Novel and recurrent PITX3 mutations in Belgian families with autosomal dominant congenital cataract and anterior segment dysgenesis have similar phenotypic and functional characteristics

Cataract is defined as opacity or cloudiness of the crystalline lens. It is the primary cause of blindness worldwide and is classified into different types based on the age of onset. Congenital cataract manifests in the first year of life; the estimated incidence of congenital cataract is 72 per 100,000 children in developed countries with a higher incidence in less-developed countries. Congenital cataract is clinically and genetically heterogeneous with about 45 loci known and 38 genes identified. Mutations in genes encoding proteins important in the development and maintenance of the structural integrity of the lens such as crystallins, connexins and aquaporins [1, 2] are typically associated with isolated congenital cataracts while mutations in the transcription factor genes PAX6[3], FOXE3[4], EYA1[5], MAF[6], and PITX3[7] have been described in congenital cataract with anterior segment dysgenesis (ASD). ASD is an umbrella term for the spectrum of developmental disorders affecting the structures of the anterior segment of the eye. The ocular anomalies typically include corneal opacity, adhesions between the iris and cornea or lens and cornea, iris hypoplasia, corectopia or polycoria, and malformation of the irido-corneal angle drainage structures [8].

PITX3 was isolated as the third gene of the PITX homeobox-containing transcription factor gene family [9, 10], along with PITX1[11, 12] and PITX2[13]. The PITX protein family is a subfamily of the paired-like class of the homeobox-containing proteins, which play a crucial role in the development of different organisms including mammals. Like the other members of this family, PITX3 contains a characteristic and strongly conserved homeodomain required for DNA binding. The conserved 14-amino acid OAR motif, named after the homeodomain proteins otp, aristaless, and rax[14], is located downstream of this homeodomain, and may function in the target specificity and transactivation of the homeodomain protein [13, 14]. Interestingly, Pitx3 mapped to the aphakia (ak) locus [9], a recessive mutation resulting in bilateral microphthalmia with lens aplasia originally described by Varnum and Stevens in 1968. The lens develops normally in ak mice until an arrest occurs around embryonic days 10.5–11 [15] corresponding to the moment of initial expression of Pitx3 in the lens [7, 9]. Consequently, Pitx3 was the top candidate for the ak phenotype but no mutation was found in the coding region of Pitx3[9]. Two different deletions in the promoter region of Pitx3 were later found to explain the ak phenotype [16, 17]. Before this finding, the first human PITX3 mutations had already been identified in two families with autosomal dominant congenital cataract (ADCC). A 17-bp duplication c.640_656dup, p.(Gly220Profs*95), was found in a family with ADCC and ASD while a missense mutation c.38G > A, (Ser13Asn), was identified in a second family [7]. Since the original gene identification study, there have been only a few additional mutations identified, bringing the total number of unique PITX3 mutations found in ADCC with or without ASD to five mutations in 13 different families [7, 18‐26]. Two of these mutations are recurrent, the most common one being the 17-bp duplication p.(Gly220Profs*95) which was reported in eight of the 13 families [7, 18‐21, 23, 24]. The other recurrent mutation is a 1-bp deletion at position 650, p.(Gly217Alafs*92), found in two families [18, 22]. Thus far, only the 17-bp duplication has been associated with ASD; all other mutations are reported with isolated cataract. Of particular note, homozygous PITX3 mutations were also described in two consanguineous pedigrees [22, 25].

The aim of this study was to analyze the PITX3 gene in five Belgian families with ADCC and ASD. We identified the recurrent 17-bp duplication c.640_656dup, p.(Gly220Profs*95), in four of these families, and a novel PITX3 mutation c.573del, p.(Ser192Alafs*117), in a fifth family. In vitro functional assays were performed for both mutations, showing similar functional characteristics. In conclusion, the similar ADCC and ASD phenotypes resulting from both mutations could be explained by our in vitro functional studies.

Congenital cataract is a clinically and genetically heterogeneous condition. Most genes are associated with isolated congenital cataracts while mutations in the transcription factor genes PAX6, FOXE3, EYA1, MAF, and PITX3 lead to congenital cataract with ASD [3‐7]. Of particular note, PITX3 mutations can cause isolated congenital cataract as well as ASD-associated cataract, even within the same family [7]. This inter- and intrafamilial phenotypic variability can likely be ascribed to modifier genes or environmental factors. PITX3 mutations represent a rare cause of congenital cataract with or without ASD. Indeed, only five unique PITX3 mutations in 13 different families were known before the onset of this study (Table 1). The most common mutation is the 17-bp duplication c.640_656dup accounting for more than half of the families [7, 18‐21, 23, 24] and ADCC with ASD has been documented only in association with this duplication. Notably, the novel PITX3 mutation c.573del, p.(Ser192Alafs*117), was found in a family with ADCC and ASD, thus being the second mutation leading to this phenotype. Like the other reported PITX3 mutations, this novel frameshift mutation is located outside the homeodomain. Interestingly, all PITX3 frameshift mutants lack the C-terminal 14-amino-acid motif (see Additional file 2), which is conserved in numerous paired-like homeodomain proteins and is named the OAR domain, after the homeodomain genes otp, aristaless, and rax[14]. The function of this OAR domain is not fully established yet but it has been suggested that it could interact with additional proteins, hence mediating the target specificity of the homeobox-containing transcription factors [13, 14]. Although the OAR domain is conserved in all of these homeodomain proteins, its function appears to vary among the proteins. Functional assays with different deletion mutants of Otp showed that mutants missing the region downstream of the homeodomain, including the OAR domain, presented with a different DNA-binding profile and a substantial reduction of transactivation activity, thus suggesting an activating regulatory role [31]. The OAR domain is also important for the transactivation activity of SHOX as a truncated SHOX mutant lacking this domain has an abolished transactivation activity [32]. In contrast, a study of the Pitx2 C-terminus demonstrates an inhibitory role for the OAR domain as it was shown that the presence of this domain inhibits DNA-binding while masking this domain through protein interaction stimulates DNA-binding and transcriptional activation [33]. This finding was supported by studies in Prx2 and Cart1 where the OAR domain inhibits transactivation and serves as an intramolecular switch of its transactivation activity, respectively [34, 35]. To explore the function of the C-terminal region in PITX3, Sakazume et al. investigated the effect of PITX3 mutants on DNA-binding and transactivation activity. In accordance with the Otp and SHOX study, an altered DNA-binding profile and diminished transcriptional activity was shown [28].

Both the novel and recurrent PITX3 mutations found here disrupt the C-terminus including the OAR domain. As they both lead to a similar phenotype, we assessed whether this similarity could be substantiated by in vitro functional assays. Since all PITX3 frameshift mutations utilize a stop codon located downstream of the normal stop codon in the last coding exon, they are not predicted to be subject to nonsense mediated decay (NMD). The novel PITX3 mutant retained its nuclear localization but displayed altered DNA-binding properties similar to the recurrent PITX3 p.(Gly220Profs*95) mutant. In addition, a similar decrease of transcriptional activity compared to wild-type PITX3 was observed for the novel and recurrent PITX3 mutants. These mutations may produce aberrant protein that can either antagonize DNA-binding activity of wild-type proteins or form inactive dimers with wild-type proteins, impairing wild-type protein function in both situations. Co-transfection assays with mutant and wild-type PITX3 could not, however, substantiate a dominant-negative effect [28]. A loss-of-function effect for the PITX3 mutations is therefore more probable. This view is supported by the report on a large ADCC pedigree in which two brothers from a consanguineous mating are homozygous for the PITX3 mutation segregating in the family and manifest a more severe ocular and neurological phenotype, with severe microphthalmia and neurological involvement in addition to cataract [22]. This severe combination of ocular and neurological involvement has also been described in the two spontaneous Pitx3 mutant mice, aphakia and eyeless[9, 10, 15, 36]. Interestingly, besides its expression in the developing lens, Pitx3 expression was also observed in the dopamine neurons of the midbrain [10]. A severe loss of dopaminergic neurons in the substantia nigra was also observed in aphakia and eyeless mice. As these neurons play important roles in the control of movement, emotion, cognition, and reward related behavior, this neuronal loss can explain the observed locomotor and behavioural defects [36‐39]. Unlike the human PITX3 mutations, heterozygous Pitx3 mice were phenotypically normal which might be attributed to interspecies differences, as mice seem to be less sensitive to gene dosage. However, the loss-of-function hypothesis is challenged by the report on another consanguineous mating between phenotypically normal first cousins resulted in a girl homozygous for the PITX3 mutation c.640_656del, manifesting as severe bilateral microphthalmia and ASD but without cataract or a neurological phenotype [25]. This phenotype is reminiscent of the manifestations in the Texel Sheep as this breed displays microphthalmia as an autosomal recessive congenital condition, caused by a missense mutation in the conserved homeodomain of PITX3, c.338G > C, p.(Arg113Pro) [40]. In addition, a heterozygous microdeletion of 10q24.32 encompassing PITX3 in a patient with Smith–Magenis syndrome-like behavioral abnormalities was recently described. The patient also presented with intellectual disability and dysmorphic features but, surprisingly, lacked an eye phenotype. Analysis of neurotransmitters in his cerebrospinal fluid revealed an absence of L-DOPA and L-DOPA treatment led to mild improvement of his behavior [41]. Based on these observations and the selective loss of dopaminergic neurons in the Pitx3 mutant mice, it might be interesting to assess the dopaminergic function of patients with PITX3 mutations.

Taken together, these findings suggest that PITX3 associated phenotypes most likely result from a more complex mutational mechanisms than loss of function effects alone and that more extensive studies are needed to correlate the phenotypic and molecular consequences of PITX3 mutations.

In this study, we identified a novel PITX3 mutation, p.(Ser192Alafs*117) and the recurrent p.(Gly220Profs*95) mutation in one Belgo-Romanian family and four Belgian families with ADCC and ASD, respectively. The novel mutation p.(Ser192Alafs*117) is only the second PITX3 mutation to be associated with this combined phenotype, along with the recurrent mutation p.(Gly220Profs*95). The similarity in phenotypic expression could be explained by in vitro functional studies which demonstrated similar molecular consequences for the novel p.(Ser192Alafs*117) and the recurrent p.(Gly220Profs*95) mutations.