defects in development of the kidney, heart and eye ... · vasculature of the eye. these data...

INTRODUCTION

The Notch signaling pathway is an evolutionarily conservedintercellular signaling mechanism, and mutations in itscomponents disrupt cell-fate specification and embryonicdevelopment in organisms as diverse as insects, sea urchins,nematodes and mammals. Genes of the Notch family encodelarge transmembrane receptors that interact with membrane-bound ligands encoded by Delta, Serrate and Jagged familygenes. The signal induced by ligand binding is transmittedintracellularly by a process involving proteolysis of thereceptor and nuclear translocation of the intracellular domainof the Notch protein (reviewed by Artavanis-Tsakonas et al.,1999; Weinmaster, 2000).

Proteins of the Notch family share several repeated peptidemotifs. The extracellular domain of the protein contains a large

number (>30) of tandemly repeated copies of an epidermalgrowth factor (EGF)-like motif. The extracellular domain ofthese proteins also contains three repeats of another motif,unique to this protein family, termed the Notch/lin-12 repeat.The intracellular domain of Notch family proteins contains sixcopies of another conserved motif termed the cdc10/ankyrinrepeat, as well as nuclear localization signals and domainsrequired for interaction with a number of cytoplasmic andnuclear proteins.

Four Notch genes have been identified in mammals (Notch1,Notch2, Notch3and Notch4). Analysis of both inherited humandiseases and mutant mice has shown that some of these genesare essential for embryonic development and adult homeostasis(Swiatek et al., 1994; Conlon et al., 1995; Joutel et al., 1996;Hamada et al., 1999; Krebs et al., 2000). Similarly, genesencoding ligands for the Notch family of receptors are also

491Development 128, 491-502 (2001)Printed in Great Britain © The Company of Biologists Limited 2001DEV4532

The Notch gene family encodes large transmembranereceptors that are components of an evolutionarilyconserved intercellular signaling mechanism. To assessthe in vivo role of the Notch2 gene, we constructed atargeted mutation, Notch2del1. Unexpectedly, we found thatalternative splicing of the Notch2del1 mutant allele leads tothe production of two different in-frame transcripts thatdelete either one or two EGF repeats of the Notch2 protein,suggesting that this allele is a hypomorphic Notch2mutation. Mice homozygous for the Notch2del1 mutationdied perinatally from defects in glomerular developmentin the kidney. Notch2del1/Notch2del1 mutant kidneys werehypoplastic and mutant glomeruli lacked a normalcapillary tuft. The Notch ligand encoded by the Jag1 genewas expressed in developing glomeruli in cells adjacent toNotch2-expressing cells. We show that mice heterozygous

for both the Notch2del1 and Jag1dDSL mutations exhibit aglomerular defect similar to, but less severe than, that ofNotch2del1/Notch2del1 homozygotes. The co-localization andgenetic interaction of Jag1and Notch2 imply that thisligand and receptor physically interact, forming part ofthe signal transduction pathway required for glomerulardifferentiation and patterning. Notch2del1/Notch2del1

homozygotes also display myocardial hypoplasia, edemaand hyperplasia of cells associated with the hyaloidvasculature of the eye. These data identify noveldevelopmental roles for Notch2in kidney, heart and eyedevelopment.

Key words: Notch2, Glomerulogenesis, Kidney, Hypomorphicmutation, Persistent hyperplastic primary vitreous, Mouse

SUMMARY

Defects in development of the kidney, heart and eye vasculature in mice

homozygous for a hypomorphic Notch2 mutation

Brent McCright 1, Xiang Gao 1,2,*, Liya Shen 2,‡, Julie Lozier 1, Yu Lan 1,§, Maureen Maguire 2,¶, Doris Herzlinger 3,Gerry Weinmaster 4, Rulang Jiang 1,§ and Thomas Gridley 1,2,**1The Jackson Laboratory, Bar Harbor, ME 04609, USA2Roche Institute of Molecular Biology, Roche Research Center, Nutley, NJ 07110, USA3Department of Physiology and Biophysics and Department of Urology, Cornell University Medical College, New York, NY 10021,USA4Department of Biological Chemistry, UCLA School of Medicine, Los Angeles, CA 90024, USA*Present address: Institute of Molecular Medicine, Nanjing University, Nanjing 210093, China‡Present address: National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA§Present address: Center for Oral Biology and Department of Biology, University of Rochester, Rochester, NY 14642, USA¶Present address: Schering-Plough Research Institute, Kenilworth, NJ 07033, USA**Author for correspondence (e-mail: [email protected])

Accepted 22 November 2000; published on WWW 23 January 2001

492

essential in both humans and mice (Hrabé De Angelis et al.,1997; Li et al., 1997; Oda et al., 1997; Sidow et al., 1997; Jianget al., 1998; Kusumi et al., 1998; Xue et al., 1999; Bulman etal., 2000). The Notch2gene was the second of the mammalianNotch family receptors cloned (Weinmaster et al., 1992).Hamada et al. have constructed and analyzed a Notch2mutantallele in which the Escherichia coliβ-galactosidase-codingregion replaces all but one of the ankyrin repeats in theintracellular domain of the mouse Notch2 protein (Hamada etal., 1999). Embryos homozygous for this Notch2mutant alleledie before 11.5 days of gestation.

In this study, we have constructed and analyzed anothertargeted mutation of the Notch2gene. Unexpectedly,alternative splicing of our Notch2mutant allele leads to theproduction of two different in-frame transcripts that deleteeither one or two EGF repeats of the Notch2 protein. Micehomozygous for this Notch2 mutant allele die as neonates,owing to defects in kidney development, and also exhibitdefects in development of the heart and eye vasculature. Thisprobable Notch2hypomorphic allele has therefore permitted usto identify and analyze additional developmental decisions thatare dependent on Notch2 function.

Our work indicates that Notch2 function is required fordevelopment of the kidney. The vertebrate kidney is formedby a series of inductive interactions between epithelial andmesenchymal tissues (Saxén, 1987). In mice, formation of theadult kidney, or metanephros, begins at embryonic day (E) 11and is driven by reciprocal induction between the ureteric budepithelium and the metanephric mesenchyme (for recentreviews, see Davies and Bard, 1998; Clark and Bertram,1999; Kuure et al., 2000; Schedl and Hastie, 2000). Theureteric bud branches from the Wolffian duct and grows intothe metanephric mesenchyme. Branching and growth of theureteric bud is induced by signals from these mesenchymalcells. At the tips of the ureteric bud branches, themesenchyme is induced by the epithelial cells of the uretericbud to form aggregates. These aggregates then undergo amesenchymal to epithelial transition to form an epithelialvesicle. The vesicles go through a series of well-characterizedmorphological stages, which have been termed the comma-shape stage, the S-shape stage, the capillary loop (or cup-shape) stage and the maturing glomerulus stage. At the S-shape stage, the region of the S-shape body destined to formthe distal tubule fuses with the terminal tips of the uretericbud derivatives, which will form the collecting ducts andureters of the kidney; vascularization of the glomerulusalso commences with the migration of endothelial cellsinto the glomerular cleft. These endothelial cells form acapillary loop, which branches to produce the complexcapillary tuft of the mature glomerulus. Mesangial cells alsopopulate the inner part of the glomerulus and maintain thelooped configuration of the capillaries of the glomerularcapillary tuft. Formation of the capillary tuft is accompaniedby differentiation of glomerular epithelial cells intopodocytes.

We report here that Notch2function is required forglomerulogenesis in the kidney, as well as for development ofheart and eye vasculature. We also demonstrate by geneexpression analyses and genetic interaction studies that theNotch ligand encoded by the Jag1gene is required for Notch2signaling in the developing glomerulus.

MATERIALS AND METHODS

Targeting vector constructionA Notch2genomic clone was isolated from a 129/Sv mouse genomicphage library. The genomic organization of a portion of the mouseNotch2 locus was determined by restriction enzyme mapping, blothybridization and nucleotide sequencing. To construct the targetingvector, a 2.8 kb ScaI-XbaI fragment of Notch2was subclonedupstream of a PGK-neo expression cassette (Soriano et al., 1991), anda 3.5 kb HindIII-XbaI Notch2fragment was subcloned downstreamof the PGK-neo cassette. This resulted in the deletion of a 0.4 kbgenomic fragment, deleting codons for 22 amino acids in EGF repeat14 (amino acids 538-560 of the Notch2 protein), the splice donor site,and 0.3 kb of intron sequence. An HSV-tk cassette (Mansour et al.,1988) was introduced for negative selection. We refer to this allele asNotch2del1.

Electroporation, selection and screening of ES cells andmouse genotypingCJ7 embryonic stem (ES) cells were electroporated with 25 µg oflinearized targeting vector, placed under positive-negative selection inG418 and FIAU, and screened for homologous recombination bySouthern blot hybridization as previously described (Swiatek andGridley, 1993). For Southern blot analysis, 10 µg of DNA wasdigested with HindIII, fractionated in 0.8% agarose gels, transferredto Zeta-Probe GT membranes (BioRad) and hybridized with a 1.0 kbClaI-XbaI fragment. ES cells containing the expected recombinationevent were injected into blastocysts from C57BL/6J mice. Malechimeras were bred with C57BL/6J females, and germlinetransmission was obtained for three independently targeted ES cellclones. Animals were genotyped by Southern blot analysis or by PCR.PCR primers for the wild-type Notch2allele were sp3 (5′-CCAG-TGTGCCACAGGTAAGTG-3′), located in the deleted fragment inthe Notch2del1 mutant allele, and sp4 (5′-TCTCCATATTGAT-GAGCCATGC-3′), located in the genomic region 3′ from the deletionin the mutant allele; the primers for the Notch2del1allele were sp4 andsp6 (5′-TTCCTGACTAGGGGAGGAGTAG-3′), located in the neocassette.

RT-PCR analysisTotal RNA was isolated from whole embryos (E10.5). A Notch2-specific primer, N2-3 (bases 2220-2199 of Notch2 cDNA, GenBankAccession Number D32210) and AMV reverse transcriptase(Promega) were used for the reverse transcription reaction. PrimersN2-10 (bases 1337-1357) and N2-2 (bases 2156-2136) were used togenerate the PCR product. The PCR products from the wild-type andmutant RNAs were then sequenced to determine the nucleotidespresent in the alternatively splicedNotch2del1 mRNA. Genomic DNAwas sequenced in the deletion region to identify splice donor andacceptor sites.

Northern blotPoly(A+) RNA was isolated from E14.5 embryos, and 2.5 µg of RNAwas separated on a 1% formaldehyde gel and transferred to a Zeta-Probe GT membrane (Bio-Rad). An 800 bp fragment from the 3′ endof Notch2 cDNA was used for generation of a 32P-dCTP labeledprobe. Hybridization was at 42°C in a buffer containing 50%formamide, 2×SSC, 5×Denhardt’s, 100 µg/ml denatured salmonsperm DNA and 1% SDS.

Histological analysisEmbryos and organs for histological analysis were fixed in Bouin’ssolution. Fixed embryos were dehydrated through graded alcohols,embedded in paraffin wax or methacrylate, sectioned at 7 µm (paraffinwax) or 2 µm (methacrylate), and stained with Hematoxylin and Eosinor Toluidine Blue.

B. McCright and others

493Analysis of a hypomorphic Notch2 mutation

Non-radioactive in situ hybridizationEmbryonic kidneys were dissected and fixed in 4% paraformaldehydein phosphate-buffered saline (PBS) overnight at 4oC, washed withPBS, then equilibrated in PBS containing 30% glycerol. Samples wereembedded in OCT, sectioned on a cryostat at 10-14 µm and mountedon Superfrost Plus (Fisher) slides. Tissue sections were permeabilizedwith 1 µg/ml proteinase K and 0.5% Triton X-100 beforehybridization. 300 ng/ml of digoxigenin-labeled antisense RNA washybridized to slides at 60oC in a buffer containing 5% dextran sulfate,50% formamide, 2×SSC, 5×Denhardt’s, 75 µg/ml yeast tRNA and0.5% SDS. After washing off excess probe, the slides were blockedwith 10% goat serum in 50 mM Tris pH 8.0, 200 mM NaCl and 0.05%Tween 20 (TTBS). Alkaline phosphatase-conjugated anti-digoxigenin(Roche) was diluted 1:500 in 2% goat serum/TTBS, and used to detectthe digoxigenin-labeled antisense RNA. After washing, developingsolution containing NBT and BCIP was added and the slides wereallowed to develop 6-18 hours. Probes used for RNA in situ were Jag1(Mitsiadis et al., 1997); Dll1 (Bettenhausen et al., 1995); Dll3(Dunwoodie et al., 1997); Notch2, bases 6437-7289 of D32210; Bmp7(Dudley et al., 1995);Vegf, nucleotides 517-864 of IMAGE cloneM95200; Wt1, nucleotides 2449-3042 of IMAGE clone M55512;Pdgfrb (Lindahl et al., 1998); Wnt4 (Stark et al., 1994); Pax2 (Dressleret al., 1990).

ImmunohistochemistryKidneys from embryos at E18.5 were dissected, fixed, andcryosectioned as for in situ hybridization. The sections were thenpermeabilized with 0.5% Triton X-100 and endogenous peroxidaseactivity was quenched in 3% H2O2/10% methanol. Blocking was

carried out with 10% goat serum/TTBS. Antibody to mouse plateletendothelial cell-adhesion molecule 1 (PECAM, Pharmingen) wasused in 2% goat serum/TTBS at a dilution of 1:100. A peroxidase-conjugated donkey anti-rat IgG secondary antibody (JacksonImmunoResearch), at a 1:100 dilution in TTBS, was used to detectPECAM and to develop the DiaminoBenzidine (DAB, Sigma)substrate with 0.2% NiCl added. Polyclonal antibody to desmin(Sigma) was used at 1:100 dilution in 2% goat serum/TTBS andwas detected using a peroxidase-conjugated donkey anti-rabbitIgG secondary antibody, at a 1:100 dilution in TTBS (JacksonImmunoResearch).

Terminal transferase-mediated dUTP nick end labeling(TUNEL) assayKidneys from E15.5 embryos were dissected, fixed in 4%paraformaldehyde, embedded in paraffin and sectioned. Cell deathwas detected using the In Situ Cell Death Detection Kit (BoehringerMannheim), according to the manufacturer’s instructions. The averagenumber of TUNEL-positive cells was determined by making 25measurements on sections of both wild-type and homozygous mutantkidneys. The field size used for counting the TUNEL-positive cellswas 0.04 mm2, and measurements were taken for both medullary andcortical regions of the kidneys. Four individual wild-type and fourNotch2del1/Notch2del1 mutant kidneys at E15.5 were used for thisanalysis.

Bromodeoxyuridine (BrdU) labeling of proliferating cellsPregnant mice were injected intraperitoneally with 3 mg of BrdU 2-3 hours before their embryos were removed for processing. Embryo

Fig. 1.Targeted disruption of the Notch2gene.(A) The upper line shows the genomicorganization of a region of the Notch2gene.Exons are indicated by boxes. The middle linerepresents the structure of the targeting vector. A0.4 kb deletion was created that disrupts andremoves most of the exon encoding EGF repeat14. Shown at the bottom is the predicted structureof the Notch2locus following homologousrecombination of the targeting vector. The probeused for Southern blot analysis is indicated.Restriction enzymes: C, ClaI; H, HindIII; S, ScaI;X, XbaI. (B) DNA isolated from embryos of theintercross of Notch2del1/+ heterozygous mice wasdigested with HindIII, blotted and hybridizedwith the indicated probe. Genotypes of progenyare indicated at the tops of the lanes. (C) Northernblot of mRNA isolated from E14.5 embryos. Twotranscripts of approximately 9.0 and 10.5 kb wereobserved for each genotype. (D) RT-PCR analysisof RNA isolated from E10.5 embryos. A sampleof Notch2del1/Notch2del1 RNA subjected to RT-PCR without reverse transcriptase is indicated by–RT. Two alternatively spliced transcripts werepresent in both heterozygous and mutant mRNA.(E) Alternative splicing of the Notch2del1 mutantallele. RT-PCR products were sequenced andcompared with the genomic DNA sequence.Nucleotide sequence analysis revealed that thealternative transcripts are the products of splicingbetween native splice donor and acceptor sites.The intermediate-sized RT-PCR product removesexon C, which results in an in-frame 108 basepair deletion, while the smallest RT-PCR productremoves exons B and C, which results in an in-frame 237 base pair deletion.

494

kidneys were fixed in 4% paraformaldehyde and processed intoparaffin blocks as for histological analysis. 7 µm sections were thenrehydrated and peroxidase activity was quenched in 3% H2O2/10%methanol. Tissue was permeabilized with 20 µg/ml proteinase K andthe DNA was denatured using 2N HCl. An anti-BrdU monoclonalantibody (Pharmingen) was used at 1:100 dilution in 2% goatserum/TTBS. It was detected with a peroxidase-conjugated donkeyanti-mouse IgG secondary antibody at a 1:100 dilution in TTBS(Jackson ImmunoResearch). Diaminobenzidine (DAB, Sigma)substrate with 0.2% NiCl added was used for detection. Quantitationwas performed as described above for detection of TUNEL-positivecells.

Generation and analysis of double heterozygous mice Official nomenclature and references for the mutant alleles used in thesestudies are Jag1dDSL (Xue et al., 1999): Jag1tm1Grid; Notch2del1 (thisreport): Notch2tm1Grid; and Dll1tm1Go (Hrabé de Angelis et al., 1997).Heterozygous Notch2del1 mice were mated with either Jag1dDSL orDll1tm1Go heterozygous mice to generate double heterozygotes.Kidneys were dissected at postnatal day 6 and were processed forhistological analysis as described above. Double heterozygous, singleheterozygous and wild-type kidneys were analyzed.

RESULTS

Targeted disruption of the Notch2 geneTo analyze the in vivo role of the Notch2 gene, a targetingvector was constructed that deleted 0.4 kb of genomic sequenceof the Notch2gene (Fig. 1A). The deleted fragment interruptsthe exon encoding EGF repeat 14 in the extracellular domainof the protein, deleting codons for 22 amino acids in EGF14,the splice donor site, and 0.3 kb of intron. We refer to this alleleas Notch2del1. The linearized targeting vector waselectroporated into ES cells and germline transmission of theNotch2del1 mutant allele was obtained for three independentlytargeted clones (Fig. 1B). Mice heterozygous for the Notch2del1

mutant allele appeared normal and were fertile.

Notch2 del1 is a hypomorphic allele of the Notch2geneHamada et al. recently described construction of a targetedmutation of the Notch2gene (Hamada et al., 1999). Theyobserved completely penetrant embryonic lethality by E11.5 in


Fig. 2.Defects in glomerulogenesis inNotch2del1/Notch2del1 homozygotes.(A) Kidney and adrenal gland isolatedfrom Notch2del1/Notch2del1

homozygous mutant (right) and wild-type littermate control (left) at P0.Mutant kidneys are hypoplastic andexhibit vascular lesions near thecortical surface. Mutant adrenalglands appear unaffected.(B-K) Histological analysis ofNotch2del1/Notch2del1 homozygousmutant (C,E,G,H,I,K) and wild-typelittermate (B,D,F,J) kidneys.(B,C) Kidneys at E13.5. Both wildtype (B) and Notch2del1/Notch2del1

mutant (C) kidneys exhibit uretericbud (ub) branching and formation ofmetanephric vesicles (arrows).(D-I) Kidneys at E16.5.(D,E) Numerous morphologicallynormal glomeruli (arrows) are presentin the wild-type kidney (D), while inthe Notch2del1/Notch2del1 kidney (E)no mature glomeruli are present.Ureteric bud branching, mesenchymalcondensations and collecting ducts arepresent in both the mutant and wildtype kidneys. (F) Wild-typeglomerulus. (G-I) Examples ofNotch2del1/Notch2del1 mutantglomeruli. (G,H) These examplesrepresent the end stage of the majorityof glomeruli in Notch2del1/Notch2del1

mutant kidneys. The wild-typeglomerulus (F) has entered thecapillary loop stage and has becomevascularized, but the mutant glomerulihave not and are instead made up of a disorganized clump of cells. (I) In a subset of abnormal glomeruli in the Notch2del1/Notch2del1 mutants,the glomerular capillary tuft was replaced by a capillary aneurysm-like structure that filled the region of Bowman’s capsule with red bloodcells. (J,K) Kidneys at P0. Methacrylate sections of wild type (J) and Notch2del1/Notch2del1 mutant (K) kidneys. Glomeruli (arrowheads) ofwild-type kidneys have morphologically normal capillary tufts, while some glomeruli of mutant kidneys exhibit capillary aneurysms. Scalebars: 100 µm in D,E; 25 µm in F-K.


mouse embryos homozygous for their targeted allele. However,when we set up timed matings of Notch2del1 heterozygousmice, we observed survival of Notch2del1/Notch2del1

homozygous mutant embryos to gestational ages later thanE11.5 (see below). We therefore examined whether anypotentially functional Notch2transcripts could be transcribedfrom the Notch2del1 mutant allele. Northern blot hybridizationusing a probe from the 3′ end of the Notch2 cDNA identifiedsimilar transcripts present in homozygous Notch2del1/Notch2del1

embryos and in wild-type embryos (Fig. 1C). In addition,in situ hybridization analyses of homozygous Notch2del1/Notch2del1 embryos using the same 3′ Notch2 probedemonstrated the presence of Notch2transcripts in the mutantembryos (data not shown). Since transcription termination andpolyadenylation signals are present in the neo selection cassettecontained in the Notch2del1-targeted allele, the presence ofNotch2 transcripts containing sequences 3′of the targeted

region suggested that there may be splicing around the neocassette in the Notch2del1 allele. Therefore, we performed aseries of RT-PCR analyses to determine the nature of theNotch2del1 transcript (Fig. 1D). RT-PCR analysis of Notch2del1

RNA and subsequent DNA sequencing of the RT-PCRproducts identified two alternative transcripts that are the resultof in-frame splicing around both the deleted region of theNotch2del1 allele and the neo cassette (Fig. 1D,E). Thealternatively spliced transcripts should result in the deletion ofeither one or two EGF repeats from the Notch2 protein. Theshorter protein isoform joins the first half of EGF12 to thesecond half of EGF14, while the longer isoform joins the firsthalf of EGF13 to the second half of EGF14 (Fig. 1E). Thelevels of these alternative transcripts are similar to the amountof wild-type transcript. These data demonstrate that theNotch2del1 mutant allele is not a null allele, and most likely isa hypomorphic allele of the Notch2gene.

Defects in glomerulogenesis inNotch2 del1/Notch2 del1 miceTo examine whether mice homozygous for the Notch2del1

mutation were viable, heterozygous F1 animals wereintercrossed. We observed that some of the progeny ofintercrosses of Notch2del1 heterozygous mice died within thefirst 24 hours of birth. Genotypic analyses revealed that thesedying neonatal animals were homozygous for the Notch2del1

mutant allele. Gross anatomical analysis of the Notch2del1/Notch2del1 neonates revealed that their kidneys were

Fig. 3. Expression of the Notch2gene and genes encoding Notchfamily ligands during metanephric kidney development. (A,B)Notch2expression is observed in comma-shape bodies and tubules (t). Inmore mature glomeruli, expression is concentrated around the outerlayer of cells where podocytes are differentiating (arrows).(C,D) E16.5 kidney showing Dll1 expression in comma-shape bodies(arrows) and tubules (t) but not in vascularized glomeruli. (E,F)Jag1is expressed in the early glomerular structures (arrows) and incollecting ducts of E16.5 kidneys. (G,H) Comparison of Jag1expression (G) and Notch2expression (H) at E18.5 in glomeruliwhich are at the beginning of vascularization at the late S- to capillaryloop stage. The boundaries of the glomerulus are indicated by thebroken line. Note that Jag1expression is centered in the middle of theglomerulus and Notch2expression is at the periphery.

Fig. 4.Analysis of molecular markers for metanephric mesenchymeand pretubular aggregates. (A,B)Pax2is expressed in the same patternin both wild type (A) and Notch2del1/Notch2del1homozygous mutant(B) kidneys at E16.5. Pax2expression is present in the condensedmesenchyme and comma shape bodies. (C,D) The Wnt4expressionpattern is similar in both wild type (C) and mutant (D). Both expressWnt4in the epithelial vesicles and comma shape bodies. (E,F)Bmp7isexpressed in metanephric mesenchyme and in condensing aggregatesin both wild-type (E) and mutant (F) kidneys at E16.5.

496

hypoplastic and exhibited vascular lesions near the corticalsurface (Fig. 2A).

Histological analysis revealed defects in glomerulogenesis inthe Notch2del1/Notch2del1kidneys. At E13.5, mutant and controlkidneys were similar in size, and both exhibited ureteric budgrowth and branching (Fig. 2B,C). In both mutant and controlkidneys, at the tips of the bud branches the metanephricmesenchyme condensed and formed epithelial vesicles.However, by E16.5, kidneys from the Notch2del1/Notch2del1

embryos were smaller than those of wild-type and heterozygouslittermate controls, and while ureteric bud branches andmetanephric vesicles appeared morphologically normal in themutants, no morphologically normal glomeruli could beobserved in the Notch2del1/Notch2del1 homozygous mutantkidneys (Fig. 2D-K).

Two types of abnormal glomeruli were observed in themutant kidneys. Most commonly, differentiation of the mutantglomeruli appeared to arrest just before the capillary loop stage.In these mutant glomeruli, the capillary tuft of the normalmature glomerulus was absent and the glomerulus appeared tobe a disorganized clump of cells (Fig. 2G,H). In the second typeof abnormal mutant glomeruli, the glomerular capillary tuft wasreplaced by a capillary aneurysm-like structure that filled theregion of Bowman’s capsule with red blood cells (Fig. 2I,K).Similar structures have been observed in mice mutant for theplatelet-derived growth factor gene (Pdgfb; Levéen et al., 1994;Lindahl et al., 1998) or the PDGFβ receptor gene (Pdgfrb;Soriano, 1994). It has been demonstrated for the Pdgfb andPdgfrbmutants that these glomerular capillary aneurysms arisedue to defects in mesangial cell differentiation.

Expression of Notch2 and of genes encoding Notchligands during kidney developmentThe Notch2gene is expressed widely in mouse embryos andadults (Weinmaster et al., 1992; Lardelli and Lendahl, 1993;Swiatek et al., 1994; Williams et al., 1995; Lindsell et al., 1996).To study the Notch2del1/Notch2del1 kidney phenotype in moredetail, the expression patterns of Notch2and the genes encodingits potential ligands were assayed by RNA in situ hybridization.Since metanephric kidney development does not proceedsynchronously, examination of kidneys during lateembryogenesis permits examination of multiple developmentalstages (Saxén, 1987). Notch2is expressed in both the uretericbud-derived collecting duct system, and in condensations andvesicular structures (such as metanephric vesicles and comma-and S-shape bodies) derived from the metanephric mesenchyme(Fig. 3A). In maturing glomeruli Notch2is expressed primarilyat the periphery, where glomerular epithelial cells aredifferentiating into podocytes (Fig. 3A,B,H).

Several genes encoding Notch ligands were also expressed inthe developing kidney. TheDll1 gene was expressed in tubulesand in comma- and S-shape bodies at E16.5 (Fig. 3C,D; see alsoBeckers et al., 1999). The Jag1gene was expressed in themesenchymal condensations and vesicular structures of thedeveloping glomeruli (Fig. 3E-G). Jag1expression wasprimarily observed in the inner region of the developingglomerulus, in a position consistent with expression indifferentiating mesangial and/or endothelial cells (Fig. 3G). TheJag1 and Notch2genes appeared to be expressed in adjacentcells in the developing glomerulus (Fig. 3G,H). Jag1expressionwas also detected in the collecting ducts. The Dll4gene, which

encodes a recently identified Notch ligand, is expressed inendothelial cells of the glomerulus (Shutter et al., 2000). TheJag2 gene was expressed at low levels (much lower than theexpression levels of Dll1and Jag1) throughout the kidney atE16.5, and we did not detect expression of the Dll3 gene in thekidney (data not shown).

Notch2 is not required for mesenchymalcondensation and epithelialization, but is essentialfor glomerular differentiation and patterning To define the Notch2del1/Notch2del1kidney phenotype in furtherdetail, and to determine at which step in development themutants deviate from the normal pathway, we analyzedexpression of a panel of molecular markers. Both the Pax2(Dressler et al., 1990) and Wnt4(Stark et al., 1994) genes areexpressed in the condensing mesenchyme and pretubularaggregates of the developing metanephric kidney. By in situanalysis, both Pax2(Fig. 4A,B) and Wnt4 (Fig. 4C,D) wereexpressed normally in these structures in Notch2del1/Notch2del1

kidneys. The Bmp7 gene is expressed early duringnephrogenesis in the ureteric bud, and expression issubsequently observed in the metanephric mesenchyme and inthe pretubular aggregates (Dudley et al., 1995; Luo et al., 1995).In Notch2del1/Notch2del1 kidneys, Bmp7 expression in themetanephric mesenchyme and the condensing aggregates wasnormal (Fig. 4E,F). Thus, the early stages of metanephrickidney development, including ureteric bud growth andbranching and mesenchymal condensation and vesicleformation, occurred relatively normally in the Notch2del1/Notch2del1 mutant kidneys. However, there appeared to be aquantitative difference in these processes in the Notch2del1/Notch2del1mutants, in that far fewer condensations and vesiclesformed in the mutants than in wild-type and heterozygouslittermate controls.

We assessed differentiation and patterning of the glomerulususing molecular markers for the three primary cell types of themature glomerulus: endothelial cells, mesangial cells andpodocytes. Endothelial cell differentiation was assayed using amonoclonal antibody to PECAM1 (platelet-endothelial celladhesion molecule 1) (Baldwin et al., 1994). This analysisrevealed that few endothelial cells were present in the abnormalmutant glomeruli (Fig. 5A,B). However, some endothelial cellscould be observed migrating into the cleft of the comma- andS-shape bodies in the mutants (Fig. 5B). The migration ofendothelial cells into the cleft of the comma- and S-shapebodies is consistent with the development of capillaryaneurysm-like structures in a subset of mutant glomeruli (Fig.2I,K).

Mesangial cell differentiation was assayed using an antibodyto desmin (Lindahl et al., 1998; Miner and Li, 2000) and an insitu hybridization probe for Pdgfrb(Seifert et al., 1998; Lindahlet al., 1998). This analysis showed there were no glomeruli thatcontained desmin-positive cells in the mutant kidneys (Fig.5C,D). Pdgfrbexpression was also not observed in the mutantglomeruli (Fig. 5F). In the wild-type and heterozygous controlkidneys, Pdgfrb expression was present in the center of cup-staged glomeruli, consistent with its expression in mesangialcells (Fig. 5E). These data reveal the absence of mesangial cellsin the abnormal glomeruli of the Notch2del1/Notch2del1mutants.

Podocyte differentiation was assayed using in situhybridization probes for the vascular endothelial growth factor



A (Vegf) gene (Breier et al., 1992; Kitamoto et al., 1997; Tufró,2000) and the Wilms’ Tumour (Wt1) gene (Kreidberg et al.,1993). In wild-type and heterozygous littermate controlkidneys, both genes were expressed in the cup-shapedsemicircle of glomerular epithelial cells that are differentiatinginto podocytes (Fig. 5G,I). In Notch2del1/Notch2del1

homozygous mutant kidneys, Vegf- and Wt1-positive cells werepresent in the abnormal mutant glomeruli, but the cellsexpressing these markers were clumped together in the centerof the glomerulus, and did not form the cup-shaped epitheliallayer observed in the controls (Fig. 5H,J).

Analysis of cell death and proliferation inNotch2 del1/Notch2 del1 kidneysSince by the later stages of embryogenesis the kidneys ofNotch2del1/Notch2del1 mutant embryos were significantlysmaller than kidneys of control littermate embryos, we analyzedcell proliferation and cell death in developing kidneys. Cellproliferation was analyzed by BrdU labeling of kidneys isolatedat E15.5. Cell proliferation was approximately equal in bothwild-type and mutant kidneys in the highly proliferative cortexof the kidney (Fig. 6A,B). There were 51.1±5.8 (mean±s.d.)BrdU-positive cells/0.04 mm2 in wild-type sections comparedwith 47.2±8.1 BrdU-positive cells/0.04 mm2 in mutant sectionsin this region. However, while proliferating cells were observedinside the morphologically normal glomeruli of wild-type andheterozygous littermate controls, no proliferating cells wereobserved inside the abnormal glomeruli that were present in theNotch2del1/Notch2del1 mutant kidneys.

Apoptotic cell death was assessed by TUNEL assay. Wedetected an increased amount of cell death in Notch2del1/Notch2del1 mutant kidneys (Fig. 6C,D). There were 9.1±3.4TUNEL-positive cells/0.04 mm2 in wild-type sectionscompared with 17.0±4.8 TUNEL-positive cells/0.04 mm2 inmutant sections. Higher numbers of TUNEL-positive cells weredetected at the periphery of the mutant kidneys, immediatelybeneath the renal capsule. These results suggest that increasedcell death may contribute to the growth retardation observed inthe Notch2del1/Notch2del1 mutant kidneys.

Notch2 del1/Jag1dDSL double heterozygous miceexhibit glomerular defectsPhenotypes sensitive to gene dose are a common feature ofNotch pathway mutants in both vertebrates and invertebrates(Artavanis-Tsakonas et al., 1991; Li et al., 1997; Oda et al.,1997; Xue et al., 1999; Krebs et al., 2000). Since the Notch2and Jag1 genes were expressed in adjacent regions of thedeveloping glomerulus, we examined kidneys from miceheterozygous for both the Notch2del1 and Jag1dDSL (Xue et al.,1999) mutant alleles. These double heterozygous miceexhibited kidney defects (Fig. 7). The kidneys of the doubleheterozygotes (n=6) were about half the size of kidneys fromthe littermate controls (which consisted of wild-type mice, andNotch2del1 and Jag1dDSL single heterozygotes) (Fig. 7A). Thenumber of glomeruli was greatly reduced in the doubleheterozygotes, and about a quarter of the glomeruli presentlacked glomerular capillary tufts and exhibited the capillaryaneuryisms similar to those observed in Notch2del1/Notch2del1

homozygous mutant kidneys (compare Fig. 7C with Fig. 2I,K).These results demonstrate that Notch2del1/+ Jag1dDSL/+ doubleheterozygotes have a glomerular defect similar to, but less

severe than, those of the Notch2del1/Notch2del1 homozygousmutants.

Since the Dll1 gene is expressed in the developingglomerulus, we also generated mice heterozygous for both theNotch2del1and Dll1tm1Go(Hrabé de Angelis et al., 1997) mutantalleles. However, we observed no kidney defects in Notch2del1/Dll1tm1Godouble heterozygous mice (n=3).

Mice homozygous for the Notch2 del1 mutation exhibitdefects in development of the hyaloid vasculature ofthe eyeIn addition to hypoplastic kidneys, other defects were observedin Notch2del1/Notch2del1 mutant embryos and neonates. Grossanatomical analysis of Notch2del1/Notch2del1 neonates revealedthat all mutants displayed bilateral microphthalmia (Fig. 8A).Histological analysis of the eyes of Notch2del1/Notch2del1

neonates typically revealed an aberrant bulbous structure at theterminus of the hyaloid artery (Fig. 8B). Numerous smallcapillaries were observed emanating from this structure (Fig.8C). At earlier embryonic stages, a pronounced retrolenticularhyperplasia was evident in sections of eyes from theNotch2del1/Notch2del1 embryos (Fig. 8D,E). Eyes of theNotch2del1/Notch2del1 embryos exhibited a pronouncedasymmetry (Fig. 8F,G), which was apparently caused by theretrolenticular hyperplasia.

Heart defects in Notch2 del1/Notch2 del1 embryosQuantitative analysis of intercross progeny from Notch2del1

heterozygous mice revealed that, at gestational stages later thanE16.5, only 12% of the progeny were Notch2del1/Notch2del1

homozygous mutants (Table 1). We therefore isolated mutantembryos at earlier stages in order to determine the cause ofembryonic lethality prior to E16.5. Through E10.5 we detectedno obvious mutant phenotype in Notch2del1/Notch2del1

embryos. However, at E11.5, approximately 40% of thehomozygous mutant embryos exhibited growth retardation,pericardial effusion and widespread hemorrhaging (Fig. 9A andTable 1). Mutant embryos that survived this stage displayed avariety of vascular defects, including myocardial hypoplasia,hemorrhaging, and edema (Fig. 9B-E, and Table 1). Beginningat E13.5, edema and hemorrhaging vessels near the surface ofthe skin were observed in approximately 50% of Notch2del1/Notch2del1 homozygous mutant embryos (Fig. 9C, and data notshown). Myocardial hypoplasia and reduced myocardialtrabeculation were observed in all mutant embryos E12.5 andolder (Fig. 9D,E). We analyzed Notch2expression in the heartby in situ hybridization to determine whether the observedphenotype, particularly the myocardial hypoplasia, correlatedwith a domain of Notch2expression. We detected no Notch2message in the heart at either E11.5 or E14.5. However, we diddetect Notch2expression in the outflow tract of the heart at thesestages (data not shown).

DISCUSSION

The Notch2 del1 allele causes a probablehypomorphic mutation due to alternative splicingThe Notch2del1allele described here is a probable hypomorphicallele, owing to unexpected in-frame alternative splicing. Thisresults in the production of novel Notch2 mRNAs that can

498

encode Notch2 proteins that may be partially functional. Thenovel mRNAs have excised the coding sequence for either oneor two of the 36 EGF repeats present in the Notch2 protein.The shorter protein isoform joins the first half of EGF12 to thesecond half of EGF14, while the longer isoform joins the first

half of EGF13 to the second half of EGF14. These excisedEGF repeats may hinder Notch2 ligand-binding ability owingto their proximity to the EGF11 and EGF12 repeats, which areessential for ligand binding by the Drosophila Notch and


Fig. 5. Analysis of molecular markers for glomerular differentiationand patterning. (A,B) PECAM1 immunohistochemistry of E18.5wild-type (A) and Notch2del1/Notch2del1 mutant (B) kidneys. SomePECAM1-positive endothelial cells can be seen migrating into theglomerular cleft in both wild-type and mutant S-shape bodies.(C,D) Desmin immunohistochemistry of E18.5 wild-type (C) andmutant (D) kidneys. Wild-type glomeruli contain desmin-expressingmesangial cells (arrows) while no desmin-positive cells wereexpressed in mutant glomeruli. (E,F)PdgfrbRNA expression inE16.5 kidneys. In the wild-type kidney, as the glomerulus reaches thecapillary loop stage, the center becomes populated with Pdgfrbexpressing mesangial cells (arrows). In the mutant, Pdgfrbexpressing cells remain on the outside of the glomerulus. Boundariesof the glomerulus are indicated by broken lines. (G,H) VegfRNAexpression in E16.5 kidneys. In the wild-type kidney, Vegf-expressingpodocytes formed a cup-shaped semicircle (arrow in G). In themutant, kidney, Vegf-expressing cells were clustered (arrow in H).(I,J) Wt1RNA expression in E16.5 kidneys. The Wt1expressionpattern resembled the Vegfexpression pattern in both wild-type (I)and mutant (J) kidneys.

Fig. 6.Cell proliferation and cell death in Notch2del1/Notch2del1

kidneys. (A,B) Cell proliferation was assayed by detectingincorporation of BrdU into newly synthesized DNA. Proliferatingcells have brown nuclei. Large numbers of proliferating cells weredetected in both wild-type (A) and Notch2del1/Notch2del1 mutant (B)kidneys at E15.5. (C,D) Cell death was assayed by TUNEL assay toidentify apoptotic cells in E15.5 kidneys. TUNEL-positive cells havebrown nuclei. In wild-type kidneys (C), few TUNEL-positive cellswere detected (arrows). In mutant kidneys (D), large numbers ofTUNEL-positive cells (arrows) were detected. TUNEL-positive cellswere present in higher numbers at the periphery of the kidney,immediately beneath the renal capsule.

Fig. 7.Kidney defects in Notch2del1/+ Jag1dDSL/+ doubleheterozygotes. Kidneys were isolated at P6 from wild-type andNotch2del1/+ Jag1dDSL/+ double heterozygous mice. (A) The doubleheterozygous kidney (right) is smaller than the wild-type littermatekidney (left). While the control kidney glomeruli exhibit a normalcapillary tuft (arrow in B), some double heterozygous glomeruli(arrows in C) exhibit capillary aneurysm-like structures similar tothose observed in Notch2del1/Notch2del1 homozygotes.


XenopusNotch1 proteins (Rebay et al., 1991). We hypothesizethat the phenotypes we observe in Notch2del1/Notch2del1

mutants are most likely caused by a decrease in Notch2signaling. An alternative possibility is that the phenotypes weobserve in Notch2del1/Notch2del1mutants are due to deleteriousindirect effects such as ligand sequestering. However, the factthat Notch2del1/+ heterozygotes have no visible phenotypeargues against this hypothesis.

Notch2 is required for glomerular morphogenesisand patterningMice homozygous for the Notch2del1 mutation exhibit acomplex phenotype, with at least two stages of lethality. AllNotch2del1/Notch2del1 homozygous mutants that survive untilbirth die within the first 24 hours. These homozygotes probablydie from renal insufficiency due to defects in glomerulardifferentiation and patterning, although heart defects (i.e.myocardial hypoplasia) may also contribute to the lethalityobserved in the Notch2del1/Notch2del1 neonates.

In the majority of abnormal glomeruli in the Notch2del1/Notch2del1 mutants, neither mesangial cells nor endothelialcells could be detected using molecular markers. Some mutant

Fig. 8.Microphthalmia and eye vasculature defects inNotch2del1/Notch2del1 homozygotes. (A) Wild-type (left) andNotch2del1/Notch2del1 littermate (right) eyes at P0. The mutant eye ismicrophthalmic. (B-E) Histological analysis of Notch2del1/Notch2del1

(B,C,E) and wild-type (D) eyes. (B,C) At E18.5, mutant eyes exhibitan aberrant bulbous structure (arrow) just dorsal to the lens. Thisstructure, which has numerous small capillaries emanating from it, isattached to the hyaloid vessel (arrow in C). (D,E) At E16.5, eyes ofNotch2del1/Notch2del1 embryos (E) exhibit a pronouncedretrolenticular hyperplasia (arrow) compared with wild type (D).(F,G) Wild-type (F) and Notch2del1/Notch2del1 (G) embryos at E17.5.The mutant embryo exhibits an asymmetry in the eye. Note theconstriction at the dorsal side of the eye (arrow) of the mutantembryo.

Table 1. Genotypes and phenotypes of Notch2del1 mutantembryos at various gestational ages

+/+ +/− −/−E10.5 32 50 27*E11.5 48 113 41‡E12.5-E15.5 88 159 37§ (22)E16.5-birth 61 96 21¶

+/+, wild type; +/−, +/Notch2del1 heterozygotes; −/−, Notch2del1/Notch2del1

mutants.*E10.5 Notch2del1/Notch2del1 mutant embryos were all normal in

appearance. ‡18 out of 41 E11.5 Notch2del1/Notch2del1 embryos exhibited growth arrest

and hemorrhaging in the heart sac. §13 out of 37 E12.5 to E15.5 Notch2del1/Notch2del1 embryos exhibited

edema and 22 necrotic embryos were also present (in parentheses). ¶Five out of 21 of the E16.5 and older Notch2del1/Notch2del1 embryos had

widespread hemorrhaging and one exhibited edema. Eye and kidney phenotypes were 100% penetrant.

Fig. 9. Heart and vascular defects inNotch2del1/Notch2del1 mutantembryos. (A) E11.5 Notch2del1/Notch2del1 homozygote (right) andwild-type littermate (left). The mutant is growth retarded andexhibits pericardial edema with hemorrhaging. (B,C) E14.5 embryos.The Notch2del1/Notch2del1 mutant (C) exhibits edema along the backof the embryo (arrows), indicating cardiovascular dysfunction. (D,E)Transverse sections of embryos at E12.5. The mutant heart (E)exhibits a thinning of the myocardial wall (arrow) and reducedmyocardial trabeculation.

500

glomeruli appeared to consist of a disordered group ofpodocytes, as the clustered cells expressed the podocytemarkers Vegfand Wt1. However, the mutant podocytes had notformed the cup-shaped epithelial layer observed in the controllittermate glomeruli. This phenotype is reminiscent of thedefects in development of epithelial tissues in Drosophilaembryos mutant for Notchor for other components of theNotch signaling pathway (Hartenstein et al., 1992; Goode etal., 1996). In the second type of abnormal glomeruli observedin Notch2del1/Notch2del1 mutants, the glomerular capillary tuftwas replaced by a capillary aneurysm-like structure similar tothat observed in mice mutant for the Pdgfbgene (Levéen et al.,1994; Lindahl et al., 1998) or the Pdgfrbgene (Soriano, 1994).It has been demonstrated for the Pdgfband Pdgfrbmutants thatthese glomerular capillary aneurysms arise because of defectsin mesangial cell differentiation.

It is important to point out that there is also a substantialquantitative effect on glomerular differentiation in Notch2del1/Notch2del1mutants. The total number of glomeruli is markedlyreduced in the mutants, compared with heterozygous and wild-type littermates. This suggests that Notch2signaling may berequired at multiple stages during glomerular differentiation.Notch2-mediated signaling may be required for cellulardifferentiation of the mesenchymal vesicles, as well as forproviding the proper signals for migration of endothelial andmesangial precursors into the glomerular cleft, and forvascularization and epithelialization of the developingglomerulus.

The Jag1 protein is a probable ligand for the Notch2receptor during glomerulogenesisExpression analysis indicated that the Notch2and Jag1geneswere expressed in adjacent cells during glomerulardifferentiation. Notch2was expressed at the periphery of theglomerulus in differentiating podocytes, while Jag1wasexpressed in the interior of the glomerulus, in a positionconsistent with expression in differentiating mesangial and/orendothelial cells. In addition, Notch2del1/+ Jag1dDSL/+ doubleheterozygotes displayed a kidney phenotype similar to,although somewhat less severe than, that of Notch2del1/Notch2del1 homozygotes. These data indicate that the Jag1gene encodes a physiological ligand for the Notch2 receptorduring glomerulogenesis. In support of this model, in vitro cellculture studies have confirmed that the Jag1 protein can act asa signaling ligand for the Notch2 protein, although eitherprotein may bind to other ligands or receptors as well (Nofzigeret al., 1999; Shimizu et al., 1999; Shimizu et al., 2000).

Heart and vascular defects in Notch2 del1/Notch2 del1

miceThe second stage of lethality observed in Notch2del1/Notch2del1

embryos was from E11 through E16. Abnormal mutantembryos at these stages exhibited symptoms of cardiovasculardysfunction. These embryos were edematous, and histologicalanalysis revealed myocardial hypoplasia and reducedmyocardial trabeculation. Many of these embryos alsoexhibited hemorrhages of small blood vessels in the skin.

We did not detect Notch2RNA in the embryonic heart, andit has been reported by others that Notch2is not expressed inthe heart (Weinmaster et. al., 1992; Hamada et al., 1999). Wedid observe Notch2expression in the cardiac outflow tract.

Jag1 expression is also observed in the cardiac outflow tract(Loomes et al., 1999), and various heart defects are observedin humans with Alagille syndrome, an inherited diseasesyndrome caused by Jag1haploinsufficiency (Li et al., 1997;Oda et al., 1997; Krantz et al., 1999).

It is intriguing that several phenotypes observed inNotch2del1/Notch2del1 homozygotes are similar to phenotypesobserved in mice homozygous for targeted mutations of Pdgfb(Levéen et al., 1994; Lindahl et al., 1998) and Pdgfrb(Soriano,1994). These shared phenotypes include formation ofglomerular capillary aneurysms, myocardial hypoplasia andhemorrhages of blood vessels in the skin. Since Pdgfb andPdgfrb are not expressed in the same cells that express theNotch2gene during glomerulogenesis, any interaction betweenthese two signaling pathways is probably indirect, at least withrespect to glomerular differentiation. Future studies willexamine in more detail possible interactions between thesesignaling pathways.

Defects in the eye vasculature ofNotch2 del1/Notch2 del1 homozygotesThe hyaloid artery forms part of the hyaloid vascular system,a transient network of intraocular blood vessels that nourish thegrowing lens during embryogenesis (Jack, 1972; Balazs et al.,1980). During late embryonic and early postnatal stages inwild-type mice, the hyaloid vessel system atrophies. Theretrolenticular hyperplasia present in the eyes of Notch2del1/Notch2del1 embryos resembles a congenital human diseasesyndrome termed persistent hyperplastic primary vitreous(PHPV) (Reese, 1955; Haddad et al., 1978). In individuals withPHPV, the hyaloid vascular system does not atrophy but insteadpersists postnatally. This can lead to retinal detachment,cataracts, glaucoma and degeneration of the eye. While mosthuman cases of PHPV are unilateral and show no obvioushereditary predisposition, approx. 10% are bilateral, and somecases of familial inheritance have been reported (cited byHaddad et al., 1978; Stades, 1983). Moreover, a genetic basisfor the development of PHPV is supported by the frequentoccurrence of PHPV in two dog breeds, the Doberman pinscher(Stades, 1983; Boevé et al., 1988) and the Staffordshire bullterrier (Curtis et al., 1984; Leon et al., 1986). However, thegenetic loci causing predisposition for development of PHPVhas not been identified in either dog model of the disease.Notch2del1/Notch2del1 mutant mice may provide a useful,genetically defined model for study of the retrolenticularhyperplasia associated with the early embryonic stages ofPHPV.

What is the Notch2 null phenotype?It is clear that the phenotype we observe inNotch2del1/Notch2del1 homozygous mutants does not representthe Notch2 null phenotype. Hamada et al. constructed andanalyzed a Notch2mutant allele (Notch2lacZ) in which the E.coli β-galactosidase-coding region replaces all but one of theankyrin repeats in the intracellular domain of the mouseNotch2 protein (Hamada et al., 1999). Embryos homozygousfor the Notch2lacZ mutant allele exhibit a more severephenotype than Notch2del1/Notch2del1 homozygotes. AllNotch2lacZ/Notch2lacZ homozygous mutant embryos die before11.5 days of gestation (Hamada et al., 1999). The design of theNotch2lacZ mutant allele is similar to the design of constructs



that exhibit dominant-negative effects when expressed inDrosophila. Rebay et al. found that dominant negativephenotypes resulted from overexpression of a Notch proteinlacking most intracellular sequences (Rebay et al., 1993).Hamada et al., argue that they are not observing dominantnegative phenotypes in Notch2lacZ/Notch2lacZ homozygotes,since Notch2lacZ/Notch2lacZ embryos arrested at a slightly laterstage than Notch1-/- mutant embryos (Swiatek et al., 1994;Conlon et al., 1995), and no abnormalities were observed inNotch2lacZ/+ heterozygotes. However, it is not clear whetherthe phenotype of Notch2lacZ/Notch2lacZ homozygotesrepresents the Notch2null phenotype. While theNotch2lacZ

mutant allele probably prevents Notch2-mediated signaling,dominant negative effects of this allele (such as sequestrationof Notch family ligands) could contribute to the phenotypeobserved in the Notch2lacZ/Notch2lacZ homozygotes. Onlycreation of another Notch2mutant allele that totally abrogatesproduction of Notch2 protein will answer the question of whatthe Notch2null phenotype really looks like.

We thank C. Betzsholtz, A. McMahon, G. Dressler and E.Robertson for probes, and Tim O’Brien and Sue Ackerman forcomments on the manuscript. This work was supported by grants toT. G. from the NIH (NS36437) and the March of Dimes Birth DefectsFoundation. B. M. was supported by a postdoctoral fellowship fromNICHD. Y. L. was supported by a Training Grant from the NationalCancer Institute to the Jackson Laboratory. This work was alsosupported by a Cancer Center Core grant (CA34196) from theNational Cancer Institute to the Jackson Laboratory.

REFERENCES

Artavanis-Tsakonas, S., Delidakis, C. and Fehon, R. G.(1991). The Notchlocus and the cell biology of neuroblast segregation. Annu. Rev. Cell Biol.7, 427-452.

Artavanis-Tsakonas, S., Rand, M. D. and Lake, R. J.(1999). Notchsignaling: cell fate and signal integration in development. Science284, 770-776.

Balazs, E. A., Toth, L. Z. and Ozanics, V.(1980). Cytological studies on thedeveloping vitreous as related to the hyaloid vessel system. Graefes Archiv.Klin. Exp. Ophthal. 213, 71-85.

Baldwin, H. S., Shen, H. M., Yan, H. C., DeLisser, H. M., Chung, A.,Mickanin, C., Trask, T., Kirschbaum, N. E., Newman, P. J., Albelda, S.M. and Buck, C. A. (1994). Platelet endothelial cell adhesion molecule-1(PECAM-1/CD31): alternatively spliced, functionally distinct isoformsexpressed during mammalian cardiovascular development. Development120, 2539-2553.

Beckers, J., Clark, A., Wunnsch, K., Hrabe De Angelis, M. and Gossler,A. (1999). Expression of the mouse Delta1 gene during organogenesis andfetal development. Mech. Dev. 84, 165-168.

Bettenhausen, B., Hrabé de Angelis, M., Simon, D., Guenet, J.-L. andGossler, A. (1995). Transient and restricted expression during mouseembryogenesis of Dll1, a murine gene closely related to Drosophila Delta.Development121, 2407-2418.

Boevé, M. H., van der Linde-Sipman, J. S. and Stades, F. C.(1988). Earlymorphogenesis of Persistent Hyperplastic Tunica Vasculosa Lentis andPrimary Vitreous: the dog as an ontogenetic model. Invest. Ophthalmol. Vis.Sci. 29, 1076-1086.

Breier, G., Albrecht, U., Sterrer, S. and Risau, W.(1992). Expression ofvascular endothelial growth factor during embryonic angiogenesis andendothelial cell differentation. Development114, 521-532.

Bulman, M. P., Kusumi, K., Frayling, T. M., McKeown, C., Garrett, C.,Lander, E. S., Krumlauf, R., Hattersley, A. T., Ellard, S. and Turnpenny,P. D. (2000). Mutations in the human delta homologue, DLL3, cause axialskeletal defects in spondylocostal dysostosis. Nat. Genet. 24, 438-441.

Clark, A. T. and Bertram, J. F. (1999). Molecular regulation of nephronendowment. Am. J. Physiol.276, F485-F497.

Conlon, R. A., Reaume, A. G. and Rossant, J.(1995). Notch1is requiredfor the coordinate segmentation of somites. Development121, 1533-1545.

Curtis, R., Barnett, K. C. and Leon, A. (1984). Persistent hyperplasticprimary vitreous in the Staffordshire bull terrier. Vet. Rec. 115, 385-394.

Davies, J. A. and Bard, J. B.(1998). The development of the kidney. Curr.Top. Dev. Biol. 39, 245-301.

Dressler, G. R., Deutsch, U., Chowdhury, K., Nornes, H. O. and Gruss, P.(1990). Pax2, a new murine paired-box containing gene and its expressionin the developing excretory system. Development109, 787-795.

Dudley, A. T., Lyons, K. M. and Robertson, E. J.(1995). A requirement forbone morphogenetic protein-7 during development of the mammaliankidney and eye. Genes Dev. 9, 2795-2807.

Dunwoodie, S. L., Henrique, D., Harrison, S. M. and Beddington, R. S.(1997). Mouse Dll3: a novel divergent Deltagene which may complementthe function of other Deltahomologues during early pattern formation inthe mouse embryo. Development124, 3065-3076.

Goode, S., Melnick, M., Chou, T. B. and Perrimon, N.(1996). Theneurogenic genes egghead and brainiac define a novel signaling pathwayessential for epithelial morphogenesis during Drosophila oogenesis.Development122, 3863-3879.

Haddad, R., Font, R. L. and Reeser, F.(1978). Persistent HyperplasticPrimary Vitreous: a clinicopathologic study of 62 cases and review of theliterature. Surv. Ophthalmol. 23, 123-134.

Hamada, Y., Kadokawa, Y., Okabe, M., Ikawa, M., Coleman, J. R.and Tsujimoto, Y. (1999). Mutation in ankyrin repeats of the mouseNotch2gene induces early embryonic lethality. Development126, 3415-3424.

Hartenstein, A. Y., Rugendorff, A., Tepass, U. and Hartenstein, V.(1992).The function of the neurogenic genes during epithelial development in theDrosophilaembryo. Development116, 1203-1220.

Hrabé de Angelis, M., McIntyre 2nd, J. and Gossler A.(1997). Maintenanceof somite borders in mice requires the Delta homologue DII1. Nature386,717-721.

Jack, R. L. (1972). Regression of the hyaloid vascular system: anultrastructural analysis. Am. J. Ophthalmol. 74, 261-272.

Jiang, R., Lan, Y., Chapman, H. D., Shawber, C., Norton, C. R., Serreze,D. V., Weinmaster, G. and Gridley, T.(1998). Defects in limb, craniofacial,and thymic development in Jagged2mutant mice. Genes Dev. 12, 1046-1057.

Joutel, A., Corpechot, C., Ducros, A., Vahedi, K., Chabriat, H., Mouton,P., Alamowitch, S., Domenga, V., Cecillion, M., Marechal, E. et al.(1996). Notch3 mutations in CADASIL, a hereditary adult-onset conditioncausing stroke and dementia. Nature383, 707-710.

Kitamoto, Y., Tokunaga, H. and Tomita, K. (1997). Vascular endothelialgrowth factor is an essential molecule for mouse kidney development:glomerulogenesis and nephrogenesis. J. Clin. Invest. 99, 2351-2357.

Krantz, I. D., Smith, R., Colliton, R. P., Tinkel, H., Zackai, E. H., Piccoli,D. A., Goldmuntz, E. and Spinner, N. B.(1999). Jagged1 mutations inpatients ascertained with isolated congenital heart defects. Am. J. Med.Genet. 84, 56-60.

Krebs, L. T., Xue, Y., Norton, C. R., Shutter, J. R., Maguire, M., Sundberg,J. P., Gallahan, D., Closson, V., Kitajewski, J., Callahan, R., Smith, G.H., Stark, K. L. and Gridley, T. (2000). Notch signaling is essential forvascular morphogenesis in mice. Genes Dev. 14, 1343-1352.

Kreidburg, J. A., Sariola, H., Loring, J. M., Maeda, M., Pelletier, J.,Houseman, D. and Jaenisch, R.(1993). WT-1 is required for early kidneydevelopment. Cell 74, 679-691.

Kusumi, K., Sun, E. S., Kerrebrock, A. W., Bronson, R. T., Chi, D. C.,Bulotsky, M. S., Spencer, J. B., Birren, B. W., Frankel, W. N. andLander, E. S.(1998). The mouse pudgy mutation disrupts Delta homologueDll3 and initiation of early somite boundaries. Nat. Genet. 19, 274-278.

Kuure, S., Vuolteenaho, R. and Vainio, S.(2000). Kidney morphogenesis:cellular and molecular regulation. Mech. Dev. 92, 31-45.

Lardelli, M. and Lendahl, U. (1993). Motch Aand Motch B– two mouseNotchhomologues coexpressed in a wide variety of tissues. Exp. Cell Res.204, 364-372.

Leon, A., Curtis, R. and Barnett, K. C. (1986). Hereditary persistenthyperplastic primary vitreous in the Staffordshire bull terrier. J. Am. Anim.Hosp. Assoc. 22, 765-773.

Levéen, P., Pekny, M., Gebre-Madhin, S., Swolin, B., Larsson, E. andBetsholtz, C. (1994). Mice deficient for PDGF-βshow renal,cardiovascular, and hematological abnormalities. Genes Dev. 8, 1875-1887.

Li, L., Krantz, I. D., Deng, Y., Genin, A., Banta, A. B., Collins, C. C., Qi,

502

M., Trask, B.J., Kuo, W. L., Cochran, J., et al. (1997). Alagille syndromeis caused by mutations in human Jagged1, which encodes a ligand forNotch1. Nat. Genet.16, 243-251.

Lindahl, P., Hellstrom, M., Kalen, M., Karlsson, L., Pekny, M., Pekna, M.,Soriano, P. and Betsholtz, C.(1998). Paracrine PDGF-B/PDGFR-βsignaling controls mesangial cell development in kidney glomeruli.Development125, 3313-3322.

Lindsell, C. E., Boulter, J., diSibio, G., Gossler, A. and Weinmaster, G.(1996). Expression patterns of Jagged, Delta1, Notch1, Notch2and Notch3genes identify ligand-receptor pairs that may function in neuraldevelopment. Mol. Cell. Neurosci. 8, 14-27.

Loomes, K. M., Underkoffler, L. A., Morabito, J., Gottlieb, S., Piccoli, D.A., Spinner, N. B., Baldwin, H. S. and Oakey, R. J.(1999). The expressionof Jagged1 in the developing mammalian heart correlates cardiovasculardisease in Alagille syndrome. Hum. Mol. Genet. 8, 2443-2449.

Luo, G., Hofmann, C., Bronckers, A., Sohocki, M., Bradley, A. andKarsenty, G. (1995). BMP-7 is an inducer of nephrogenesis and is alsorequired for eye development and skeletal patterning. Genes Dev. 9, 2808-2820.

Mansour, S. L., Thomas, K. R. and Capecchi, M. R.(1988). Disruption ofthe proto-oncogene int-2 in mouse embryo-derived stem cells: a generalstrategy for targeting mutations to non-selectable genes. Nature336, 348-352.

Miner, J. H. and Li, C. (2000). Defective glomerulogenesis in the absence ofLaminin α5 demonstrates a developmental role for the kidney glomerularbasement membrane. Dev. Biol. 217, 278-289.

Mitsiadis, T. A., Henrique, D., Thesleff, I. and Lendahl, U.(1997). MouseSerrate-1(Jagged1): Expression in the developing tooth is regulated byepithelial-mesenchymal interactions and fibroblast growth factor-4.Development124, 1473-1483.

Nofziger, D., Miyamoto, A., Lyons, K. M. and Weinmaster, G.(1999).Notch signaling imposes two distinct blocks in the differentiation of C2C12myoblasts. Development126, 1689-1702.

Oda, T., Elkahloun, A. G., Pike, B. L., Okajima, K., Krantz, I. D., Genin,A., Piccoli, D. A., Meltzer, P. S., Spinner, N. B., Collins, F. S. andChandrasekharappa, S. C.(1997). Mutations in the human Jagged1geneare responsible for Alagille syndrome. Nat. Genet. 16, 235-242.

Rebay, I., Fleming, R. J., Fehon, R. G., Cherbas, L., Cherbas, P. andArtavanis-Tsakonas, S.(1991). Specific EGF repeats of Notch mediateinteractions with Delta and Serrate: Implications for Notch as amultifunctional receptor. Cell67, 687-699.

Rebay, I., Fehon, R. G. and Artavanis-Tsakonas, S.(1993). Specifictruncations of Drosophila Notch define dominant activated and dominantnegative forms of the receptor. Cell74, 319-329.

Reese, A. B.(1955). Persistent hyperplastic primary vitreous. Am. J. Ophthal.40, 317-331.

Saxén, L. (1987). Organogenesis of the Kidney. Cambridge: CambridgeUniversity Press.

Schedl, A. and Hastie, N. D.(2000). Cross-talk in kidney development. Curr.Opin. Genet. Dev.10, 543-549.

Seifert, R. A., Alpers, C. E. and Bowen-Pope, D. F.(1998). Expression ofplatelet-derived growth factor and its receptors in the developing and adultmouse kidney. Kid. Int. 54, 731-746.

Shizumi, K., Shigeru, C., Kumano, K., Hosoya, N., Takahashi T., Kanda,Y., Hamada Y., Yazaki, Y. and Hirai, H. (1999). Mouse Jagged1physically interacts with Notch2and other notch receptors. J. Biol. Chem.274, 32961-32969.

Shizumi, K., Shigeru, C., Hosoya, N., Kumano, K., Saito, T., Kurokawa,M., Kanda, Y., Hamada Y. and Hirai, H. (2000). Binding of Delta1,Jagged1, and Jagged2 to Notch2 rapidly induces cleavage, nucleartranslocation, and hyperphosphorylation of Notch2. Mol. Cell. Biol. 20,6913-6922.

Shutter, J. R., Scully, S., Wei, F., Richards, W. G., Kitajewski, J.,Deblandre, G. A., Kintner, C. R. and Stark, K. L. (2000). Dll4, a novelNotch ligand expressed in arterial endothelium. Genes Dev. 14, 1313-1318.

Sidow, A., Bulotsky M. S., Kerrebrock, A. W., Bronson R. T., Daly, M. J.,Reeve, M. P., Hawkins, T. L., Birren, B. W., Jaenisch, R. and Lander,E. S. (1997). Serrate2 is disrupted in the mouse limb development mutantsyndactylism. Nature389, 722-725.

Soriano, P. (1994). Abnormal kidney development and hematologicaldisorders in PDGFβ-receptor mutant mice. Genes Dev. 8, 1888-1896.

Stades, F. C.(1983). Persistent hyperplastic Tunica Vasculosa Lentis andPersistent Hyperplastic Primary Vitreous in the Doberman pinscher: geneticaspects. J. Am. Anim. Hosp. Assoc. 19, 957-964.

Stark, K., Vainio, S., Vassileva, G. and McMahon, A. P.(1994). Epithelialtransformation of metanephric mesenchyme in the developing kidneyregulated by Wnt-4. Nature372, 679-683.

Swiatek, P. and Gridley, T.(1993). Perinatal lethality and defects in hindbraindevelopment in mice homozygous for a targeted mutation of the zinc fingergene Krox20. Genes Dev.7, 2071-2084.

Swiatek, P. J., Lindsell, C. E., Franco del Amo, F., Weinmaster, G. andGridley, T. (1994). Notch1is essential for postimplantation development inmice. Genes Dev. 8, 707-719.

Tufró, A. (2000). VEGF spatially directs angiogenesis during metanephricdevelopment in vitro. Dev. Biol. 226(in press).

Weinmaster, G.(2000). Notch signal transduction: a real Rip and more. Curr.Opin Genet. Dev. 10, 363-369.

Weinmaster, G., Roberts, V. J. and Lemke, G.(1992). Notch2: A secondmammalian Notchgene. Development 116, 931-941.

Williams, R., Lendahl, U. and Lardelli, M. (1995). Complementary andcombinatorial patterns of Notchgene family expression during early mousedevelopment. Mech. Dev. 53, 357-368

Xue, Y., Gao, X., Lindsell, C. E., Norton, C. R., Chang, B., Hicks, C.,Gendron-Maguire, M., Rand, E. B., Weinmaster, G. and Gridley T.(1999). Embryonic lethality and vascular defects in mice lacking the Notchligand Jagged1. Hum. Mol. Genet. 8, 723-730.


defects in development of the kidney, heart and eye ... · vasculature of the eye. these data...

Documents