Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- GER, gastroesophageal reflux
- HSCR, Hirschsprung disease
- PAS, pulmonary artery sling
- PDA, patent ductus arteriosus
- TOF, tetralogy of Fallot
Hirschsprung disease (HSCR), a clinically complex syndrome often associated with a combination of mental retardation, microcephaly, and characteristic facial features, is genetically heterogeneous.1–10 Recently, mutations in ZFHX1B, which encodes Smad-interacting protein 1 (SIP1), were identified by our group11 and Cacheux et al12 in patients with t(2;13)(q22;q22) and t(2;11)(q22.2;q21), respectively. These patients had clinical profiles consistent with a dominant form of the disease (Mowat-Wilson syndrome: MIM 235730).9,13 These and subsequent articles have demonstrated a link between patients with nonsense and frameshift mutations of the gene and the presence in these patients of the following clinical findings: profound mental retardation, delayed motor development, and specific facial features, including hypertelorism, a broad nasal bridge, strabismus, a pointed chin, prognathism, a nose with a prominent columella, and posteriorly rotated ears. Often associated with these features were microcephaly, epilepsy, congenital heart disease, HSCR, and midline defects such as agenesis or hypoplasia of the corpus callosum and hypospadias.13–19 This variety of clinical features is observed not only in patients with nonsense and frameshift mutations but also in patients harbouring deletions involving ZFHX1B at 2q22,9,11,15,20 including a previously reported isolated case.6 Thus, ZFHX1B clearly plays a critical role in the manifestation of clinical features of Mowat-Wilson syndrome. However, several points remain to be elucidated: (1) How broad are the clinical features in patients with ZFHX1B mutations and deletions? (2) What are the underlying molecular mechanisms by which the deletion of ZFHX1B at 2q22 affects the clinical phenotype? (3) What kind of clinical features might appear if the deletions were extended to include 2q23–q24.1? To address these questions, we have characterized the clinical features and underlying molecular basis of 27 cases with mutations or deletions in ZFHX1B, including 12 previously reported cases comprising 10 mutations, one deletion, and another case that had previously been found to have neither a deletion nor a mutation in the gene.
Informed consent was obtained from all participants in accordance with institutional guidelines for human materials and subjects. Evaluation was by review of medical records and physical examination, including findings for mental retardation, developmental delay, microcephaly, facial dysmorphism, epilepsy, and Hirschsprung disease, as well as CT and/or MRI brain images. The subjects participating in this study were named S1 to S42. Subjects who are not listed in the tables, such as S12, have not been identified with an abnormality in ZFHX1B, or the analysis of that patient has not yet been completed.
We identified five novel nonsense and frameshift mutations in one allele of ZFHX1B from patients with Mowat-Wilson syndrome. One of the mutations was an insertion of 382 bp of the AluYa5 sequence at nucleotide number 1169.
Two novel symptoms, pulmonary artery sling and septum of the vagina, were observed in patients with nonsense or frameshift mutations in ZFHX1B. Repeated vomiting attacks observed in five patients might be related to one aspect of epilepsy, as these attacks can be controlled with anticonvulsants.
Using a multiplex PCR method, we identified 0.2–10.42 Mb deletions at 2q22–q24.1, including ZFHX1B, in eight patients. Nucleotide sequence analyses demonstrated no sequence similarity for the regions on the two sides of the breakpoints in five patients.
All patients examined thus far have a paternal deletion and present clinical features that closely overlap those of cases with nonsense and frameshift mutations resulting from deletions of up to about 5 Mb. However, two patients with large deletions (10.42 and 8.83 Mb) had significantly delayed psychomotor development. One of them also presented with a cleft of the hard palate and complicated heart disease, features that have not been reported in patients with Mowat-Wilson syndrome.
Mutational analysis of ZFHX1B
To screen for mutations, DNA fragments including each exon of ZFHX1B from patients and normal controls were amplified by PCR14 and subjected to direct sequencing analysis.21 To confirm mutations detected in patients, PCR products were subcloned into pGEM-T Easy (Promega) and sequenced.
Chromosome 2 microsatellite analysis
Microsatellite analyses were performed as described previously.11 The polymorphic loci used for haplotype analysis around ZFHX1B on chromosome 2q were the following: D2S129, D2S2149, D2S122, D2S132, D2S381, D2S2301, D2S151, D2S222, D2S356, D2S2299, and D2S142 (in order from centromere to telomere, according to the NCBI Homo sapiens Map View).
Chromosomal deletion analysis using multiplex PCR
To determine regions of deletion on chromosome 2q22–q24.1 in patients, we established a quantification method for determining the dosage of genomic DNA using multiplex PCR. CCNT2, which encodes Cyclin T2 and maps about 10 Mb centromeric to ZFHX1B, was used as a quantitative control gene with both alleles present, because deletion of CCNT2 was not observed in any of the patients examined. Specific primers (sense: 5′-GACAGGTATCAGAGACACCA-3′ and antisense: 5′-ATTTTAGAGGTACTGGCGCAG-3′) were designed to amplify a 140 bp DNA fragment in exon 9 of CCNT2. Another pair of primers was also designed to amplify approximately 200 bp genomic DNA fragments from every region at 2q22–q24.1 so that repetitive sequences such as Alu and L1 were not selected. Aliquots (50 ng) of genomic DNA from each patient and a normal control subject were amplified by PCR in a total volume of 20 μl, each containing 0.3 μM in total of two pairs of primers, one for a specific region at 2q22–q24.1 and the other for CCNT2, 30 μM of each dNTP, 2.5 mM MgCl2, 10 mM Tris–HCl (pH 8.3), 50 mM KCl, 1 U of AmpliTaq-Gold (Applied Biosystems), and 0.2 μl of [α-32P]dCTP (80 kBq: 111 TBq/mmol). The ratios of the concentrations of the two primer pairs targeted to the specific region at 2q22–q24.1 and CCNT2 varied from 1 to 3, depending on the sequences chosen. PCR samples were preheated at 94°C for 10 min, and from 18 to 24 PCR cycles were performed. Each cycle consisted of 30 s denaturation at 94°C, 30 s annealing at 58°C, and 30 s extension at 72°C. The PCR products were electrophoresed on 8% polyacrylamide gels and quantification of radioactivity was performed with a BAS 1800 image analyzer (Fujifilm). Ratios of radioactivity of genomic DNA at 2q22–q24.1 versus CCNT2 were determined for patients (Rp) with respect to normal controls (Rc). We previously reported a patient (S1) with an unbalanced reciprocal translocation t(2;13)(q22;q22) and an approximately 5 Mb deletion containing ZFHX1B at 2q22.11 Using genomic DNA from this patient and five normal controls as templates, we performed multiplex PCR and determined the gene dosage of KYNU, which encodes kynureninase, and ZFHX1B, both located in the deletion region at 2q22 of this patient. When PCR was performed for between 18 and 24 cycles, the ratios (Rp/Rc) for both genes from S1 ranged from 0.3 to 0.6 (data not shown). Thus, when amplified genomic DNA was included in the region of deletion, Rp/Rc ratios typically varied from 0.3 to 0.6, whereas values for DNA without deletion ranged from 0.9 to 1.1. When ratios were between 0.6 and 0.9, new primer pairs were chosen for an adjacent region and multiplex PCR was repeated to determine whether the amplified genomic DNA was deleted. ZFHX1B spans approximately 132 kb of the human reference genomic contig NT_005058 in the reverse orientation, and consists of 10 exons and nine introns. The following primer pairs were designed for screening the deletion of ZFHX1B by multiplex PCR: exon 1 (sense: 5′-GTCCGTCCCCAGGTTTGTGT-3′ and antisense: 5′-AAAGCATGAAGAAGCCGCGA-3′); exon 3 (sense: 5′-TGTAGTGGACACAGGTTCTG-3′ and antisense: 5′-TAGAATCTCGTTGTTGTGCCA-3′); exon 6 (sense: 5′-TGCCACCTGGAACTCCAGAT-3′ and antisense: 5′-ATATCGTTTCTCTAAGGGGTTA-3′); and exon 10 (sense: 5′-ACTCGCAGCACATGAATCAC-3′ and antisense: 5′-CTCCTTCTCGCTCTCGCCA-3′).
Determination of the breakpoints of chromosomal deletions
Multiplex PCR was repeatedly performed to limit the region of deletion breakpoints in eight patients. Finally, each breakpoint was confined to a region containing a few kb of genomic DNA within a single BAC clone. Sense and antisense primers were designed for centromeric and telomeric adjacent regions of the breakpoint (fig 1A) and 50 ng of genomic DNA from each patient and a normal control subject were amplified by PCR using these primers to detect deletions.
Clinical features and ZFHX1B mutations detected in patients with Mowat-Wilson syndrome
We identified five novel nonsense and frameshift mutations (390fs430X, 465fs467X, 286fs293X, R302X, Q497X) and a nonsense mutation, R343X, reported recently19 in one allele of ZFHX1B, in addition to those in the 10 patients described in detail earlier (table 1). It should be noted that a 382 bp AluYa5 sequence was inserted at nucleotide number 1169 in S15. The previously reported R695X mutation11,14,17 was also found in three cases (S21, S32, and S38). Clinical features for all 19 patients are summarized in table 1. All subjects, except for S18, who has been described previously and harboured a single amino acid deletion (N99del),22 demonstrated severe mental retardation, facial dysmorphism, and delayed motor development. Hirschsprung disease was present in about half the cases (9/19), microcephaly in 16/19, and epilepsy in 15/19. Congenital heart disease and a partial or complete defect of the corpus callosum were observed in nine and six patients, respectively. Defects of midline structures, such as hypospadias, cryptorchism, or hypertrophic pyloric stenosis (HPS)15 were found in eight patients. Pulmonary artery sling (PAS), a rare anomaly in which the left pulmonary artery, arising from the right pulmonary artery, forms a vascular ring around the trachea, causing tracheal compression,23 was observed in patient S28. This is the first case of PAS associated with a ZFHX1B mutation, and was complicated by patent ductus arteriosus (PDA) and congenital tracheal stenosis. Likewise, a septum of the vagina was observed in S38, the first such defect seen together with a ZFHX1B mutation. The affected individual also had periodically repeated vomiting attacks since 7 months of age. Initially, we believed her vomiting to be caused by a surgical problem with gastroesophageal reflux (GER) after Ramstedt pyloromyotomy. Therefore, Nissen’s fundoplication and percutaneous endoscopic gastrostomy were performed at 8 months of age. Despite this treatment, her vomiting did not decrease. Thus, we suggested the repeated vomiting attacks might be related to her epilepsy, although detailed examination of her EEG was not performed during the attacks. Currently, the attacks are well controlled with sodium valproate and chlorpromazine. It should be noted that S4, S15, S19, and S31 suffered from the same repeated vomiting attacks.
Clinical features associated with deletions of 2q22–q24.1
The clinical features of eight patients with deletions including ZFHX1B are summarized in table 2. In six cases (S1, S3, S19, S27, S41, and S42), the features were quite similar to those of patients with ZFHX1B mutations (table 1); patient S3 had been previously shown to be negative for any abnormality in ZFHX1B.11 Two patients (S23 and S30) who were shown to have chromosomal deletions by cytogenetic analysis, had midline facial defects (a broad nasal bridge for S23 and a cleft of the hard palate for S30), complicated congenital heart disease (tetralogy of Fallot for S23 and DORV, PA, AVSD, hypoLV, left-PDA, and RAA for S30), and severe developmental delay, which is expected to preclude them from sitting without support (fig 2A, C). Tetralogy of Fallot (TOF) has been observed in three patients having nonsense or frameshift mutations in ZFHX1B.19 One of our cases (S23) has experienced several anoxic spells since the age of 6 months. These could be well controlled with calteolol (0.2 mg/kg) and phenobarbital (5 mg/kg) for some time, but the spells eventually returned. Finally, a right-modified Blalock-Taussig shunt for TOF was performed at the age of 1 year and 10 months. Another patient (S30) underwent a left-modified Blalock-Taussig shunt and PDA ligation at the age of 34 days, because of PDA-dependent pulmonary high flow. This patient also exhibited severe constipation and abdominal distension at birth. However, during treatment for congenital heart disease, the patient’s abdominal symptoms spontaneously disappeared. Thus, further examination of HSCR was not performed due to the patient’s severe general condition.
Determination of BAC clones containing deletion breakpoints
Using the multiplex PCR method, five patients (S3, S19, S27, S41, and S42) with features consistent with ZFHX1B truncation mutations were newly shown to have deletions including a part or all of ZFHX1B, in addition to S1, who had been previously reported.11 By repeating many sets of multiplex PCR, we finally determined both sides of the breakpoint in a single BAC clone from each patient as follows (BAC clones containing the centromeric side of the breakpoint, and the telomeric side of the breakpoint): S1 (RP11-257N13, RP11-699O7); S3 (RP11-95O9, RP11-107E5); S19 (RP11-107E5, RP11-699O7); S27 (CTD-2252P21, RP11-474G7); S41 (RP11-4C18, RP11-514A9); and S42 (RP11-95O9, RP11-357J9) (data not shown). We determined next the breakpoints in two patients (S23 and S30) with large deletions detectable by cytogenetic analysis and severe developmental delay (table 2) compared to patients with nonsense and frameshift mutations in ZFHX1B using the same method. BAC clones containing the centromeric and telomeric side of the breakpoint of deletion in one allele were (RP11-107E5, RP11-157M22) for S23 and (RP11-186C15, RP11-484E6) for S30 (data not shown). As seen previously for S1,11 microsatellite analysis demonstrated that the deletion occurred in the paternal allele in the four newly analyzed patients (fig 3).
Determination of nucleotide sequences at breakpoints
To evaluate the molecular mechanisms underlying the deletions at 2q22–q24.1, the precise sequences at the breakpoints were determined. Further sets of multiplex PCR were conducted to limit the regions of both breakpoints as close as possible for seven patients other than S1, who has an unbalanced reciprocal translocation. We performed PCR amplification of the DNA fragments including the breakpoint using primer pairs located on each side of the breakpoint. In five cases, fragments could be amplified and sequenced directly, and the precise deletion sizes were determined to be 1.96, 2.25, 0.33, 10.42, and 8.83 Mb (fig 1). Nucleotide sequence analysis around the breakpoint demonstrated that one deletion in patient S27 occurred within an L1 repeat-sequence region. However, the other patients’ breakpoints occurred in regions without any specific repeating sequences. Moreover, there was no sequence similarity for the regions on the two sides of the breakpoints in any of the five patients.
In the present study, we focused on 27 patients with heterozygous changes in ZFHX1B, including mutations and chromosomal deletions of part or all of the gene. Those with nonsense or frameshift mutations in ZFHX1B consistently presented with severe mental retardation, delayed motor development, and facial dysmorphism in various combinations with microcephaly, epilepsy, HSCR, congenital heart diseases, and abnormal brain findings, including agenesis of the corpus callosum. In addition to these clinical features, non-specific findings including vaginal septum, HPS, and hypospadias were encountered, likely caused by failure of midline structures to form.15 Repeated vomiting attacks suggestive of epilepsy were observed in five cases. Moreover, PAS was first identified in a patient with a ZFHX1B mutation. Despite this wide range of clinical features, all of the patients had heterozygous nonsense and frameshift mutations only in ZFHX1B. Zfhx1b (popularly known as Sip1) was recently isolated from mouse embryo using a yeast two-hybrid system as a Smad-interacting protein. It is a member of the vertebrate ZFHX1 family, and contains a Smad-binding domain, a homeobox-like domain, a transcriptional corepressor CtBP (carboxyl-terminal binding protein) interacting domain, and two separate clusters of zinc fingers (ZF) at its amino and carboxy terminals. Each ZF cluster binds independently, as a DNA-binding transcriptional repressor, to the promoter regions of genes like CDH1 and Cdh1, encoding human and mouse E-cadherin.24,25 A recent study demonstrated that full-length ZFHX1B can repress CDH1 in a CtBP-independent manner.26 This suggests that the CtBP interacting domain located at amino acid numbers 757–863 in ZFHX1B may have a different function than that of the ZF domains. Because all nonsense and frameshift mutations identified in patients are located on the amino-terminal side of the CtBP interacting domain, the prematurely truncated ZFHX1B in patients likely lack the CtBP-binding and ZF-mediated transcriptional repression activities encoded in the carboxy end of the protein. Thus, if the prematurely truncated protein is stable in certain patients, it is important to analyze its function in association with the clinical features of the disease. Defects in homozygous Zfhx1b-null mice show a more severe phenotype, such as a failure of neural tube closure and delaminating arrest of cranial neural crest cells,27 demonstrating that ZFHX1B is essential for the normal development of brain and neural-crest precursors as well as midline structures in mammals.
Six patients with deletions located around 2q22, which included some of 2q23 in S1, S19, and S27, and encompassed part or all of ZFHX1B in all cases, demonstrated mental retardation, microcephaly, and hypertelorism in common. However, HSCR, epilepsy, congenital heart diseases, and abnormality of the corpus callosum were also observed in five, five, three, and two patients, respectively. Clearly, a wide variety of phenotypes is associated with deletions in these patients. These observations indicate that the clinical features in patients with deletions around 2q22 that encompass part or all of ZFHX1B are quite similar to those in cases with nonsense or frameshift mutations in ZFHX1B. This is consistent with cases reported recently.28 Therefore, our data suggest that screening for deletions of ZFHX1B is warranted in patients presenting with the Mowat-Wilson phenotype, characterized by isolated cases, severe mental retardation, characteristic facial features, and microcephaly, even in the absence of a ZFHX1B mutation. S1 was previously considered to have more severe delayed psychomotor development than the three patients with nonsense and frameshift mutations, due to findings in S1 including DQ and brain atrophy.11,17 Now, her DQ is 7 and she is able to walk with support. Moreover, a recent FISH study showed that the chromosomal breaking point at 13q22 in S1 was within two BAC clones of RP11-452B18 and RP11-115N13 (Ono et al, unpublished data). The interval spanned by these two BAC clones is about 1.1 Mb, and currently no deletions have been identified in this region. The differences between the clinical features caused by the 5.1 Mb deletion around 2q22 in S1 and those seen in other patients with mutations (table 1) or deletions up to 3 Mb (table 2) in ZFHX1B do not seem nearly as great as they did in the past. Thus, her clinical features are likely to be just phenotypic variations of the disease. Indeed, her clinical features as well as clinical course are quite distinguishable from those of S23 and S30, who had significantly delayed psychomotor development in association with large deletions at 2q22–q24.1. S30 also had a cleft of the hard palate and complicated congenital heart diseases such as DORV and hypoLV as additional features. These clinical features have not been reported in patients associated with ZFHX1B mutations and deletions, even though a cleft soft palate has been noted in cases with ZFHX1B mutations.19 Moreover, a patient with an 11 Mb deletion reported by Zweier et al28 presented with the remarkable clinical features of early seizures with a lethal course and hypoplasia of the big toes. Thus, the genes located at the region shared in the deletion interval at 2q23 in these three patients, such as ARHE, which encodes RhoE and functions in integrin-based focal adhesions and actin polymerization,29 are candidates to affect brain development, midline structures of the face, and formation of the heart and extremities. However, these findings are not observed in another patient with a large deletion30 including 2q23. Detailed clinical and molecular analyses of a number of patients with deletions at 2q23 that do not include ZFHX1B are important to consolidate our hypotheses concerning the role of the 2q23 deletion in this clinical phenotype.
We have also clarified the nucleotide sequence around the breakpoints of deletions in five patients. In a further two cases, S3 and S41, the breakpoints could be limited within a 1 kb region by repeating the multiplex PCR method, although we did not unequivocally determine the breakpoint sequences. The results suggest that the deletion mechanisms at 2q22–q24.1 are not simple, and that complexity might be associated, for example, with rearrangement of portions of genomic DNA. It should be noted that deletions in five cases (S19, S23, S27, S30, and S42) were not caused by mechanisms featuring inappropriate recombination between homologous sequences. The findings for all patients, including previously reported cases15,28 examined thus far, have paternal deletions, also suggesting that chromosomal rearrangements occurred non-specifically around chromosome 2q22 during spermatogenesis. However, more detailed analyses of many patients are required to confirm this conclusion.
We are grateful to the patients and their families for agreeing to participate in this study. We also thankfully acknowledge the help of Dr John Tolmie (Yorkhill Hospitals).
This study was supported by a grant from Research on Psychiatric and Neurological Diseases and Mental Health, the Ministry of Health, Labor, and Welfare, Japan (to NW).
Conflicts of interest: none declared.