Pedigree of the family described in this study. Small black dotsrepresent individuals whose DNA was analyzed.
HYMAN L, Mortemousque B, Amati-Bonneau P, Couture F, Graffan R, Dubois S, Colin J, Bonneau D, Morissette J, Lacombe D, Raymond V. Axenfeld-Rieger AnomalyA Novel Mutation in the Forkhead Box C1 (FOXC1) Gene in a 4-Generation Family. Arch Ophthalmol. 2004;122(10):1527-1533. doi:10.1001/archopht.122.10.1527
Copyright 2004 American Medical Association. All Rights Reserved.Applicable FARS/DFARS Restrictions Apply to Government Use.2004
To characterize DNA mutations in a pedigree of Axenfeld-Rieger anomaly(ARA) (Online Mendelian Inheritance of Man 601631), a clinically and geneticallyheterogeneous, autosomal dominantly inherited disorder associated with anteriorchamber abnormalities and glaucoma.
Observational case-control and DNA linkage and screening studies.
Affected (10 cases) and unaffected (5 controls) members of a familywith ARA.
Clinical characteristics of ARA were documented by history or physicialexamination of symptomatic individuals. With their informed consent, a bloodsample was collected from each of 10 affected and 5 unaffected family members.DNA was tested for linkage to the IRID1 locus atchromosome 6p25, a known locus for ARA/Rieger syndrome. A candidate gene previouslymapped at this locus, FOXC1, was screened for mutationsin cases and controls.
Main Outcome Measure
Linkage of the ARA phenotype at the 6p25 locusand mutation detected in FOXC1.
Direct sequencing of FOXC1 detected a new mutation, T272C, that segregated with the ARA phenotype in this familyand was not detected in DNA from family members without ARA. This mutation,a T→C transition, is predicted to result in a change of isoleucine tothreonine (Ile9lThr) in a highly conserved location within the first helixof the forkhead domain.
Characterization of the FOXC1 mutation in familymembers with ARA furthers our understanding of the molecular origin of developmentalglaucoma and other anterior segment disorders.
Axenfeld-Rieger anomaly (ARA); (Online Mendelian Inheritance of Man601631) is a clinically and genetically heterogeneous disorder with an autosomaldominant mode of transmission. Clinically, ARA is characterized by iris stromalhypoplasia, prominent Schwalbe line, adhesion between the iris and Schwalbeline, microcornea, corneal opacity, and increased intraocular pressure (IOP)that leads to glaucoma in about half of the cases.1- 4 Whennonocular clinical features are present, the disorder is named Rieger syndrome(RS). In addition to the anterior segment anomalies just described, patientswith RS often have maxillary hypoplasia, dental anomalies,5,6 umbilicalhernia,7 and/or hypospadias. More rarely, theymay have hydrocephalus, hearing loss, cardiac and kidney abnormalities, andcongenital hip dislocation in addition to ocular abnormalities.
Axenfeld-Rieger anomaly is genetically as well as clinically heterogeneous.Many chromosomal aberrations involving chromosomes 4, 6, 9, 13, 18, and 21have been identified in patients affected with ARA,8 andlinkage studies have identified at least 3 loci for the abnormalities in ARA.The first locus, at chromosome 4q25, which displayed the ARA phenotype, wasmapped in 1992 by Murray et al9 and named RIEG1. Subsequently, the RIEG1 abnormalitywas identified as a mutation in the PITX2 gene ina patient with RS.10PITX2 is a bicoid homeobox gene that is expressed in the anterior structuresof the eye and regulates the expression of other genes during embryonic development.To date, 9 mutations of the PITX2 gene have beenreported,11 all resulting in various anteriorsegment disorders such as RS,10 iris hypoplasia,12 iridogoniodysgenesis syndrome type 2,13 andPeters anomaly (or anterior chamber cleavage syndrome).14
A second locus, for the RS phenotype labeled RIEG2 (Online Mendelian Inheritance of Man 601499), was linked at chromosome13q14 by Phillips et al15 in 1996 using a large4-generation family.The gene responsible for RS at this locus has not yetbeen identified.
A third locus for ARA has been mapped to chromosome 6p25 at the IRID1 locus.16 Subsequently,mutations in the forkhead box C1 (FOXC1) gene wereidentified in some patients with ARA by Mears et al16 andNishimura et al.17FOXC1 (previously called FKHL7 or FREAC3) is a member of the forkhead winged/helix transcription-factorfamily and is a monomeric DNA-binding protein consisting of 553 amino acidsthat is encoded by a single exon of 1659 base pair. More than 100 proteinsencoding this evolutionarily conserved domain have been identified so farin species ranging from yeast to man.18 Includingthe present study, 15 mutations of the FOXC1 genehave been reported to date, all resulting in a variety of anterior segmentdisorders (Table 1).16,17,19,20 Onearticle published in 2001 described a chromosomal duplication involving 6p25in 2 families.19 In both families, the duplicatedregion contained FOXC1, FOXF2, and FOXQ1, previously known as HFH1, all of whichare genes of the forkhead family.19 Moreover,a chromosomal duplication involving the 6p25 region, including FOXC1, has been reported in a large pedigree with iris hypoplasia andglaucoma.21 We studied the clinical characteristicsof ARA, tested linkage at 6p25, and screened for mutations the FOXC1 gene in an 18-month-old girl with ARA we have followed up sincebirth and in members of 4 generations of the proband's family.
Clinical evaluations included history, examination of the anterior andposterior chambers (funduscopy), testing of visual acuity and the visual fields,and measurement of IOP. For genotype analysis, a 20-mL sample of blood wascollected in an EDTA-coated tube from each family member who was availablefor testing and who gave informed consent. DNA was isolated and analyzed usingstandard techniques, as follows. Oligonucleotide primers were obtained fromthe Centre de Recherches du Centre Hospitalier et Universitaire Laval (CHUL),Quebec, Quebec.
Six (CA)n microsatellite markers,including 4 markers (Généthon, Paris, France), identified byDib et al22—AFMa350zc9 at D6S1600, AFM092xb7 at D6S344, AFM205xh4m at D6S1617, and AFM088yh3 at D6S1685,and 2 new markers, AM01 and CA43, developed by searching the Human GenomeWorking Draft sequence for (CA)n and (TG)n repeats—were tested for linkage at 6p25.
Polymerase chain reaction (PCR) amplification was performed with 100-ngDNA to which was added 200nM each of primer; 200µM each of [α-33P]-deoxyadenosine triphosphate ([α-33P] (dATP), 2′-deoxycytidine-5′-triphosphate(dCTP), deoxyguanosine triphosphate (dGTP), and deoxythymidine triphosphates(dTTP); 1×PCR buffer: 10mM Tris (hydroxymethyl) aminomethane (pH 9.0at room temperature); 50mM potassium chloride; 1.5mM magnesium chloride; 0.1%Triton X-100; and 0.001% gelatin in water to achieve a total sample volumeof 40 µL. Each reaction was overlaid with 25 µL of light mineraloil to prevent evaporation.
Polymerase chain reaction specificity was increased by including a "hot-start"step in which samples were denatured for 5 minutes at 95°C. After thehot-start step, l U of Thermus aquaticus DNA polymerasein 1×PCR buffer was added to achieve a final PCR sample volume of 50µL. Each sample was put through 35 cycles of denaturing at 95°Cfor 30 seconds, annealing at 55°C for 30 seconds, and extending at 72°Cfor 30 seconds.
Polymerase chain reaction products were separated on 6% denaturing polyacrylamidegels. After electrophoresis, the gel was transferred to a positively chargednylon membrane (Boehringer-Mannheim, Quebec, Canada) and hybridized with adigoxigenin–11-2′,3′–dideoxy-uridine-5′-triphosphate(DIG-11-ddUTP) 3′ end-labeling CA probe, according to the instructionsin the DIG System User's Guide for filter hybridization(Roche Diagnostic, Basel, Switzerland). The probe was made chemoluminescentwith Lumigen PPD (Roche, Montreal, Quebec), which is the 4-methoxy-4-(3-phosphatephenyl)spiro-(1,2-dioxetane-3,2′-admantane)substrate for alkaline phosphatase, before being exposed to the single emulsionfilm, BioMax MR-1 (Eastman Kodak Co, Rochester, NY) that provides maximumresolution and sensitivity for 12 to 24 hours.
For the AMO1 (forward, S′-CTGGTAAGGAGGGTTGAGG-3′; reverse,5′-AGTTCCAATAGTCAACTTGCC-3′) and CA43 (forward, 5′-AGGTGGAAACAACTCACG-3′;reverse, 5′-AGTGTCCACAAGGTGCAT-3′) markers, PCR amplificationwas performed with 50 ng of DNA to which was added 200nM of each primer; 200µMeach of dCTP, dGTP, and dTTP; 10µM of dATP; 1.5 µCi of deoxyadenosine5′[α-35S]-thiotriphosphate, triethylammonium salt–dATP;and 1×PCR buffer to achieve a total sample volume of 15 mL. Each reactionwas overlaid with 25 µL of light mineral oil to prevent evaporation.
Polymerase chain reaction specificity was increased through a hot-startstep (denaturing for 5 minutes at 95°C) before 1 U of T aquaticus DNA polymerase, in 5 mL of 1×PCR buffer was addedto reach a final volume of 20 µL. As a control for the determinationof allele size, the Centre d'Etude du Polymorphisme Humain, Paris, France,control DNA 134702 sequence was amplified and loaded on the gel together withthe patient's DNA sequences.
Betaine (Sigma-Aldrich Co, St Louis, Mo) was used for amplificationof DNA sequences. Oligonucleotide primers used for DNA sequencing were obtainedfrom the Centre de Recherches du CHUL and from Synovis Life Technologies Inc,St Paul, Minn, and are given in Table 2.
For each analysis, 100-ng DNA was added to a solution containing 200nMeach of primer; 200µM each of dATP, dCTP, dGTP, and dTTP; 1×PCRbuffer (10mM Tris; pH 9.0 at room temperature); 50mM potassium chloride; 1.5mMmagnesium chloride; 0.1% Triton X-100; and 0.01% gelatin), and 1.2M betaine.The reaction was overlaid with 25 µL of light mineral oil and denaturedfor 5 minutes at 95°C (the hot-start step). Then 1 U of T aquaticus DNA polymerase, 1×PCR buffer, was added for a finalPCR reaction sample volume of 50 µL. The sample was put through 35 cyclesof denaturing at 95°C for 40 seconds, annealing at 57°C for 50 seconds,and extending at 72°C for 50 seconds.
All PCR products were purified using QIAGEN columns (Mississauga, Ontario)(ie, convenient spin-columns that provide selective binding for small DNAmolecules) prior to sequencing. The mutation described here was first identifiedby manual sequencing using the dideoxyribonucleotide 5′-[α-33P]-triphosphates, triethylammonium salts–radiolabeled TerminatorThermosequenase Cycle Sequencing Kit (Amersham-Pharmacia, Mississauga). Familymembers were subsequently analyzed by automated sequencing using an ABI PRISM3700 DNA Analyzer (Perkin-Elmer Applied Biosystems, Foster City, Calif). Automatedsequencing was performed on both strands.
The PedCheck software program was used to identify genotype incompatibilitiesin linkage analysis.23 Two-point analysis wasthen performed using the MLINK option of the linkage package.24 TheARA/RS phenotype was modeled as an autosomal dominant trait with penetranceof 90% in persons from birth to 110 years old. Phenocopy rate was estimatedat 0.001%. Recombination rates were assumed to be equal in male and femalesubjects. Because the frequencies of marker alleles in this population areunknown, the frequencies for each marker were set at 1/N where N was the number of alleles observedin the pedigree.
Based on clinical information, 13 members of this pedigree were consideredto have ARA (Table 3). None ofthem displayed nonocular symptoms associated with RS. DNA analysis was completedfor 15 members (10 with and 5 without ARA) (Figure 1). As depicted on the Figure1, segregation of the disorder was clearly autosomal dominant with5 affected males and 8 symptomatic females as well as presence of male-to-maletransmission.
In the proband (IV:1), a posterior embryotoxon (PE) was visible in theclear zone of each cornea (Table 3)and both irises were abnormal (IA in Table3). On echography, the axial length measured 20 mm and the cornealdiameter measured 12 mm in both eyes. The IOP averaged 13 mm Hg in both eyesand reached maximal values of 26 and 27 mm Hg (Table 3).
The proband's 27-year-old father (III:4) was seen with Rieger-type anomalies.Although his visual acuity was 20/20 in both eyes, both eyes had severe PEand corectopia and the iris of the right eye consisted only of a temporalatrophic zone. However, there was no polycoria and the IOP and fundus werenormal in both eyes.
The proband's 24-year-old aunt (III:3) had a moderate PE in each eyebut normal IOP.
The proband's 50-year-old grandfather (II:3) had first been examinedby an ophthalmologist at the age of 25 years because of progressively decreasingvisual acuity over several months. Slitlamp examination at that time revealedsevere bilateral temporal PE, anterior synechiae extending over a 200°area, and iris atrophy. Funduscopy showed glaucomatous atrophy of the opticnerve in the right eye and glaucomatous papillar excavation in the left eye.Visual field testing showed a Bjerrum scotoma in the left eye and a smallnasal islet scotoma in the right eye. This individual underwent bilateraltrabeculectomy in 1973. At the time of the current study, his visual acuitywas 20/20 OD and 20/25 OS. The right eye had tubular retraction and a temporalislet scotoma as well as the nasal islet scotoma present before surgery. Intraocularpressure was 15 mm Hg bilaterally.
Individual II:2, a 58-year-old man, has congenital ARA. He underwentsurgery twice (at ages 30 and 42 years) for bilateral glaucoma. His visualacuity is light perception OD and 20/20 OS with alteration in the visual field.Both eyes have transparent cornea, PE, corectopia (more pronounced in theright eye), and corneo-iridal adhesions that may be responsible for the corectopia.Intraocular pressure remains difficult to control medically, even after 2operations on each eye for glaucoma.
Individual II:6 is a 45-year-old woman who has been followed up sinceher childhood for trabeculo-irido-corneal dysgenesis with corectopia. Shehas polycoria and a history of retinal detachment in the right eye. The lefteye does not have polycoria but does have areas of iris atrophy associatedwith a PE and posterior synechiae of the Rieger type. Her best–correctedvisual acuity is 20/25 OD and light perception OS. Although IOP reached maximaof 30 and 26 mm Hg, respectively, it has been normal in both eyes since surgeryfor glaucoma.
Individual III:7 has congenital bilateral glaucoma for which he underwentsurgery at the age of 18 months. His visual acuity is 20/20 OD and 20/30 OS.He has bilateral irido-corneal dysgenesis with PE, goniodysgenesis, and corectopiaaggravated by filtering procedures performed during his childhood. His IOPis being maintained in the normal range by topical β-adrenergic blockingmedication and funduscopy shows no excavation of the papillae.
Individuals II:8, II:9, III:10, and III:11 all have clear corneas withPE and goniodysgenesis without glaucoma. Individuals I:1 and I:3 were deadat the time of the study. They both were reported to be blind owing to glaucoma.
Examination of 6 microsatellite markers (D6S1600, AMO1, D6S344, CA43, D6S1617, and D6S1685) flanking the FOXC1 gene at locus 6p25 showed linkage of this locus with ARA in this family. Nine ofthe 10 affected individuals shared a complete haplotype for markers spanningthe region when a recombination event occurred between markers D6S344 and CA43 in affected individual FR021.A maximum lod score (Zmax) of 2.82 at a recombination fraction (thetas;)of 0.00 was observed with marker D6S344. With referenceto BAC 118B18 of the working draft sequence, D6S344 islocated approximately 10 kb from FOXC1.D6S1617 also had a positive lod score of 2.39 at a thetas; of 0.00.
The maximum lod score of 3.00 required to indicate significant linkagewas not reached in this study because the family size was small and some ofthe markers were not very polymorphic. Nevertheless, we judged the lod scoresto be high enough to look for an ARA-associated mutation in FOXC1.
The FOXC1 genes of 2 affected and 2 unaffectedfamily members were sequenced manually. This direct DNA sequencing of FOXC1 showed a T272C alteration in 1 allele of FOXC1 in both of the affected patients. The association between thismutation and the ARA phenotype was confirmed by the results of automatic sequencingof the FOXC1 gene for all 15 family members fromwhom blood samples had been obtained. Indeed, the T272C alteration was presentin all 10 family members with ARA and absent in the 5 unaffected members.No sequence variation was observed at this position in FOXC1 in 54 healthy French-Canadian individuals representing a totalof 108 chromosomes sequenced. The FOXC1T272C mutationis predicted to result in a change of an isoleucine to threonine at codon91 of the polypeptide, a highly conserved amino acid of the first helix ofthe FOXC1 forkhead domain.
The forkhead/winged helix transcription factors are characterized bya 100–amino acid, monomeric DNA-binding domain and play critical rolesduring early embryonic development, cell differentiation and specialization,tumorigenesis, and tissue-specific gene expression in both vertebrates andinvertebrates.25 To date, more than 100 forkheadfamily genes have been cloned and characterized in species ranging from yeastto man.
Fourteen of the 15 mutations of FOXC1 reportedto date affect the forkhead box. Missense mutations occur directly in theforkhead domain, whereas frameshift mutations caused by insertion or deletiontake place 5′ of the forkhead box and result in truncated proteins withmissing or abnormal forkhead domains.
To date only 1 mutation causing anomalies of the anterior chamber hasbeen found outside the forkhead box. This mutation, a 1–base pair deletionof the nucleotide 1512, is located in the C-terminalregion. In addition to these mutations, distinct duplications encompassing FOXC1 has recently been reported in 3 different familieswith iris hypoplasia and glaucoma.19,21
Most of the FOXC1 mutations described so faremphasized the critical role of the forkhead domain for DNA binding and nuclearlocalization. Nishimura et al19 also reporteda mutation 3′ of the forkhead box, which suggested the presence of functionallyimportant elements at the end of the FOXC1 protein. This assumption was enforcedby the fact that the DNA sequence near the end of the FOXC1 gene is conserved in man, mouse, rat, Xenopus species,and chicken, and by in vitro studies showing that the C-terminus of the FOXC1protein contained an activation domain. In addition to missense and frameshiftmutations affecting FOXC1, it seems that underexpressionor overexpression (as with duplication of the gene) of FOXC1 can cause defects in the anterior chamber of the eye. For example,Smith et al26 demonstrated that haploinsufficiencyof the transcription factors FOXC1 in the mouse (Foxc1+/−) resulted in histologically evident anterior segment abnormalitiesin every affected mouse. Moreover, Foxc1 homozygousmutants (Foxc1−/−) died during the perinatalperiod, indicating that this gene plays a key role in embryonic development.
In the 10 patients with ARA we sequenced, we found a missense mutation(T272C) in the first helix of the forkhead domain of FOXC1. The alignment of amino acids in forkhead domain proteins (Table 4) indicates that the isoleucineat codon 91 in the forkhead domain is highly conserved.25 The mutationwe report herein occurs in this conserved motif. This suggests that the Ile9lThrmutation we describe occurs in a nuclear localization sequence of FOXC1 and contributes to anterior chamber abnormalities by hinderingnuclear localization. Saleem et al,27 recentlystudied the effect of 5 missense mutations of the winged/helix domain foundin patients with AR malformations. Although these authors did not investigatethe Ile91Thr variation, they demonstrated that mutations in the FOXC1 forkhead domain reduced stability, DNA binding, or transactivation,all causing a decrease in the ability of the polypeptide to transactivategenes. Further experimentation should reveal the exact mechanism(s) by whichthe Ile91Thr mutation alter FOXC1 transactivation.
The most important feature of ARA is the high risk of developing glaucoma,which causes progressive narrowing of the visual field and, when uncontrolled,blindness.27 It was estimated, for 2000, thatalmost 6 million people worldwide have developed glaucoma.28 Glaucomais often insidious and rarely hurts, and it is because its severe consequencesmay be minimized if it is diagnosed early, it becomes important to understandthe genetic bases of disorders of the anterior chamber of the eye. Our resultsserve to improve the understanding of the role FOXC1 playsin developmental glaucoma and expands the knowledge of the genetic causesof anterior segment disorders.
Correspondence: Bruno Mortemousque, MD, Centre Hospitalier et UniversitaireBordeaux, Service d'Ophtalmologie, Place Amélie Raba-Léon, 33076Bordeaux, France.
Submitted for publication September 5, 2002; final revision receivedMarch 25, 2004; accepted May 10, 2004.
This study was supported by grants MOP-13428 from the Canadian Institutesfor Health Research Ottawa, Ontario, and 548 from the Canada Foundation forInnovation, Ottawa.
Dr Raymond is a Fonds de la Recherche en Santé du QuébecNational Investigator.
We thank all the families and patients who participated in this study.