Genotype-Phenotype Correlations in Axenfeld-Rieger Malformation and Glaucoma Patients with FOXC1 and PITX2 Mutations

Abstract

purpose. To improve the understanding of Axenfeld-Rieger Malformation (ARM)–associated glaucoma and to determine the best glaucoma treatment for patients with ARM who have known genetic defects in FOXC1 or PITX2.

methods. Clinical data were collected from patients with diagnosed ARM, in whom we had previously identified disease-causing mutations in either the FOXC1 or PITX2 genes, by examination of patient records and use of clinical questionnaires. One hundred twenty-six patients with ARM, representing 20 different probands, with FOXC1 and PITX2 alterations were included in the study.

results. ARM-associated glaucoma is a bilateral anterior segment dysgenesis disease that affects males and females equally. Seventy-five percent of the patients with ARM who participated in this study had glaucoma that had developed in adolescence or early adulthood. Of note, the patients with nonocular findings were more likely to have PITX2 defects than FOXC1 defects. Glaucoma in only 18% of patients with either PITX2 or FOXC1 genetic defects responded to medical or surgical treatment (used solely or in combination).

conclusions. Patients with FOXC1 mutations have the mildest prognosis for glaucoma development, whereas patients with PITX2 defects and patients with FOXC1 duplication have a more severe prognosis for glaucoma development than do patients with FOXC1 mutations. In the present study, current medical therapies do not successfully lower intraocular pressure or prevent progression of glaucoma in patients with ARM who have FOXC1 or PITX2 alterations. This clinical study also provides useful diagnostic criteria to identify the gene responsible for ARM.

Axenfeld-Rieger malformation (ARM) is a rare autosomal dominant disorder that affects anterior eye structures derived from the periocular mesenchyme. ^{1 2} The diagnosis of ARM refers to a heterogeneous constellation of dominantly inherited ocular findings, ^{3 4 5 6 7 8 9} including anomalies of the anterior chamber angle and aqueous drainage structures. These patients may present with iridogoniodysgenesis (IGD; iris hypoplasia and goniodysgenesis with excess tissue in the angle and anomalous angle vascularity ¹⁰ ), iris hypoplasia (IH), corectopia (eccentric pupil), polycoria (iris tears), peripheral anterior synechiae (PAS; iridocorneal tissue adhesions traversing the anterior chamber), and posterior embryotoxon (PE; prominent, centrally displaced Schwalbe’s line) ^{11 12 13 14} (Fig. 1) . The corneal endothelium and Descemet’s membrane may also be absent in affected eyes. ¹⁵ The corneal stroma is sometimes centrally opaque perhaps due to edema associated with the permeability of the endothelial barrier. ^{3 15 16} Patients with ARM may also present with systemic malformations with incomplete penetrance and variable expressivity. ¹⁷ Nonocular features typically include facial (maxillary hypoplasia), dental (hypodontia, microdontia), and umbilical (failure of the involution of the periumbilical skin) defects. ^{18 19} In rare cases, patients with ARM may also have hypertelorism, sensory hearing loss, congenital heart defects, gastrointestinal (GI) defects and growth retardation (GR). ^{11 19 20 21}  

The changes in eye morphogenesis in ARM are highly penetrant and, anecdotally, have been associated with an approximately 50% risk of the development of glaucoma. ¹¹ Glaucoma, which can lead to blindness, is the major consequence of ARM. ^{11 19} Glaucoma in patients with ARM can develop in infancy, but more usually in adolescence or early adulthood. In rare cases, glaucoma has been observed after middle age. As a result, patients with ARM remain at risk for the development of glaucoma throughout their lives. 

ARM is genetically heterogeneous. Genetic linkage analysis has revealed five loci associated with ARM: the pituitary homeobox 2 gene (PITX2) located at 4q25, the forkhead box C1 gene (FOXC1, formally called FKHL7) located at 6p25, and two unidentified genes at 13q14 and 16q24. ^{4 5 6 7 9 22 23} ARM due to deletion of the paired-box transcription factor PAX6 has been reported in a single case. ^{23 24} In approximately 60% of patients, ARM is not caused by chromosomal anomaly or mutations of any identified gene ²³ (Mirzayans F, Walter MA, unpublished observations, 2006). 

Tight control of FOXC1 and PITX2 appears to be necessary for normal development, as either too much or too little activity of these transcription factors results in anterior segment defects and glaucoma. ²⁴ Approximately 40% of patients with ARM have mutations or duplication of FOXC1 or mutations or deletion of PITX2. ^{3 4 7 8 9 24} Mutation or deletion of PITX2 appears to produce equivalent haploinsufficiency phenotypes. ²⁴ Thirty mutations of PITX2 have been described to date that either encode a truncated product or produce point mutations into the homeodomain. ^{4 6 8 24 25 26 27 28 29 30 31} Chromosomal anomalies involving the PITX2 locus have also been described, in the forms of both cytologically visible deletions and translocations involving 4q25. ^{31 32 33 34 35 36} 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. ³⁷ A single hypermorphic allele of PITX2 (V45L in the homeodomain) has been identified in ARM, suggesting an upper limit for PITX2 activity in normal ocular development. ²⁸ Recent mouse transgenic experiments support this hypothesis as well. ³⁸ Of note, both mutations and duplication of the FOXC1 locus produce a similar AR phenotype. ^{39 40 41} Frameshift or missense mutations and telomeric and segmental deletions of FOXC1 produce a spectrum of anterior segment disorders associated with glaucoma, ^{3 7 9 22 39 42 43 44 45 46 47 48 49} whereas duplications of the 6p subtelomeric region of various extent cause the related anterior segment dysgenesis IGD. ^{41 42} These findings, taken together, suggest both lower and upper thresholds for the appropriate level of FOXC1 activity in vivo. ^{39 40} A correlation between the dosage of normal PITX2 protein and the severity of the phenotype has also been noted, as a dominant-negative mutation in PITX2 was shown to result in a more severe ocular developmental phenotype. ^{14 25} However, amounts of FOXC1 activity do not correlate with the severity of the anterior segment dysgenesis. The recent discovery that PITX2 and FOXC1 directly interact to regulate FOXC1 activity negatively provides a possible explanation for this observation. ¹ In cells expressing both FOXC1 and PITX2 proteins, PITX2 target genes are expressed, whereas FOXC1 target genes are inhibited by FOXC1–PITX2 complexes. When there are PITX2 loss-of-function mutations, PITX2 target gene expression is reduced, whereas FOXC1 target genes are inappropriately activated. 

The FOXC1 and PITX2 mutations and the PITX2 deletion found in patients with ARM (see Table 1 ) have been analyzed in our laboratory for their effects on the structure and function of these two transcription factors. ^{3 8 9 14 22 24 42 43 44 45} In this article, we describe the results of a retrospective study on a large cohort of patients with ARM in whom our laboratory identified FOXC1 and PITX2 defects to gain a better understanding of the ARM-associated glaucoma and insight into the best glaucoma treatment for patients with ARM with known genetic defects in FOXC1 or PITX2. 

Materials and Methods

This study adhered to the tenets of the Declaration of Helsinki and was approved by the University of Alberta Ethics Board. Informed consent was obtained from each subject. 

Patients

The retrospective study included 126 patients with diagnosed ARM, in whom our laboratory had identified disease-causing genetic defects in either FOXC1 or PITX2. Polymerase chain reaction (PCR) sequence analysis was used to find PITX2 mutations, PITX2 deletion, and FOXC1 mutations. ^{9 10 13 23 42 43 44 47} Fluorescence in situ hybridization (FISH) was used to identify FOXC1 duplication. ^{40 41} Of the 126 patients, we collected complete clinical data from 55 and incomplete clinical data from 45; 26 were lost to follow-up or did not wish to participate in our study. 

Clinical Data

Clinical data were collected through the examination of patients’ records and through clinical questionnaires. The patients’ records were examined for local patients. Outside Edmonton, Alberta, we sent a questionnaire to physicians who referred the patients to us. Complete ophthalmic examination information, including visual acuity (VA) with refraction, slit-lamp biomicroscopy, applanation tonometry, gonioscopy, dilated fundus examination, and photography, when appropriate, was asked for each patient. We obtained information regarding the incidence of glaucoma among patients with ARM, the age of diagnosis of ARM and of glaucoma, the distribution of ARM and glaucoma in females and males, which ocular and nonocular malformations were present, whether the ARM-associated glaucoma was present in one eye (unilateral disease) or in both eyes (bilateral disease), which glaucoma treatment(s) was used, and whether the treatment was successful in managing the ARM-associated glaucoma. 

Definitions

Glaucoma.

The diagnosis of glaucoma was based on the observation of at least two of the following criteria: glaucomatous optic disc damage, glaucomatous visual fields defects, or high intraocular pressure (≥22 mm Hg). 

Glaucomatous Visual Fields Defects.

Typical glaucomatous visual field defects are paracentral scotoma, nasal step, arcuate scotoma, and temporal and/or central island fields or a localized sensitivity decrease equal to or greater than 6 dB in at least one location of the central 10°, two locations of the central 20°, or three locations of the central 30°. ⁵⁰  

Cup-to-Disc Ratio.

Abnormalities of the optic disc are defined as excavation with the vertical cup-to-disc ratio (CDR) of 0.5 or more and asymmetric disc excavation with a difference in vertical CDR of >0.2 between the two eyes. ^{50 51}  

Intraocular Pressure.

IOP data in this study was collected before applying any treatment. IOP was measured for each eye and scored as either low (<22 mm Hg) or high (≥22 mm Hg) pressure. 

Visual Acuity.

VA was reported with the best correction in place and was classified into three groups: good (20/20–20/40), fair (20/40–20/100), and poor (≤20/200, counting fingers, hand motion, light perception or no light perception). ⁵²  

Blindness.

Blindness or visual loss was defined as VA ≤ 20/200. 

Risk Factors Associated with Glaucoma.

Factors known to be associated with glaucoma development in the general population were tested, including family history, hypertension, ischemic heart disease, myopic eye, previous ocular trauma, and topical steroid use on eyes. ^{53 54}  

Type of Glaucoma Treatment.

Medical treatment was defined as topical ocular hypotensive drugs, including the following classes of medications: β-blockers, α-2 agonists, carbonic anhydrase inhibitors, prostaglandins, miotics, and combination medication. Surgical treatment was defined as invasive procedures including conventional surgery and laser. Conventional surgery was defined as incisional procedures, such as trabeculotomy, trabeculectomy with or without adjunctive antifibrosis therapy, glaucoma drainage surgery, or cyclodestructive procedures. Laser surgeries used to treat glaucoma included laser trabeculoplasty, laser peripheral iridotomy, laser peripheral iridoplasty, and laser cyclophotocoagulation. 

Result of Treatment.

Results of treatment were assessed based on the success in reducing IOP and maintaining a good VA. Results were “not successful” when IOP increased after treatment (≥ 22 mm Hg) or when IOP was <21 mm Hg, but with poor VA. Results were “successful” when IOP decreased after treatment (IOP < 21 mm Hg) and were “stable” when IOP was maintained at the same level, without any increase in values. 

Statistical Analysis

Clinical data were tabulated and compared by using statistical analyses. For ophthalmic tests, data from both eyes were analyzed. If no significant difference was found between the two eyes, then data for the left eye were chosen for further statistical analysis. Frequencies and cross-tabulations were constructed for categorical variables. For both univariate and bivariate analyses, we used various tests, including asymptotic and exact versions, to protect against small counts. For the univariate analysis, the Pearson χ² test or z-test assessed the difference in proportions. Two-sided probabilities were reported for tests with two-sided hypotheses, and one-sided probabilities were reported for tests with one-sided hypotheses. For the bivariate analysis, we identified several situations. When both variables were nominal, we used a Pearson χ² test to check for a significant relationship between the two variables (χ² test with corresponding number of degrees of freedom). When one of the variables was nominal, the response was ordinal, and the measurements were performed on the same patient (for example, VA measured for both eyes of a patient), we used the marginal homogeneity test to assess the difference in responses between the nominal categories. ⁵⁵ When one of the variables was nominal, the response was binary, and the measurements were performed on the same patient (for example, IOP measured in both eyes of a patient), we used the McNemar test to assess the difference in responses between the nominal categories. ⁵⁵ When both variables were ordinal, we used the Jonckheere-Terpstra (JT) trend test, ⁵⁶ to assess whether there was a monotonic relationship between the two variables. To account for correlations in binary outcomes when patients were from the same family, we used the generalized linear mixed models and the procedure GLIMMIX. ^{56 57 58} Because we did not use the same patients for two comparisons, it was not necessary to make adjustments for multiple comparisons. Commercial software (Statistical Analysis System and Statistical Package for the Social Sciences; SPSS, Chicago, IL) was used for the statistical analyses. 

Results

Subjects Demographics

Our laboratory identified FOXC1 and PITX2 alterations in 20 different probands, representing 126 patients with ARM (Table 1) . Of these patients, 91 were found to have FOXC1 defects (13 different probands), whereas 35 had PITX2 defects (7 different probands). Of 91 patients with FOXC1 defects, 57, representing two families, had FOXC1 gene duplication; and 34, representing seven families and 4 isolated cases, had FOXC1 mutations. Of the 35 patients with PITX2 defects, 26, representing two families and 2 isolated cases, had PITX2 mutations; and 9, representing two families and 1 isolated case, had PITX2 gene deletion. 

There was no significant difference in gender among patients with ARM (P > 0.32; 48% females, 52% males). 

Clinical Features of the Patients with ARM: Ocular and Systemic Malformations

The patients with FOXC1 duplications, FOXC1 mutations, PITX2 deletion, and PITX2 mutations were tested for differences in clinical presentation. The absolute number of probands and/or different families with ARM available for our study precluded interfamilial comparison of clinical presentation; therefore, the absolute number of patients was used in this study. Unfortunately, there was an insufficient number of patients in the phenotypic groups to allow comparisons between phenotypic parameters and the biochemical analysis of mutations. 

Patients who participated in this study usually had ARM diagnosed in childhood (0–15 years, P < 0.01). 

The ocular malformations found in patients with ARM are presented in Table 2 . Notably, the patients with FOXC1 mutations were more likely to have iris hypoplasia, corectopia, peripheral anterior synechiae (PAS), and posterior embryotoxon (PE) than were the patients with FOXC1 duplication, whereas the patients with FOXC1 duplication were more likely to have IGD than were the patients with FOXC1 mutations. The patients with PITX2 deletion were more likely to have PE than were the patients with PITX2 mutations. The patients with PITX2 defects (mutations and deletion) were more likely to have corectopia than were those with FOXC1 defects (mutations and duplication). Overall, the patients with FOXC1 duplications were more likely to have IGD than were the patients with PITX2 defects, whereas those with PITX2 defects were more likely to have IH, corectopia, PAS, and PE. No other statistically significant differences in ocular clinical presentation were found. In all the patients with ARM, irrespective of the type of defect, the disease affected both eyes in all cases (P < 0.05). The small number of the patients with polycoria or “other” ocular malformations (microcornea, macrocornea, dyscoria) for any category of defect precluded statistical analysis for differences in frequencies between the different categories. The systemic malformations found in the patients with ARM are presented in Table 3 . The patients with FOXC1 mutations were more likely to have systemic malformations than were the patients with FOXC1 duplication (P < 0.0001). The patients with PITX2 defects were more likely to have systemic malformations than were the patients with FOXC1 defects (P = 0.006). The patients with PITX2 defects were more likely to have systemic malformations of the teeth, umbilicus, and facies than were the patients with FOXC1 defects (P < 0.04). No other statistically significant differences in nonocular clinical presentation were found. There were too few patients with heart malformations and hearing loss with PITX2 mutations and PITX2 deletion to allow for statistical analysis. 

Clinical Features of the Patients with ARM: Glaucoma

The incidence of glaucoma in the patients with ARM who participated in this study was 100% (29/29) in the patients with duplications of FOXC1, 74% (17/23) in the patients with mutations of PITX2, and 75% (71/95) in all the patients with ARM. The patients with FOXC1 duplication were more likely to have an increased incidence of glaucoma than were the patients with FOXC1 mutations (P = 0.00). The patients with FOXC1 duplication also were more likely to have an increased incidence of glaucoma than were the patients with PITX2 defects (P = 0.00). With respect to the age when glaucoma was diagnosed, the patients with ARM were more likely to have glaucoma develop in adolescence or early adulthood (<30 years old; P = 0.00). However, in the patients with FOXC1 duplication, glaucoma was more likely to develop at early age, in childhood (0–15 years, P < 0.01). The patients with FOXC1 duplication were more likely to have glaucoma develop at an earlier age than were those with PITX2 defects (P = 0.01). No other statistically significant differences in the group of age were found. 

Regarding the risk factors contributing to development of glaucoma in the patients with ARM, the patients with FOXC1 defects, either mutations or duplication; the patients with PITX2 mutations; the total patients with PITX2 defects; and all the patients with ARM were more likely to have family history of glaucoma than not to have one (P = 0.00). Because of the small number of patients with PITX2 deletion included in this study, family history was not found to be a risk factor for these patients (P = 0.70). The patients with FOXC1 defects were more likely to have a family history than were those with PITX2 defects (P = 0.02). For the rest of the risk factors (hypertension, ischemic heart disease, myopic eye, previous ocular trauma and topical steroid use on eyes), the patients with ARM and glaucoma were more likely to have at least one risk factor than not to have one (P = 0.01). 

Ophthalmic Examination of the Patients with ARM with Glaucoma

VA results of each group of genetic defects are displayed in Figure 2 . In the patients with glaucoma, the patients with FOXC1 mutations had VA significantly lower and were more likely to have lost vision due to glaucoma than were the patients with FOXC1 duplication (P < 0.02). In the patients with glaucoma, the patients with PITX2 defects were more likely to have fair or poor vision and bilateral vision loss than were those with FOXC1 defects (P < 0.03). VA in the patients with glaucoma with PITX2 mutations, total PITX2, and all the patients with ARM was significantly lower than was VA in those without glaucoma (P < 0.03). No other statistically significant differences in VA were found. 

IOP results in each group of genetic defects are displayed in Figure 3 . The range of IOP in the patients with FOXC1 defects was 15 to 68 mm Hg and in the patients with PITX2 defects was 14 to 58 mm Hg. The patients with FOXC1 duplication, total FOXC1, or PITX2 mutations and all the patients with ARM were more likely to have an elevated IOP (≥22 mm Hg) than to have a low IOP (<21 mm Hg, Fig. 3 ). In both eyes, in all categories of patients, independent of the type of defect, elevated IOP (≥22 mm Hg) was significantly associated with glaucoma (P < 0.05). However, there was a significant difference in IOP between the eyes and in the patients with glaucoma (P = 0.031). This difference of IOP in the patients with glaucoma may be due to differences in central corneal thickness between the two eyes. 

CDR results in each defect group are displayed in Figure 4 . The mean of vertical CDR in the patients with FOXC1 mutations, FOXC1 duplication, total FOXC1 defects, PITX2 mutations, and total PITX2 was significantly higher than the mean of vertical CDR in the general population ⁵¹ (P < 0.05). However these patients did not have a difference in CDR of >0.2 between the two eyes (P < 0.04). 

For ophthalmic tests and treatment categories of data, there were too few patients in the PITX2 deletion group to apply any statistical tests or to compare the patients with PITX2 deletion with those with PITX2 mutations. No statistical test was applied for corneal thickness and visual fields, because information on these parameters was obtained from too few patients. 

Treatment

The only effective treatment of glaucoma is reducing the IOP. There are two ways to lower IOP: medication and surgical treatment. Congenital glaucoma is primarily a surgical disease, with medical management serving as a temporary measure before surgery or as a postoperative adjunctive treatment. Treatment data are displayed in Figure 5 . More than half of all the patients with ARM received both treatments (57%, 29/51), whereas 24% (12/51) received only medication, 12% (6/51) received only surgical treatment, and 8% (4/51) received no treatment. The patients who did not receive treatment were either blind or were being monitored (the patients with glaucomatous visual fields defects, but low IOP). The patients with FOXC1 mutations, total FOXC1, and all the patients with ARM received medication as a first treatment (P < 0.04). Glaucoma in only 18% of the patients with ARM responded to medical or surgical (used solely or in combination) treatment. In the patients with ARM who participated in our study, treatment was not successful in 59% (23/39) and was successful in 18% (7/39); 23% were stable (9/39). However, no significant differences in the results of treatment were found in the patients with either FOXC1 or PITX2 defects, independent of the type of defect (P > 0.06). 

Discussion

In the present study, we conducted a retrospective examination of the glaucoma-related clinical presentation of individuals with PITX2 or FOXC1 mutations. Our investigation included the largest cohort of patients with ARM ever investigated. We found that ARM-associated glaucoma is a bilateral disease, equally prevalent in males and females. Of all the patients with ARM who participated in this study 75% had glaucoma. This finding is not consistent with previous reports ^{10 59 60} that suggested that glaucoma develops in approximately 50% of patients with ARM. This discrepancy could be due to the larger number of patients in our study or the fact that we included only patients with ARM with known defects of PITX2 or FOXC1 or may also be a consequence of the high proportion of the patients with FOXC1 duplication in comparison with the proportion of patients with other types of genetic defects included in this study. However, similar to the previous reports, ^{10 59 60} glaucoma in these patients developed in adolescence or early adulthood. As found in other glaucoma studies, ^{53 54} at least one risk factor (family history, myopia, hypertension, heart disease, trauma, or optical steroids) was found to be significantly associated with the development of glaucoma in the patients with ARM. 

A closer examination of the patients with ARM who had glaucoma indicates that VA of these patients was decreased due to glaucoma (20/40–20/100). The patients with ARM had an elevated IOP and a higher CDR than the CDR in the general population. An important observation is that the glaucoma in only 18% of the patients with ARM responded to medical or surgical (used solely or in combination) treatment. The finding that more than half of the patients with ARM received both treatments suggests that neither medication (which acts to reduce aqueous production to lower IOP) nor surgery appeared to be effective. This finding suggests that the glaucoma in these patients, which may arise due to a progressive impairment of outflow facility, aqueous production refractive to medication and surgery, especially pressure-sensitive retinal ganglion cells, or a combination of these factors, was resistant to current glaucoma treatments. However, failure to respond to the surgical treatment may be due to surgical complications such as early fibrosis after trabeculectomy. Another explanation of the differences in responses to treatment may be, in part, the presence of modifier genes. Research is currently under way to attempt to identify such loci. 

Genotype-Phenotype Correlations

The patients with duplication of FOXC1 typically had IGD malformations and an higher incidence of elevated IOP and of glaucoma (usually childhood-onset) than did the patients with FOXC1 mutations. In contrast, the patients with mutations of FOXC1 presented with iris hypoplasia, corectopia, peripheral anterior synechiae, and posterior embryotoxon. Thus, the known diversity in clinical presentation found in the patients with FOXC1 mutations ^{9 61 62} appeared to be more diverse than that in the patients with FOXC1 duplication. However, the incidence of elevated IOP and glaucoma was lower in these patients than in the patients with FOXC1 duplication, and glaucoma developed later in life than in patients with FOXC1 duplication. Therefore, on the basis of significant differences of glaucoma incidence and age of onset of glaucoma, we suggest that patients with FOXC1 duplication have a more severe prognosis in glaucoma development than do patients with FOXC1 mutations (Fig. 6) . Of interest, patients with nonocular findings appeared likely to have FOXC1 mutations rather than FOXC1 duplication. This suggests that the eye is particularly sensitive to duplication of FOXC1. Because of the small number of patients with the PITX2 deletion, we could not statistically compare the patients with PITX2 mutations and PITX2 deletion. The patients with PITX2 defects typically had corectopia. VA in the patients with PITX2 defects was worse than that in the patients with FOXC1 defects, and they were more likely to have bilateral vision loss due to glaucoma than were the patients with FOXC1 defects. However, no patients with FOXC1 defects had polycoria; VA in the patients with FOXC1 defects was better, and they were more likely to have unilateral vision loss than were patients with PITX2 defects. The absence of polycoria in patients with FOXC1 defects is consistent with previous studies. ^{30 61 62} Patients with PITX2 defects required multiple surgeries to achieve the same results of treatment as patients with FOXC1 defects. Therefore, glaucoma in patients with PITX2 defects was more difficult to treat than that in patients with FOXC1 defects. Patients with nonocular findings are more likely to have PITX2 defects than FOXC1 defects. ⁶³ Taken together, on the basis of significant differences in VA and treatment outcome, the results suggest that patients with PITX2 defects have a more severe prognosis for glaucoma development than do patients with FOXC1 defects (Fig. 6) . In this study also, the patients with PITX2 defects had a more severe prognosis for glaucoma development than did the patients with FOXC1 duplication (Fig. 6) . Taking into consideration the recent finding that PITX2 inhibits FOXC1 activity, ¹ our model predicts that PITX2 defects may result in both an inability to activate PITX2 targets and a gain-of-function activation of FOXC1 targets. Therefore, the severity of the ocular phenotype in the patients with PITX2 defects may be the simultaneous consequence of PITX2 haploinsufficiency and a gain of function of FOXC1. We suggest that patients with PITX2 defects may benefit from more frequent periodic ophthalmic examinations, closer monitoring of the disease, and more aggressive treatment—both medical and surgical—when glaucoma appears. This study may also help guide mutation screening process for a patient with newly diagnosed ARM (Fig. 6) . A potential limitation of our study is the absence of sufficient information regarding visual fields and central corneal thickness to allow statistical testing. Central corneal thickness is a factor with possible impact on IOP measurements by applanation tonometry. Eyes with thin corneas have an underestimation of IOP, and eyes with thick corneas have an overestimation. Recently, a study of corneal thickness in patients with FOXC1 duplication has shown that patients with FOXC1 duplication presented with increased central corneal thickness, leading to overestimation of IOP, independent of the tonometric method used. ⁶⁴ However, because diagnosis of glaucoma in our study was based on the observation of at least two of the following criteria—glaucomatous optic disc damage, glaucomatous visual fields defects, or high IOP (≥22 mm Hg)—it is very unlikely that altered corneal thickness explains the increased rate of glaucoma diagnosis in the patients with FOXC1 duplication (100%) compared with other gene defects. Nevertheless, a prospective clinical study, following up on the findings reported herein, that includes central corneal thickness would be worthwhile. An obvious limitation of the present study is the large number of patients coming from a small number of families, which may influence some of the results. Nevertheless, this analysis includes the largest cohort of patients with ARM analyzed to date. Our study has revealed that current medical therapies do not successfully lower IOP or prevent progression of glaucoma in patients with ARM with FOXC1 or PITX2 alterations. Further comparisons of the underlying genetic defects with glaucoma treatment outcomes, together with the testing of novel glaucoma therapies in cohorts of patient with ARM, could result in improved glaucoma treatment for these patients. 

 Supported by Canadian Institutes of Health Research (CIHR) Grants G118160422, G118160550, G118160216 (HS, MAW) and Alberta Heritage Foundation for Medical Research (AHFMR) Grant 200401491 (ID).

 Submitted for publication April 25, 2006; revised July 17, 2006; accepted October 18, 2006.

 Disclosure: M.H. Strungaru, None; I. Dinu, 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.

 

Ocular characteristics of patients with Axenfeld-Rieger Malformation. (A) Left eye of a patient with FOXC1 duplication shows marked iris stroma hypoplasia with exposure of the sphincter muscle. (Reprinted, with permission from Gould DB, Mears AJ, Pearce WG, Walter MA. Autosomal Dominant Axenfeld-Rieger anomaly maps to 6p25. Am J Hum Genet. 1997;61:765–768. © American Medical Association). (B) Right eye of a patient with FOXC1 mutation shows marked iris stroma hypoplasia with exposure of the sphincter muscle (generously provided by Elise Héon, University of Toronto, Ontario, Canada). (C) Left eye of a patient with PITX2 mutation shows displaced pupil, iris atrophy, polycoria, prominent and displaced Schwalbe’s line (posterior embryotoxon) (generously provided by Ordan Lehmann, University of Alberta, Edmonton, Alberta, Canada).

Table 1.

 

Summary of Patients Who Participated in the Study and the Defects Found in These Patients

	Probands	Affected Individuals (n)	Defect
FOXC1 mutation	Family 1	7	S82T 9 42 43
FOXC1 mutation	Family 2	12	Q23Stop 47
FOXC1 mutation	Family 3	2	Δ363C*
FOXC1 mutation	Family 4	2	L130F*
FOXC1 mutation	Family 5	2	L86F 41
FOXC1 mutation	Family 6	3	Δ10bp 9
FOXC1 mutation	Family 7	2	187M 9 41 42
FOXC1 mutation	Patient 1	1	P297S*
FOXC1 mutation	Patient 2	1	L86F 41 44
FOXC1 mutation	Patient 3	1	G165R 43 44
FOXC1 mutation	Patient 4	1	R169P 43 44
FOXC1 duplication	Family 8	46	Gene duplication 10 40 41
FOXC1 duplication	Family 9	11	Gene duplication 40 41
PITX2 mutation	Family 10	12	R63H 10 13
PITX2 mutation	Family 11	12	ΔC416FS 23
PITX2 mutation	Patient 5	1	Intron 2, G→T 23
PITX2 mutation	Patient 6	1	ΔG114FS 23
PITX2 deletion	Family 12	6	Gene deletion 23
PITX2 deletion	Family 13	2	Gene deletion 23
PITX2 deletion	Patient 7	1	Gene deletion 23

*  Unpublished data (M.H.S., 2005).

Table 2.

 

Ocular Malformations Found in Patients with ARM

Defect	Affected Individuals (n)	IGD*	IH*	Corectopia* , ‡	Polycoria	PAS*	PE* , † , §	Other, ∥
FOXC1 mutations	23	5 (22%)	17 (74%)	6 (26%)	0 (0%)	19 (83%)	19 (83%)	3 (13%)
FOXC1 duplication	25	25 (100%)	7 (28%)	0 (0%)	0 (0%)	3 (12%)	0 (0%)	0 (0%)
Total FOXC1	48	30 (63%)	24 (50%)	6 (13%)	0 (0%)	22 (46%)	19 (40%)	3 (6%)
PITX2 deletion	4	0 (0%)	4 (100%)	2 (50%)	1 (25%)	2 (50%)	3 (75%)	1 (25%)
PITX2 mutations	17	12 (71%)	11 (65%)	5 (29%)	1 (6%)	6 (35%)	2 (12%)	1 (6%)
Total PITX2	21	12 (57%)	15 (71%)	7 (33%)	2 (10%)	8 (38%)	5 (24%)	2 (10%)
All ARM	69	42 (61%)	39 (57%)	13 (19%)	2 (3%)	30 (43%)	24 (35%)	5 (7%)

 IGD, iridogoniodysgenesis; IH, iris hypoplasia; PAS, peripheral anterior synechiae; PE, posterior embryotoxon.

*  Statistically significant difference between patients with FOXC1 duplication and FOXC1 mutations (P < 0.003).

†  Statistically significant difference between patients with PITX2 deletion and PITX2 mutations (P = 0.004).

‡  Statistically significant difference between patients with Total FOXC1 and Total PITX2 (P = 0.04).

§  Statistically significant difference between patients with FOXC1 duplication and Total PITX2 (P < 0.02).

∥  Other: microcornea, macrocornea, dyscoria.

Table 3.

 

Systemic Malformations in Patients with ARM

Defect	Patients with Systemic Malformation* , † , ‡	Dental, † , ‡	Umbilical, †	Heart, † , ‡	Facial, ‡	GI	GR	HT	Other, §
FOXC1 Mutations	16 (67%)	6 (25%)	3 (13%)	3 (13%)	10 (42%)	1 (4%)	3 (13%)	6 (25%)	2 (8%)
FOXC1 Duplications	2 (8%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	1 (4%)	0 (0%)	1 (4%)
Total FOXC1	18 (37%)	6 (12%)	3 (6%)	3 (6%)	10 (20%)	1 (2%)	4 (8%)	6 (12%)	3 (6%)
PITX2 Deletion	4 (100%)	4 (100%)	1 (25%)	0 (0%)	3 (75%)	0 (0%)	1 (25%)	0 (0%)	0 (0%)
PITX2 Mutations	14 (78%)	11 (61%)	10 (56%)	0 (0%)	12 (67%)	4 (22%)	0 (0%)	0 (0%)	0 (0%)
Total PITX2	18 (82%)	15 (68%)	11 (50%)	0 (0%)	15 (68%)	4 (18%)	1 (5%)	0 (0%)	0 (0%)
All ARM	36 (51%)	21 (30%)	14 (20%)	3 (4%)	25 (35%)	5 (7%)	5 (7%)	6 (17%)	3 (4%)

  n = 45. GI, gastrointestinal defects; GR, growth Retardation; HT, hypertelorism.

*  Statistically significant difference between patients with FOXC1 duplication and FOXC1 mutations (P < 0.0001).

†  Statistically significant difference between patients with Total FOXC1 and Total PITX2 (P < 0.04).

‡  Statistically significant difference between patients with FOXC1 duplication and Total PITX2 (P < 0.025).

§  Other: hearing loss.

Figure 2.

 

The percentage of left eyes of patients with ARM (A) and patients with ARM with glaucoma (B) with good (20/20–20/40), fair (20/40–20/100), or poor (≤20/200, counting fingers, hand motion, light perception, or no light perception) VA (VA) for each type of defect. There were too few patients in the PITX2 deletion group to apply any statistical tests. *Statistically significant difference, P < 0.05.

Figure 3.

 

Percentage of left eyes of patients with ARM with low IOP (IOP < 21 mm Hg) and elevated IOP (IOP ≥ 22 mm Hg) for each type of gene defect. There were too few patients in the PITX2 deletion group to apply any statistical tests. *Statistically significant difference, P < 0.05.

Figure 4.

 

CDR in patients with ARM with glaucoma (circle) and patients with ARM without glaucoma (triangle). The mean of Cup/Disc ratio in patients with ARM was statistically higher than the mean of Cup/Disc ratio in general population (dotted lines). The mean and 95% confidence intervals (solid lines) are displayed for each data set. There were too few patients in the PITX2 deletion group to apply any statistical tests. *Statistically significant difference, P < 0.05.

Figure 5.

 

Treatment received by patients with ARM. (A) Patients with FOXC1 mutations, FOXC1 duplication and FOXC1 defects were statistically more likely to have both medical and surgical treatments than not (P < 0.05). (B) Patients with PITX2 mutations were more likely to have multiple surgeries than have one surgery (P < 0.05), whereas patients with FOXC1 duplication were significantly more likely to not have multiple surgeries than to have (P < 0.05). Patients with PITX2 defects were more likely to have multiple surgeries than patients with FOXC1 defects (P < 0.05). (C) No statistically significant differences in the results of treatment were found in patients with either FOXC1 or PITX2 defects (P > 0.06).

Figure 6.

 

Severity of the prognosis of glaucoma and diagnostic criteria for patients with ARM. Top: patients with FOXC1 mutations had the mildest prognosis for glaucoma development, whereas patients with PITX2 defects and patients with FOXC1 duplication had more severe prognosis. Bottom: Patients presenting with polycoria and systemic malformations were more likely to have PITX2 defects; patients with IGD without systemic malformations were more likely to have FOXC1 duplication, and patients with peripheral anterior synechiae (PAS), posterior embryotoxon (PE), and systemic malformations were more likely to have FOXC1 mutations. Glaucoma progression was defined as advance of the disease, with increasing IOP, cupping of the optic nerve head, and VA defects, despite intensified medical treatment or surgery.

The authors thank all the referring physicians, without whose participation this study could not be done, especially Ian MacDonald, Ordan Lehmann (University of Alberta, Edmonton, Alberta, Canada), and Robert Ritch (NY Eye and Ear Infirmary, New York, NY) for their help in collecting clinical information and providing anterior segment photography of local patients. The authors also thank Farideh Mirzayans for her help with laboratory patient files.