Am. J. Hum. Genet. 74:721–730, 2004
Mutations in LRP5 or FZD4 Underlie the Common Familial Exudative Vitreoretinopathy Locus on Chromosome 11q
Carmel Toomes,1 Helen M. Bottomley,1 Richard M. Jackson,2 Katherine V. Towns,1 Sheila Scott,1 David A. Mackey,3 Jamie E. Craig,3,4 Li Jiang,5 Zhenglin Yang,5 Richard Trembath,6 Geoffrey Woodruff,7 Cheryl Y. Gregory-Evans,8 Kevin Gregory-Evans,8 Michael J. Parker,9 Graeme C. M. Black,10,11 Louise M. Downey,1 Kang Zhang,5 and Chris F. Inglehearn1
1Molecular Medicine Unit and 2School of Biochemistry and Molecular Biology, University of Leeds, Leeds; 3Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Melbourne; 4Department of Ophthalmology, Flinders Medical Centre, Adelaide; 5Department of Ophthalmology & Visual Science and Program in Human Molecular Biology and Genetics, University of Utah, Salt Lake City; Departments of 6Clinical Genetics and 7Ophthalmology, Leicester Royal Inﬁrmary, University of Leicester, Leicester, United Kingdom; 8Faculty of Medicine, Imperial College, London; 9Shefﬁeld Children’s Hospital, Shefﬁeld, United Kingdom; 10Academic Unit of Ophthalmology, University of Manchester, Manchester Royal Eye Hospital, and 11University Department of Medical Genetics and Regional Genetics Service, St. Mary’s Hospital, Manchester, United Kingdom
Familial exudative vitreoretinopathy (FEVR) is an inherited blinding disorder of the retinal vascular system. Autosomal dominant FEVR is genetically heterogeneous, but its principal locus, EVR1, is on chromosome 11q13-q23. The gene encoding the Wnt receptor frizzled-4 (FZD4) was recently reported to be the EVR1 gene, but our mutation screen revealed fewer patients harboring mutations than expected. Here, we describe mutations in a second gene at the EVR1 locus, low-density-lipoprotein receptor–related protein 5 (LRP5), a Wnt coreceptor. This ﬁnding further underlines the signiﬁcance of Wnt signaling in the vascularization of the eye and highlights the potential dangers of using multiple families to reﬁne genetic intervals in gene-identiﬁcation studies.
Familial exudative vitreoretinopathy (FEVR [MIM 133780]) is a well-deﬁned inherited disorder of retinal vessel development (Benson 1995). It is reported to have a penetrance of 100%, but clinical features can be highly variable, even within the same family. Severely affected patients may be legally blind during the 1st decade of life, whereas mildly affected individuals may not even be aware of symptoms and may receive a diagnosis only by use of ﬂuorescein angiography (Ober et al. 1980). The primary pathological process in FEVR is believed to be a premature arrest of retinal angiogenesis/vascu-logenesis or retinal vascular differentiation, leading to incomplete vascularization of the peripheral retina (van Nouhuys 1991). This failure to vascularize the periph-
Received November 14, 2003; accepted January 15, 2004; electronically published March 11, 2004.
Address for correspondence and reprints: Dr. Carmel Toomes, Molecular Medicine Unit, Level 6, Clinical Sciences Building, St. James’s University Hospital, Leeds LS9 7TF, United Kingdom. E-mail: c.toomes @leeds.ac.uk
� 2004 by The American Society of Human Genetics. All rights reserved. 0002-9297/2004/7404-0012$15.00
eral retina is the unifying feature seen in all affected individuals, but, by itself, it usually causes no clinical symptoms. The visual problems in FEVR result from secondary complications due to the development of hyperpermeable blood vessels, neovascularization, and vitreoretinal traction (ﬁg. 1). These features cause a reduction in visual acuity and, in 20% of cases, can lead to partial or total retinal detachment (van Nouhuys 1991).
FEVR is genetically heterogeneous, with autosomal dominant (Gow and Oliver 1971; Laqua 1980), X-linked (Plager et al. 1992; Shastry et al. 1997a), and autosomal recessive (Shastry and Trese 1997b; De Crecchio et al. 1998) modes of inheritance described, with autosomal dominant inheritance being the most common (Mu¨ller et al. 1994; Shastry et al. 2000; Kondo et al. 2001). To date, two autosomal dominant loci have been mapped. EVR1 on chromosome 11q was the ﬁrst FEVR locus to be identiﬁed (Li et al. 1992a), and, subsequently, other groups have published further families with EVR1 linkage (Li et al. 1992b;Mu¨ller et al. 1994; Price et al. 1996; Shastry et al. 2000; Kondo et al. 2001), suggesting that mutations
Figure 1 Clinical appearance of FEVR. Fundus photograph shows a fold of retinal tissue due to vitreoretinal traction originating from the temporal periphery. This obscures the posterior pole of the right eye and therefore compromises the visual acuity.
at this locus are a common cause of FEVR. A second dominant locus, EVR3 on chromosome 11p, has also been described in a single large pedigree (Downey et al. 2001). Although the gene underlying EVR3 remains unidentiﬁed, the gene encoding the Wnt receptor frizzled-4, FZD4 (MIM 604579), was recently reported as the EVR1 gene (Robitaille et al. 2002).
Following the identiﬁcation of FZD4 mutations in patients with FEVR, we screened an FEVR patient panel of 40 index cases for mutations in this gene. Although we did ﬁnd mutations, conﬁrming the results of Robitaille et al. (2002), these were identiﬁed in only eight (20%) patients (Toomes et al., in press). A similar low number of mutations was reported by Robitaille et al. (2002), who identiﬁed mutations in two (40%) of the ﬁve families they screened, and a recent study by Kondo et al. (2003) also identiﬁed mutations in only 5 (20%) of 24 probands screened. These results are inconsistent
Am. J. Hum. Genet. 74:721–730, 2004
with previous reports suggesting that EVR1 is the common locus for autosomal dominant FEVR (Mu¨ ller et al. 1994; Price et al. 1996; Shastry et al. 2000; Kondo et al. 2001). Moreover, our screen also revealed the absence of a mutation in a family with proven EVR1 linkage (Price et al. 1996). To investigate this family further, we analyzed the linked haplotype in the EVR1 region of chromosome 11q13, and we found that crossovers reﬁned the disease gene in this family to an interval of ∼15 cM between the markers D11S1368 and D11S937, 10 cM centromeric to FZD4. These data strongly suggested the presence of a second gene mutated in FEVR, close to FZD4 and within the previously deﬁned EVR1 locus (Toomes et al. 2004).
Despite the fact that 1250 genes are located within the deﬁned interval, the discovery that FEVR can be caused by mutations in the Wnt receptor FZD4 highlighted proteins involved in the Wnt-signaling pathway as candidate FEVR genes (for an up-to-date list of all the proteins involved in Wnt signaling, see the Wnt Gene Homepage maintained by the Roel Nusse lab). We identiﬁed LRP5 (MIM 603506) as one such candidate gene in the region. LRP5 and its closely related homologue LRP6 encode single-pass transmembrane receptors that partner with members of the frizzled family of seven-pass transmembrane receptors to bind Wnt proteins, forming a functional ligand-receptor complex that activates the canonical Wnt-b-catenin pathway (Pinson et al. 2000; Tamai et al. 2000; Wehrli et al. 2000). LRP5 consists of 23 exons and encodes a 1,615–amino acid protein. Its extracellular segment contains four domains, each composed of six YWTD repeats (which form a b-propeller structure) and an epidermal growth factor–like domain (YWTD-EGF domain). These are followed by three low-density-lipo-protein receptor–like (LDL-R–like) ligand-binding domains (ﬁg. 2). The speciﬁc function of LRP5 remains unknown, but loss of this protein causes osteoporosispseudoglioma syndrome (OPPG [MIM 259770]), a recessive disorder characterized by very low bone mass and blindness (Gong et al. 2001). Individuals with OPPG are
Figure 2 Schematic diagram of the LRP5 protein showing the location of the mutations within the protein domains. The six YWTD repeats are shown as a hexagon to represent their b-propeller structure, whereas the EGF domains are represented as gray blocks.
prone to developing bone fractures and deformations and can have various eye abnormalities, including phthisis bulbi, retinal detachments, falciform folds, or persistent vitreal vasculature (Frontali et al. 1985). Dominant missense mutations in LRP5 have also been described in patients with high-bone-mass disorders (high bone mass [MIM 601884], endosteal hyperostosis [MIM 144750], and osteopetrosis [MIM 607634]) (Boyden et al. 2002; Little et al. 2002; Van Wesenbeeck et al. 2003). It is thought that these missense mutations result in the formation of a mutant LRP5 protein that is functionally abnormal (Boyden et al. 2002).
In light of these ﬁndings, LRP5 was a strong candidate for involvement in FEVR and was therefore screened in the remaining 32 patients from our FEVR patient panel. Informed consent was obtained from all subjects tested, and ethical approval was obtained from the Leeds Teaching Hospitals Trust research ethics committee. We designed primers (table 1) and screened all 23 exons and ﬂanking intronic sequences by SSCP-heteroduplex analysis (SSCP-HA) and direct sequencing, using established protocols (Toomes et al. 2001). In total, we dis-
Table 1 LRP5 SSCP-HA Primers
covered six LRP5 mutations not present in control individuals (ﬁgs. 2 and 3) and a further 13 benign sequence variants not previously described (table 2).
We ﬁrst identiﬁed a splice-donor mutation that segregated with FEVR in the family with EVR1 linkage in which FZD4 mutations had been excluded by genetics (ﬁg. 3B [family 1]). This mutation is a substitution of the second nucleotide of intron 21, changing the GT splice-donor site to GG, c4488�2trg. The precise effect of this mutation has not been determined in this family, since no RNA was available. However, the most common outcome of splice-donor mutations is the deletion of the preceding exon. This would lead to the deletion of exon 21, resulting in a frameshift after codon 1449, followed by 52 incorrect amino acids and a premature termination at codon 1502.
In a second family originating from the United States, we identiﬁed a 1-bp insertion in exon 20, c41194120insC. This mutation segregates with the disease in that family but was also found in two asymptomatic family members (ﬁg. 3B [family 2]). However, neither of these individuals had been examined by ﬂuorescein angiog-
|Exona||Primer Name||5r3� Sequence||Primer Name||5r3� Sequence||(bp)|
|2 start||2Fb||CATCCCAGGGCTGTGTATCT||SSCP 2-R||GTGATGACCACGTTCTGCAC||288|
|2 end||SSCP 2-F||CAAGCAGACCTACCTGAACC||2Rb||ACTTGGGCTCATGCAAATTC||309|
|3 start||3F3b||GAAACCATACCTGTTGGTTATTTCC||SSCP 3-R||GGATGAAGCTGAGCTTGGCGTC||311|
|3 end||SSCP 3-F||CGGATTGAGCGGGCAGGGAT||3R3b||CACAGACCCTGACGCTGTTC||323|
|6 start||6F1b||TGGCTGAGTATTTCCCTTGCCC||SSCP 6-R||GTCGACCGCGATGCCATCGG||352|
|6 end||SSCP 6-F||CGACCCGCTAGAGGGCTATGT||SSCP 6-2R||CAATCTCCCTCTCGCCTGTGC||338|
|8 start||8F1b||TGAAGTTGCTGCTCTTGGGCA||SSCP 8-R||CACCCGCTCGATGCTGCGGC||256|
|8 end||SSCP 8-F||CGCACATTTTCGGGTTCACGC||8R1b||AACACTTATGCCCAGGCATGGA||252|
|9 start||9F1b||TGCTGGGCTGTTGATGTTTAGACT||SSCP 9-R||CCGTGAGCGGGATGGCCACG||263|
|9 end||SSCP 9-F||GTGCCTGAGGCCTTCTTGGTCT||9R1b||CTTGAACTGCGTTACAATAAATACGA||291|
|12 start||12Fb||ATTCATGTGGTCGCTAGGCT||SSCP 12-R||CATCACGAAGTCCAGGTGG||281|
|12 end||SSCP 12-F||CTAGCGGCCGGAACCGCA||12Rb||GAAGCTCCTTTCAGCGTCAG||287|
|14||SSCP 14-F||GTCTCCGCCAGTGCTCAG||SSCP 14-R||CTGTGAGAGGCTGGCATTC||334|
|23||SSCP 23-F||CGGATGTGCCTACCGAATC||SSCP 23-R||TTACAGGGGCACAGAGAAGC||347|
Large exons were ampliﬁed in two overlapping segments, designated “start” and “end.”
Primer sequences obtained from Gong et al. (2001).
raphy to exclude a very mild phenotype, so their undiagnosed condition was not unexpected (Ober et al. 1980). This mutation causes a frameshift resulting in 175 incorrect amino acids after codon 1373, followed by a premature termination at codon 1549 (K1374fsX1549). The third mutation identiﬁed was a 1-bp deletion in exon 18, c3804delA, in a single Indian patient. This mutation causes a frameshift resulting in the substitution of 168 amino acids after codon 1269 and a premature stop in codon 1438 (R1270fsX1438).
The remaining three potential mutations were missense changes. In exon 3, we identiﬁed T173M (c518CrT) in an elderly British woman with abnormal retinal vasculature and retinal folds. In exon 16, we identiﬁed Y1168H (c3502TrC) in a British woman with a total retinal detachment and retinoschisis. The patient was born after 36 wk of pregnancy but was never treated with oxygen therapy. The patient had no family history of FEVR, although the patient’s asymptomatic father was also found to have the mutation. The ﬁnal mutation, C1361G (c4081TrG), was identiﬁed in exon 19 in a 6-year-old Australian boy with classic features of FEVR. This patient has very poor vision and has undergone surgery on his left eye, which is phthisical and sunken. There is no family history of FEVR, but, again, the asymptomatic mother was found to have the mutation. Neither of the asymptomatic parents had been examined by ﬂuorescein angiography, but they showed no signs of retinopathy upon fundus examination. To exclude the possibility that these changes were common polymorphisms, we screened 200 ethnically matched control individuals (400 chromosomes) for each of these changes (Collins and Schwartz 2002). We also checked each of the mutated amino acids for conservation within homologues of LRP5 from human and other species (ﬁg. 3C). Both Y1168H and C1361G are changes to highly conserved residues. T173M is not well conserved, although the methionine residue at this position is not seen in any FZD4 homologue in another species. Whereas these results suggest that these missense changes are indeed pathogenic, without further examination by means of a functional assay, we are unable to prove categorically that they are disease-causing mutations.
Table 2 Summary of Novel Benign Sequence Variants Detected in LRP5
Location cDNA Change Protein Change Frequency
|Exon 2||324CrG||V108V||C 98%; G 2%|
|Intron 2||488�53trc||…||t 52%; c 48%|
|Exon 3||639CrT||D213D||C 98%; T 2%|
|Intron 3||686�61crt||…||c 99%; t 1%|
|Intron 8||1801�62 crt||…||c 99%; t 1%|
|Intron 11||2503�37 crt||…||c 99%; t 1%|
|Intron 11||2503�78 gra||…||g 52%; a 48%|
|Exon 18||3918GrA||A1306A||G 99%; A 1%|
|Intron 20||4348�13arg||…||a 99%; g 1%|
|Exon 21||4380CrT||S1460S||C 98%; T 2%|
|Exon 21||4431CrT||H1477H||C 98%; T 2%|
|Intron 21||4488�54cra||…||c 99%; a 1%|
|Intron 21||4488�58gra||…||g 98%; a 2%|
|Exon 22||4574TrC||V1525A||T 99%; C 1%|
NOTE.—Amino acid and nucleotide numbering follows the cDNA sequence, with nucleotide position 1 assigned to the ﬁrst nucleotide of the ATG initiation codon in exon 1. Bases in exons are denoted by uppercase letters, bases in introns by lowercase letters. (NCBI assay ID numbers for these SNPs appear in the dbSNP entry of the “Electronic-Database Information” section.)
On the basis of these results, we propose that FEVR results from heterozygous mutations that cause haploinsufﬁciency of LRP5. The insertion, deletion, and splicing mutations all lead to transcripts with premature termination, which are likely to be targeted by nonsense-me-diated decay. Similar mutations, when homozygous, cause OPPG. Indeed, c3804delA has been reported elsewhere as a homozygous mutation in a patient with OPPG (Gong et al. 2001). It is therefore likely that the missense mutations identiﬁed also knock out the function of LRP5. In support of this theory, the C1361G mutation, which affects the ﬁfth cysteine within the third LDL-R– like ligand-binding domain, is an invariant cysteine required to form a disulﬁde bond necessary for correct protein folding (Bieri et al. 1995; Fass et al. 1997).
The remaining two missense mutations identiﬁed in patients with FEVR occur within the ﬁrst and fourth YWTD-EGF domains (ﬁg. 2). To determine the functional signiﬁcance of these mutations, we constructed a model of the YWTD-EGF domain 1 of LRP5 based on the
Figure 3 FEVR caused by mutations in LRP5. A, Sequence traces of the six mutations identiﬁed and the corresponding wild-type alleles. The mutation-sequence trace for exon 18 was obtained from a cloned PCR product, so only the disease allele is shown. Other mutations were identiﬁed from directly sequenced PCR products and therefore appear as heterozygous changes. B, Mutations segregating in two pedigrees with FEVR. Unaffected individuals (unblackened); affected individuals (blackened); individuals asymptomatic for FEVR (blackened circle in square). Family 1 shows the results of a restriction digest assay through use of HphI, the recognition site of which is abolished by the mutation. The arrow indicates the undigested product present in affected individuals. Family 2 shows an SSCP trace (arrow indicates the aberrant shift). In family 2, an asterisk (*) indicates individuals affected by low bone mass. C, Protein-sequence alignment of human LRP5 (h) with homologues from human and other species: Mus musculus (m), X. laevis (x), Drosophila melanogaster (“arrow”), and Anopheles gambiae str. (“mosquito”). (GenBank accession numbers for all species appear in the “Electronic-Database Information” section.) Only 20 amino acid residues surrounding each mutation are shown. Conserved amino acid residues are highlighted. The positions of the missense mutations T173M, Y1168H, and C1361G are indicated. Both Y1168H and C1361G affect highly conserved amino acids, whereas T173M occurs at a less well-conserved residue.
crystal structure of the YWTD-EGF domain of the LDLR (Rudenko et al. 2002). To construct the model, the sequences of the four YWTD-EGF domains of LRP5 were aligned with the LDL-R sequence taken from the Protein Data Bank entry 1n7d (chain A), through use of the multiple-sequence-alignment program ClustalW (Chenna et al. 2003) (ﬁg. 4A). The protein homology-modeling server Swiss-Model (Schwede et al. 2003) was then used to construct the model based on this alignment and the coordinates of the crystal structure of the LDLR YWTD-EGF domain (1n7d chain A) (Rudenko et al. 2002) (ﬁg. 4B). We used this model to map the missense mutations found in this study, as well as those missense mutations identiﬁed in patients with OPPG and high bone mass in other studies (Gong et al. 2001; Boyden et al. 2002; Little et al. 2002; Van Wesenbeeck et al. 2003), onto the three-dimensional structure of the LRP5 protein (ﬁg. 4B).
Many of the mutations occur in the core of the protein and are likely to cause destabilization of the protein fold. This includes the FEVR mutation Y1168H from this study and some of the mutations shown to cause increased bone density (A242T and T253I) and OPPG (R570W). The remaining mutations occur at the protein surface (ﬁg. 4B). It is less clear why mutations at these non–sequence-conserved positions (ﬁg. 3C) would cause destabilization of the fold. However, six of the eight are clustered at one location on the protein surface (ﬁg. 4B). It is interesting that all these mutations (T173M [causing FEVR] and D111Y, G171V, G171R, A214T, and A214V [causing high bone density]) are in the YWTD-EGF domain 1 of LRP5. Furthermore, all are in the same region that corresponds to the protein-protein–binding interface between modules R4 and R5 of the ligand-binding domain of LDL-R and the b-propeller domain of LDLR (Rudenko et al. 2002). By analogy with LDL-R, it is possible that the surface region of the YWTD-EGF domains of LRP5 is also involved in protein-protein interactions and that mutating these nonconserved residues at the protein surface disrupts functionally important protein interactions.
Studies of LRP6 suggest that the ﬁrst and second YWTD-EGF domains are required for LRP6 to interact functionally with the Wnt-frizzled complex, whereas the third and fourth YWTD-EGF repeats are required to bind Dickkopf-1 (Dkk-1), a secreted antagonist of the canonical Wnt-b-catenin pathway (Mao et al. 2001). A more recent study in Xenopus has also identiﬁed a new LRP6interacting protein known as “Wise.” Depending on its environment, Wise can act as both an activator and an inhibitor of Wnt signaling and interacts with the ﬁrst and second YWTD-EGF repeats of LRP6, presumably by competing with the Wnt-frizzled complex (Itasaki et al. 2003). The high similarity between LRP5 and LRP6 suggests that equivalent interactions will occur in LRP5. Consistent with this notion, LRP5 has been shown to interact with different Wnt/frizzled complexes (Tamai et al. 2000; Seme¨nov et al. 2001; Kato et al. 2002; Caricasole et al. 2003), although the binding sites have not been determined. Similarly, LRP5 has been shown to interact with Dkk-1 (Baﬁco et al. 2001; Caricasole et al. 2003), and in vitro studies have shown that the antagonizing effect of Dkk-1 on Wnt signaling is almost abolished when the wild-type LRP5 protein is replaced with one harboring the G171V mutation known to cause high bone mass (Boyden et al. 2002). Furthermore, LRP5 has also been reported to interact with other members of the Dickkopf (Dkk) family of proteins, Dkk-2 and Dkk-3 (Caricasole et al. 2003), and evidence suggests that these proteins, like the Wise protein, can function as either agonists or antagonists, depending on their cellular context (Wu et al. 2000; Mao and Niehrs 2003). These preliminary experiments indicate that comparable results can be obtained for both LRP5 and LRP6 and strengthen the conclusions, based on the modeling data, that the missense mutations disrupt a protein-binding site.
Expression studies have shown that LRP5 and LRP6 are widely expressed in both embryonic and adult tissues, including the retina (Brown et al. 1998; Hey et al. 1998; Kim et al. 1998). However, whereas mice lacking Lrp6 show a wide range of developmental defects in
Figure 4 Homology model of LRP5 YWTD-EGF domain 1. A, ClustalW alignment of the sequence from the YWTD-EGF domain, from LDL-R against the four YWTD-EGF domains found in LRP5. Amino acids are numbered in accordance with Swiss-Prot accession number P01130 (LDL-R) and GenBank accession number NP_002326 (LRP5). The amino acids highlighted in cyan correspond to the location of the T173M and Y1168H missense changes identiﬁed in the present study. The residues highlighted in pink correspond to the location of the missense mutations shown elsewhere to cause increased bone density (D111Y, G171V, G171R, A214T, A214V, A242T, and T253I) (Boyden et al. 2002; Little et al. 2002; Van Wesenbeeck et al. 2003) and OPPG (R570W, R494Q, and V667M) (Gong et al. 2001). The boxes correspond to the six YWTD-repeats present in the b-propeller module. The LDL-R and model LRP5 domain 1 share 37% sequence identity over 310 residues. B, Surface representation of the LRP5 YWTD-EGF domain 1. The model is based on the alignment shown in panel A. The location of mutations listed in panel A are highlighted on the surface in cyan and pink (numbers correspond to the amino acid residues). The blue backbone ribbon represents the ligand-binding domains R4 and R5 of LDL-R and illustrates how these interact with the propeller domain in LDL-R (Rudenko et al. 2002). Six of the missense mutations are located on the surface of the LRP5 YWTD-EGF domain 1 structure at a site corresponding to the protein-protein interface surface between the R4 and R5 ligand-binding domains and the propeller domain of LDL-R. (Figure prepared using GRASP [Nicholls et al. 1991])
multiple organs (Pinson et al. 2000), mice lacking Lrp5 have a restricted phenotype. Two groups have independently created mice with a targeted disruption of Lrp5 (Kato et al. 2002; Fujino et al. 2003), but it is surprising that these mice had different phenotypes. Kato et al. (2002) disrupted Lrp5 at amino acid 373 (exon 6), and immunohistochemistry showed that this allele resulted in the formation of a truncated polypeptide. The resultant homozygous mice had low bone mass and persistent embryonic eye vascularization, a phenotype similar to that observed in patients with OPPG. Fujino et al. (2003) disrupted Lrp5 at exon 18, but, this time, both immunoblotting and northern analysis indicated that the homozygous mice totally lacked Lrp5. These mice appeared normal, with no eye phenotype documented and only a very mild low-bone-mass phenotype, similar to osteoporosis, present in females aged 16 mo. These mice were, however, shown to have defects in their cholesterol metabolism and glucose-induced insulin secretion (Fujino et al. 2003). In both studies, the eyes of the heterozygotes appeared normal, but detailed analysis of the retinal vasculature by ﬂuorescein angiography was not reported, so an FEVR phenotype cannot be ruled out.
In light of the results of the present study, we would have expected obligate OPPG carriers to have FEVR; it is surprising, therefore, that no eye defects have been reported. This is presumably a result of the subtle phenotype observed in some patients with FEVR (Ober et al. 1980). OPPG carriers have, however, been found to have reduced bone mass, when compared with age-and sex-matched control individuals (Gong et al. 2001). A closer inspection of affected members of family 2 who had the c4119–4120insC mutation showed that they, too, had signs of low bone mass (ﬁg. 3B). Individual I:2 is osteoporotic. Her bone mineral density (BMD) is 60% of the normal sex-matched value and 65% of the age-and sex-matched value. No BMD ﬁgures were available for other affected family members, but their medical histories highlighted a large number of fractures. Individuals III:2 and III:3 have had multiple radius and humerus bone fractures since age 6 mo. Individual III:4 had a radial bone fracture at age 5 years, and individual I:3 has had six breaks in total (left wrist, right wrist, foot, femur, clavicle in two places, and right arm). These results suggest that patients with FEVR who have LRP5 mutations show signs of osteoporosis, a clinical feature not previously noted in FEVR.
The microheterogeneity observed in this study highlights potential dangers in using multiple families for disease-gene reﬁnement and identiﬁcation. It is possible that one or both of these neighboring FEVR genes could have been excluded by combining haplotype data from families with mutations in both FZD4 and LRP5, a worrying scenario for disease gene–mapping projects. Likewise, if the genetic analysis in family 1 had not excluded
Am. J. Hum. Genet. 74:721–730, 2004
FZD4 as the mutated gene in the family, the search for their mutation would have been abandoned, since it would have been presumed to lie outside the coding sequence of the gene. As a result, the second mutated gene at this locus might not have been identiﬁed.
In summary, we have shown that mutations in either LRP5 or FZD4, two distinct genes within the EVR1 locus, can cause FEVR. Together, these two loci account for 35% of FEVR cases in our patient series (15% LRP5 and 20% FZD4), indicating that other unidentiﬁed FEVR genes may be a more signiﬁcant cause of the disease than previously thought. The primary pathological process in FEVR is believed to be a premature arrest of retinal angiogenesis/vasculogenesis or retinal vascular differentiation, leading to incomplete vascularization of the peripheral retina. The implication of two components of the Wnt-signaling pathway in this disease process highlights the importance of its role in the development of the vasculature of the eye and, potentially, other organs.
We thank the families with FEVR, for their help in this study; The Wellcome Trust, for grants 069718/Z/02, 055145/Z/98, and 035535/Z/96; and the Jack Brockhoff Foundation, for ﬁnancial support. K.Z. is supported by National Institutes of Health grants RO1EY14438, RO1EY14448, and GCRC M01RR00064; the American Health Assistance Foundation; the Ruth and Milton Steinbach Fund; and the Ronald McDonald House Charities.
Accession numbers and URLs for data presented herein are as follows:
dbSNP, http://www.ncbi.nlm.nih.gov/SNP/ (for NCBI assay IDs ss16359666, ss16359667, ss16359668, ss16359669, ss16359670, ss16359671, ss16359672, ss16359673, ss16359674, ss16359675, ss16359676, ss16359677, ss16359678, and ss16359679)
GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ (for hLRP5 [accession numbers NP_002326 and NM_002335], mLRP5 [accession number NP_032539], xLRP5 [accession number AAN09806], hLRP6 [accession number NP_002327], mLRP6 [accession number NP_032540] xLRP6 [accession number AAN09807], arrow [accession number AAF91072], mosquito [accession number EAA00402], and hLRP4 [accession number XP_035037])
Online Mendelian Inheritance in Man (OMIM), http://www .ncbi.nlm.nih.gov/Omim/ (for FEVR, FZD4, LRP5, OPPG, high bone mass, endosteal hyperostosis, and osteopetrosis)
Protein Data Bank, http://www.rcsb.org/pdb/
Swiss-Prot, http://ca.expasy.org/sprot/sprot-top.html (for LDLR [accession number P01130])
Wnt Gene Homepage, http://www.stanford.edu/˜rnusse/ wntwindow.html
Baﬁco A, Liu G, Yaniv A, Gazit A, Aaronson SA (2001) Novel mechanisms of Wnt signalling inhibition mediated by Dick-kopf-1 interaction with LRP6/arrow. Nat Cell Biol 3:683– 686
Benson WE (1995) Familial exudative vitreoretinopathy. Trans Am Ophthalmol Soc 93:473–521
Bieri S, Djordjevic JT, Daly NL, Smith R, Kroon PA (1995) Disulﬁde bridges of a cysteine-rich repeat of the LDL receptor ligand-binding domain. Biochemistry 34:13059–13065
Boyden LM, Mao J, Belsky J, Mitzner L, Farhi A, Mitnick MA, Wu D, Insogna K, Lifton RP (2002) High bone density due to a mutation in LDL-receptor-related protein 5. N Engl J Med 346:1513–1521
Brown SD, Twells RCJ, Hey PJ, Cox RD, Levy ER, Soderman AR, Metzker ML, Caskey CT, Todd JA, Hess JF (1998) Isolation and characterization of LRP6, a novel member of the low density lipoprotein receptor gene family. Biochem Biophys Res Commun 248:879–888
Caricasole A, Ferraro T, Iacovelli L, Barletta E, Caruso A, Melchiorri D, Terstappen GC, Nicoletti F (2003) Functional characterization of WNT7A signalling in PC12 cells. J Biol Chem 278:37024–37031
Chenna R, Sugawara H, Koike T, Lopez R, Gibson TJ, Higgins DG, Thompson JD (2003) Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res 31:3497– 3500
Collins JS, Schwartz CE (2002) Detecting polymorphisms and mutations in candidate genes. Am J Hum Genet 71:1251– 1252
De Crecchio G, Simonelli F, Nunziata G, Mazzeo S, Greco GM, Rinaldi E, Ventruto V, Ciccodicola A, Miano MG, Testa F, Curci A, D’Urso M, Rinaldi MM, Cavaliere ML, Castelluccio P (1998) Autosomal recessive familial exudative vitreoretinopathy: evidence for genetic heterogeneity. Clin Genet 54:315–320
Downey LM, Keen TJ, Roberts E, Mansﬁeld DC, Bamashmus M, Inglehearn CF (2001) A new locus for autosomal dominant familial exudative vitreoretinopathy maps to chromosome 11p12-13. Am J Hum Genet 68:778–781
Fass D, Blacklow S, Kim PS, Berger JM (1997) Molecular basis of familial hypercholesterolaemia from structure of LDL receptor module. Nature 388:691–693
Frontali M, Stomeo C, Dallapiccola (1985) Osteoporosis-pseu-doglioma syndrome: report of three affected sibs and an overview. Am J Med Genet 22:35–47
Fujino T, Asaba H, Kang M-J, Ikeda Y, Sone H, Takada S, Kim D-H, Ioka RX, Ono M, Tomoyori H, Okubo M, Mu-rase T, Kamataki A, Yamamoto J, Magoori K, Takahashi S, Miyamoto Y, Oishi H, Nose M, Okazaki M, Usui S, Imaizumi K, Yanagisawa M, Sakai J, Yamamoto TT (2003) Low-density lipoprotein receptor-related protein 5 (LRP5) is essential for normal cholesterol metabolism and glucose-in-duced insulin secretion. Proc Natl Acad Sci USA 100:229– 234
Gong Y, Slee RB, Fukai N, Rawadi G, Roman-Roman S, Reginato AM, Wang H, et al (2001) LDL receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell 107:513–523
Gow J, Oliver GL (1971) Familial exudative vitreoretinopathy: an expanded view. Arch Ophthalmol 86:150–155
Hey PJ, Twells RCJ, Phillips MS, Nakagawa Y, Brown SD, Kawaguchi Y, Cox R, Xie G, Dugan V, Hammond H, Metzker ML, Todd JA, Hess JF (1998) Cloning of a novel member of the low-density lipoprotein receptor family. Gene 216: 103–111
Itasaki N, Jones CM, Mercurio S, Rowe A, Domingos PM, Smith JC, Krumlauf R (2003) Wise, a context-dependent activator and inhibitor of Wnt signalling. Development 130: 4295–4305
Kato M, Patel MS, Levasseur R, Lobov I, Chang BH-J, Glass DA, Hartmann C, Li L, Hwang T-H, Brayton CF, Lang RA, Karsenty G, Chan L (2002) Cbfa1-independent decrease in osteoblast proliferation, osteopenia, and persistent embryonic eye vascularization in mice deﬁcient in Lrp5, a Wnt coreceptor. J Cell Biol 157:303–314
Kim D-H, Inagaki Y, Suzuki T, Ioka RX, Yoshioka SZ, Magoori K, Kang M-J, Cho Y, Nakano AZ, Liu Q, Fujino T, Suzuki H, Sasano H, Yamamoto TT (1998) A new low density lipoprotein receptor related protein, LRP5, is expressed in hepatocytes and adrenal cortex, and recognizes apoliprotein E. J Biochem 124:1072–1076
Kondo H, Hayashi H, Oshima K, Tahira T, Hayashi K (2003) Frizzled 4 gene (FZD4) mutations in patients with familial exudative vitreoretinopathy with variable expressivity. Br J Ophthalmol 87:1291–1295
Kondo H, Ohno K, Tahira T, Hayashi H, Oshima K, Hayashi K (2001) Delineation of the critical interval for the familial exudative vitreoretinopathy gene by linkage and haplotype analysis. Hum Genet 108:368–375
Laqua H (1980) Familial exudative vitreoretinopathy. Albrecht Von Graefes Arch Klin Exp Ophthalmol 213:121–133
Li Y, Fuhrmann C, Schwinger E, Gal A, Laqua H (1992a) The gene for autosomal dominant exudative vitreoretinopathy (Criswick-Schepens) on the long arm of chromosome 11. Am J Ophthalmol 113:712–713
Li Y, Mu¨ller B, Fuhrmann C, van Nouhuys CE, Laqua H, Humphries P, Schwinger E, Gal A (1992b) The autosomal dominant familial exudative vitreoretinopathy locus maps on 11q and is closely linked to D11S533. Am J Hum Genet 51:749–754
Little RD, Carulli JP, Del Mastro RG, Dupuis J, Osborne M, Folz C, Manning SP, et al (2002) A mutation in the LDL receptor–related protein 5 gene results in the autosomal dominant high–bone-mass trait. Am J Hum Genet 70:11–19
Mao B, Niehrs, C (2003) Kremen2 modulates Dickkopf2 activity during Wnt/LRP6 signalling. Gene 302:179–183
Mao B, Wu W, Li Y, Hoppe D, Stannek P, Glinka A, Niehrs C (2001) LDL-receptor-related protein 6 is a receptor for Dickkopf proteins. Nature 411:321–325
Mu¨ ller B, Orth U, van Nouhuys CE, Duvigneau C, Fuhrmann C, Schwinger E, Laqua H, Gal A (1994) Mapping of the autosomal dominant exudative vitreoretinopathy locus (EVR1) by multipoint linkage analysis in four families. Genomics 20:317–319
Nicholls A, Sharp KA, Honig B (1991) Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins 11:281–296
Ober RR, Bird AC, Hamilton AM, Sehmi K (1980) Autosomal dominant exudative vitreoretinopathy. Br J Ophthalmol 64: 112–120
Pinson KI, Brennan J, Monkley S, Avery BJ, Skames WC (2000) An LDL receptor-related protein mediates Wnt signalling in mice. Nature 407:535–538
Plager DA, Orgel IK, Ellis FD, Hartzer M, Trese MT, Shastry BS (1992) X-linked recessive familial exudative vitreoretinopathy. Am J Ophthalmol 114:145–148
Price SM, Periam N, Humphries A, Woodruff G, Trembath RC (1996) Familial exudative vitreoretinopathy linked to D11S533 in a large Asian family with consanguinity. Ophthalmic Genet 17:53–57
Robitaille J, MacDonald MLE, Kaykas A, Sheldahi LC, Zeisler J, Dube MP, Zhang LH, Singaraja RR, Guernsey DL, Zheng B, Siebert LF, Hoskin-Mott A, Trese MT, Pimstone SN, Shastry BS, Moon RT, Hayden MR, Goldberg YP, Samuels ME (2002) Mutant frizzled-4 disrupts retinal angiogenesis in familial exudative vitreoretinopathy. Nat Genet 32:326–330
Rudenko G, Henry L, Henderson K, Ichtchenko K, Brown MS, Goldstein JL, Deisenhofer J (2002) Structure of the LDL receptor extracellular domain at endosomal pH. Science 298: 2353–2358
Schwede T, Kopp J, Guex N, Peitsch MC (2003) SWISSMODEL: an automated protein homology-modelling server. Nucleic Acids Res 31:3381–3385
Seme¨nov MV, Tamai K, Brott BK, Ku¨hl M, Sokol S, He X (2001) Head inducer Dickkopf-1 is a ligand for Wnt coreceptor LRP6. Curr Biol 11:951–961
Shastry BS, Hejtmancik JF, Hiraoka M, Ibaraki N, Okubo Y, Okubo A, Han DP, Trese MT (2000) Linkage and candidate gene analysis of autosomal-dominant familial exudative vitreoretinopathy. Clin Genet 58:329–332
Shastry BS, Liu X, Hejtmancik JF, Plager DA, Trese MT (1997a)
Am. J. Hum. Genet. 74:721–730, 2004
Evidence for genetic heterogeneity in X-linked familial exudative vitreoretinopathy. Genomics 44:247–248
Shastry BS, Trese MT (1997b) Familial exudative vitreoretinopathy: further evidence for genetic heterogeneity. Am J Med Genet 69:217–218
Tamai K, Semenov M, Kato Y, Spokony R, Liu C, Katsuyama Y, Hess F, Saint-Jeannet JP, He X (2000) LDL-receptor-re-lated proteins in Wnt signal transduction. Nature 407:530– 535
Toomes C, Bottomley HM, Scott S, Mackey DA, Craig JE, Appukuttan B, Stout JT, Zhang K, Black GCM, Fryer A, Downey LM, Inglehearn CF. Spectrum and frequency of FZD4 mutations in familial exudative vitreoretinopathy (FEVR). Invest Ophthalmol Vis Sci (in press)
Toomes C, Downey LM, Bottomley HM, Scott S, Woodruff G, Trembath RC, Inglehearn CF (2004) Identiﬁcation of a fourth locus (EVR4) for familial exudative vitreoretinopathy (FEVR). Mol Vis 10:37–42
Toomes C, Marchbank NJ, Mackey DA, Craig, JE, Newbury-Ecob, RA, Bennett, CP, Vize, CJ, Desai, SP, Black, GCM, Patel, N, Teimory, M, Markham, AF, Inglehearn, CF, Churchill, AJ (2001) Spectrum, frequency and penetrance of OPA1 mutations in dominant optic atrophy. Hum Mol Genet 10: 1369–1378
van Nouhuys CE (1991) Signs, complications, and platelet aggregation in familial exudative vitreoretinopathy. Am J Ophthalmol 111:34–41
Van Wesenbeeck L, Cleiren E, Gram J, Beals RK, Be´nichou O, Scopelliti D, Key L, Renton T, Bartels C, Gong Y, Warman ML, de Vernejoul M-C, Bollerslev J, Van Hul W (2003) Six novel missense mutations in the LDL receptor-related protein 5 (LRP5) gene in different conditions with an increased bone density. Am J Hum Genet 72:763–771
Wehrli M, Dougan ST, Caldwell K, O’Keefe L, Schwartz S, Vaizel-Ohayon D, Schejter E, Tomlinson A, DiNardo S (2000) Arrow encodes an LDL-receptor-related protein essential for Wingless signalling. Nature 407:527–530
Wu W, Glinka A, Delius H, Niehrs, C (2000) Mutual antagonism between Dickkopf1 and Dickkopf2 regulates Wnt/b-catenin signalling. Curr Biol 10:1611–1614