Article Text


Characterisation of deletions of the ZFHX1B region and genotype-phenotype analysis in Mowat-Wilson syndrome
  1. C Zweier1,
  2. I K Temple2,
  3. F Beemer3,
  4. E Zackai4,
  5. T Lerman-Sagie5,
  6. B Weschke6,
  7. C E Anderson7,
  8. A Rauch1
  1. 1Institute of Human Genetics, Friedrich-Alexander University of Erlangen-Nuremberg, Erlangen, Germany
  2. 2Wessex Clinical Genetics Service, Southampton University NHS Hospital Trust, Southampton, UK
  3. 3Department of Biomedical Genetics, University Medical Centre Utrecht, The Netherlands
  4. 4Clinical Genetics Center of The Children’s Hospital of Philadelphia, USA
  5. 5Metabolic Neurogenetic Clinic, E Wolfson Medical Centre, Holon, Israel
  6. 6Department of Paediatric Neurology, Charité Campus Virchow-Klinikum, Humboldt University, Berlin, Germany
  7. 7SCHC Pediatrics, Philadelphia, USA
  1. Correspondence to:
 Dr A Rauch, Institut für Humangenetik, Schwabachanlage 10, 91054 Erlangen, Germany; 

Statistics from

In 1998, Mowat et al1 delineated a syndrome with Hirschsprung disease (HSCR) or severe constipation, microcephaly, mental retardation, and a distinctive facial appearance.1 Because two of the patients had a cytogenetically visible deletion of 2q22-q23,1,2 and all patients were sporadic cases, a contiguous gene syndrome or a dominant single gene disorder involving this locus were suggested.1 Two similar patients with cytogenetically balanced translocation t(2;13)(q22;q22) and t(2;11)(q22.2;q21), respectively, allowed Wakamatsu et al3 and Cacheux et al4 to narrow down the critical interval to 5 Mb and to one single gene respectively, which led both groups independently to the detection of intragenic mutations in the gene coding for Smad interacting protein-1 (formerly SIP1, now called zinc finger homeobox 1B (ZFHX1B)) in patients with so called “syndromic HSCR”. However, because HSCR is not an obligatory symptom and patients with and without HSCR can be recognised by other features, especially their distinct facial gestalt,5,6 we suggested that “Mowat-Wilson syndrome” (MWS) is a more appropriate name.6

Although the developmental ZFHX1B expression pattern fully explains the clinical spectrum observed in patients with Mowat-Wilson syndrome by haploinsufficiency of this gene alone,5,7 Wakamatsu et al3 initially stated that their deletion patient would have a more severe phenotype and therefore would have a contiguous gene syndrome. Amiel et al8 reported that the phenotype was similar in patients with “syndromic HSCR” caused by mutations and cytogenetically non-visible large scale deletions of the ZFHX1B locus, respectively, but the deletion sizes were not delineated. We therefore analysed deletion size and genotype-phenotype correlation in four new patients with cryptic deletions of the ZFHX1B locus.

Key points

  • Mowat-Wilson syndrome (MWS) is a distinct multiple congenital anomalies-mental retardation syndrome characterised by severe mental retardation, recognisable facial gestalt, pre- or postnatal microcephaly, and postnatal growth retardation, as well as seizures (82%) and malformations such as Hirschsprung disease (67.6%), congenital heart defects (47%), and agenesis of the corpus callosum (35%), caused by mutations or large scale deletions of the ZFHX1B gene in 2q22.

  • Deletion sizes and breakpoints in Mowat-Wilson syndrome patients vary widely from 300 kb to at least 11 Mb, thus ruling out a true microdeletion syndrome.

  • So far parental origin has only been determined in four patients and has always been paternal.

  • In general, patients with deletions are very similar to those with truncating mutations. There was no correlation between the phenotype and size of deletion up to 5 Mb. However, one patient with a larger deletion (~11 Mb) had early seizures with a lethal course and hypoplasia of the big toes as additional features.



The diagnosis of Mowat-Wilson syndrome was made in patients 3 and 4 (fig 1C) because of HSCR and associated features and in patients 1 and 2 because of mental retardation associated with the distinct facial gestalt (fig 1A, B) in the absence of HSCR. Clinical details are provided in table 1; patient 2 will be described in more detail elsewhere.9

Table 1

Phenotype of previously published patients with Mowat-Wilson syndrome and mutation, deletion, or translocation breakpoint in the ZFHX1B gene and of the present patients

Figure 1

(Top row) Facial appearance of patient 1 in early childhood (A) and aged 8 years (B), and of patient 4 aged 10 years (C). (Bottom row) Note similarity to patients with ZFHX1B point mutation reported elsewhere.6 (D, E) Patient with nt553-554insTG mutation aged 6 months and 6 years 10 months, respectively. (F) Patient with nt1892delA mutation aged 3 years 10 months.


Conventional chromosome analysis was performed from cultivated peripheral blood cells after GTG and CBG banding at a 550–850 band level10 according to standard protocols. FISH analysis was performed with directly labelled BAC probes on metaphase spreads as described previously.11 In patient 3 and his parents, additional polymorphic markers D2S2184 (location 147.5 Mb), D2S2335 (147.2 Mb), D2S2277 (148.3 Mb), D2S2324 (148.6 Mb), D2S2275 (150.3 Mb), D2S2299 (152.2 Mb), D2S2241 (153.0 Mb), and a newly created marker from the ZFHX1B locus were analysed with an ABI 310 capillary sequencer as described previously.11 The ZFHX1B marker is localised between bp 149003 and 149221 of BAC RP11-107E5 and was amplified with the following primers: ZFHX1Bms1f-gctgcagtagttgcctttga and ZFHX1Bms1r-gtcctttcgaggtccagttg.


Results of FISH and marker analysis are shown in fig 2. In patient 3 the distal border of the deletion was determined with polymorphic markers, which showed a distal breakpoint between markers D2S2275 (150.3 Mb) and D2S2299 (152.2 Mb), and a paternal origin of the deletion. Deletions were of differing sizes, approximately 300 kb in patient 4, 700 kb in patient 2, 5 Mb in patient 1, and 11 Mb in patient 3. The mothers of patients 2 and 4 and both parents of patient 3 were available for FISH analysis, which showed normal results. The phenotype observed in patient 3, with the largest deletion, showed early seizures, hypoplastic big toes, and premature death at the age of 4 months as additional features (table 1).

Figure 2

Results of FISH analysis with several BAC clones (RP11 library) in present patients P1-P4, informative results of analysis of polymorphic markers in patient 3, and deletion size in published cases (W = Wakamatsu et al,3 M = Mowat et al,1 A = S.203 from Amiel et al8). − lack of signal on one chromosome 2; + regular signals on both chromosomes 2; × 2: two alleles; pLOH: non-transmission of a paternal allele.


Our results indicate that deletion sizes and breakpoints in Mowat-Wilson syndrome patients vary widely, ruling out a true microdeletion syndrome with recurring breakpoints mediated by low copy repeat regions. There was generally no obvious correlation between the phenotype and the size of the deletion and the phenotypic spectrum was similar to that observed in patients with truncating mutations within ZFHX1B (table 1). The only remarkable difference was noticed in patient 3 with the 11 Mb deletion, who presented with seizures much earlier, had marked hypoplasia of the big toes, and who died in early infancy. Thus, genes within the close vicinity of the ZFXH1B gene seem not to be subject to gross haploinsufficiency. Parental origin was only determined in one of the present and three published patients,3,8 and was of paternal origin in all cases investigated. As all investigated patients had HSCR and congenital heart defects, it is not possible to draw any conclusion about these symptoms, but agenesis of the corpus callosum was present and absent in two patients each, and thus shows no correlation with parental origin of the deletion. Similarly, the early onset of seizures in patient 3 is also not attributable to the parental origin of the deletion.

The most frequently observed major malformation in Mowat-Wilson syndrome is HSCR, which occurred in 21 of 30 (70 %) patients reported so far (table 1). As has been described for patients with ZFHX1B truncating mutations, two of our patients with deletions of approximately 700 kb and 5 Mb, respectively, did not have HSCR, while the two with the smallest and largest deletion (300 kb and 11 Mb deletions, respectively) did have it. Thus, our results suggest that the manifestation of HSCR is not influenced by deletion size. As ZFHX1B knockout mice also do not exhibit HSCR,12 a non-allelic modifier might contribute to the manifestation of HSCR. The high rate of HSCR in humans is probably the result of recognition bias, as in our cohort (four patients reported earlier6 and the present four patients) HSCR occurs in only 50%. Less frequent malformations include various congenital heart defects (for example, septal defects, pulmonary stenosis, or atresia), agenesis of the corpus callosum, urogenital anomalies, talipes, and strabismus.

Similarly, there is no difference in degree of mental retardation, facial appearance, and growth parameters. Regardless of the underlying defect, which may be a truncating mutation in ZFHX1B or a large scale deletion, psychomotor retardation is severe with a mean walking age of 4–5 years and speech starting at the age of 5–6 years, being restricted to single words. Personality is generally happy and affectionate. Although shortness of stature and low weight are characteristic in school age children, birth measurements are usually normal or even in the upper normal range. Only microcephaly was already evident at birth in eight out of 19 patients with reported measurements (table 1), and it has a tendency to occur before the decline of body length in our patients. Therefore, our findings do not support the initial statement by Wakamatsu et al3 about the more severe phenotype in their deletion patient. Nevertheless, severe cerebral atrophy remains remarkable in this patient, but might be related to the other translocation breakpoint on chromosome 13.

A seizure disorder with varying age of onset is a very common feature which is found in 82% of all 34 patients (table 1). Severe neonatal seizures, however, have been reported only in our patient 3 and a patient with a cytogenetically visible deletion.2 Thus a gene(s) responsible for early seizures with a lethal course and hypoplastic big toes might be located between BAC RP11-207O14 at 145.3 Mb and marker D2S2299 at 152.2 Mb where at least one gene related to epilepsy, CACNB4 (OMIM 601949), is known to be located. However, detailed analysis in further patients is required for confirmation of this putative association.

The characteristic facial appearance was evident in all patients with deletion or truncating mutations and allows the distinction between Mowat-Wilson syndrome and other types of “syndromic HSCR” such as Goldberg-Shprintzen syndrome. The facial features are probably diagnosable in the neonatal period in the presence of HSCR, but the sunken eyes, broad, flared eyebrows, pointed nasal tip, short philtrum, and upturned ear lobes become more obvious in early childhood. Of 12 patients with the distinct facial gestalt of Mowat-Wilson syndrome analysed in our laboratory so far (data not shown), eight had truncating mutations and four had large scale deletions, thus giving a ZFHX1B defect in 100% of patients and a deletion rate of 33%. However, it is possible that less severe cases are being missed following the work of Yoneda et al,13 who described an atypically mild phenotype with late adult onset severe constipation and mild mental retardation in the absence of specific facial anomalies, seizures, and other malformations owing to a non-truncating 3 bp in frame deletion.

It seems likely that more patients will soon be described with Mowat-Wilson syndrome now that the clinical features are becoming increasingly recognised by clinical geneticists and in time it will be possible to elucidate the true clinical spectrum.


We thank Michaela Kirsch and Leonora Bille for their excellent technical assistance.

Data access. BAC position was obtained from map viewer at Position of polymorphic markers within the sequence map were obtained from uniSTS at


View Abstract

Request permissions

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.

Linked Articles