Disregulation of Ocular Morphogenesis by Lens-Specific Expression of FGF-3/Int-2 in Transgenic Mice

DEVELOPMENTAL BIOLOGY 198, 13–31 (1998)ARTICLE NO. DB988879

Disregulation of Ocular Morphogenesis byLens-Specific Expression of FGF-3/Int-2in Transgenic Mice

Michael L. Robinson,*,† Chiaki Ohtaka-Maruyama,‡,1 Chi-Chao Chan,‡Susan Jamieson,§,2 Clive Dickson,§ Paul A. Overbeek,Ø,Hand Ana B. Chepelinsky‡*Children’s Hospital Research Foundation and †Department of Pediatricsand the Comprehensive Cancer Center, The Ohio State University,700 Children’s Drive, Columbus, Ohio 43205; ‡National EyeInstitute, National Institutes of Health Bethesda, Maryland20892; §Laboratory of Viral Carcinogenesis, ImperialCancer Research Fund, London WC2A 3PX, UnitedKingdom; and ØDepartment of Cell Biology andHDepartment of Ophthalmology, BaylorCollege of Medicine, Houston,Texas 77030

FGF-3, originally named int-2, was discovered as an oncogene frequently activated in mammary carcinomas resulting fromthe chromosomal integration of the mouse mammary tumor virus (MMTV). Int-2 was later designated FGF-3 based onsequence homology with other members of the fibroblast growth factor (FGF) family. FGF-1 is the prototypical memberof the FGF family, and is the only family member which activates all known FGF receptor isoforms. Transgenic miceexpressing in the lens a form of FGF-1 engineered to be secreted show premature differentiation of the entire lens epithelium.In contrast, transgenic mice engineered to secrete FGF-2 in the lens do not undergo premature differentiation of the lensepithelium (C. M. Stolen et al., 1997, Development 124, 4009–4017). To further assess the roles of FGFs and FGF receptorsin lens development, the aA-crystallin promoter was used to target expression of FGF-3 to the developing lens of transgenicmice. The expression of FGF-3 in the lens rapidly induced epithelial cells throughout the lens to elongate and to expressfiber cell-specific proteins including MIP and b-crystallins. This premature differentiation of the lens epithelium wasfollowed by the degeneration of the entire lens. Since FGF-1 and FGF-3 can both activate one FGF receptor isoform (FGFR2IIIb) that is not activated by FGF-2, these results suggest that activation of FGFR2 IIIb is sufficient to induce fiber celldifferentiation throughout the lens epithelium in vivo. Furthermore, transgenic lens cells expressing FGF-3 were able toinduce the differentiation of neighboring nontransgenic lens epithelial cells in chimeric mice. Expression of FGF-3 in thelens also resulted in developmental alterations of the eyelids, cornea, and retina, and in the most severely affected transgeniclines, the postnatal appearance of intraocular glandular structures. q 1998 Academic Press

INTRODUCTION ferentiation, and migration. Despite the recent progress inidentifying transcription factors such as Chx-10 (Burmeisteret al., 1996), Pax-6 (Grindley et al., 1995; Hill et al., 1991),The architecture of the mammalian eye results from aRx (Mathers et al., 1997), and others which play essentialcomplex and well-orchestrated symphony of induction, dif-roles in the regulation of ocular gene expression, extracellu-lar or cell surface molecules responsible for inducing prolif-eration, migration, and/or differentiation of many ocular1 Current address: The Institute of Physical and Chemical Re-tissues in vivo remain to be conclusively identified. Theresearch, Riken, Wako, Saitama 351-01, Japan.is considerable evidence to suggest that FGFs and FGF re-2 Current address: University Department of Surgery, Western

Infirmary, Glasgow, G11 6NT, UK. ceptors (FGFRs) play important roles in ocular develop-

13

0012-1606/98 $25.00Copyright q 1998 by Academic PressAll rights of reproduction in any form reserved.

AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

14 Robinson et al.

ment. The FGF family currently consists of at least 15 mem- protein under the direction of the aA-crystallin promotermight influence development of either the lens or adjacentbers (FGF 1–15) (McWhirter et al., 1997). Different FGF

family members display overlapping expression patterns ocular structures or both. The normal mouse lens consistsof a spherical mass of differentiated lens fiber cells coveredduring embryogenesis and are known to be involved in a

wide variety of developmental processes including postim- on the anterior hemisphere by a single layer of undifferenti-ated cuboidal epithelial cells. As lens epithelial cells prolif-plantation growth (Feldman et al., 1995), patterning of em-

bryonic mesoderm (Deng et al., 1994; Yamaguchi et al., erate, the epithelial cells at the equatorial region undergodifferentiation into lens fiber cells. The source of the fiber1994), inner ear development (Mansour et al., 1993; McKay

et al., 1996), and the regulation of hair growth (Hebert et cell differentiation signal in the developing eye is unknown,but lens reversal experiments in chicks (Coulombre andal., 1994; for reviews see Wilkie et al., 1995; Yamaguchi

and Rossant, 1995). The mediation of FGF induced signal Coulombre, 1963) and mice (Yamamoto, 1976) have sug-gested that the differentiation signal is present in the vitre-transduction occurs through the binding and activation of

high-affinity cell surface receptor tyrosine kinases (FGFRs). ous. This hypothesis has been strengthened by experimentsdemonstrating that bovine vitreous, but not aqueous fluid,Four different FGFR genes have been identified (FGFR 1–

4), three of which give rise to multiple isoforms resulting induces the differentiation of lens epithelial explants in cul-ture (Lovicu et al., 1995). The continued proliferation offrom differential splicing of primary RNA transcripts (Chel-

laiah et al., 1994; Johnson et al., 1991). lens epithelial cells and the differentiation of equatorial epi-thelial cells into fiber cells is responsible for growth of theFGFs and FGF receptors are widely expressed during ocu-

lar development and are likely to play essential roles in the vertebrate lens (Piatigorsky, 1981). While the developmentof the ocular lens has been used as a model system fordevelopmental process. FGF-1 is present in several ocular

tissues throughout ocular morphogenesis (de Iongh and developmental biologists for decades, the endogenous fac-tor(s) which direct the differentiation of lens epithelial cellsMcAvoy, 1993). FGF-2 expression is first detected in associa-

tion with the developing lens capsule and remains associated into lens fiber cells is(are) unknown. Transgenic mice ex-pressing a secreted FGF-1 under the regulatory control of thewith the developing lens and retina throughout the last third

of rodent embryogenesis (de Iongh and McAvoy, 1993). FGF- aA-crystallin promoter undergo complete differentiation ofthe anterior lens epithelium to fiber cells prior to birth, and3 expression appears in the developing mouse retina by

E14.5 where it is associated with the early differentiation of often develop postnatal retinal folding and subtle alter-ations in corneal morphogenesis (Robinson et al., 1995b).post-mitotic cells (Wilkinson et al., 1989). FGF-5 has been

detected in the photoreceptors, ganglion cells and retinal Although FGF-1 is capable of inducing lens differentiationin vivo, there is still no conclusive evidence that FGF-1 ispigment epithelium of the adult rhesus macaque retina (Ki-

taoka et al., 1994). FGF-7 expression is restricted to the mes- indeed the endogenous lens differentiation factor.The properties of FGF-3 differ from those of FGF-1 inenchyme of the developing eye socket and eyelid (Finch et

al., 1995). The eye also expresses three different FGFR genes, several respects. In contrast to FGF-1 which has no knownsignal peptide sequence, the FGF-3 coding sequence con-FGFR1, FGFR2, and FGFR3 (Orr-Urtreger et al., 1991; Peters

et al., 1993). FGFs have been shown to induce cellular prolif- tains a signal peptide and is normally secreted in vivo (Kieferet al., 1991). Furthermore, while FGF-1 shows a broad recep-eration, migration, and differentiation of lens epithelial cell

explants (McAvoy and Chamberlain, 1989), to stimulate tor specificity, FGF-3 is only known to bind with high affin-ity to two specific FGF receptor isoforms (Mathieu et al.,photoreceptor differentiation in cultured newborn rat retinal

cells (Hicks and Courtois, 1992), and to stimulate ganglion 1995; Ornitz et al., 1996). FGF-3 was originally known asint-2, based on its discovery as a gene commonly activatedcell differentiation in fetal rat retinal explants (Guillemot

and Cepko, 1992). Furthermore, organ culture experiments in mammary carcinomas by the proximal chromosomal in-tegration of mouse mammary tumor virus (MMTV) (Mooreon developing chick optic vesicles indicate that FGFs are

required for the differentiation of the neural retina (Pittack et al., 1986). Later, the amino acid sequence homology ofint-2 to FGF-1 and FGF-2 (Dickson and Peters, 1987) led toet al., 1997). In vivo FGFs have been shown to induce the

transdifferentiation of retinal pigment epithelium into neu- its inclusion into the fibroblast growth factor family. Ec-topic expression of FGF-3 in the mammary epithelium ofral retina (Park and Hollenberg, 1989), promote wound heal-

ing in deepithelialized corneas (Fredj-Reygrobellet et al., transgenic mice resulted in pronounced mammary epithe-lial cell hyperplasia, confirming its ability to act as an onco-1987), promote the survival of retinal ganglion cells follow-

ing optic nerve transection (Sievers et al., 1987) or ischemic gene (Muller et al., 1990; Stamp et al., 1992). To answerseveral questions regarding the function of FGFs in eye de-injury (Unoki and LaVail, 1994; Zhang et al., 1994), and

protect photoreceptors from apoptosis in some forms of in- velopment, we have directed expression of FGF-3 to thelens using an aA-crystallin promoter in transgenic mice.herited retinal degeneration (Faktorovich et al., 1990) or fol-

lowing light damage (Faktorovich et al., 1992). While FGF-3 does not play an essential role in the morpho-genesis of the lens (Mansour et al., 1993), we were interestedThe use of the aA-crystallin promoter in transgenic mice

makes it possible to direct gene expression to the lens at or to know if the FGFRs responsive to FGF-3 were capable oftransducing a lens differentiation signal. We were particu-before 12.5 days of embryonic development (E12.5), well

before the final architecture of the eye is established (Over- larly interested to know whether ectopic expression of FGF-3 would result in lens differentiation or cause ocular hyper-beek et al., 1985). The expression of a secreted transgenic

Copyright q 1998 by Academic Press. All rights of reproduction in any form reserved.

AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

15Overexpression of FGF-3 in the Eye

25 mM NaCl at 657C. Washed membranes were then exposed toplasia or tumorigenesis. In contrast to the consequences ofKodak X-ray film for 4 h at 0707C with an intensifying screen.FGF-3 expression in the mammary gland, FGF-3 expression

in the lens did not result in lens cell hyperplasia, but ratherled to the differentiation of the entire lens epithelium, andsubsequent degeneration of the lens. Surprisingly, FGF-3 PCR Analysesexpression in the lens also induced developmental alter-

Primers used for PCR include: 5*-GCATTCCAGCTGCTGA-ations of several other ocular tissues including the retinaCGGT-3* (termed CR4), a sense primer that hybridizes to the mu-and cornea.rine aA-crystallin promoter; 5*-CCCAGAGGCTCCTGTCTGAC-TCACT-3* (termed CR5), a sense primer that hybridizes to the 5*

untranslated region of the murine aA-crystallin transcript (Rob-METHODS inson et al., 1995b); 5*-GTACTTGGTAGCGCAGTAGAGC-3*

(termed F4), an antisense primer that hybridizes to sequences inthe first exon of the murine FGF-3 gene (see Fig. 1). TransgenicGeneration of Transgenic Micemice were screened by PCR amplification with primers CR4 and

A Bgl I I fragment containing murine FGF-3 sequences from F4 using the following conditions: an initial melt at 947C for 3pKC3.2 (Dixon et al., 1989) was inserted between the BamHI and min followed by 30 cycles of denaturation at 947C for 30 s, primerBclI sites of paA-366aT (Mahon et al., 1987), replacing the coding annealing at 587C for 30 s, and polymerase extension at 727C for 1sequences for the SV40 T-antigen. The resultant plasmid paA/FGF- min. A 2-min extension at 727C was performed after the final cycle3, contained the entire coding sequence for the secreted form of of PCR. Each PCR reaction consisted of approximately 250 ng ofFGF-3, followed by 1581 bp of FGF-3 noncoding sequence inserted tail DNA in a buffer containing 10 mM Tris–HCl, pH 8.3, 50 mMbetween the 0366//46 aA-crystallin promoter and a 237-bp poly- KCl, 1.5 mM MgCl2, 0.01% gelatin, 10% DMSO, 2.5 mM eachadenylation signal from SV40 virus. The microinjection construct dNTP, 800 nM each oligonucleotide primer, 2.5 units of Taq DNAwas released from paA/FGF-3 by digestion with NaeI, AatII, and polymerase (Perkin–Elmer) in a 50-ml reaction volume. TransgenicPvuI, followed by gel electrophoresis to isolate the 3081-bp micro- DNA amplified with primers CR4 and F4 yielded a DNA fragmentinjection fragment. The aA/FGF-3 fragment was injected into pro- of 336 bp which was visualized by agarose gel electrophoresis andnuclear stage FVB/N embryos (Taketo et al., 1991) at a concentra- ethidium bromide staining.tion of 2 ng/ml. Injected embryos were transferred into pseudopreg- For detection of transgene transcription, total ocular RNA wasnant ICR strain females. Potential aA/FGF-3 transgenic mice were isolated from embryonic day 17.5 (E17.5) embryos. Eyes were dis-screened by isolating genomic DNA (Hogan et al., 1994) from tail sected from embryos and quick-frozen in a dry ice/ethanol bath.biopsies and testing for transgenic sequences by Southern hybrid- Total RNA was isolated using RNA STAT-60 according to theization (Southern, 1975) and/or by the use of polymerase chain manufacturer’s protocol (Tel-Test B, Inc.). The resultant RNA wasreaction (PCR) (Saiki et al., 1988) as described below. treated with FPLCpure DNase I (Pharmacia) for 30 min at 377C to

remove contaminating genomic DNA. The DNase I was subse-quently heat inactivated by a 10-min incubation at 657C and re-

Southern Analysis moved by two 1:1 phenol:chloroform extractions. The ocular RNAwas then precipitated with ethanol, reverse transcribed into cDNA,Southern hybridization was performed on each aA/FGF-3 trans-and amplified by primers CR5 and F4 as described previously (Rob-genic family. Ten micrograms of BamHI-digested genomic tailinson et al., 1995b). In all cases PCR was performed on RNA sam-DNA was separated by gel electrophoresis on a 1% agarose gel. Asples which had and had not been converted to cDNA by reversea positive control for hybridization, one lane was loaded with 0.05transcriptase since primers CR5 and F4 do not distinguish betweenng of the 3081-bp microinjection construct. After electrophoresis,expressed transgenic cDNA and contaminating transgenic genomicthe gel was soaked in 0.25 M HCl for 15 min followed by neutraliza-DNA. Transgene expression was confirmed by the presence of ation in 0.4 M NaOH for 60 min. The DNA was then transferred207-bp band specifically in the PCR-amplified samples which hadfrom the gel to a nylon membrane under alkaline conditions (Sam-been treated with reverse transcriptase.brook et al., 1989). DNA was fixed to the membrane by UV irradia-

tion. The nylon membrane containing the DNA was prehybridizedfor 1 h at 427C in 75 mM sodium citrate, 1.25 M NaCl, 2 mg/mlFicoll (type 400 Pharmacia), 2 mg/ml polyvinylpyrrolidone (Sigma), Production of Chimeras2 mg/ml bovine serum albumin (Fraction V, Sigma), 50 mM Tris–HCl, pH 7.4, 0.1% sodium pyrophosphate, 1.2% sodium dodecyl Chimeric mice were produced by aggregating early morula stage

aA/FGF-3 transgenic or FVB/N control embryos with similarlysulfate (SDS), 100 mg/ml sheared herring sperm DNA (Sigma), 5%dextran sulfate, 50% formamide. The entire microinjection frag- staged embryos from strain Rosab-geo26 (Friedrich and Soriano,

1991) as described previously (Robinson et al., 1995b). All of thement (50 ng) from paA/FGF-3 was radioactively labeled with [32P]-dCTP using an oligolabeling kit (Pharmacia) according to the manu- aA/FGF-3, and control FVB/N mice are albino, while the Rosab-

geo26 mice are pigmented. Potential chimeric mice were initiallyfacturer’s instructions. The probe (specific activity of 1 1 109 cpm/mg) was denatured and added to the hybridization mixture which identified by the presence of a mosaic pattern of pigmentation in

the retinal pigment epithelium. Because the Rosab-geo26 micediffered from the prehybridization buffer as follows: 0.2 mg/ml Fi-coll, 0.2 mg/ml polyvinylpyrrolidone, 0.2 mg/ml bovine serum al- used to produce chimeras were homozygous for the b-geo

transgene, the presence of pigmentation in potential chimeric off-bumin, 10% dextran sulfate. Hybridization was carried out at 427Cfor 12 h. Posthybridization washes were performed consecutively spring confirmed the presence of bacterial b-galactosidase-express-

ing cells. The presence of the aA/FGF-3 transgene in the chimericas follows: two 15-min washes in 30 mM sodium citrate, 0.5 MNaCl, 0.5% SDS at 207C followed by one wash in the same buffer aA/FGF-3}Rosab-geo26 mice was confirmed by the amplification

of a 336-bp band by PCR using primers CR4 and F4 from genomicfor 30 min at 657C; a final 30-min wash in 1.5 mM sodium citrate,


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

16 Robinson et al.

DNA. Embryonic heads and chimeric eyes were removed, pro- Sense riboprobes were prepared from EcoRI cut pMMIPR1 usingT3 RNA polymerase. Antisense riboprobes were likewise preparedcessed, and analyzed as described (Robinson et al., 1995b).from NotI cut pMMIPR1 using T7 RNA polymerase. Heads fromE13.5 mouse embryos were fixed and paraffin embedded as de-scribed above. Paraffin sections were deparaffinized and nonradioac-Histological Analysestive in situ hybridization was performed essentially as described

For adult eye histology, mice were sacrificed by cervical disloca- (Sasaki and Hogan, 1994). Hybridizations were performed at 507Ction prior to the removal of ocular tissues. Embryos were obtained for 17 h, and washes were performed at high stringency. Digoxi-from timed matings with noon of the day of vaginal plug discovery genin signals were detected by alkaline phosphatase coloring reac-designated as 0.5 days gestation (E 0.5). Adult eyes and embryonic tion using the DIG-Nucleic Acid detection kit (Boehringer Mann-heads were fixed overnight either in 10% neutral buffered Formalin heim) with antidigoxigenin–alkaline phosphatase conjugate and de-(Richard Allan) at ambient temperature for immunohistochemical veloped with NBT/BCIP as indicated by the manufacturer.analysis or in phosphate buffered 4% paraformaldehyde (pH 7.4) at47C for in situ hybridization. All tissues were embedded in PolyFinwax (Triangle Biomedical Sciences), sectioned at 5 mm, and applied RESULTSto Superfrost Plus (Curtin Matheson Scientific) glass slides. Forroutine histology, sections were stained with hematoxylin and eo-

Generation of Transgenic Micesin. Immunohistochemistry for b-crystallin was performed usingpolyclonal antibodies kindly provided by Dr. S. Zigler (National

The plasmid paA/FGF-3 was made by replacing the SV40Eye Institute, Bethesda, MD). Deparaffinized sections were firstT-antigen coding sequences of paA-366aT (Mahon et al.,incubated in a blocking solution consisting of 3% nonimmune se-1987) with the murine FGF-3 coding and noncoding se-rum diluted in phosphate-buffered saline (PBS) for 30 min at 377C.quences contained in the pKC3.2 clone (Dixon et al., 1989).Rabbit anti-b-crystallin antibodies were diluted 1:100 in theThe resulting injection construct contained the coding se-blocking solution and were incubated with the tissue sections for

2 h at 377C. Sections were subsequently rinsed five times in PBS quences for the secreted form of FGF-3 downstream of thefor 3 min per rinse. Sections were then incubated for 1 h at 377C murine aA-crystallin promoter (Fig. 1). This construct didwith a fluorescein-conjugated donkey anti-rabbit Ig antibody (Am- not contain the alternative upstream CUG initiation codonersham N1034, Arlington Heights, IL) diluted 1:30 in the blocking which results in a proportion of FGF-3 being directed tosolution. Sections were then rinsed three times in PBS for 5 min the cell nucleus (Acland et al., 1990; Kiefer et al., 1994).per rinse. Slides were finally treated with one drop of FluoroGuard

Furthermore, sequences surrounding the AUG initiation co-anti-fade reagent (Bio-Rad, Hercules, CA), before aqueous mountingdon were optimized to increase the translational efficiencyand visualization by ultraviolet epifluorescence. Negative controlsof the construct (Dixon et al., 1989). The aA/FGF-3 DNAincluded sections which were not incubated with the primary anti-fragment was injected into pronuclear stage FVB/N embryosb-crystallin antibody.to obtain transgenic mice. Three initial founder animalswere identified by PCR and designated OVE 391, 392, and

In Situ Hybridization 393, respectively. The founder animal for OVE 393 wasfound to contain two transgenic integration sites which seg-

FGF-3. A 498-bp EcoRI–ScaI fragment from the 3* untranslated regated in subsequent generations to give rise to transgenicportion of the murine FGF-3 cDNA was liberated from paA/FGF-lines OVE 393A and OVE 393B. These two sublines were3 and subcloned between the EcoRI and SmaI sites of pBluescriptdistinguished both by ocular phenotype and by SouthernKS0 (Stratagene) to generate the riboprobe vector pBSFGF3a. [35S]-blot analysis of genomic DNA (Fig. 2). Southern hybridiza-UTP (Amersham) was used to generate 35S-labeled riboprobes by intion also revealed that all transgenic families containedvitro transcription (Robinson et al., 1995b). Sense riboprobes were

generated from BamHI cut pBSFGF3a using T3 RNA polymerase. multiple tandem copies of the transgene based on the ap-EcoRI cut pBSFGF3a was transcribed by T7 RNA polymerase to pearance of a hybridizing band of unit length present in allgenerate antisense riboprobes. These riboprobes are essentially transgenic samples (Fig. 2). After visual inspections of sev-identical to those previously used to determine the endogenous eral Southern blots containing known amounts of transgeneexpression pattern of FGF-3 in mouse embryos (Wilkinson et al., DNA (Fig. 2, lane 6), it was estimated that transgenic fami-1988). Embryonic heads, or adult eyes were fixed, processed, sec- lies OVE 391, 392, 393A, and 393B contained 5–10, ú10,tioned, and mounted on Superfrost Plus (Curtin Matheson Scien-

2–4, and 5–10 copies of the transgene, respectively.tific) glass slides followed by hybridization and washing as de-scribed (Robinson et al., 1995b). Hybridized slides were exposed toNTB-2 emulsion (Kodak) for 60 h at 47C, then developed with Ko-

Phenotype of aA/FGF-3 Transgenic Micedak D-19 developer and counterstained with hematoxylin.MIP. A 556-bp fragment of murine MIP coding sequence was Normal mice are born with closed eyelids, but all trans-

amplified from mouse lens cDNA using sense and antisense primers genic mice from lines OVE 392 and OVE 393A can be identi-5*-GGTCAGCCTCCTTCTGGAGGGCCAT-3* and 5*-AGCAAA- fied at birth by bilateral blood crusts within the eye socketsGGAGCGGGCAGGGTTCATGCCTGCACC-3* corresponding to

between the opened eyelids (see Fig. 3D). All transgenicbovine MIP coding sequences 14/38 and 570/538, respectivelymice with the exception of those from OVE 393B are se-(Gorin et al., 1984). The PCR amplified fragment was subclonedverely microphthalmic, and no obvious differences exist be-into pCR-Script SK/ (Stratagene) to create the riboprobe vectortween mice hemizygous and homozygous for the transgene.pMMIPR1. Digoxigenin-11-UTP (Boehringer Mannheim) was used

to prepare dioxigenin-labeled riboprobes by in vitro transcription. Hemizygous mice from OVE 393B often appear to have eyes


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal


FIG. 1. Schematic depiction of the 3081-bp fragment used to generate the aA/FGF-3 transgenic lines. The microinjection constructcontained a 415-bp murine aA-crystallin promoter (aA-cry) fused to a 2328-bp murine FGF-3 clone and a 237-bp 3* polyadenylation sequencederived from SV40 virus. The FGF-3 clone contained a 9-bp artificial 5* non-coding region (to facilitate cloning), followed by 738 bp ofFGF-3 coding sequences, and 1581 bp of 3* FGF-3 noncoding sequences. The entire construct was flanked on the 5* and 3* ends by 26 and75 bp, respectively, of plasmid-derived sequence. The construct contained the entire coding region for the secreted form of murine FGF-3, but did not contain the endogenous upstream CUG codon which can act as an alternative translation initiation codon in vivo.

of normal size and clear lenses at weaning, but usually de- and parturition, resulting in the loss of the lens and theocular injury obvious at birth (Fig. 3D). Embryos fromvelop unilateral or bilateral cataracts within 2 to 6 months

of age. Mice homozygous for the 393B transgenic insertion transgenic family OVE 391 exhibited less severe ocular pro-ptosis which typically did not progress to corneal rupture.have bilateral cataracts and are slightly microphthalmic at

weaning.To better characterize the transgenic phenotype, embryos

aA/FGF-3 Transgene Expressionwere collected from pregnant females to test for transgeneexpression and to use for histological analysis. Transgenic Total ocular RNA was collected from transgenic and con-embryos from all lines except OVE 393B could be visually trol embryos at E17.5 and was reverse transcribed intodistinguished from control embryos or wild-type lit- cDNA to perform reverse transcriptase–PCR (RT–PCR).termates as early as 12.5 days postcoitus (E12.5) by the ap- The cDNA was then amplified by PCR using a sense primerpearance of ocular proptosis which became more severe as complementary to aA-crystallin sequences in the 5* un-development progressed (Fig. 3B). In embryos from trans- translated region of transgenic transcripts (CR5), and angenic lines OVE 392 and 393A, the exophthalmos typically antisense primer complementary to sequences from the firstprogressed to the point of corneal rupture between E15.5 exon of the murine FGF-3 gene (F4). The technical difficulty

in collecting intact lenses from severely affected transgenicfamilies precluded performing RT–PCR on isolated lensRNA. The predicted 207-bp band was amplified from ocularcDNA from all four transgenic lines (Fig. 4). Because thereare no introns in the transgene, it was necessary to preventtransgenic mouse genomic DNA from serving as a templatefor PCR amplification. Therefore all RNA samples weretreated with DNase I to destroy any contaminating genomicDNA. No bands were amplified when primers CR5 and F4were used with nontransgenic cDNA (Fig. 4, lane 1) or withtransgenic ocular RNA samples not treated with reversetranscriptase (data not shown).

Although RT–PCR confirmed the ocular expression ofthe transgenic sequences, it failed to reveal the spatial pat-tern of transgene expression within the eye. This informa-FIG. 2. Southern blot of aA/FGF-3 transgenic families. The entiretion was provided by in situ hybridization of transgenicmicroinjection construct was used as a hybridization probe on a

Southern blot containing 10 mg of BamHI-digested genomic DNA embryos using a probe specific to murine FGF-3 (Wilkinsonfrom nontransgenic (lane 1) and transgenic (lanes 2–5) mice. As a et al., 1988). For these studies transgenic embryos from thepositive control for hybridization, the 3081-bp microinjection frag- severely affected line OVE 392 were examined at E12.5,ment was included at a concentration of approximately five copies 13.5, and 14.5. While this FGF-3 probe was unable to distin-per haploid genome (lane 6). Because there is only one BamHI site guish between transgenic and endogenous FGF-3 tran-present in the microinjection construct, the band in lane 6 represents scripts, transgene expression was determined by comparingthe unit length expected in transgenic mice containing multiple tan-

stage-matched transgenic and non-transgenic embryos (Fig.dem copies of the transgene. The bands present in the nontransgenic5). FGF-3 transcripts were present in transgenic lenses andlane represent hybridization to the endogenous murine aA-crystallinabsent from nontransgenic lenses at all embryonic stagespromoter, and endogenous FGF-3 gene. The mobilities of relevantexamined. Comparatively weak endogenous expression ofDNA molecular weight standards (1-kb ladder, Gibco) run on the

same DNA separation gel are indicated by arrows. FGF-3 was detected in transgenic and control retinas by


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

18 Robinson et al.

FIG

.3.

Ext

ern

alph

enot

ype

ofa

A/F

GF-

3tr

ansg

enic

mic

e.T

ran

sgen

icem

bryo

sfr

omal

lli

nes

exce

ptO

VE

393B

can

bedi

stin

guis

hed

from

non

tran

sgen

icli

tter

mat

esby

extr

eme

bulg

ing

ofth

etr

ansg

enic

eyes

.N

ontr

ansg

enic

eyes

atE

17.5

are

cove

red

byth

eey

elid

(arr

ow,A

),w

hil

eth

eey

esin

the

litt

erm

ate

tran

sgen

icem

bryo

sar

en

otco

vere

dby

aney

elid

,an

dpr

otru

defr

omth

eh

ead

(arr

ow,

B).

Th

eem

bryo

nic

hea

ds(A

,B

)ar

evi

ewed

from

abov

ew

ith

the

snou

tspo

inti

ng

tow

ard

the

upp

erri

ght

corn

er.

At

birt

h,

non

tran

sgen

icm

ouse

eyes

are

cove

red

wit

ha

seal

edey

elid

(arr

ow,

C).

Bet

wee

nE

17.5

and

birt

h,

thea

A/F

GF-

3tr

ansg

enic

eyes

from

fam

ilie

sO

VE

392

and

393A

rupt

ure

resu

ltin

gin

anop

enw

oun

dor

‘‘sca

bby’

’ey

eat

birt

h(a

rrow

,D

).


AID DB 8879 / 6x3c$$8879 05-18-98 21:46:19 dbal


Unless otherwise noted, the following description of thetransgenic phenotype applies to the most severely affectedfamilies OVE 392 and OVE 393A.

Deviations from normal ocular development are obvious inaA/FGF-3 transgenic mice as early as E12.5 (data not shown),but are particularly dramatic by E14.5 (Fig. 6). At this latterstage in development the primary fiber cells become filledwith vacuoles and all lens epithelial cells exhibit an elongatedmorphology similar to that of normal primary fiber cells. Thetransgenic lens is considerably larger than the normal lensat this stage and fails to separate from the surface ectodermresulting in a nipple-like extension of the anterior lens to thesurface ectoderm (Fig. 6B, asterisk). It is through this defectin the developing cornea that the embryonic lens is expelledbefore birth. Normally, by E14.5 the eyelid folds have devel-oped and begun the process of covering the eye, which istypically completed by E16.5 (Pei and Rhodin, 1970). In theE14.5 nontransgenic embryo, the lid folds are evident and have

FIG. 4. Transgenic transcripts were detected in ocular RNA from begun to extend over the outer surface of the cornea (Fig.all aA/FGF-3 transgenic lines. Total ocular RNA was collected6A). In contrast, eyelid development is perturbed (Fig. 6B) infrom E17.5 eyes and reverse transcribed into cDNA. The resultanttransgenic families OVE 392 and 393A. In these families, thecDNA was amplified nonquantitatively by PCR with a senseeyelids fail to cover the eye until several days after birth (seeprimer corresponding to the 5* untranslated sequence contained inFig. 3D). At E14.5, the normal cornea consists of a singlethe aA-crystallin promoter and an antisense primer corresponding

to FGF-3-coding sequences. A diagnostic band of 207 bp was ampli- layer of cuboidal epithelial cells which overlies several closelyfied in all transgenic samples (lanes 2–5), but did not appear in the packed layers of primitive stromal fibroblasts (Fig. 6C). Thenontransgenic sample (lane 1). The mobilities of relevant DNA aA/FGF-3 transgenic corneas are abnormal in several respects.molecular weight standards (1 kb ladder, Gibco) run on the same The transgenic corneal epithelium is columnar rather thanDNA separation gel are indicated by arrows. cuboidal (Fig. 6D), and in some places appears multilayered.

There appear to be two distinct types of mesenchyme presentposterior to the corneal epithelium in transgenic embryos.The first type is found directly posterior to the corneal epithe-

E13.5. While it cannot be ruled out that longer exposures lium, where the corneal stroma normally forms. This mesen-may have revealed some nonlenticular expression of the chyme is characterized by widely spaced, irregularly orientedtransgene within the eye, the overwhelming majority of stellate cells embedded in what appears to be a loosely orga-transgenic transcripts are found within the lens (Fig. 5 D), nized extracellular matrix (Figs. 6B and 6D, arrowheads). Theand far exceed the number of endogenous ocular FGF-3 tran- second type of mesenchyme is much more compact and sur-scripts at these stages. rounds the anterior outer optic cup and lens (Fig. 6B, arrow).

At present, it is unclear which cell types give rise to thesemesenchymal populations, but their locations suggest thatAbnormal Ocular Morphogenesis in aA/FGF-3the first type may derive from corneal stroma and the secondTransgenic Embryostype from perioptic mesenchymal cells. The transgenic retinashave more subtle, but consistent changes in architectureHistological analyses of the aA/FGF-3 transgenic lines

revealed several interesting deviations from normal ocular which may be primary or secondary effects of the transgenicexpression of FGF-3. In nontransgenic embryos, the anteriordevelopment. While there was some variation in transgenic

phenotype between lines of transgenic mice, the mice margin of the optic cup, which will normally differentiateinto the ciliary body and iris, curves around the developingwithin a given transgenic line (summarized in Table 1) ex-

hibited a consistent phenotype. Transgenic lines OVE 392 lens enclosing most of the lens mass (Figs. 6A and 6E). In thetransgenic mice, this region of the optic cup characteristicallyand 393A were phenotypically indistinguishable and OVE

391 exhibited a similar, but slightly delayed onset of devel- curves away from the developing lens resulting in a muchmore widely open optic cup and anterior displacement of theopmental disregulation. Transgenic FGF-3 expression in the

eye directly or indirectly influenced the development of all lens (Figs. 6B and 6F). Within this curvature of the optic cup,the neural retina derived from the inner wall of the optic cupocular tissues in these three severely affected transgenic

lines. Mice from OVE 393B characteristically developed a joins with the outer wall of the optic cup which gives rise tothe retinal pigment epithelium (RPE). In the normal retina, asmuch milder, but consistent ocular phenotype involving

only the cornea and lens. At the present time the cause of the anterior retina curves posteriorly, the multilayered neuralretina immediately gives way to the single layered cuboidalthe phenotypic variations between the different transgenic

families is not known, but may result from differences in appearance more typical of the outer optic cup (Fig. 6E, aster-isk). This transition is less distinct in the transgenic retinaonset or quantitative level of transgenic FGF-3 expression.


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

20 Robinson et al.

FIG. 5. In situ hybridization of nontransgenic control (A, B) and transgenic OVE 392 (C, D) E13.5 embryos using an antisense 35S-labeledriboprobe for murine FGF-3. Silver grains revealing the presence of FGF-3 transcripts appear dark under brightfield (A, C) and light underdarkfield (B, D) illumination. After a 60-h exposure, a very strong FGF-3 signal was present over the transgenic lens (l), but completelyabsent from the nontransgenic lens. At this stage, a weak FGF-3 signal was detected in nontransgenic and transgenic retinas (r). No FGF-3 transcripts were detected in the transgenic or nontransgenic cornea (c). Scale bar, 100 mm.

where the transition from neuronal-like columnar cells to cu- ers, MIP and b-crystallin. MIP is a lens-specific membranechannel protein which is normally found exclusively in lensboidal RPE-like cells takes place well posterior to the anterior

margin of the retina (Fig. 6F, asterisk). fiber cells. b-Crystallins are also lens fiber cell-specific pro-teins thought to be essential for maintaining lens transpar-ency. In situ hybridization was used to localize MIP expres-

Premature Differentiation of the Lens Epithelium sion in transgenic and control embryonic tissue sections.in aA/FGF-3 Transgenic Mice In contrast to the control embryos, where MIP transcripts

could only be detected in the lens fiber cells, MIP transcriptsThe elongation of the epithelial cells in the transgenicwere found in both the primary fiber cells and the elongatinglenses was suggestive of premature differentiation into fiberepithelial cells of transgenic lenses as early as E13.5 (Fig. 7).cells. To confirm this hypothesis, transgenic lenses were

examined for the expression of two fiber cell-specific mark- Similar results were obtained in an immunohistochemical


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal


TABLE 1Ocular Phenotype of aA/FGF-3 Transgenic Families

Family E14 Newborn Adult

OVE 391 Ocular proptosis Externally normal MicrophthalmiaCorneal abnormalities Elongation of lens epithelial cells Intraocular glandular structuresAnterior retinal eversion Vacuolization of primary fiber cells

OVE 392 Ocular proptosis External ‘‘scabby’’ eye MicrophthalmiaCorneal abnormalities Microphthalmia Intraocular glandular structuresAnterior retinal eversion Retinal detachment and disorganizationFiber cell vacuolizationLens epithelial elongationAnterior lens rupture

OVE 393A Identical to OVE 392 Identical to OVE 392 Identical to OVE 392

OVE 393B Not determined Externally normal CataractsGoblet cells in corneal epitheliumMild microphthalmia in homozygotes

localization of b-crystallin at E14.5 (Fig. 8). In control em- sion induces the differentiation of the entire lens epithe-lium prior to the loss of the lens near the time of birth.bryos b-crystallin expression was confined to the lens fiber

cells with virtually no signal present in the central or equa- While lenses from OVE 392 and 393A mice typically arelost through the cornea before birth, eyes from OVE 391torial lens epithelium (Figs. 8B and 8F). In transgenic lenses

from OVE 392 or OVE 393A, there is no normal lens epithe- mice typically do not rupture. The epithelial cells from OVE391 lenses do differentiate prematurely, albeit at a slowerlium at E14.5. The elongating cells on the lens surface,

anterior to the lens equator, are cells which would have pace, such that lenses from newborn OVE 391 mice containno epithelial cells and are larger than those of littermatebeen epithelial cells in the normal lens and are referred to

as elongating lens epithelial cells (Figs. 8C and 8G). While controls. These exclusively fiber cell lenses undergo rapidvacuolization and degeneration leaving mice from OVE 391the overall intensity of anti-b-crystallin immunofluores-

cence was lower in the transgenic lens than in the non- severely microphthalmic within 2 weeks after birth.transgenic lens, b-crystallin protein was detected in boththe transgenic fiber cells and the elongating epithelial cells

Chimeric Mice Reveal Paracrine Effect of aA/FGF-3(Figs. 8D and 8E). Negative controls using both transgenicTransgene Expressionand nontransgenic sections in which the primary antibody

incubation was omitted demonstrated no fluorescence The aA/FGF-3 transgene was designed to drive the ex-pression of a secreted form of FGF-3 in the lens. To confirmabove background (data not shown). Therefore, based on

both morphological and biochemical criteria, FGF-3 expres- that transgenic FGF-3 was being secreted from the lens in

FIG. 6. FGF-3 transgene expression by the lens alters the morphogenesis of several ocular tissues. Hematoxylin and eosin-stained tissuesections through comparable regions of nontransgenic (A, C, E) and OVE 392 transgenic (B, D, F) eyes at E14.5 are shown. Sectionsencompassing the entire nontransgenic and transgenic eye are displayed at low magnification in A and B, respectively. Regions boxed inwhite in A and B are shown at higher magnification in C and D, respectively. Regions boxed in black in A and B are shown at highermagnification in E and F, respectively. In the nontransgenic eye (A), the lens (l) is roughly circular and is separated from the cornea (c)by an acellular space which will become the anterior chamber (a). The nontransgenic eyelids (ld) have begun to extend over the outersurface of the cornea. In contrast, the transgenic lens (B) is large and pear-shaped and extends through the cornea (asterisk), preventingthe formation of an anterior chamber. The normal cornea at E14.5 (C) consists of several layers of flattened, condensed mesenchymalcells forming the corneal stroma (s) covered by a single layer of cuboidal corneal epithelium (ce). The transgenic corneal stroma consistsof widely spaced mesenchymal cells which appear to be stellate in morphology and randomly oriented (B, D, arrowheads). The transgeniccorneal epithelium is columnar rather than cuboidal (D) and in places, particularly around the protruding lens, is multilayered. Anothermesenchymal population, more condensed than the corneal stroma is found surrounding the transgenic lens (B, arrow), but is absent fromthe corresponding location in the nontransgenic eye (A). In the nontransgenic eye (A), the majority of the lens is enclosed within the opticcup and the transition from neural retina (r) to retinal pigment epithelium (rpe) is very distinct, occurring at the anterior tip of the retina(E, asterisk). In the transgenic eye, the majority of the lens lies anterior to the optic cup (B), and the tips of the anterior retina appear toprematurely curve to the posterior resulting in a transitory segment of cells between the morphologically distinct neural retina and retinalpigment epithelium (F, asterisk). Scale bar, 210 mm (A, B) or 21 mm (C–F).


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

22 Robinson et al.

05-18-98 21:46:19 dbal


FIG. 7. In situ hybridization sections of nontransgenic (A) and OVE 392 transgenic (B) eyes at E13.5 with an antisense probe to murineMIP. In the nontransgenic eye, MIP expression is confined to differentiated lens fiber cells (lf) and is absent from the lens epithelium (le).In the transgenic lens, the epithelium has elongated (ee) and expresses MIP. The thickness of the lens epithelium in (A) or the elongatingepithelium in (B) is indicated by double arrows. The neural retina and cornea are indicated by r and c, respectively. Scale bar, 32 mm.

05-18-98 21:46:19 dbal

24 Robinson et al.

FIG

.8.

Len

sb

-cry

stal

lin

imm

un

ohis

toch

emis

try

iney

ese

ctio

ns

from

non

tran

sgen

ic(A

,B

,E

,F)

and

OV

E39

3Atr

ansg

enic

(C,

D,

G,

H)

len

ses

atE

14.5

.T

he

phot

ogra

phs

inth

eto

pro

war

elo

w-p

ower

mag

nifi

cati

ons

ofey

ese

ctio

ns

wit

hth

ebo

xed

regi

ons

show

nat

hig

her

mag

nifi

cati

onin

the

corr

espo

ndi

ng

pan

els

inth

ebo

ttom

row

.A

djac

ent

non

tran

sgen

ican

dtr

ansg

enic

sect

ion

sar

esh

own

stai

ned

wit

hh

emat

oxyl

inan

deo

sin

(A,

C,E

,G)t

osh

owoc

ula

rm

orph

olog

yor

wit

hu

ltra

viol

etep

iflu

ores

cen

ce(B

,D,F

,H)t

ovi

sual

izeb

-cry

stal

lin

expr

essi

on.I

nth

en

ontr

ansg

enic

len

s,b

-cry

stal

lin

expr

essi

onis

evid

ent

inth

ele

ns

fibe

rce

lls

(lf)

,bu

tis

abse

nt

from

the

len

sep

ith

eliu

m(l

e).

Inth

etr

ansg

enic

len

s,th

ere

isn

on

orm

alle

ns

epit

hel

ium

,an

db

-cry

stal

lin

expr

essi

onis

evid

ent

thro

ugh

out

the

enti

rele

ns,

incl

udi

ng

the

supe

rfici

alel

onga

tin

gce

lls

ante

rior

toth

ele

ns

equ

ator

(ee)

.T

he

len

gth

ofth

eel

onga

tin

gep

ith

eliu

mis

indi

cate

dby

ado

ubl

ear

row

(C).

Th

etr

ansg

enic

len

sfi

ber

cell

sal

soco

nta

inn

um

erou

sla

rge

vacu

oles

(v).

Th

em

esen

chym

alce

lls

(m)s

urr

oun

din

gth

etr

ansg

enic

len

sdo

not

expr

essb

-cry

stal

lin

.Th

eh

ole

evid

ent

inth

en

ontr

ansg

enic

len

sre

pres

ents

apr

oces

sin

gar

tifa

ct(*

),an

dsc

atte

red

flu

ores

cen

tce

lls

outs

ide

the

len

sre

pres

ent

auto

flu

ores

cen

tbl

ood

cell

s(a

rrow

hea

ds).

Th

eco

rnea

and

reti

na

are

repr

esen

ted

by(c

)an

d(r

)re

spec

tive

ly.

Scal

eba

r,15

0m

m(A

–D

)or

37.5

mm

(E–

H).

05-18-98 21:46:19 dbal


FIG. 9. Eye sections from newborn chimeric mice stained with X-gal to reveal the presence of bacterial b-galactosidase. In the Rosab-geo26}FVB/N chimeric lens (A), patches of blue (Rosab-geo26-derived) cells (b) are interspersed with pink (FVB/N-derived) cells (p) in thelens epithelium (le) and cornea (c). In the Rosab-geo26}OVE 391 chimeric lens (B), blue streaks (arrows) can be seen in the elongatedepithelium (ee), indicating that Rosab-geo26-derived lens epithelial cells underwent premature fiber cell differentiation as did the neigh-boring OVE 391 cells. In the normal lens X-gal fails to penetrate into the lens fiber cell mass; therefore, none of the lens fiber cells (lf)stained blue in the Rosab-geo26}FVB/N chimeric lens. Scale bar, 21 mm.

vivo, chimeric mice were made by aggregating aA/FGF-3 though secretion of FGF-3 by the transgenic lens cells ofthe chimera provides a straightforward mechanism to ex-transgenic embryos with embryos derived from Rosab-

geo26 mice. Rosab-geo26 mice were created in a promoter plain the premature differentiation in nontransgenic cells,other interpretations are possible. For example, it is a formaltrap experiment (Friedrich and Soriano, 1991) and ubiqui-

tously express bacterial b-galactosidase during develop- possibility that FGF-3 expression within transgenic lenscells leads to the secretion of a non-FGF-3 product whichment. If FGF-3 was indeed being secreted by transgenic lens

cells, both aA/FGF-3 transgenic-derived cells and Rosab- in turn induces the differentiation of neighboring Rosab-geo26 lens epithelial cells.geo26-derived cells should respond similarly in a developing

chimeric embryo. Cells from the two different genotypeswere distinguished by blue staining with X-gal, which spe-

Ectopic Glandular Structures in Adult Transgeniccifically stains cells expressing bacterial b-galactosidase.MiceSeveral chimeras were generated and examined at various

developmental stages. Cells derived from the Rosab-geo26 Subsequent to the expulsion (in OVE 392 and 393A) ordegeneration (in OVE 391) of the transgenic lens, the retinaembryos were consistently found in the prematurely differ-

entiating lens epithelium (Fig. 9), demonstrating that cells typically detaches and ocular architecture becomes disorga-nized. Despite having had a very abnormal embryonic archi-not carrying the aA/FGF-3 transgene were able to respond

to transgenic cells in a paracrine manner. This result sug- tecture, the structural abnormalities in adult transgenic cor-neas are comparatively mild (Fig. 10A). An interesting post-gests that the transgenic FGF-3 is indeed secreted and binds

to FGF receptors on the surface of neighboring Rosab-geo26 natal feature which is common to the eyes of all thesetransgenic lines is the appearance of intraocular glandularderived cells in the aA/FGF-3}Rosab-geo26 chimeras. Al-


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

FIG. 10. Eye sections from an adult OVE 393 (before the separation of sublines A and B) X C57BL/6 F1 hybrid mouse exhibiting intraocularglandular structures viewed at low (A) and high (B) magnification. The microphthalmic transgenic eye is bordered by a cornea (c) to theanterior and a mass of pigmented cells (p) to the posterior. Fragments of the lens (l) remain, but most of the intraocular space is occupiedby glandular structures (g). These acinar structures contain both secretory cells (sc) with prominent cytoplasmic lipid vacuoles, andcuboidal epithelial cells arranged in a duct-like pattern (dc) interspersed with pigmented cells (p). The apparent absence of neural retinaseen in these photographs is atypical and may represent a sectioning artifact. Scale bar, 100 mm (A) or 25 mm (B).FIG. 11. Sections of nontransgenic (A) and OVE 393B transgenic (B) adult corneas. Both corneas display the characteristic endothelium(en), stroma (s), and epithelium (ce), but in the transgenic cornea there are patches of cells exhibiting morphology similar to that of mucus-producing goblet cells (B, arrows). Scale bar, 38 mm.

26

05-18-98 21:46:19 dbal


structures. Although these glandular structures can be directly or indirectly alters the morphogenesis of severalocular tissues. Alterations in corneal morphogenesis in-found in the microphthalmic adult eyes of all mice from

these transgenic families, the effect is particularly pro- clude the elongation of the corneal epithelium and the mal-formation of the corneal stroma (compare Figs. 6C and 6D).nounced in transgenic F1 hybrids resulting from a cross

between C57BL/6 inbred mice and transgenic mice main- These alterations in corneal architecture, coupled with thelarge unusually shaped lens, likely contribute to ocular rup-tained on an FVB/N genetic background (Fig. 10). In these

mice, well-differentiated acinar structures appear within ture leading to the prenatal loss of the lens in the mostseverely affected aA/FGF-3 transgenic families. It is cur-the ocular cavity. The glandular tissue includes cells with

cytoplasmic lipid vacuoles and/or secretory vesicles as well rently unknown if the abnormal corneal development re-sults from a direct effect of the secreted FGF-3 protein, or anas epithelial cells organized into intralobular ducts (Fig.

10B). Examination of aA/FGF-3}Rosab-geo26 chimeric indirect effect due to the disruption of an as yet undefinedinductive signal from the lens to the developing cornea.mice indicates that both transgenic and nontransgenic cells

can participate in the development of this glandular mate- Several pieces of data suggest a direct effect of FGF-3 on thecornea. First, while there are many spontaneous mutationsrial (data not shown). The effect of the transgene on glandu-

lar cell types is even demonstrated in the most mildly af- which are known to disrupt the early development of thelens (for example Harch et al., 1978; Hill et al., 1991; Zwaan,fected transgenic line, OVE 393B, where cells having the

morphological appearance of secretory goblet cells consis- 1975), none of these result in a disruption of corneal mor-phogenesis similar to that caused by the expression of FGF-tently appear in the corneal epithelium (Fig. 11). Mucus-

secreting goblet cells are normally found in the conjunctival 3. In addition, corneal architecture is somewhat restored inpostnatal development after the expulsion or degenerationepithelium, but are not present in the normal cornea.of the lens (see Fig. 10A), indicating that once the source ofFGF-3 (the lens) is removed and in the absence of an intactlens, the cornea is able to establish a more typical develop-DISCUSSIONmental pathway.

The development of the eyelids and anterior optic cupFGF-3 expression in the lens leads to the premature differ-entiation of the entire lens epithelium. This conclusion is are also altered in the prenatal aA/FGF-3 transgenic eyes.

Formation and closure of the eyelids is normally a rapidsupported by several independent observations. First, thelens epithelium undergoes dramatic elongation similar to process, initiated by E14.5 and completed by E16.5 (Kauf-

man, 1992; Pei and Rhodin, 1970). In severely affectedthat of differentiating lens fiber cells. Second, these elongat-ing lens epithelial cells express mRNA for the fiber cell- transgenic families, eyelid development is perturbed such

that eyelid closure is not complete until several days afterspecific gene MIP. Finally, the elongating lens epitheliumalso expresses fiber cell-specific b-crystallin proteins. The birth. Altered eyelid development is evident in transgenic

eyes from these families by E14.5 (compare Figs. 6A andresponsiveness of the lens epithelium for FGF-3 suggeststhat the embryonic lens epithelium expresses one or both 6B). Currently, it is unknown whether the delay in eyelid

closure is a direct effect of transgene expression or an indi-fibroblast growth factor receptor isoforms most responsiveto FGF-3, namely IIIb isoforms (Johnson and Williams, 1993) rect effect resulting from the stearic hindrance of the bulg-

ing prenatal eye.of FGFR1 and FGFR2 (Mathieu et al., 1995; Ornitz et al.,1996). Although it is known that the embryonic lens ex- In the normal embryonic eye there is a clear histological

distinction between the cells of the inner and outer opticpresses FGFR1 (de Iongh et al., 1996; Orr-Urtreger et al.,1991), it is not known which FGFR1 isoforms are expressed. cup which form the neural retina and the retinal pigment

epithelium, respectively. These two optic cup layers areThe lens expresses both the IIIb and IIIc isoforms of FGFR2(de Iongh et al., 1997; Orr-Urtreger et al., 1993). In an in joined at the anterior margin of the retina. In the aA/FGF-

3 transgenic eye, the neural retina appears to extend beyondsitu hybridization analysis of embryonic rat lenses, FGFR2IIIc transcripts were found uniformly distributed in the em- the inner optic cup into the outer optic cup (compare Figs.

6E and 6F). While the cause of this apparent neural retinabryonic lens epithelium, while FGFR2 IIIb transcripts mark-edly increased with the onset of lens differentiation at the eversion or expansion is not entirely clear, we have formu-

lated two hypotheses. Either the retina is forced backwardtransitional zone (de Iongh et al., 1997). Recently, trans-genic mice expressing a secreted form of FGF-2 in the lens by physical stress induced by the rapid enlargement of the

lens or the anterior portion of the outer optic cup is beingwere reported where premature differentiation of the lensepithelium did not occur (Stolen et al., 1997), in agreement induced to differentiate into neural retina by the FGF-3 se-

creted from the lens. There is some precedent in support ofwith unpublished observations of similar mice produced inour laboratory. FGF-2 is capable of activating FGFR1 IIIb, the second hypothesis. In chick embryos, subsequent to the

removal of the neural retina, FGF-2 has been shown to in-but is unable to activate FGFR2 IIIb (Ornitz et al., 1996).Taken together, these results suggest that FGFR2 IIIb is duce the regeneration of neural retina from the cells of the

RPE (Park and Hollenberg, 1989). Furthermore, FGF-2 hascapable of mediating FGF-induced fiber cell differentiationin vivo. been shown to specifically suppress RPE development and

promote neural retina development of the outer optic cupIn addition to inducing the differentiation of the entirelens epithelium, FGF-3 expression by the developing lens cells in chick embryonic eye cultures (Pittack et al., 1997).


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

28 Robinson et al.

In any case, the alterations in retinal development induced degeneration are not entirely clear. It is likely that lensfiber cells are dependent on the lens epithelium in vivoby the expression of FGF-3 by the lens require further inves-

tigation. and the degeneration of the primary fiber cells is secondaryto the loss of a functional lens epithelium. Alternatively,Perhaps the most unusual and distinctive feature of the

aA/FGF-3 transgenic mice is the postnatal appearance of synthesis of secreted FGF-1 or FGF-3 by the lens fiber cellsmay directly interfere with their normal function and phys-intraocular secretory epithelium. This effect is most evident

in hybrid mice resulting from the cross of aA/FGF-3 iology. However, overexpression of nonsecreted FGF-1 didnot result in premature differentiation of the lens epithe-transgenic mice on an FVB/N background to C57BL/6 inbred

mice, where the appearance of well-differentiated acinar and lium or in degeneration of the primary fiber cells (Robinsonet al., 1995b).ductal epithelium is a characteristic finding (Fig. 10). At

present, the origin of the intraocular glandular tissue is spec- Although both FGF-3 and secreted FGF-1 are capable ofinducing premature differentiation of the lens epitheliumulative, but it apparently results from a migration of extraoc-

ular glandular tissue into the eye, perhaps facilitated by cor- in transgenic mice, FGF-3 expression induced many otherocular changes not observed in transgenic mice expressingneal wounding, or an ectopic glandular transdifferentiation

of a preexisting ocular cell type. The FGF-3 gene was ini- FGF-1 (Robinson et al., 1995b). The mechanisms that causethe phenotypic differences between these transgenic micetially identified by virtue of its activation in mouse mam-

mary carcinoma (Moore et al., 1986), and both FGF-3 and are not entirely clear. FGF-1 is the only member of the FGFfamily which is capable of binding and activating all FGFFGF-7 induce hyperplasia of the prostate and mammary

gland in vivo (Muller et al., 1990; Stamp et al., 1992; Tutrone receptor isoforms (Ornitz et al., 1996). Therefore the moredramatic phenotype induced by FGF-3 is unlikely to resultet al., 1993). While it is possible that the FGF-3 expressed

by the lens acts as a trophic factor for the development of from activating a specific subset of FGF receptors insensi-tive to FGF-1. The phenotypic differences observed be-the glandular epithelium within the eye, it is curious that

in most transgenic families the glandular appearance fol- tween secreted FGF-1 and FGF-3 expression may relate todifferences in the quantitative amount of active ligandlows the expulsion or destruction of the majority of express-

ing lens cells. It is also interesting to note that in the most reaching the relevant ocular receptors. FGF-3 contains anendogenous signal peptide consensus sequence and ismildly affected aA/FGF-3 transgenic family (OVE 393B) the

most consistent phenotype is the appearance of secretory known to be secreted in vivo (Kiefer et al., 1991), whilethe transgenic FGF-1 construct had to be engineered forgoblet cells embedded within the corneal epithelium (Fig.

11). While goblet cells are not unusual in the conjunctiva, secretion by the addition of a signal peptide sequence fromFGF-4 (Robinson et al., 1995b). Therefore the FGF-1 trans-they do not normally appear in the cornea.

Despite being initially discovered in mice as an oncogene genic transcripts or proteins may be produced less effi-ciently than those from the aA/FGF-3 construct. Alterna-activated by proviral integration of MMTV (Moore et al.,

1986), overexpression of FGF-3 by the lens does not result tively, FGF-3 may diffuse more rapidly or may be a moreeffective FGFR agonist in the eye.in tumorigenesis of the lens or any other ocular cell type.

While induction of proliferation and hyperplasia in certain It is still unknown which FGF or FGF receptor is responsi-ble for lens development and differentiation in vivo. Genecell types within the eye may result from FGF-3 transgene

expression, these effects are transient and have not been targeted mice have provided some insights, but have failedto reveal which FGFs or FGF receptors are the essential rolequantified in this report. With respect to the lens, BrdU

labeling studies have shown that all lens cell proliferation players in lens development. There is abundant evidencesuggesting that signals from outside the lens regulate nor-ceases by E14.5 in the most severely affected transgenic

lines OVE 392 and OVE 393A (M.L.R. and P.A.O., unpub- mal lens fiber differentiation (Coulombre and Coulombre,1963; Yamamoto, 1976). Since FGF-3 is expressed in thelished observations). We conclude that the major effect of

FGF-3 overexpression in the eye is the alteration of morpho- developing retina (Wilkinson et al., 1989), this growth factoris well placed to influence lens differentiation. Despite this,genesis and differentiation.

With respect to the lens, these results are similar to those mice lacking FGF-3 have no reported ocular abnormalities(Mansour et al., 1993) indicating that any possible endoge-achieved by expressing a secreted FGF-1 clone in the lens

(Robinson et al., 1995b), but in the case of the aA/FGF-3 nous role for FGF-3 in lens development is redundant. Also,mice lacking FGFR3 have not been reported to suffer frommice, lens differentiation occurs much earlier in em-

bryogenesis. The lens epithelium in three of the four aA/ ocular defects (Colvin et al., 1996). Mice homozygous for adeletion of FGFR1 die before the initiation of lens develop-FGF-3 transgenic families is completely differentiated prior

to birth, and in two of these three families the process is ment (Deng et al., 1994; Yamaguchi et al., 1994), and micelacking FGFR2 have not been reported. Transgenic miceessentially completed by E14.5. In contrast, only one of

five secreted FGF-1 families shows evidence of premature expressing a dominant negative version of FGFR1 in thelens have been produced by two independent laboratoriesdifferentiation of the lens epithelium prior to E16.5 (M. L.

Robinson, personal observation). Overexpression of either (Chow et al., 1995; Robinson et al., 1995a). The dominantnegative FGFR1 used in these studies was previously shownsecreted FGF-1 or FGF-3 within the lens leads to the de-

struction of the primary lens fiber cells (see central lens to block the activity of multiple FGF receptor isoforms(Ueno et al., 1992). While the transgenic mice produced byvacuoles in Figs. 6B, 8C, and 8D). The reasons for this


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal


de Iongh, R., and McAvoy, J. W. (1993). Spatio-temporal distribu-both groups exhibited abnormal lenses, in each case expres-tion of acidic and basic FGF indicates a role for FGF in rat lenssion of the dominant negative receptor was restricted to themorphogenesis. Dev. Dyn. 198, 190–202.lens fiber cells. Therefore, while these mice suggest that

de Iongh, R. U., Lovicu, F. J., Chamberlain, C. G., and McAvoy,FGF signaling is required to maintain healthy fiber cells,J. W. (1997). Differential expression of fibroblast growth factorthe requirement or role of FGF receptor activation in thereceptors during rat lens morphogenesis and growth. Invest. Oph-

initiation of fiber cell differentiation could not be addressed thalmol. Visual Sci. 38, 1688–1699.by these mice. de Iongh, R. U., Lovicu, F. J., Hanneken, A., Baird, A., and McAvoy,

The identification of individual roles for different FGFs J. W. (1996). FGF receptor-1 (flg) expression is correlated withand FGF receptors in eye development remains a challenge fibre differentiation during rat lens morphogenesis and growth.for future investigations. The results reported here with the Dev. Dyn. 206, 412–426.

Deng, C. X., Wynshaw-Boris, A., Shen, M. M., Daugherty, C., Or-aA/FGF-3 transgenic mice support the hypothesis thatnitz, D. M., and Leder, P. (1994). Murine FGFR-1 is required forFGFs are important molecules in the development of theearly postimplantation growth and axial organization. Geneseye. Two different FGF family members have now beenDev. 8, 3045–3057.shown to induce the premature differentiation of the lens

Dickson, C., and Peters, G. (1987). Potential oncogene product re-epithelium in the developing mouse embryo. Furthermore,lated to growth factors [letter]. Nature 326, 833.deregulation of FGF-3 expression also has wide reaching

Dixon, M., Deed, R., Acland, P., Moore, R., Whyte, A., Peters, G.,effects on the development and/or differentiation of other and Dickson, C. (1989). Detection and characterization of theocular cell types, suggesting that these other ocular lineages fibroblast growth factor-related oncoprotein INT-2. Mol. Cellcan respond to differentiation signals produced by the lens. Biol. 9, 4896–4902.

Faktorovich, E. G., Steinberg, R. H., Yasumura, D., Matthes, M. T.,and LaVail, M. M. (1990). Photoreceptor degeneration in inher-ited retinal dystrophy delayed by basic fibroblast growth factor.ACKNOWLEDGMENTSNature 347, 83–86.

Faktorovich, E. G., Steinberg, R. H., Yasumura, D., Matthes, M. T.,The authors thank Dr. Joan E. Durbin, Dr. Sue Hammond, and and LaVail, M. M. (1992). Basic fibroblast growth factor and local

Dr. Mark H. Luquette for helpful discussions about the pathology injury protect photoreceptors from light damage in the rat. J.of the aA/FGF-3 transgenic eyes, Dr. Natarajan Muthusamy and Neurosci. 12, 3554–3567.Julia E. Robinson for critical review of the manuscript, and Theresa Feldman, B., Poueymirou, W., Papaioannou, V. E., DeChiara, T. M.,Cox for secretarial assistance. This work was supported in part by and Goldfarb, M. (1995). Requirement of FGF-4 for postimplanta-Grants EY-10448 and EY-10803 and an NRSA Fellowship EY-06511 tion mouse development. Science 267, 246–249.from the National Eye Institute, Grant P30 CA16058 from the

Finch, P. W., Cunha, G. R., Rubin, J. S., Wong, J., and Ron, D. (1995).National Cancer Institute and additional funding from the Chil-

Pattern of keratinocyte growth factor and keratinocyte growthdren’s Hospital Research Foundation.

factor receptor expression during mouse fetal development sug-gests a role in mediating morphogenetic mesenchymal-epithelialinteractions. Dev. Dyn. 203, 223–240.

Fredj-Reygrobellet, D., Plouet, J., Delayre, T., Baudouin, C., Bour-REFERENCESret, F., and Lapalus, P. (1987). Effects of aFGF and bFGF on woundhealing in rabbit corneas. Curr. Eye Res. 6, 1205–1209.Acland, P., Dixon, M., Peters, G., and Dickson, C. (1990). Subcellu-

Friedrich, G., and Soriano, P. (1991). Promoter traps in embryoniclar fate of the int-2 oncoprotein is determined by choice of initia-stem cells: a genetic screen to identify and mutate developmentaltion codon. Nature 343, 662–665.genes in mice. Genes Dev. 5, 1513–1523.Burmeister, M., Novak, J., Liang, M. Y., Basu, S., Ploder, L., Hawes,

Gorin, M. B., Yancey, S. B., Cline, J., Revel, J. P., and Horwitz, J.N. L., Vidgen, D., Hoover, F., Goldman, D., Kalnins, V. I., Roder-(1984). The major intrinsic protein (MIP) of the bovine lens fiberick, T. H., Taylor, B. A., Hankin, M. H., and McInnes, R. R.membrane: Characterization and structure based on cDNA clon-(1996). Ocular retardation mouse caused by Chx10 homeoboxing. Cell 39, 49–59.null allele: impaired retinal progenitor proliferation and bipolar

Grindley, J. C., Davidson, D. R., and Hill, R. E. (1995). The role ofcell differentiation. Nat. Genet. 12, 376–384.Pax-6 in eye and nasal development. Development 121, 1433–Chellaiah, A. T., McEwen, D. G., Werner, S., Xu, J., and Ornitz,1442.D. M. (1994). Fibroblast growth factor receptor (FGFR) 3. Alterna-

Guillemot, F., and Cepko, C. L. (1992). Retinal fate and gangliontive splicing in immunoglobulin-like domain III creates a recep-cell differentiation are potentiated by acidic FGF in an in vitrotor highly specific for acidic FGF/FGF-1. J. Biol. Chem. 269,assay of early retinal development. Development 114, 743–754.11620–11627.

Harch, C., Chase, H. B., and Gonsalves, N. I. (1978). Studies on anChow, R. L., Roux, G. D., Roghani, M., Palmer, M. A., Rifkin, D. B.,anophthalmic strain of mice. VI. Lens and cup interaction. Dev.Moscatelli, D. A., and Lang, R. A. (1995). FGF suppressesBiol. 63, 352–357.apoptosis and induces differentiation of fibre cells in the mouse

Hebert, J. M., Rosenquist, T., Gotz, J., and Martin, G. R. (1994).lens. Development 121, 4383–4393.FGF5 as a regulator of the hair growth cycle: Evidence from tar-Colvin, J. S., Bohne, B. A., Harding, G. W., McEwen, D. G., and Or-geted and spontaneous mutations. Cell 78, 1017–1025.nitz, D. M. (1996). Skeletal overgrowth and deafness in mice lack-

Hicks, D., and Courtois, Y. (1992). Fibroblast growth factor stimu-ing fibroblast growth factor receptor 3. Nature Genet. 12, 390–lates photoreceptor differentiation in vitro. J. Neurosci. 12, 2022–397.2033.Coulombre, J. L., and Coulombre, A. J. (1963). Lens development:

fiber elongation and lens orientation. Science 142, 1489–1490. Hill, R. E., Favor, J., Hogan, B. L., Ton, C. C., Saunders, G. F., Han-


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

30 Robinson et al.

son, I. M., Prosser, J., Jordan, T., Hastie, N. D., and van Hey- Muller, W. J., Lee, F. S., Dickson, C., Peters, G., Pattengale, P., andLeder, P. (1990). The int-2 gene product acts as an epithelialningen, V. (1991). Mouse small eye results from mutations in a

paired-like homeobox- containing gene [published erratum ap- growth factor in transgenic mice. EMBO J. 9, 907–913.pears in Nature 1992 Feb. 20; 355(6362),750]. Nature 354, 522– Ornitz, D. M., Xu, J., Colvin, J. S., McEwen, D. G., MacArthur,525. C. A., Coulier, F., Gao, G., and Goldfarb, M. (1996). Receptor

specificity of the fibroblast growth factor family. J. Biol. Chem.Hogan, B., Beddington, R., Costantini, F., and Lacy, E. (1994). ‘‘Ma-nipulating the Mouse Embryo: A Laboratory Manual.’’ Cold 271, 15292–15297.Spring Harbor Laboratory Press, Plainview, NY. Orr-Urtreger, A., Bedford, M. T., Burakova, T., Arman, E., Zimmer,

Y., Yayon, A., Givol, D., and Lonai, P. (1993). DevelopmentalJohnson, D. E., Lu, J., Chen, H., Werner, S., and Williams, L. T.(1991). The human fibroblast growth factor receptor genes: A localization of the splicing alternatives of fibroblast growth factor

receptor-2 (FGFR2). Dev. Biol. 158, 475–486.common structural arrangement underlies the mechanisms forgenerating receptor forms that differ in their third immunoglobu- Orr-Urtreger, A., Givol, D., Yayon, A., Yarden, Y., and Lonai, P.lin domain. Mol. Cell. Biol. 11, 4627–4634. (1991). Developmental expression of two murine fibroblast

growth factor receptors, flg and bek. Development 113, 1419–Johnson, D. E., and Williams, L. T. (1993). Structural and functionaldiversity in the FGF receptor multigene family. Adv. Cancer Res. 1434.60, 1–41. Overbeek, P. A., Chepelinsky, A. B., Khillan, J. S., Piatigorsky, J.,

and Westphal, H. (1985). Lens-specific expression and develop-Kaufman, M. H. (1992). ‘‘The Atlas of Mouse Development.’’ Aca-mental regulation of the bacterial chloramphenicol acetyltrans-demic Press, San Diego, CA.ferase gene driven by the murine alpha A- crystallin promoter inKiefer, P., Acland, P., Pappin, D., Peters, G., and Dickson, C. (1994).transgenic mice. Proc. Natl. Acad. Sci. USA 82, 7815–7819.Competition between nuclear localization and secretory signals

Park, C. M., and Hollenberg, M. J. (1989). Basic fibroblast growthdetermines the subcellular fate of a single CUG-initiated formfactor induces retinal regeneration in vivo. Dev. Biol. 134, 201–of FGF3. EMBO J. 13, 4126–4136.205.Kiefer, P., Peters, G., and Dickson, C. (1991). The Int-2/Fgf-3 onco-

Pei, Y. F., and Rhodin, J. A. (1970). The prenatal development ofgene product is secreted and associates with extracellular matrix:the mouse eye. Anat. Rec. 168, 105–125.implications for cell transformation. Mol. Cell. Biol. 11, 5929–

Peters, K., Ornitz, D., Werner, S., and Williams, L. (1993). Unique5936.expression pattern of the FGF receptor 3 gene during mouse or-Kitaoka, T., Aotaki-Keen, A. E., and Hjelmeland, L. M. (1994). Dis-ganogenesis. Dev. Biol. 155, 423–430.tribution of FGF-5 in the rhesus macaque retina [published erra-

Piatigorsky, J. (1981). Lens differentiation in vertebrates. A reviewtum appears in Invest. Ophthalmol. 1995 May; 36(6), 976]. In-of cellular and molecular features. Differentiation 19, 134–153.vest. Ophthalmol. Visual Sci. 35, 3189–3198.

Pittack, C., Grunwald, G. B., and Reh, T. A. (1997). FibroblastLovicu, F. J., Chamberlain, C. G., and McAvoy, J. W. (1995). Differ-growth factors are necessary for neural retina but not pigmentedential effects of aqueous and vitreous on fiber differentiation andepithelium differentiation in chick embryos. Development 124,extracellular matrix accumulation in lens epithelial explants. In-805–816.vest. Ophthalmol. Visual Sci. 36, 1459–1469.

Robinson, M. L., MacMillan-Crow, L. A., Thompson, J. A., and Ov-Mahon, K. A., Chepelinsky, A. B., Khillan, J. S., Overbeek, P. A.,erbeek, P. A. (1995a). Expression of a truncated FGF receptor re-Piatigorsky, J., and Westphal, H. (1987). Oncogenesis of the lenssults in defective lens development in transgenic mice. Develop-in transgenic mice. Science 235, 1622–1628.ment 121, 3959–3967.Mansour, S. L., Goddard, J. M., and Capecchi, M. R. (1993). Mice

Robinson, M. L., Overbeek, P. A., Verran, D. J., Grizzle, W. E.,homozygous for a targeted disruption of the proto-oncogene int-Stockard, C. R., Friesel, R., Maciag, T., and Thompson, J. A.2 have developmental defects in the tail and inner ear. Develop-(1995b). Extracellular FGF-1 acts as a lens differentiation factorment 117, 13–28.in transgenic mice. Development 121, 505–514.Mathers, P. H., Grinberg, A., Mahon, K. A., and Jamrich, M. (1997).

Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R.,The Rx homeobox gene is essential for vertebrate eye develop-Horn, G. T., Mullis, K. B., and Erlich, H. A. (1988). Primer-di-ment. Nature 387, 603–607.rected enzymatic amplification of DNA with a thermostableMathieu, M., Chatelain, E., Ornitz, D., Bresnick, J., Mason, I.,DNA polymerase. Science 239, 487–491.Kiefer, P., and Dickson, C. (1995). Receptor binding and mito-

Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). ‘‘Moleculargenic properties of mouse fibroblast growth factor 3. ModulationCloning: A Laboratory Manual.’’ Cold Spring Harbor Laboratoryof response by heparin. J. Biol. Chem. 270, 24197–24203.Press, Plainview, NY.McAvoy, J. W., and Chamberlain, C. G. (1989). Fibroblast growth

Sasaki, H., and Hogan, B. L. (1994). HNF-3 beta as a regulator offactor (FGF) induces different responses in lens epithelial cellsfloor plate development. Cell 76, 103–115.depending on its concentration. Development 107, 221–228.

Sievers, J., Hausmann, B., Unsicker, K., and Berry, M. (1987). Fibro-McKay, I. J., Lewis, J., and Lumsden, A. (1996). The role of FGF-3 inblast growth factors promote the survival of adult rat retinalearly inner ear development: An analysis in normal and kreislerganglion cells after transection of the optic nerve. Neurosci. Lett.mutant mice. Dev. Biol. 174, 370–378.76, 157–162.McWhirter, J. R., Goulding, M., Weiner, J. A., Chun, J., and Murre,

Southern, E. (1975). Detection of specific sequences among DNAC. (1997). A novel fibroblast growth factor gene expressed in thefragments separated by gel electrophoresis. J. Mol. Biol. 98, 503–developing nervous system is a downstream target of the chime-517.ric homeodomain oncoprotein E2A-Pbx1. Development 124,

3221–3232. Stamp, G., Fantl, V., Poulsom, R., Jamieson, S., Smith, R., Peters,G., and Dickson, C. (1992). Nonuniform expression of a mouseMoore, R., Casey, G., Brookes, S., Dixon, M., Peters, G., and Dick-mammary tumor virus-driven int-2/Fgf-3 transgene in preg-son, C. (1986). Sequence, topography and protein coding potentialnancy-responsive breast tumors. Cell Growth Differ. 3, 929–938.of mouse int-2: A putative oncogene activated by mouse mam-

mary tumour virus. EMBO J. 5, 919–924. Stolen, C. M., Jackson, M. W., and Griep, A. E. (1997). Overex-


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal


pression of FGF-2 modulates fiber cell differentiation and sur- pattern of the FGF-related proto-oncogene int-2 suggests multipleroles in fetal development. Development 105, 131–136.vival in the mouse lens. Development 124, 4009–4017.

Taketo, M., Schroeder, A. C., Mobraaten, L. E., Gunning, K. B., Wilkinson, D. G., Peters, G., Dickson, C., and McMahon, A. P.(1988). Expression of the FGF-related proto-oncogene int-2 duringHanten, G., Fox, R. R., Roderick, T. H., Stewart, C. L., Lilly, F.,

Hansen, C. T., and Overbeek, P. A. (1991). FVB/N: an inbred gastrulation and neurulation in the mouse. EMBO J. 7, 691–695.Yamaguchi, T. P., Harpal, K., Henkemeyer, M., and Rossant, J.mouse strain preferable for transgenic analyses. Proc. Natl. Acad.

Sci. USA 88, 2065–2069. (1994). fgfr-1 is required for embryonic growth and mesodermalpatterning during mouse gastrulation. Genes Dev. 8, 3032–3044.Tutrone, R. F., Jr., Ball, R. A., Ornitz, D. M., Leder, P., and Richie,

J. P. (1993). Benign prostatic hyperplasia in a transgenic mouse: Yamaguchi, T. P., and Rossant, J. (1995). Fibroblast growth factorsin mammalian development. Curr. Opin. Genet. Dev. 5, 485–a new hormonally sensitive investigatory model. J. Urol. 149,

633–639. 491.Yamamoto, Y. (1976). Growth of lens and ocular environment: roleUeno, H., Gunn, M., Dell, K., Tseng, A., and Williams, L. (1992).

A truncated Form of fibroblast growth factor receptor 1 inhibits of neural retina in the growth of mouse lens as revealed by ansignal transduction by multiple types of fibroblast growth factor implantation experiment. Dev. Growth Differ. 18, 273–278.receptor. J. Biol. Chem. 267, 1470–1476. Zhang, C., Takahashi, K., Lam, T. T., and Tso, M. O. (1994). Effects

Unoki, K., and LaVail, M. M. (1994). Protection of the rat retina of basic fibroblast growth factor in retinal ischemia. Invest. Oph-from ischemic injury by brain-derived neurotrophic factor, ciliary thalmol. Visual Sci. 35, 3163–3168.neurotrophic factor, and basic fibroblast growth factor. Invest. Zwaan, J. (1975). Immunofluorescent studies on aphakia, a muta-Ophthalmol. Visual Sci. 35, 907–915. tion of a gene involved in the control of lens differentiation in

Wilkie, A. O., Morriss-Kay, G. M., Jones, E. Y., and Heath, J. K. the mouse embryo. Dev. Biol. 44, 306–312.(1995). Functions of fibroblast growth factors and their receptors.

Received for publication September 16, 1997Curr. Biol. 5, 500–507.Wilkinson, D. G., Bhatt, S., and McMahon, A. P. (1989). Expression Accepted February 19, 1998


AID DB 8879 / 6x3c$$$221 05-18-98 21:46:19 dbal

Disregulation of Ocular Morphogenesis by Lens-Specific Expression of FGF-3/Int-2 in Transgenic Mice

Documents

Transcript of Disregulation of Ocular Morphogenesis by Lens-Specific Expression of FGF-3/Int-2 in Transgenic Mice