMRL-1066 medium (Gibco al., 1999). BRL) supplemented with heat-inactivated 10% fetal calf serum, 2% At embryonic day (E)12.5, the P6 5.0 lacZ transgene is rat serum, glutamine, non-essential amino acids and an antibiotic- antimycotic (Gibco BRL). After culture for 48 hours, tissues were expressed in a ring of ectodermal cells surrounding the fixed and stained with X-gal according to established protocols (Song immature cornea (Williams et al., 1998). This includes the et al., 1996). Mesenchyme-free explants were generated from E16.5- region destined to form the epithelial component of the E17.0 lacrimal glands according to established procedures (Shannon lacrimal gland (Fig. 2A, red arrowhead). The first gross et al., 1999). The tips of developing buds were cultured in defined medium (Zuniga et al., 1999) alone or defined medium supplemented with FGF2 or FGF10 at 80 ng/ml. FGF-loaded beads were used as previously described (Cohn et al., 1995). Briefly, heparin acrylic beads (Sigma) of 80-120 µm diameter were washed in PBS and immersed in PBS containing BSA with or without recombinant FGFs. Beads were incubated with BSA or FGF solution (1 mg/ml) at 4°C overnight. Human recombinant FGF7 and 10 were obtained from R&D systems while human recombinant FGF2 was a gift from D. Rifkin. For bead implantation, an explant was punctured through the ectoderm near the conjunctival epithelium with a tungsten needle, and a bead was inserted into the mesenchyme using blunt forceps. Antisense experiments The sense and antisense oligonucleotides used have been described previously (Miralles et al., 1999) and were used at final concentration Fig. 1. Chemical structures of SU5402 and SU9597. FGF10 and Pax6 in lacrimal gland development 2565 indication of lacrimal gland development arises at E13.5, lacrimal gland was also observed in Sey heterozygous animals where an ectodermal bud appears at the temporal edge of the that did not carry the P6 5.0 lacZ reporter (data not shown). eye (Fig. 2B,C). In section at E13.5, the developing lacrimal bud is observed at the fornix (the deepest rim of invaginating FGF10 is expressed in mesenchyme adjacent to the conjunctival epithelium) on the temporal side. The P6 5.0 lacZ lacrimal gland primordium reporter is expressed in the lacrimal bud (Fig. 2D, red Where budding morphogenesis has been characterized, arrowhead), but on the nasal side the fornix is negative for members of the FGF family have been implicated (Hogan, reporter expression (Fig. 2E, black arrowhead). The Harderian 1999). In particular, FGF10 stimulates proliferation of gland primordium projects from the retinal side of the nasal endodermal cells in the lung (Bellusci et al., 1997b) and of conjunctival epithelium (Fig. 2E, blue arrowhead) and this too ectodermal cells in the limb (Ohuchi et al., 1997; Xu et al., does not express the P6 5.0 lacZ reporter. 1998; Sekine et al., 1999). At E15.5, extension of the lacrimal bud has given a tubular By whole-mount in situ hybridization at E13.5, the FGF10 structure with a thickening at the tip (Fig. 2F). At this stage, mRNA was detected in the periocular mesenchyme with the bud has extended caudally in the subcutaneous tissue concentrations on the nasal and temporal side of the eye (Fig. between skin and cranial bones and underneath the supraorbital 4A, arrows) in positions adjacent to the budding of lacrimal branch of the stapedial artery (Wakusawa, 1968) (Fig. 2F, sas). (temporal) and Harderian (nasal) glands. The distance between Branching in the developing gland appears between E15.5 and the conjunctival epithelium and FGF10 expressing E16.5 (Fig. 2G). By E17.5, the exorbital lobe is located close mesenchyme was greater in the ventral region. At E14.5, to the pinna (Fig. 2H). Flattened whole-mount specimens from expression of FGF10 was detected adjacent to the tip of the both reporter (Fig. 2I) and wild-type mice (Fig. 2J) reveal that extending lacrimal bud (Fig. 4B). Hybridization signal for the branching at E17.5 is unaffected by the reporter transgene. At FGF7 transcript in the eye region was less distinct but appeared this stage a proximal branch projects ventrally (marking the to be present in both the dorsal and ventral periocular future intraorbital lobe of the gland; Fig. 2K, arrow) while the mesenchyme (Fig. 4C). No distinctive expression domain of exorbital lobe lies superficial to the stapedial artery (Fig. 2K, FGF7 was associated with outgrowth of the lacrimal gland bud. sa). By E19.5, further branching is apparent in both the intra- Sey heterozygous animals retained the pattern of FGF10 and exorbital lobes (Fig. 2L). As would be anticipated, in the transcript distribution (Fig. 4D). Indeed, even in the embryos mature gland, expression of the P6 5.0 lacZ transgene is that phenotypically were Sey homozygotes, FGF10 expression restricted to the epithelially derived component (Fig. 2M). was observed in a location that was probably equivalent to lacrimal bud mesenchyme (Fig. 4E). A probe to FGFR2 Pax6 is required for normal development of the showed that the gene is expressed in the conjunctival lacrimal gland epithelium, the whisker follicles, and the pinnae of wild-type Given the expression of Pax6 in presumptive lacrimal gland animals (Fig. 4F). The pattern of expression appeared epithelium, we determined whether lacrimal gland formation unchanged in Sey heterozygotes (data not shown). The was defective in heterozygous Sey mice (Hogan et al., 1986). distinctive location of FGF10 and FGFR2 expression The Sey allele is a point mutation in Pax6 that results in suggested a role in formation of both the lacrimal and truncation of the protein prior to the homeodomain (Hill et al., Harderian glands. 1991). SeyNeu is a mutation in a splice donor that also results in premature translation termination but in the transcription FGF7 and FGF10 can induce ectopic lacrimal bud activation domain (Callaerts et al., 1997). Mice that are formation heterozygous for either of these Sey alleles have an eye that is In order to determine how FGFs 7 and 10 might influence small but generates the conjunctival epithelium that gives rise lacrimal gland development, we established an explant culture to the lacrimal gland (Hill et al., 1991; Ton et al., 1992). The system. The tissue explanted included the eye together with early arrest of eye formation precluded a meaningful surrounding ectoderm and mesenchyme. Explants were examination of lacrimal gland development in homozygote Sey established at E13.0 on Millipore filters placed on metal grids mice. and were cultured for 48 hours. Between 40% and 75% of the Interbreeding of Sey and SeyNeu with mice carrying the P6 explants developed endogenous lacrimal gland buds (Fig. 5, 5.0 lacZ transgene gave offspring heterozygous for the Sey Table 1). To test the action of FGFs 7 and 10, heparin- alleles and hemizygous for P6 5.0 lacZ. Lacrimal gland Sepharose beads were loaded with each factor and inserted at development was assessed at E13.5-19.5 and found to be various locations into the periocular mesenchyme adjacent to defective with either Sey allele (Fig. 3). Specifically, the the conjunctival epithelium. Explants were cultured for a lacrimal gland bud was absent at E13.5 when it would appear period of 48 hours and then fixed and stained for lacZ in wild-type mice (Fig. 3). At E15.5 too, when the wild type expression. lacrimal gland has extended some distance from the Both FGF7 and FGF10 could stimulate the formation of conjunctiva (Fig. 3A), there was no evidence of this structure ectopic lacrimal buds (Table 1). Ectopic buds extended in the in either Sey (Fig. 3B) or SeyNeu (data not shown) direction of the implanted bead from the conjunctival heterozygotes. When the primordia of both exorbital and epithelium. This was evident either in whole mounts (Fig. intraorbital lacrimal gland lobes were apparent in wild-type 5A,C) or in histological section (Fig. 5B,E,F). In many cases, mice at E17.5 (Fig. 3C), Sey heterozygotes show the first ectopic buds (Fig. 5B) had morphology very similar to that indication of lacrimal gland budding (Fig. 3D, red arrow). observed in endogenous buds (Fig. 5D). There was no E19.5 Sey+/− animals had a vestigial gland compared with wild discernible difference in the morphology of buds induced by type (Fig. 3E,F). The absence of the exorbital lobe of the FGF7 compared with those induced by FGF10. In a few cases, 2566 H. P. Makarenkova and others

Fig. 2. The morphology of lacrimal gland development. (A) Section of the eye in an E12.5 P6 5.0 lacZ reporter mouse. lacZ expression is indicated by the blue X-gal labeling in the lens epithelium (le) and the future conjunctival epithelium on the temporal side of the eye (red arrowhead). (B) Head region of E13.5 P6 5.0 lacZ reporter mouse showing X-gal-labeled corneal epithelium and lens. The lacrimal bud (red arrowhead) is observed projecting towards the auditory canal (ac). (C) As in B, but at higher magnification. The X-gal labeled lens epithelium (l) can be seen through the cornea, the epithelium (c) of which is also labeled. The lacrimal gland bud (red arrowhead) projects form the X-gal-labeled conjunctival epithelium. (D,E) Histological sections through the eye in an E13.5 P6 5.0 lacZ reporter mouse. (D) Blue X-gal labeling is apparent in the continuous layer of the embryonic corneal and conjunctival epithelium (ce) and the early lacrimal gland bud (red arrowhead) is seen extending from the deepest aspect of the conjunctival epithelium (the fornix). (E) Reporter construct expression is absent from the fornix on the nasal side (black arrowhead). Early budding of the Harderian gland primordium is apparent on the surface of the epithelium on the retinal side and is also negative for reporter expression (blue arrowhead). (F) The tip of the lacrimal bud has extended to a position dorsal to the stapedial artery (sa) and caudal to the bifurcation of its supraorbital (dashed lines – sas) and infraorbital (dashed lines – sai) branches by E15.5. (G) By E16.5 the first branching events are apparent. (H) By E17.5, the exorbital lobe of the lacrimal gland (ex) is nearing its final position adjacent to the pinna (p). Flattened exorbital lobe preparations either from an X-gal-labeled reporter mouse (I) or a wild-type mouse (J) at E17.5 show the degree of branching at this stage. In addition, at E17.5 (K) and E19.5 (L), the first bud of the intraorbital lobe of the lacrimal gland (small arrow) projects ventrally in a proximal position. (K) The exorbital lobe is superficial to the stapedial artery (sa). (L) Both intraorbital and exorbital lobes have branched further by E19.5. (M) Expression of the reporter in the epithelial component of the mature exorbital gland acini is apparent in histological section.

multiple buds were observed (Fig. 5C). Of those explants where an endogenous lacrimal bud developed and an FGF10 bead was inserted, 24% gave an ectopic bud in response (Table 1). For FGF7, the response rate was lower at 13% (Table 1). In addition to clear cases of budding, in some explants, the response to an FGF7- or FGF10-loaded bead was an epithelial overgrowth (Table 1), where bud morphology was less certain. We presume that these events indicate incomplete bud formation. The total response discernible by these criteria (ectopic buds and epithelial overgrowth) for FGF7 was 40%

Fig. 3. Pax6 is required for normal development of the lacrimal gland. Lacrimal gland development at E15.5 in wild-type (A) and Sey+/− (B) embryos showing absence of lacrimal budding at the appropriate time. (C) At E17.5, both exorbital and intraorbital (arrow) components of lacrimal gland are apparent in wild-type mice. (D) In Sey+/− mice at the same stage, a lacrimal bud is just becoming visible (pink arrowhead). By E19.5, extensive branching is apparent in both lobes of the lacrimal gland in wild-type mice (E, arrow) while only one lobe exists in Sey heterozygotes (F). FGF10 and Pax6 in lacrimal gland development 2567

Fig. 4. FGF7 and FGF10 are expressed in mesenchyme adjacent to the lacrimal gland primordium. (A) Whole-mount in situ hybridization for FGF10 in a wild-type mouse embryo at E13.5. The conjunctival boundary is indicated by the dashed red line. Hybridization signal can be seen in periocular mesenchyme. The most intense labeling is seen at the temporal aspect of the eye where the lacrimal bud is emerging (red arrowhead). An intense region of labeling is also apparent on the nasal side (black arrow), on the pinna (p) and in whisker follicles (wf). (B) Whole-mount in situ hybridization for FGF10 in excised eye region tissue at E14.5. The view is from the retinal side of the eye. The outline of the eye is indicated by the dashed white line. FGF10 hybridization signal (red arrowhead) is apparent in the mesenchyme surrounding the elongating lacrimal bud (lb). (C) FGF7 probe hybridization signal in Fig. 5. FGF7 and FGF10 can induce ectopic lacrimal gland buds. E14.5 whole-mount wild-type embryo. Labeling is apparent in the (A) Whole mount of the explanted eye region from P6 5.0 lacZ lens (le) and within or adjacent to the conjunctival epithelium, but is reporter mouse showing X-gal labeling of corneal epithelium (ce), diminished in the region of lacrimal gland budding (red arrowhead). endogenous lacrimal bud (end) and ectopic lacrimal bud (ect) +/− (D) FGF10 whole-mount hybridization in and E13.5 Sey embryo. adjacent to the FGF10-loaded bead (b). The red line through the The distribution of labeling on the temporal side of the eye appears image approximates the plane of section for the image shown in (B) similar to that observed in wild-type mice (red arrowhead). p, pinna. where the conjunctival epithelium (co) has budded and branched The brown pigment granules of the immature retinal pigment adjacent to the bead (b). Adjacent is the presumptive neural retina epithelium are apparent. The region of salivary gland development (nr). (C) As in A, except that two ectopic buds (ect1, ect2) are (black arrow) also expresses FGF10. (E) FGF10 in situ hybridization observed adjacent to the endogenous bud (end). The three red lines in a Sey homozygote shows expression in salivary gland region indicate section planes for each of the buds that are shown in (black arrow) and in a location probably equivalent to lacrimal bud subsequent panels. (D) Section through endogenous lacrimal bud mesenchyme (red arrowhead), despite the absence of an eye in the from the example in C. (E,F) Sections through the ﬁrst (E; ect1) and anticipated location (asterisk). (F) Whole-mount in situ hybridization second (F; ect2) ectopic buds formed in response to an FGF10- for the FGFR2 in wild-type embryo at E13.5. The FGFR2 signal is loaded bead (F; b). (G-L) Mesenchyme-free lacrimal gland epithelial observed in the conjunctival epithelium and additional facial explants stimulated with either no factor (G,J), or with FGF2 (H,K) structures including the pinna (p), the whisker follicles (wf) and the or with FGF10 (I,L). Explants were photographed at the same salivary gland primordium (black arrow). le, lens. magniﬁcation on day 1 (G-I) and day 4 (J-L) of culture. and for FGF10 was 39%. Interestingly, when FGFs 10 and 7 low affinity, and signals primarily through both isoforms of were applied to explants from stage E11.5-12.0, conjunctival FGFR1 and the IIIc isoforms of FGFRs 2 and 3. Thus, to epithelium did not form ectopic lacrimal buds or show determine whether activation of these different classes of FGF epithelial overgrowth (data not shown). This may indicate that receptors could mediate lacrimal gland induction, we inserted there is a temporally restricted window of competence. FGF2-loaded beads into the periocular mesenchyme of eye FGFs 7 and 10 bind with high affinity to FGFR2IIIb (Lu et region explants. Only one case of ectopic lacrimal bud al., 1999). In contrast, FGF2 can bind FGFR2IIIb only with induction was observed in response to FGF2 used at the same 2568 H. P. Makarenkova and others

Table 1. Ectopic lacrimal gland bud formation in response Table 2. Suppression of FGF signaling inhibits lacrimal to FGF7 and FGF10 gland formation Explants Concentration Budding µ Number Number Number Agent ( M) (%) with with with Number of Control 0* 18/24 (64) endogenous ectopic epithelial Total independent SU5402 100 0/9 (0) Bead Number bud bud overgrowth response experiments 20 0/9 (0) reagent generated (%) (%)* (%)* %* performed 2.5 0/8 (0) BSA 42 17 (40) 0 (0) 0 (0) 0 17 SU9597 20 0/6 (0) FGF2 29 15 (52) 1 (7) 0 (0) 7 3 5 0/8 (0) FGF7 75 56 (75) 7 (13) 15 (27) 40 10 2.5 1/8 (13) FGF10 52 34 (65) 8 (24) 5 (15) 39 5 Sense 10 14/21‡ (67) Antisense 10 5/20‡ (25) *Expressed as the percentage of those with an endogenous bud. Beads were placed adjacent to ventral conjunctival epithelium. *Control explants contained the inhibitor vehicle DMSO. This did not inhibit lacrimal gland budding at even the highest concentration tested (0.02%). ‡Statistical significance was assessed using the non-parametric Wilcoxon concentration as FGF7 and FGF10 (1 mg/ml). Furthermore, signed-rank test (Ostle and Mensing, 1975). Using this test for 20 pairs of although beads were implanted in all quadrants of periocular sense and antisense treated explants, a T value of 2.4 was obtained. Since this mesenchyme, ectopic budding in response to FGF7 or FGF10 was less than the critical T value of 3, the hypothesis that the sense and antisense responses were the same was rejected at a confidence level of was not observed in a dorsally located sector. Control BSA- P=0.01. loaded beads showed no effect on lacrimal bud induction. To determine whether the action of FGF10 on lacrimal gland epithelium was direct, we generated explants of the gland in receptor or kinase activity of epidermal growth factor (EGF) which the mesenchyme had been removed. When cultured over receptor, and only weak inhibition of the platelet-derived 4 days in defined medium in the absence of an FGF or in the growth factor (PDGF) receptor was described (Mohammadi et presence of FGF2, the explants did not grow and did not form al., 1997) for SU5402 at high concentrations. additional buds (Fig. 5H-K). By contrast, in the presence of Explants of the eye region from the P6 5.0 lacZ reporter FGF10, epithelial explants grew dramatically in size (Fig. transgenic mice were cultured from E13.5 for 48 hours in the 5L,M) and continued to express the P6 5.0 lacZ reporter (data presence or absence of SU5402 and SU9597 (Table 2). In not shown). This indicates that FGF10 can act directly on the control experiments where explants were treated with the epithelium to stimulate growth. inhibitor vehicle, lacrimal gland budding occurred at typical frequency (64%). Treatment with SU5402 resulted in a FGF10 is necessary for both lacrimal and Harderian complete inhibition of lacrimal bud formation at gland formation concentrations of 2.5-100 µM. Similarly, SU9597 prevented The necessity for FGF10 in ocular gland formation was tested lacrimal bud formation completely, except at the lowest by examining the FGF10-null mice (Min et al., 1998). concentration tested of 2.5 µM where one out of eight explants Sectioning showed that in wild-type E18.5 embryos, the formed a bud. These data indicated that FGF receptor developing lacrimal gland was bordered by the carotid artery, activation was required for the formation of the endogenous the masseter muscle and the dermis (Fig. 6A). In the FGF10- lacrimal gland bud. null embryos, the epithelial component of the gland was To determine whether, as would be anticipated, the FGF10- completely absent, though the mesenchyme remained (Fig. binding FGFR2IIIb was required for formation of the lacrimal 6B). An examination of serial sections through the eye region gland, we used antisense oligonucleotides as a specific of FGF10-null mice indicated the absence of any lacrimal bud inhibitor. An increasing number of developmental systems formation from the conjunctival epithelium (data not shown). have used antisense oligonucleotides successfully to probe the Similarly, the Harderian gland epithelium, which in wild-type action of a given gene product. In particular, antisense animals is located within mesenchyme adjacent to the orbit oligonucleotides to FGFR2IIIb have been used to examine the (Fig. 6C), was absent from the FGF10-null mice (Fig. 6D). mechanism of lung branching (Post et al., 1996) and pancreas Taken together, these data indicate that FGF10 is required for development (Miralles et al., 1999). development of both the lacrimal and Harderian glands. Explants were cultured either in the presence of control oligonucleotide or an antisense oligonucleotide Ectopic lacrimal bud formation requires signaling complementary to exon 8 of mouse FGFR2, and the through FGFR2IIIb appearance of lacrimal gland buds monitored. Explants were The expression pattern and activity of FGF10 in inducing performed either at E13.5, to monitor FGFR2IIIb function in ectopic lacrimal buds implied that normal lacrimal gland bud induction (Table 2), or at E14 in order to monitor bud formation would depend on signaling through FGF receptors. extension (Fig. 7). In each experiment one side of embryo was To test whether this was the case, we employed a new family treated with sense and another side with antisense. of FGF signaling inhibitors that bind specifically to the active Oligonucleotides were dissolved in pluronic gel and placed on site of FGF receptor kinase domains (Mohammadi et al., the explant in the culture dish as a means of maintaining locally 1997). The compounds SU5402 and SU9597 are specific high concentrations of oligonucleotide. The formation of inhibitors for the FGF receptors. In the range of 10-200 µM lacrimal gland buds was significantly inhibited when explants they did not inhibit tyrosine phosphorylation of the insulin were treated with 10 µM antisense oligonucleotide (Table 2) FGF10 and Pax6 in lacrimal gland development 2569 while the same concentration of sense oligonucleotide inferred from the observation that dorsally located Pax6- permitted lacrimal bud formation at control levels (Table 2). expressing conjunctival epithelium does not produce ectopic Antisense oligonucleotide was also shown to suppress the lacrimal buds in response to FGF10 (Fig. 7). Though we had elongation of lacrimal gland buds (Fig. 7) under conditions considered the possibility that Pax6 expression in conjunctival where sense oligonucleotides had no effect (Fig. 7). The results epithelium was required for induction of FGF10 expression in of these experiments argue that FGFR2IIIb is critical for the adjacent mesenchyme, the normal pattern of FGF10 mRNA induction of the lacrimal bud. distribution in Sey mutants argues against this. For this reason too, it is likely that perturbed lacrimal gland formation in Sey heterozygotes reflects an aspect of epithelial programming that DISCUSSION is independent of or downstream of the initial epithelial- mesenchymal signal exchange. Using a variety of strategies, we have investigated the mechanism of lacrimal gland development. The experiments Superimposed domains of epithelial competence performed show that both the soluble mediator FGF10 and the and inducer expression determine ocular gland transcription factor Pax6 are required for normal lacrimal gland location formation. We also show that in the context of periocular The implantation of FGF10-loaded beads throughout the mesenchyme, FGF10 is sufficient for induction of the lacrimal periocular mesenchyme resulted in formation of ectopic bud, an observation consistent with the formation of ectopic lacrimal buds only in ventral, nasal and temporal domains. This glands in transgenic gain-of-function analyses (V. G., M. I., H. showed that a dorsal sector of conjunctival epithelium was not P. M., R. A. L. and P. A. O., unpublished data). FGF10 has an competent to form the lacrimal gland (Fig. 7). The FGF10 expression pattern which strongly suggests that it is an transcript is present at high levels and immediately adjacent to endogenous inducer of the lacrimal gland. Combined with the conjunctival epithelium in an arc that extends over the evidence for a restricted domain of lacrimal gland-competence anterior, dorsal and ventral aspects (Fig. 7). In a ventral sector, in conjunctival epithelium, these data provide new information FGF10 transcripts are at low levels and separated from the on the mechanism of lacrimal gland induction and on the conjunctival epithelium. The FGF10 transcript is at its highest mechanism by which the location of the ocular glands is levels immediately adjacent to the location of Harderian and determined. lacrimal gland outgrowth. Since in culture FGF10 can induce ectopic lacrimal buds ventrally, this suggests that although Ectopic glandular structures represent developing competent epithelium extends throughout the ventral domain, lacrimal glands in vivo there is insufficient FGF10 in the central sector to The P6 5.0 lacZ transgene is expressed in the lens placode, the stimulate lacrimal gland budding. In turn, these data suggest a lens epithelium (Williams et al., 1998) and the epithelial model in which superimposed domains of epithelial component of the lacrimal gland, but not in the Harderian competence and FGF10 expression give precisely localized gland. This argues that any glandular tissue expressing the gland induction (Fig. 7). reporter represents lacrimal gland. Ectopic application of FGF7 or FGF10 in an explant of the eye region results in the FGF10 is an essential component of the formation of ectopic gland-like buds and tubes. Morphological endogenous lacrimal gland induction system assessment of this response indicates that the buds arise from Both FGF7 and FGF10 can elicit ectopic lacrimal bud the conjunctival epithelium and that they strongly express the development from the conjunctival epithelium in explant P6 5.0 lacZ transgene. The morphology of the ectopic culture and from corneal epithelium in transgenic gain-of- outgrowth, combined with Pax6 reporter gene expression, function analysis (Lovicu et al., 1999; V. G., M. I., H. P. M., establishes the identity of the ectopic structures as developing R. A. L. and P. A. O., unpublished data). Based on the lacrimal gland. expression domains of both factors, FGF10 is the best candidate for an endogenous inducer as it is expressed with Pax6 is a competence factor for normal lacrimal appropriate spatial pattern. In situ hybridizations with an FGF7 gland development probe indicate that it has a more uniform pattern of expression The lack of normal lacrimal gland formation in the Sey (Hill around the eye, which does not imply a role in stimulating et al., 1991) heterozygous mice indicates that the Pax6 endogenous lacrimal or Harderian gland development. transcription factor is required. Pax6 is expressed in the The response of conjunctival epithelium to FGF7 and FGF10 conjunctival epithelium but not in the neural crest-derived is presumably a reflection of their high affinity for FGFR2IIIb periocular mesenchyme (Grindley et al., 1995; Callaerts et al., (Orr-Urtreger et al., 1993a; Xu et al., 1998). Consistent with this 1997). This suggests that the requirement for Pax6 in gland responsiveness is the expression of FGFR2 in the precursor formation is autonomous to the cells of the precursor conjunctival epithelium (Finch et al., 1989) and the observation conjunctival epithelium and is consistent with an autonomous that FGFR tyrosine kinase inhibitors (Mohammadi et al., 1998) requirement for Pax6 in the formation of the lens (Fujiwara et and an antisense oligonucleotide specific for the FGFR2IIIb al., 1994; Quinn et al., 1996; Altmann et al., 1997; Chow et mRNA (Miralles et al., 1999) can prevent lacrimal gland al., 1999). It is likely that Pax6 is one factor that establishes budding. When combined with the observation that lacrimal competence and permits gland development from conjunctival gland development is absent in FGF10-null mice, these data epithelium in response to an FGF ligand. indicate that mesenchymal FGF10 stimulates lacrimal gland It is unlikely that Pax6 is sufficient to confer gland- morphogenesis by directly activating FGFR2IIIb in the competence on conjunctival epithelium, however. This can be conjunctival epithelium. We would predict that the FGFR2IIIb 2570 H. P. Makarenkova and others

Fig. 6. FGF10 is required for development of the ocular glands. (A) Section through the head of a wild- type E18.5 mouse showing the immature lacrimal gland (arrowheads) bordered by the carotid artery (ca) the masseter muscle (mm) and the dermis (der). (B) In an FGF10-null mouse at the same age, the epithelial component of the lacrimal gland is absent although the triangular region of mesenchyme (arrowheads) is present. (C) The Harderian gland in a wild-type E18.5 mouse is found adjacent to the nasal retina. (D) In an FGF10-null mouse at the same age, the Harderian gland is absent.

isoform-speciﬁc gene-targeted mice (De Moerlooze et al., formation of ectopic Harderian glands (V. G., M. I., H. P. M., 2000) would display lacrimal gland agenesis. Since it is R. A. L. and P. A. O., unpublished data), and (3) FGF10 is expressed at the appropriate time and place, is necessary for expressed in Harderian gland mesenchyme, we can also suggest gland formation and is sufficient for formation of an ectopic that FGF10 is an inducer of both ocular glands (Fig. 7). bud, FGF10 fulﬁlls the criteria for an endogenous inducer (Slack, 1993). Given that (1) Harderian glands are absent in FGF10-null mice, (2) misexpression of FGF10 leads to the

Fig. 7. FGFR2IIIb signaling is required for lacrimal gland development. The red arrowheads indicate either the extending lacrimal bud or the position where this structure would be anticipated. (A) Multiple ocular region explants cultured for 48 hours from E13.5 in the presence of sense oligonucleotide show normal extension of the lacrimal gland buds (red arrowheads). (B) As in A, but at higher magnification. The extending lacrimal gland bud is clearly visible, extending through the mesenchyme. (C) As in A, but treated with antisense oligonucleotide specific to FGFR2IIIb. This results in the inhibition of bud extension. (D) As in C, but at higher magnification.

Fig. 8. The induction and placement of ocular glands is regulated at multiple levels. The diagram represents the left eye of an E13.5 mouse. The blue shading indicates the region of FGF10 expression in periocular mesenchyme. The epithelia of the eye that express Pax6 are depicted in green and include the conjunctival, lens and lacrimal bud epithelia. The Harderian gland bud is depicted in purple on the nasal side of the eye. The sensory hair follicles (shf) also express FGF10. The model proposes that FGF10 is an inductive signal for the lacrimal and Harderian glands and acts directly on the conjunctival epithelium to stimulate proliferation (purple arrows). Additional speciﬁcation signals, either autonomous or paracrine, are presumably also required to give the lacrimal and Harderian glands their unique features (lacrimal gland, green; Harderian gland, purple). The model also proposes a mechanism by which the location of the ocular glands is determined. The dorsal conjunctival epithelium that is unresponsive to FGF10 in producing ectopic lacrimal buds is indicated by the sector outlined in gray. The lacrimal gland-competent epithelium is a reciprocal region designated by the red shaded sector. The ventral domain in which the FGF10 mRNA is distant from the conjunctival epithelium and is expressed at low levels is also indicated by a sector outlined in gray. We propose that in this domain, there is insufficient FGF10 to induce the lacrimal or Harderian gland buds. FGF10 and Pax6 in lacrimal gland development 2571

Though FGF10 appears to be an inducer of both lacrimal Cohn, M. J., Izpisúa-Belmonte, J. C., Abud, H., Heath, J. K. and Tickle, and Harderian glands, a number of observations argue that it C. (1995). Fibroblast growth factors induce additional limb development does not provide an instructive signal. First and most obvious, from the flank of chick embryos. Cell 80, 739-746. De Moerlooze, L., Spencer-Dene, B., Revest, J., Hajihosseini, M., Rosewell, FGF10 has an essential role in the development of many I. and Dickson, C. (2000). An important role for the IIIb isoform of different organs that are quite distinct in structure. Second, fibroblast growth factor receptor 2 (FGFR2) in mesenchymal-epithelial isolated lacrimal gland epithelium responds to FGF10 by signalling during mouse organogenesis. Development 127, 483-492. growing, but in the absence of branching morphogenesis. This Deng, C., Wynshaw-Boris, A., Zhou, F., Kuo, A. and Leder, P. (1996). Fibroblast growth factor receptor 3 is a negative regulator of bone growth. latter observation suggests that the primary function of FGF10 Cell 84, 911-21. on precursor epithelium may be to stimulate cell division. This Finch, P. W., Rubin, J. S., Miki, T., Ron, D. and Aaronson, S. (1989). response is consistent with that noted in other systems where Human KGF is FGF-related with properties of a paracrine effector of FGF10 is required for budding morphogenesis (Bellusci et al., epithelial cell growth. Science 245, 752-755. 1997b; Ohuchi et al., 1997). Fujiwara, M., Uchida, T., Osumi-Yamashita, N. and Eto, K. (1994). Uchida rat (rSey): a new mutant rat with craniofacial abnormalities resembling those In extending this argument, we can suggest that the of the mouse Sey mutant. Differentiation 57, 31-8. distinctiveness of FGF10-induced structures is a consequence Graff, J. M. (1997). Embryonic patterning: to BMP or not to BMP, that is the of (1) unique competence in precursor epithelial cells, (2) question. Cell 89, 171-4. FGF10 acting in concert with other soluble mediators to Grindley, J. C., Davidson, D. R. and Hill, R. E. (1995). The role of Pax-6 provide specification signals or (3) both. In the case of the in eye and nasal development. Development 121, 1433-42. Halder, G., Callaerts, P. and Gehring, W. J. (1995). Induction of ectopic ocular glands, the absence of Pax6 expression in the Harderian eyes by targeted expression of the eyeless gene in Drosophila. Science 267, gland epithelium might argue that Pax6 autonomously confers 1788-1792. features unique to the lacrimal gland. In the lung, the soluble Hill, R. E., Favor, J., Hogan, B. L., Ton, C. C., Saunders, G. F., Hanson, mediators FGF10 and BMP4 act together in regulating I. M., Prosser, J., Jordan, T., Hastie, N. D. and van Heyningen, V. (1991). Mouse small eye results from mutations in a paired-like homeobox- branching morphogenesis and the proximo-distal containing gene. Nature 354, 522-5. differentiation that is characteristic of this organ (Bellusci et Hogan, B. L. (1999). Morphogenesis. Cell 96, 225-33. al., 1997b; Sekine et al., 1999; Weaver et al., 1999). A model Hogan, B. L., Horsburgh, G., Cohen, J., Hetherington, C. M., Fisher, G. to summarize these arguments is presented in Fig. 8. and Lyon, M. F. (1986). Small eyes (Sey): a homozygous lethal mutation FGF10 has to date been implicated in the outgrowth of the on chromosome 2 which affects the differentiation of both lens and nasal placodes in the mouse. J. Embryol. Exp. Morph. 97, 95-110. limb buds (Ohuchi et al., 1997; Sekine et al., 1999) and in Hogan, B. L. and Yingling, J. M. (1998). Epithelial/mesenchymal budding of the lung endoderm (Bellusci et al., 1997b; Sekine interactions and branching morphogenesis of the lung. Curr. Opin. Genet. et al., 1999) the prostate epithelium (Thomson and Cunha, Dev. 8, 481-6. 1999) and the exocrine pancreas (Miralles et al., 1999). The Iseki, S., Wilkie, A. O., Heath, J. K., Ishimaru, T., Eto, K. and Morriss- Kay, G. M. (1997). Fgfr2 and osteopontin domains in the developing skull data presented here are a further illustration of the conserved vault are mutually exclusive and can be altered by locally applied FGF2. activity of FGFs in the development of structures that form Development 124, 3375-84. through a process of budding (Hogan, 1999), and we show that Johnson, D. E. and Williams, L. T. (1993). Structural and functional diversity FGF10 is an endogenous inducer of the lacrimal gland. In in the FGF receptor multigene family. Adv. Cancer Res. 60, 1-41. addition, the data indicate that FGF10 is a non-instructive Johnston, M. C., Noden, D. M., Hazelton, R. D., Coulombre, J. L. and Coulombre, A. J. (1979). Origins of avian ocular and periocular tissues. signal and one of presumably several required for induction. In Exp. Eye Res. 29, 27-43. the future, it will be interesting to determine the identity of the Kammandel, B., Chowdhury, K., Stoykova, A., Aparicio, S., Brenner, S. additional paracrine signals that combine with FGF10 to and Gruss, P. (1999). Distinct cis-essential modules direct the time-space stimulate development of a structure that is unique. pattern of the Pax6 gene activity. Dev. Biol. 205, 79-97. Lovicu, F. J., Kao, W. W. and Overbeek, P. A. (1999). Ectopic gland We are indebted to David Ornitz for a gift of tissue from FGF10- induction by lens-specific expression of keratinocyte growth factor (FGF-7) in transgenic mice. Mech. Dev. 88, 43-53. null mice, Dan Rifkin for recombinant FGF2, Clive Dickson for the Lu, W., Luo, Y., Kan, M. and McKeehan, W. L. (1999). Fibroblast growth cDNA to mouse FGF7, Noboyuki Itoh for the rat FGF10 cDNA and factor-10. A second candidate stromal to epithelial cell andromedin in to Chuxia Deng for the FGFR2 cDNA. This work was supported by prostate. J. Biol. Chem. 274, 12827-34. grants from the National Institutes of Health (R. A. L., no. EY11234, Makarenkova, H. and Patel, K. (1999). Gap junction signalling mediated P. O. nos EY10448 and EY10803) and from the National Defense through connexin-43 is required for chick limb development. Dev. Biol. 207, Medical College of Japan (M. I). 380-92. Mansouri, A., Stoykova, A. and Gruss, P. (1994). Pax genes in development. J. Cell. Sci. Suppl. 18, 35-42. Martin, G. R. (1998). The roles of FGFs in the early development of REFERENCES vertebrate limbs. Genes Dev. 12, 1571-86. Mason, I. J., Fuller-Pace, F., Smith, R. and Dickson, C. (1994). FGF-7 Altmann, C. R., Chow, R. L., Lang, R. A. and Hemmati-Brivanlou, A. (keratinocyte growth factor) expression during mouse development suggests (1997). Lens induction by Pax-6 in Xenopus laevis. Dev. Biol. 185, 119-23. roles in myogenesis, forebrain regionalisation and epithelial-mesenchymal Bellusci, S., Furuta, Y., Rush, M. G., Henderson, R., Winnier, G. and interactions. Mech. Dev. 45, 15-30. Hogan, B. L. (1997a). Involvement of Sonic hedgehog (Shh) in mouse Miki, T., Bottaro, D. P., Fleming, T. P., Smith, C. L., Burgess, W. H., Chan, embryonic lung growth and morphogenesis. Development 124, 53-63. A. M.-L. and Aaronson, S. A. (1992). Determination of ligand-binding Bellusci, S., Grindley, J., Emoto, H., Itoh, N. and Hogan, B. L. (1997b). specificity by alternative splicing: Two distinct growth factor receptors Fibroblast growth factor 10 (FGF10) and branching morphogenesis in the encoded by a single gene. Proc. Natl. Acad. Sci. USA 89, 246-250. embryonic mouse lung. Development 124, 4867-78. Min, H., Danilenko, D. M., Scully, S. A., Bolon, B., Ring, B. D., Tarpley, Callaerts, P., Halder, G. and Gehring, W. J. (1997). PAX-6 in development J. E., DeRose, M. and Simonet, W. S. (1998). Fgf-10 is required for both and evolution. Annu. Rev. Neurosci. 20, 483-532. limb and lung development and exhibits striking functional similarity to Chow, R. L., Altmann, C. R., Lang, R. A. and Hemmati-Brivanlou, A. Drosophila branchless. Genes Dev. 12, 3156-61. (1999). Pax6 induces ectopic eyes in a vertebrate. Development 126, 4213- Miralles, F., Czernichow, P., Ozaki, K., Itoh, N. and Scharfmann, R. 4222. (1999). Signaling through fibroblast growth factor receptor 2b plays a key 2572 H. P. Makarenkova and others

role in the development of the exocrine pancreas. Proc. Natl. Acad. Sci. USA Tanswell, A. K. (1996). Keratinocyte growth factor and its receptor are 96, 6267-72. involved in regulating early lung branching. Development 122, 3107-15. Mohammadi, M., Froum, S., Hamby, J. M., Schroeder, M. C., Panek, R. Quinn, J. C., West, J. D. and Hill, R. E. (1996). Multiple functions for Pax6 L., Lu, G. H., Eliseenkova, A. V., Green, D., Schlessinger, J. and in mouse eye and nasal development. Genes Dev. 10, 435-46. Hubbard, S. R. (1998). Crystal structure of an angiogenesis inhibitor bound Quiring, R., Walldorf, U., Kloter, U. and Gehring, W. J. (1994). Homology to the FGF receptor tyrosine kinase domain. EMBO J. 17, 5896-904. of the eyeless gene of Drosophila to the Small eye gene in mice and Aniridia Mohammadi, M., McMahon, G., Sun, L., Tang, C., Hirth, P., Yeh, B. K., in humans. Science 265, 785-9. Hubbard, S. R. and Schlessinger, J. (1997). Structures of the tyrosine Sekine, K., Ohuchi, H., Fujiwara, M., Yamasaki, M., Yoshizawa, T., Sato, kinase domain of fibroblast growth factor receptor in complex with T., Yagishita, N., Matsui, D., Koga, Y., Itoh, N. et al., (1999). Fgf10 is inhibitors. Science 276, 955-60. essential for limb and lung formation. Nat. Genet. 21, 138-41. Nieto, M. A., Patel, K. and Wilkinson, D. G. (1996). In situ hybridization Shannon, J. M., Gebb, S. A. and Nielsen, L. D. (1999). Induction of alveolar analysis of chick embryos in whole mount and tissue sections. Methods Cell type II cell differentiation in embryonic tracheal epithelium in mesenchyme- Biol. 51, 219-35. free culture. Development 126, 1675-88. Ohuchi, H., Nakagawa, T., Yamamoto, A., Araga, A., Ohata, T., Ishimaru, Slack, J. M. W. (1993). Embryonic induction. Mech. Dev. 41, 91-107. Y., Yoshioka, H., Kuwana, T., Nohno, T., Yamasaki, M. et al., (1997). Song, D. L., Chalepakis, G., Gruss, P. and Joyner, A. L. (1996). Two Pax- The mesenchymal factor, FGF10, initiates and maintains the outgrowth of binding sites are required for early embryonic brain expression of an the chick limb bud through interaction with FGF8, an apical ectodermal Engrailed-2 transgene. Development 122, 627-635. factor. Development 124, 2235-44. Sun, L., Tran, N., Tang, T., App, H., Hirth, P., McMahon, G. and Tang, Ornitz, D. M., Xu, J., Colvin, J. S., McEwen, D. G., MacArthur, C. A., C. (2000). 3-[(substituted pyrrol-2yl)methylindenyl]indolin-2-ones that Coulier, F., Gao, G. and Goldfarb, M. (1996). Receptor specificity of the block receptor tyrosine kinases. J. Med. Chem., in press. fibroblast growth factor family. J. Biol. Chem. 271, 15292-7. Thomson, A. A. and Cunha, G. R. (1999). Prostatic growth and development Orr-Urtreger, A., Bedford, M. T., Burakova, T., Arman, E., Zimmer, Y., are regulated by FGF10. Development 126, 3693-701. Yayon, A., Givol, G. and Lonai, P. (1993a). Developmental localization of Ton, C. C., Miwa, H. and Saunders, G. F. (1992). Small eye (Sey): cloning the splicing alternatives of fibroblast growth factor receptor-2 (FGFR2). Dev. and characterization of the murine homolog of the human aniridia gene. Biol. 158, 475-486. Genomics 13, 251-6. Orr-Urtreger, A., Bedford, M. T., Burakova, T., Arman, E., Zimmer, Y., Wakusawa, S. (1968). [The development of the arterial system of the mouse Yayon, A., Givol, D. and Lonai, P. (1993b). Developmental localization of eye] [Japanese]. Nippon Ganka Gakkai Zasshi – Acta Societatis the splicing alternatives of fibroblast growth factor receptor-2 (FGFR2). Dev. Ophthalmologicae 72, 1803-1828. Biol. 158, 475-86. Weaver, M., Yingling, J. M., Dunn, N. R., Bellusci, S. and Hogan, B. L. Ostle, B. and Mensing, R. W. (1975). Statistics in Research. Ames: The Iowa (1999). Bmp signaling regulates proximal-distal differentiation of endoderm State University Press. in mouse lung development. Development 126, 4005-15. Pepicelli, C. V., Lewis, P. M. and McMahon, A. P. (1998). Sonic hedgehog Williams, S. C., Altmann, C. R., Chow, R. L., Hemmati-Brivanlou, A. and regulates branching morphogenesis in the mammalian lung. Curr. Biol. 8, Lang, R. A. (1998). A highly conserved lens transcriptional control element 1083-6. from the Pax-6 gene. Mech. Dev. 73, 225-9. Peters, H. and Balling, R. (1999). Teeth. Where and how to make them. Xu, X., Weinstein, M., Li, C., Naski, M., Cohen, R. I., Ornitz, D. M., Leder, Trends Genet. 15, 59-65. P. and Deng, C. (1998). Fibroblast growth factor receptor 2 (FGFR2)- Peters, K., Werner, S., Liao, X., Wert, S., Whitsett, J. and Williams, L. mediated reciprocal regulation loop between FGF8 and FGF10 is essential (1994). Targeted expression of a dominant negative FGF receptor blocks for limb induction. Development 125, 753-65. branching morphogenesis and epithelial differentiation of the mouse lung. Zuniga, A., Haramis, A. P., McMahon, A. P. and Zeller, R. (1999). Signal EMBO J. 13, 3296-301. relay by BMP antagonism controls the SHH/FGF4 feedback loop in Post, M., Souza, P., Liu, J., Tseu, I., Wang, J., Kuliszewski, M. and vertebrate limb buds. Nature 401, 598-602.