Characterization and Prevalence of PITX2 Microdeletions and Mutations in Axenfeld-Rieger Malformations

Abstract

purpose. Mutations of the homeodomain protein PITX2 produce Axenfeld-Rieger (AR) malformations of the anterior chamber, an autosomal dominant disorder accompanied by a 50% risk of glaucoma. Twenty-nine mutations of PITX2 have been described, with a mutational prevalence estimated between 10% and 60% in AR. In the current study, the possible role of altered PITX2 gene dosage in the etiology of AR was investigated. Gross gene deletions and duplications should alter PITX2 activity analogously to hypomorphic and hypermorphic mutations, respectively.

methods. Sixty-four patients with AR, iridogoniodysgenesis (IGD), iris hypoplasia (IH), or anterior segment dysgenesis (ASD) were screened for PITX2 mutations by sequencing. PITX2 gene dosage was concurrently examined in these patients by real-time quantitative PCR. Microsatellite markers were used to map 4q25 microdeletions at a contig scale, as well as for haplotype analysis in an extended AR kindred. An additional 27 patients with other assorted ocular phenotypes were evaluated by similar methods, amounting to a total of 91 cases analyzed.

results. Three novel mutations of PITX2 (4.7%) were identified among 64 patients with AR, IGD, IH, or ASD. Deletions of PITX2 were as frequent as mutations in our sample. Chromosome 4q25 microdeletions were physically mapped relative to several microsatellite markers in each patient. Cosegregation of AR and a PITX2 deletion was demonstrated in an extended kindred.

conclusions. Point mutations and gross deletions of PITX2 appear to produce an equivalent haploinsufficiency phenotype. Quantitative PCR is an efficient means of detecting causative PITX2 deletions in patients with AR and may increase the detection rate at this locus.

A xenfeld-Rieger (AR) malformation comprises a group of autosomal dominant clinical disorders affecting anterior eye structures derived from constituents of the embryonic neural crest (NC). ¹ Classic ocular features of AR include iridocorneal synechiae, iris hypoplasia, corectopia, polycoria, and/or prominent Schwalbe’s line. ^{2 3 4} AR is often associated with small, malformed, or missing teeth and/or excessive periumbilical skin. ^{5 6 7} The phenotype occasionally includes cardiac septal defects and/or sensorineural hearing loss. ^{8 9} AR is situated at the severe end of a continuum of anterior segment phenotypes that includes iridogoniodysgenesis (IGD) and iris hypoplasia (IH). ^{10 11} In approximately 50% of patients with AR, secondary high-tension glaucoma develops due to obstruction of the aqueous drainage pathway of the iridocorneal angle. ¹ Genetic causes of AR include mutations, deletions, or duplications of the forkhead-related transcription factor FOXC1, as well as mutations of the homeodomain (HD) protein PITX2. ^{12 13 14 15 16 17} AR due to deletion of the paired-box transcription factor PAX6 has been reported in a single case. ¹⁸  

Pituitary homeobox 2 (PITX2) is a paired-bicoid HD protein that is expressed during ocular development. ^{12 19} There are four differentially expressed isoforms of PITX2 that vary in their N-termini but share common HD and C-terminal sequences. Of these, PITX2a and/or PITX2b, as well as PITX2c, are expressed in the embryonic periocular mesenchyme in mouse. ^{12 20 21} The periocular mesenchyme is a neural-crest-derived, migratory cell population that contributes widely to development of anterior segment structures including the iris, cornea and trabecular meshwork, any of which are potentially affected in AR. ^{1 22} The variable systemic anomalies in AR are also found in a subset of PITX2-expressing tissues, as additional domains of PITX2 expression exist in the umbilicus, dental epithelium, heart, abdominal organs, and limb buds. ^{12 21 23}  

Recent work has placed PITX2 in a signaling cascade that appears to modulate cell proliferation, differentiation, and morphogenesis. PITX2 induces cell proliferation in response to β-catenin signals raised during Wnt signaling in C2C12 cells. ²⁴ In contrast, PITX2 expression has antiproliferative effects in HeLa cells, causing arrest at the G₀/G₁ checkpoint. ²⁵ Activation of the small guanosine triphosphatases (GTPases) RhoA and/or Rac appears necessary for PITX2 signaling accompanied by cytoskeletal changes in both cell lines. ^{24 25} Transfection of a dominant negative RhoA mutant into perfused human anterior segment cultures increases aqueous outflow facility, suggesting that this PITX2 pathway may be essential for intraocular pressure homeostasis in the prevention of glaucoma. ²⁶  

Twenty-nine mutations of PITX2 have been described to date, nearly all of which either encode a truncated product or produce point alterations of the HD ^{10 11 12 27 28 29 30 31 32 33} (Richards JBB, et al. IOVS 2001;42:ARVO Abstract 3041). Mutations of PITX2 have been associated with various anterior segment dysgeneses (AR, IGD, and IH) ^{10 11 12 28 29 30 31 32} (Idrees F, et al. IOVS 2002;43:ARVO E-Abstract 3402) and, rarely, with additional features resembling Peters’ anomaly. ^{27 28} Chromosomal aberrations involving the PITX2 locus have also been described, both in the form of cytologically visible deletions and as a result of translocations involving 4q25. ^{33 34 35 36 37 38} Such cases support haploinsufficiency as a mechanism for PITX2-related ocular maldevelopment. An increased PITX2 copy number may also be pathologic, as duplication of a distal region of 4q2 (including 4q25) has been noted in one patient with hypoplastic left heart. ³⁹ A single hypermorphic allele of PITX2 has been identified in AR, suggesting an upper limit for PITX2 activity in normal ocular development. ²⁹ To the best of our knowledge, no subcytologic deletions of this locus have yet been described. 

We conducted a mutational screen and real-time quantitative PCR (qPCR) analysis of PITX2 in a panel of 64 unrelated clinical cases of AR, IGD, IH, or ASD, as well as a secondary panel of 27 cases of other anterior segment malformations and/or glaucoma. We identified three novel mutations of PITX2, as well as three gross deletions, in patients with AR. In our sample, deletions and mutations of PITX2 were equally prevalent, together comprising only 10% of cases of anterior segment dysgenesis. These findings agree with previous estimates supporting a limited involvement of PITX2 in AR ²⁶ and provide a second avenue for diagnostics at this locus. 

Methods

Human Subjects

Written, informed consent was obtained from all studied subjects before enrollment in the study. The use of human subjects in this study was approved by the University of Alberta Health Research Ethics Board, in accordance with the tenets specified within the Declaration of Helsinki. DNA banking at The Hospital for Sick Children (HSC; cases 4 and 5) was performed in accordance with the HSC Research Ethics Board’s approved consent policies. 

Sequence Analysis

Sequencing was performed by PCR followed by a combination of manual ³³P-labeled terminator (Amersham Biosciences, Little Chalfont, UK) and labeled primer fluorescent sequencing, (Li-Cor, Lincoln, NE), using previously published amplimers. ¹²  

Realtime qPCR

We used a fluorogenic PCR assay (TaqMan; Applied Biosystems [ABI], Foster City, CA) to quantitate the relative abundance of an exon IV target sequence 3′ of the PITX2 HD in each sample on our panel. All equipment and reagents indicated were obtained from ABI. 

Each reaction contained 10 picomoles of forward and reverse primers (CAGTTCAATGGGCTCATGCA, CGGCCCAGTTGTTGTAGGAA) and 4 pmol of a dual-label PCR (TaqMan; ABI) probe (VIC-CCCTACGACGACATGGTACCCAGGC-TAMRA). A commercial assay (TaqMan; ABI) for quantitation of the human connexin (Cx)40 locus was also included for normalization. Each sample was amplified in triplicate, 15-μL reactions containing 25 ng of lymphocyte DNA. Reactions were cycled in a thermocycler (model 9700; ABI) with the 2× assay kit as per the manufacturer’s instructions (TaqMan; ABI). Each 384-well plate contained triplicate reactions of two unrelated normal samples and a DNA-free control. Output data were analyzed by a relative quantification method. Briefly, rate of change in fluorescence (ΔF) in each well, during each cycle, at each reporter wavelength was charted. Threshold values (T) of ΔF were selected for the PITX2 and Cx40 probes, so that a graph of ΔF was linear in all wells as ΔF approached T. The threshold cycle (C_t) at which ΔF reached T was determined for each reporter in each well. A C_t ratio (C_t[PITX2]/C_t[Cx40]) was generated for each reaction. A graph of C_t[PITX2] versus C_t[Cx40] yielded the expected linear correlation on a panel of 50 unrelated normal control samples. Deletions formed a distinct line in which the normalized C_t[PITX2] was one cycle greater than that of normal control samples. We compared the C_t ratios obtained from each sample with both PITX2-deleted and normal control samples by using a t-test statistic. A ratio of these values represents the relative likelihood that a patient’s sample has one rather than two copies of PITX2. The PITX2 target amplicon is contained within exon IV, hence sequence analysis ruled out the potential confound of an underlying probe or primer site mutation. The qPCR dataset and normal control data. 

Microsatellite Markers

We examined eight microsatellite markers in the immediate genomic vicinity of PITX2 with the following amplimers: 320d3-1 (CAGAGGTAGGGTCCAGGTTG/TGCAGAGCAATTCCTGTACCT, T_A = 60°C), 320d3-2 (TCAGTTGCATGAATGGAGGA/ACCCTGGGACTTTGATGGAT, T_A = 60°C), 320d3-6 (TGTTTGGGTTCCCCAAGTAT/CGAGATTGCCCCACTAAACC, T_A = 60°C), 320d3-7 (TGGGTGACAGAGCAAGACAA/GGCTTATCAGGAGGGTCCA, T_A = 60°C), 320d3-14 (AAACACAAAGCCTCAACAGGA/AAACACAAAGCCTCAACAGGA, T_A = 53°C), 320d3-15 (TGAATGGATAGCCTTCTCAG/AAAGCACCAAGGACAACCAG, T_A = 52°C), 320d3-16 (GAAATGAATGGGTTCAGTGGA/TCTGCAACATAAGTGGAGTCTCA, T_A = 50°C), and 320d3-17 (TCCAGAGAGTGGGTTTCTGA/GCCTGGGTGACAAGAACAAG, T_A = 52°C). Each of the above reactions were performed in 2 mM MgCl₂ with 35 cycles of 30 seconds 95°C, 30 seconds T_A, and 30 seconds 72°C. Heterozygosity of each marker was estimated by genotyping six unrelated normal samples. We also genotyped affected individuals with the established markers GATA10G07, D4S2361, D4S1647, D4S2623, D4S2301, D4S2945, D4S193, D4S406, D4S1651, D4S2394, D4S1644, and D4S1625. Products were size separated on 6% denaturing polyacrylamide gel after PCR incorporation of α³⁵S-dATP (Amersham Biosciences). Microsatellite gels were optically scanned, and the autoradiographs. 

Results

To estimate the contribution of PITX2 mutations and deletions in AR, we screened PITX2a coding regions and dosage in a cohort of 91 cases with mixed clinical phenotypes involving dysgenesis of the anterior segment and/or glaucoma. Our primary (anterior segment mesenchymal dysgenesis) panel contained 64 cases with various PITX2-associated anterior ocular phenotypes: AR (39 cases), IGD (19 cases), IH (2 cases), or an unspecified anterior segment dysgenesis (ASD; 4 cases). The remainder of our patient cohort contained cases of either ectodermal anterior eye malformations, or simple glaucoma, as follows: Peters’ anomaly (seven cases), juvenile open-angle glaucoma (two cases), congenital glaucoma (two cases), and primary open angle glaucoma (one case). We screened a further five patients who had a variety of other ocular findings: sclerocornea (one case), congenital cataract with glaucoma (one case), corneal endothelial dystrophy (one case), short stature, joint hyperextensibility, ocular depression, Rieger anomaly, and teething delay (SHORT syndrome, Mendelian Inheritance in Man [MIM] 269880, one case), or aniridia plus microdontia (one case). 

We examined coding regions and splice junctions present in the PITX2a transcript. Comprehensive screening of the 91 individuals in our panel identified three novel mutations (Fig. 1) that are predicted to disrupt severely the primary structure of all four PITX2 isoforms. For clarity, nucleotide and codon positions given in this manuscript are relative to the initiation codon of PITX2a (GenBank: NM153427; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). The four exons screened are similarly numbered according to the PITX2a transcript. 

Patient 1 displayed full ocular manifestations of AR (IH, prominent Schwalbe’s line, corectopia, and glaucoma). No systemic findings were evident in this individual, who was found to harbor a heterozygous deletion of nucleotide G114 in PITX2 exon III. This frameshift mutation is predicted to ablate most of PITX2 beginning at Q39, the first residue of the HD. A large nonsense peptide of 114 further amino acids is encoded. Complete loss of the HD, coupled with the likelihood of protein or transcript instability, together suggest that this allele may be a functional null. 

Patient 2 had severe AR malformations (bilateral polycoria; displaced, pinpoint left pupil; and irregular left iris contour) in conjunction with underdeveloped maxilla and redundant umbilical skin. A G]→T transversion at the invariant −1 position of the PITX2 exon III splice acceptor site was identified in this patient, abolishing helices I and II of the HD, which are encoded by exon III. This mutation is also therefore likely to constitute a null allele. An adjacent splice-site mutation with a similar phenotype has been reported. ²⁸  

Patient 3 was a member of a four-generation kindred whose ocular findings have been described (1, family of case 14). Findings in both patient 3 and an affected grandson included iridocorneal adhesions, corectopia, polycoria, IH, iris atrophy, prominent Schwalbe’s line, and glaucoma. Patient 3 also had endothelial lesions of the cornea (guttata). Gonioscopy demonstrated an abnormal angle. Sequencing identified a single-base deletion of nucleotide C416 within exon IV of PITX2. This mutation causes a frameshift from T139 onward, encoding 15 mutant residues followed by a premature stop codon. 

Realtime qPCR and Microsatellite Studies of PITX2 Dosage

Realtime qPCR analysis of the same panel of 91 individuals identified three samples that met the statistical criteria consistent with hemizygosity of the PITX2 target amplicon (see the Methods section and supplementary data for details of the analysis). To confirm and determine the approximate extent of each deletion, we developed eight microsatellite markers flanking the PITX2 locus. Genotyping of these markers, and of 12 established markers on the long arm of chromosome 4 markers (see the Methods section) produced a contig-level map of each deletion (Fig. 2 and supplementary data). 

Patient 4 had a variant AR phenotype in which corectopia, prominent line of Schwalbe, and iridocorneal adhesions are accompanied by the additional feature of bilateral anterior polar cataracts. An unaffected brother had normal PITX2 gene dosage, as measured by qPCR. The maximum deleted interval in patient 4 is bounded by the heterozygous markers 320d3-2 and D4S2394, which flank a region extending 19.3 Mb telomeric of PITX2. Because 320d3-2 is only 4.1 kb centromeric of PITX2, this individual may carry a partial deletion potentially sparing the extreme 3′ end of the gene. 

Patient 5 had a pronounced AR phenotype with ocular findings including bilateral iris atrophy, posterior embryotoxon, iridocorneal adhesions, miosis, corectopia, and left polycoria. Other features included pointed and maloccluded teeth and redundant periumbilical skin. This patient had markedly short stature, being, at age 6, only 3 inches taller than a 2-year-old sibling. Patient 5 had deep-set eyes and a prominent mandible, which are also feature of SHORT syndrome. His hearing and speech are normal. This patient’s deletion spans a region containing PITX2, with both breakpoints located in the 1.24-Mb interval between D4S2623 and D4S1651. This interval contains PITX2, human epidermal growth factor (EGF), long-chain fatty-acyl elongase (LCE), glutamyl aminopeptidase (ENPEP), and two predicted expressed sequence tag (EST) transcripts, denoted ENSG00000164092 and ENSG00000168999 in the Ensembl database (http://www.ensembl.org). ⁴⁰ An unrelated SHORT patient sample displayed normal PITX2 dosage. It is at present unclear whether patient 5’s phenotype constitutes a bona fide case of SHORT syndrome. Both pituitary anomalies and short stature due to pituitary growth hormone insufficiency have been described in patients with ocular and dental findings of AR, implying that short stature may be a rare feature of the AR phenotype that occurs at reduced penetrance in some pedigrees. ^{1 5 41}  

Patient 6 was an affected member of a large pedigree in which AR cosegregates with a hemizygous deletion of PITX2. Familial haplotyping (Fig. 3 and supplementary data) delineated a maximal deleted region of 417 kb, bounded telomerically by 320d3-6 and centromerically by D4S2945. This interval includes PITX2, ENPEP, and ENSG00000164092 and does not include EGF or LCE. 

Discussion

PITX2 Mutations

Our sequence analysis of 91 patients with anterior segment dysgeneses and/or glaucoma identified three novel, truncating PITX2 alleles that are likely to disrupt severely the DNA binding and/or regulation of the mutant protein. Each mutation and deletion presented herein is likely to affect all four known PITX2 transcripts, as exons III and IV are common to each. With the exception of patient 3, each patient mutation described is predicted to abolish DNA-binding functions of all isoforms of the affected allele. Although the mutant PITX2 allele of patient 3 encodes a normal HD, truncation of the C-terminus may lead to reduced stability of the product peptide. Alternatively, this allele could produce a stable protein with altered autoregulatory and/or protein-interaction functions. The C-terminal tail of PITX2 contains a conserved 14-amino-acid OAR (Otx, aristaless, rax) domain that is common to multiple HD proteins in several species. ¹² This region has been proposed to mediate synergistic protein-protein interactions with PIT-1, as well as self-inhibitory interactions with the PITX2 N-terminus. ⁴² The C-terminus is essential for PITX2’s wild-type function and/or stability, as a small number of mutations sparing the HD but ablating the C-terminus have been described ^{12 28} (Richards JBB, et al. IOVS 2001;42:ARVO Abstract 3041). 

The pathologic mechanism of PITX2 haploinsufficiency in AR is currently unknown, as it remains to be determined which ocular genes are subject to regulation by PITX2. PITX2 target genes plausibly involved in human dental development include distal-less homeobox 2 (DLX2) and procollagen lysyl hydroxylase 1 (PLOD1) genes. ^{43 44 45 46} As DLX genes are key regulators of jaw ontogeny in mice, this pathway may be relevant to the maxillary hypoplasia of patient 2. ⁴⁷ Delineation of the ocular pathogenesis of AR requires identification of relevant PITX2 target genes in the ocular anterior segment. 

Estimates of Prevalence

Among the AR, IGD, IH, and ASD cases which composed our primary panel, 6 (9.4%) of 64 individuals carried an identifiable PITX2 deletion or mutation. The incidence of microdeletions of PITX2 was equal to that of point mutations in our sample. The prevalence of PITX2 mutations in anterior segment dysgenesis has been placed at about 10%, although estimates in smaller cohorts have ranged to as much as 60%. ^{12 28} Our findings support a limited prevalence of PITX2 mutations in AR. Moreover, these data point to the existence of a subset of PITX2-associated cases of AR not detectable through sequence analysis alone. 

Altered balance of transcription factor activities appears to be a theme in the etiology of AR. Our findings suggest that PITX2 hemizygosity produces a phenotype similar to that of PITX2 mutations, providing further support for haploinsufficiency as a general pathologic mechanism in AR. Increased PITX2 dosage can also be pathologic, as both hyper- and hypomorphic alleles of PITX2 lead to AR malformations. ²⁹ Similarly, both deletions and duplications of the FOXC1 locus produce a similar AR phenotype. ^{15 16} Studies of AR-causing alleles of FOXC1 also indicate that normal ocular development requires a strictly enforced threshold level of transactivity. ⁴⁸ Such findings imply that development of the anterior chamber involves finely tuned regulatory networks that are sensitive to even modest changes in activity. 

 Supported by Canadian Institutes of Health Research (CIHR) Grant MT12916. MAW is an Alberta Heritage Foundation for Medical Research (AHFMR) senior scholar and a CIHR investigator. MAL is supported by an AHFMR graduate studentship.

 Submitted for publication March 25, 2003; revised September 17, 2003; accepted October 22, 2003.

 Disclosure: M.A. Lines, None; K. Kozlowski, None; S.C. Kulak, None; R.R. Allingham, None; E. Héon, None; R. Ritch, None; A.V. Levin, None; M.B. Shields, None; K.F. Damji, None; A. Newlin, None; M.A. Walter, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Sequencing autoradiograph displaying novel AR-associated mutations in PITX2 exons III and IV. All nucleotide and exon positions refer to the PITX2a isoform.

Figure 2.

 

Microsatellite hemizygosity mapping of PITX2-containing deletions. A maximal deletion interval (boxed region) was defined between the most proximal centromeric and telomeric markers shown to be heterozygous in each affected individual. Haplotyping of the extended family of patient 6 (Fig. 3) was used to define further the deleted interval in this family. Nucleotide positions given are derived from the Ensembl chromosome 4 supercontig (http://www.ensembl.org). Estimated heterozygosity is provided for markers generated for this study. Diamonds: microsatellite markers; stars: position of the PITX2 qPCR probe. Individual marker genotypes are coded as follows. Open symbols: heterozygous; gray symbols: homo- or hemizygous; black symbols: null allele evident by qPCR or by haplotype analysis; N, haplotype analysis rules out deletion of this position.

Figure 3.

 

Haplotype analysis illustrating cosegregation of del(PITX2) and AR in the kindred of patient 6. Haplotype phases were inferred from microsatellite genotypes where possible. Individuals III-9 and III-13 have ambiguous haplotypes due to unavailable parental samples. Null alleles (▴) were identified on the ΔPITX2 haplotype due to aberrant segregation of the indicated markers. ( ) patient 6; N, ophthalmically examined normal individuals; ?, at-risk individuals not clinically ascertained.

Supplementary Materials

qPCR dataset - (300 KB) Excel worksheet containing PITX2 realtime PCR dataset for the patient screen and normal panel. Data format is explained within. 

Microsatellite gels - (42.9 KB) Autoradiographs of the microsatellite gels used to derive the haplotype information in Figure 3. 

Gonioscopy demonstrated an abnormal angle - (188 KB) Clinical photographs of AR malformations in patient 3. Images are of the unoperated (less severe) eye (OD). Panels A and B: Diffuse iris atrophy (note prominent pupillary sphincter) and mild corectopia. Panel C: Gonioscopy reveals an abnormal iris root insertion. Normal angle structures are not clearly distinguishable. 

Realtime qPCR analysis - (53.3 KB) Typical sample data showing simultaneous amplification of PITX2 and Cx40 in two samples, one of which has a PITX2 deletion. The vertical logarithmic scale is given in arbitrary units and represents the change in fluorescence emission at a reporter wavelength during each cycle. Reaction progress (in cycles) is represented on the horizontal axis. The threshold cycle occurs during the exponential phase of the reaction, during the cycle in which the curve crosses the red line shown. 

The authors thank patients and family members, without whose participation this work could not be attempted, and the University of Alberta Genetics Clinic for substantial help with pedigree construction and sample collection.