Introduction

The Aristaless-related homeobox (ARX) gene (MIM 300382) is a homeobox-containing gene; it maps to Xp22.13, encodes a protein of 562 amino acids, and acts as a transcription factor (Stromme et al. 2002). The expression pattern suggests that ARX is involved in the differentiation and maintenance of specific neuronal cell types in the central nervous system (Ohira et al. 2002; Poirier et al. 2004; Suri 2005).

ARX mutations are responsible for a wide spectrum of phenotypes, including various forms of epilepsy, myoclonic seizures, dystonia, autism, lissencephaly, as well as syndromic and nonsyndromic X-linked mental retardation (Bienvenu et al. 2002; Kitamura et al. 2002; Stromme et al. 2002; Poirier et al. 2006). Previous studies estimated that ARX mutations are highly prevalent among X-linked mental retardation (XLMR) families (9.5%), with approximately 6.5% due to a c.428_451dup(24 bp) mutation, which suggests an important role for this gene in XLMR (Mandel and Chelly 2004; Poirier et al. 2006). In this work, we report the mutational screening of the entire ARX gene in Brazilian males with mental retardation (MR) of unknown etiology.

Materials and methods

We screened a total of 143 unrelated males from Rio de Janeiro, Brazil, aged 2–29 years (10.2±5.3), with mild to severe MR associated or not with other clinical features (Table 1) for ARX mutations. The patients were classified as XLMR families (with at least two affected males in a kindred who are related through females), BP families (with at least two affected male siblings), and sporadic cases (SC). This included 66 syndromic (11 XLMR, 6 BP, and 49 SC) and 77 nonsyndromic (10 XLMR, 8 BP, and 59 SC) mentally retarded individuals. Chromosomal analysis and fragile-X syndrome test had normal results. The regional Ethics Committee approved the research protocols, and informed consent was obtained from the parents of patients.

Table 1 Clinical description of the study sample
Full size table

Genomic DNA was obtained from peripheral blood (Miller et al. 1988), and mutation screening of all five exons and flanking intronic regions of ARX was performed using PCR, followed by single-strand conformation polymorphism (SSCP) analysis (exons 1, 3, 4, and 5) and direct sequencing (exon 2). PCR conditions are available on request. The primer pairs used were previously designed by the group of Dr. Jozef Gécz and are available on request (Stromme et al. 2002).

The splice site predictions were carried out using the electronic tools: http://www.fruitfly.org/seq_tools/splice.html (Berkley Drosophila Genome Project); http://www.tigr.org/tdb/GeneSplicer/gene_spl.html (Gene Splicer Web Interface); http://www.cbs.dtu.dk/services/NetGene2/ (NetGene2 Server); http://www.l25.itba.mi.cnr.it/∼webgene/wwwspliceview.html (Institute of Advanced Biomedical Technologies). For exonic splicing enhancers (ESE) predictions we used the http://www.rulai.cshl.edu/tools/ESE/ (Cartegni et al. 2003), and RESCUE-ESE at http://www.genes.mit.edu/burgelab/rescue-ese (Fairbrother et al. 2002).

Results

Mutational analysis of the ARX gene in 143 Brazilian mentally retarded males revealed one case of the mutation c.428_451dup(24 bp) in a male with a familial history of XLMR (Fig. 1), one exonic variant in two sporadic MR cases, and a novel intronic variant in a male with sporadic MR. No variants were identified in exons 1, 3, and 5 (Table 2).

Fig. 1a, b
figure 1

DNA sequence electropherograms from exon 2 of ARX showing the 428_451dup(24 bp) mutation. a Normal control, b affected male

Full size image
Table 2 Summary of the ARX sequence variants detected in 143 patients with MR from Brazil
Full size table

The patient with the mutation c.428_451dup(24 bp) has moderate MR, language delay, macrocephaly, and prominent forehead. This mutation was also observed in the proband’s older brother, and in his cousin, both mentally retarded males. The family’s clinical features have been previously detailed by us (Gestinari-Duarte et al. 2006).

In two other patients we found a silent variant in exon 4 (c.1347C>T, p.G449G): the first, a 7-year-old boy, presenting with MR, seizures/epilepsy, and autism, and the second, a 15-year-old, presenting with MR and language delay. We also identified the c.1347C>T mutation in 2.6% (2/78) of our sample of normal controls.

A novel intronic variant (c.1074-3T>C) was identified in one patient, a 9-year-old boy with moderate MR, seizures, and language delay. This variant is located 1 bp upstream of the 3′ splice site of ARX intron 2. His mother was found to be a carrier of the variant; however, we did not observe this substitution in his nonaffected sister. This variation was predicted not to alter splicing, and it could not be associated with the etiology of MR. We found the same intronic variant in approximately 1% (1/102) of normal controls.

Discussion

ARX gene mutations have already been considered as an important cause of syndromic and nonsyndromic XLMR (Mandel and Chelly 2004; Ropers and Hamel 2005). Our study is the first ARX screening in MR males from Latin America, and contributes new clues about the significance of this gene for genetic counseling of individuals with MR of unknown etiology.

The most common ARX mutation is the c.428_451dup(24 bp), which leads to an elongation of the second poly A tract in ARX protein (Stromme et al. 2002). This mutation was reported in Partington syndrome, West syndrome, and in approximately 70% of MRX families linked to Xp22.1 (Bienvenu et al. 2002; Ropers et al. 2003). The data showed a high frequency (6.1%) of the c.428_451dup(24 bp) in XLMR families and a low rate in BP families (1.5%) (Poirier et al. 2006). In comparison, we found one male, a familial case, with this duplication among 21 XLMR families (4.8%). We did not find this mutation in 14 BP families or in 108 SC. To date, 31 families with the c.428_451dup(24 bp) mutation including those from this report, have been described (Bienvenu et al. 2002; Frints et al. 2002; Stromme et al. 2002; Turner et al. 2002; Gronskov et al. 2004; Kato et al. 2004; Partington et al. 2004; van Esch et al. 2004; Gestinari-Duarte et al. 2006; Poirier et al. 2005, 2006; Stepp et al. 2005; Nawara et al. 2006).

The silent variant c.1347C>T was identified in two boys with MR, and in 2.6% of our normal population, and has been considered a polymorphism. Although this variant was firstly identified in a boy with XLAG (Kato et al. 2004) and in a boy with MR (Gronskov et al. 2004), it was also found in 4% of the normal controls tested in these studies. Gronskov et al. (2004) performed in-silico analysis through ESEfinder program and showed that c.1347C>T creates a strong binding site for SRp55, a kind of splicing binding factor to exonic splice enhancers (ESE). However, in our study, using the RESCUE-ESE program, this same analysis did not show clearly possible ESE sites. The ability of both programs to predict natural point mutations that disrupt genuine ESE motifs has yet to be fully assessed (Baralle and Baralle 2005).

We excluded a pathogenic role of the novel intronic variant c.1074-3T>C. This substitution was identified in a boy with MR and in 1% of our normal controls. Our results did not shown any significant alterations in the splicing site values compared to the wild-type sequence, which suggests that it is a polymorphism. To date, 11 intronic variants have been reported in the ARX gene, 9 as polymorphisms and 2 have been identified in XLAG individuals, and described as premature stop codon mutations (Kato et al. 2004).

In conclusion, our data showed that the 24-bp duplication in exon 2 of the ARX gene has a considerable frequency among XLMR families, but the incidence of this mutation in sporadic males with MR appears to be low, reinforcing previous suggestions (Poirier et al. 2006; Stepp et al. 2005) that ARX mutation screening of exon 2 should be considered for any male with MR of unknown etiology, particularly among familial cases. Recently, mutations in CDKL5, also located within the Xp22, have been associated with a broad range of phenotypes, such as seizures/epilepsy, autism, and MR (Kalscheuer et al. 2003; Mari et al. 2005; Scala et al. 2005), which overlap the cases for ARX mutations. We believe that CDKL5 is, perhaps, another gene that should be screened in patients presenting MR associated with seizures/epilepsy and autism, particularly in XLMR families, who have negative results for ARX mutations.