|Home | About | Journals | Submit | Contact Us | Français|
The Williams–Beuren syndrome (WBS) locus, at 7q11.23, is prone to recurrent chromosomal rearrangements, including the microdeletion that causes WBS, a multisystem condition with characteristic cardiovascular, cognitive, and behavioral features. It is hypothesized that reciprocal duplications of the WBS interval should also occur, and here we present such a case description. The most striking phenotype was a severe delay in expressive speech, in contrast to the normal articulation and fluent expressive language observed in persons with WBS. Our results suggest that specific genes at 7q11.23 are exquisitely sensitive to dosage alterations that can influence human language and visuospatial capabilities.
The underlying genetic bases for the majority of cases of language impairment have been postulated to be complex, involving several loci that interact with one another and the environment to produce an overall susceptibility to disease onset.1 Clues to the discovery of which genes potentially influence language ability may be found in mendelian disorders that have distinctive language components to their clinical phenotype. The Williams–Beuren syndrome (WBS) is one such neurodevelopmental disorder, in which persons show considerable strength in expressive language relative to their overall level of intellectual ability.2 WBS is also associated with a recognizable facies, supravalvular aortic stenosis, hypersensitivity to sound, visual impairment, dental problems, growth deficiency, infantile hypercalcemia, musculoskeletal abnormalities, and a hoarse voice.3 The syndrome is caused by the recurrent deletion of a specific set of genes, so it provides a unique opportunity to identify genes that are directly involved in language ability.4
The chromosomal locus that is deleted in WBS (on chromosome 7, band q11.23) is prone to deletion because it is flanked by blocks of DNA that have a very high degree of similarity to one another (called low copy repeats [LCRs]).4 The deletions, which almost invariably span a common interval, are caused by nonallelic homologous recombination within the LCRs of either the same chromosome 7 (i.e., intrachromosomal) or different chromosome 7s (i.e., interchromosomal). In each case, the chromosomes are envisaged to form loops, thereby allowing the alignment of the two LCRs, the occurrence of recombination, and the excision of the DNA contained within the intervening loop.4 The syndrome occurs at a frequency of approximately 1 in 7500 live births, with approximately two thirds of the deletion events being interchromosomal.5 Other microdeletion disorders — including the velocardiofacial syndrome, the Smith–Magenis syndrome, the Prader–Willi and Angelman syndromes, and hereditary neuropathy with liability to pressure palsies — are also mediated by nonallelic homologous recombination.6 For each of these microdeletions, a reciprocal duplication disorder has also been identified: dup22q11.2, dup17p11.2, dup15q11-q13, and Charcot–Marie–Tooth type 1A, respectively.6
Theoretically, duplications of the WBS region should occur at the same frequency as deletions occurring through interchromosomal nonallelic homologous recombination6 (i.e., at a frequency of 1 in 13,500). However, such duplication has not yet been described, even though there has been ample opportunity for the observation of such rearrangements in studies of patients with WBS-like clinical features but no deletion.4,7,8 That duplications have not been found in these patients suggests that the effect of duplication of the WBS locus is either lethal or has no observable consequence. Alternatively, the duplication could result in a clinical presentation that does not overlap with that of WBS and is therefore not represented in this study population. The patient we describe here, who has an exact duplication of the WBS region, has a phenotype that includes a severe delay in expressive language —a characteristic that is distinct from any of the typical clinical features seen in WBS.
The proband (Patient 1) was born after a full-term pregnancy complicated by maternal Crohn’s disease and poor weight gain. (Patient 1’s mother died when he was four years three months old.) His birth weight was 2520 g (below the 5th percentile), his length was 44.5 cm (below the 5th percentile), and his head circumference was 33 cm (at the 10th percentile). He initially required gavage feeding and was evaluated for failure to thrive and hypotonia at 13 months. According to a parental report, he sat at one year and walked at two years. He was evaluated by a speech pathologist and given diagnoses of moderate-to-severe language delay at the age of two years and severe delay in receptive and expressive language at the age of three years two months. When he was four years two months of age, attention deficit–hyperactivity disorder (ADHD), an overall developmental delay, an unspecified sleep disorder, and a severe delay in speech and language and in fine motor skills were diagnosed. Results of behavioral audiometry and otoacoustic emissions were consistent with the presence of normal hearing. At the age of eight years, Patient 1 was again evaluated by a speech pathologist, and childhood apraxia of speech was diagnosed. Additional clinical, biochemical, and genetic analyses did not reveal any abnormalities (see the Supplementary Appendix, available with the full text of this article at www.nejm.org).
Patient 1’s sister (Patient 2) was also given a diagnosis of ADHD, and both were receiving 5 mg of dextroamphetamine sulfate (Dexedrine) twice daily. The history revealed that both parents had also had attention and academic difficulties but that Patient 1 was the only member of the family who had difficulty with language acquisition.
Written informed consent was obtained from the guardians of the patients for genetic and psychological studies, which were approved by institutional ethics review boards. At the time of our assessment, Patient 1 was 8 years 10 months old, and Patient 2 was 11 years 1 month old.
Patient 1 had growth retardation and mild dysmorphism. His physical examination showed height, weight, and head circumference at the 2nd, 5th, and 30th percentiles, respectively. Dysmorphic features included dolichocephaly, a high and narrow forehead, long eyelashes, a high and broad nose, a short philtrum, a high-arched palate, dental malocclusion (specifically, an anterior open bite), retrognathia, and asymmetric crying face (Fig. 1 and Table 1). Patient 1 did not have any of the facial features of WBS.3 He had bilateral simian creases, and the left hand was smaller than the right. On neurologic examination, he was noted to have very mild dysmetria and mild difficulty with tandem gait and unipedal stance.
Patient 1’s difficulties with expressive language were immediately apparent. He was able to pronounce only a very small number of words correctly; most words were approximations composed of the first consonant (or a related consonant) and the first vowel (or a related or neutral vowel) of the target word. The results of the intellectual and vocabulary assessments are summarized in Table 2. Patient 1’s performance on the Differential Ability Scales9 differed from his sister’s only on the verbal cluster, in which verbal responses, manual signs, gestures, pantomime, and drawing were all considered acceptable responses from him. If only verbal responses had been accepted, his verbal-cluster standard score would have been considerably lower.
To provide a direct comparison of Patient 1’s receptive and expressive language abilities, the Pea-body Picture Vocabulary Test,10 which measures receptive vocabulary, and the Expressive Vocabulary Test11 were administered (Table 2). Patient 1’s standard score on receptive vocabulary was in the low-average range (age equivalent, 6 years 10 months). In contrast, his standard score on expressive vocabulary was in the severe-impairment range (age equivalent, two years three months), even though both word approximations and manual signs were considered acceptable responses. Similar results were obtained with the Clinical Evaluation of Language Fundamentals test12 (see the Supplementary Appendix).
Although Patient 1’s standard scores on the non-verbal reasoning and spatial clusters of the Differential Ability Scales were in the range of mild mental deficiency, his performance was very similar to his sister’s and was consistent with the type of difficulties his mother was also reported to have had, indicating that the duplication was not the primary basis for these difficulties. Both children’s scores on the Conners’ scales13 (found in the Diagnostic and Statistical Manual of Mental Disorders, 4th ed.) were consistent with their previous diagnosis of ADHD.
Patient 1 was initially referred for testing for the velocardiofacial syndrome and was screened with the use of a real-time method on the basis of a polymerase-chain-reaction (PCR) assay, which detected the duplication of markers within the WBS critical region.14 Fluorescence in situ hybridization (FISH) with probes from 7q11.237 showed that the in-tandem duplication was limited to the region commonly deleted in WBS (Fig. 2A). Probes from within the WBS common deletion (CTA-208H19 and RP5-1186P10) gave three signals on interphase FISH, whereas probes flanking the deletion region (RP11-815K3 and CTB-139P11) gave only two signals (Fig. 2A; some data not shown). Cosmid LL07NCO1-207g3, which lies between the medial and telomeric LCRs, also gave only two signals, indicating that the duplication was restricted to the region spanning the centromeric and medial LCRs, which corresponds to the region commonly deleted in WBS. No duplication was present in Patient 1’s father or sister or in more than 250 other controls with or without WBS who were also examined.
Analysis of single-copy microsatellite markers from within the WBS region revealed that Patient 1 carried three distinct alleles at loci D7S2476, D7S3194, and D7S1870 (Table 1 of the Supplementary Appendix). In each case, only one of the alleles was present in the father, indicating that Patient 1 had inherited two different copies of the WBS region from his mother. Analysis of the maternal grandparents revealed that the duplicated chromosome in the proband contained segments of chromosome 7 that were inherited independently from each maternal grandparent (Table 1 and Fig. 1 of the Supplementary Appendix). These analyses indicate that the meiotic interchromosomal recombination that led to the duplication took place in the mother’s germ cells, making it a new rearrangement.
Further narrowing of the duplication breakpoints could not be accomplished with FISH or microsatellite analysis, owing to the high sequence identity of the LCRs. We used site-specific nucleotide (SSN) dosage analysis to define the relative number of B block–type copies (centromeric, medial, or telomeric) at each position analyzed, an approach that enabled us to infer the position where the recombination had occurred.4 Analysis of seven SSNs showed that in Patient 1 the transition between blocks Bm and Bc occurred between SSN 4 and SSN 6, within the NCF1 gene (Fig. 2B). The majority of WBS deletion breakpoints (>90 percent) also occur within the B block,4 indicating that the duplication is the exact reciprocal of the common WBS deletion.
Gene-expression analysis in lymphoblastoid cell lines by real-time PCR demonstrated that five of six genes that were examined within the duplicated region (GTF2I, LIMK1, WBSCR1, RFC2, and BAZ1B) showed increased expression in Patient 1 and reduced expression in persons with WBS (Table 2 of the Supplementary Appendix). The exception was WBSCR5, which showed reduced expression in persons with WBS but levels in Patient 1 that were consistent with levels in controls. WBSCR16, located just outside the telomeric deletion or duplication, did not show altered expression in either Patient 1 or in persons with WBS, indicating that the rearrangement breakpoints do not affect its transcription. The nearest gene outside the centromeric breakpoint, the gene encoding calneuron 1, was not expressed in lymphoblasts, but it is at least 300 kb from the proximal B block4 and is separated from the breakpoint by the complex and actively transcribed LCRs; hence, its expression unlikely to be affected.
We describe a syndrome that is associated with a reciprocal duplication of the WBS microdeletion region. Patient 1 has intellectual strengths and weaknesses that are in direct contrast to those of children with WBS. Expressive language, especially syntax and phonology, is the area of greatest weakness for Patient 1, whereas for children with WBS, expressive language is a relative strength.2 When Patient 1 had difficulty making himself understood, he often successfully resorted to drawing what he was trying to express. Visuospatial construction (including drawing) is the area of greatest weakness for children with WBS, and most children with WBS who are Patient 1’s age are able to draw only a few recognizable objects.15
Previous descriptions of persons with larger duplications of the region (supernumerary ring chromosome 7) have noted delay in expressive language or impairment accompanied by articulation problems, but none of these reports contain standardized assessment results or comparative data on expressive and receptive language.16 Karyotypes were derived from G-banding, not from molecular analysis, so the extent of each duplication remains unknown. Our findings specify the expressive-language phenotype associated with dup7q11.23 and define the precise region of chromosome 7 contributing to it as the 1.5-Mb interval commonly deleted in WBS.
There appears to be a subtle but recognizable facial phenotype that is shared by both Patient 1 and previously described persons with supernumerary ring chromosome 7, one that consists of a high and broad nose, posteriorly rotated ears, a high-arched palate, and a short philtrum.16–18 When accompanied by delay in expressive language, this gestalt warrants testing for duplication of the WBS critical region.
Although there is a strong genetic component to language impairment, so far only the transcription factor FOXP2 has been implicated as a cause of the problem, and only in a few cases.19 Disruption of FOXP2 results in reduced functional dosage and leads to deficits in both expressive and receptive language, in addition to orofacial dyspraxia that impairs the coordination of complex fine-motor movements of the lower face.19,20 The remaining cases of language impairment were predicted to involve complex causes.1 The identification of a second locus associated with language impairment is therefore unexpected and opens up the possibility of linking the expressive-language phenotype to a specific gene or genes from within the WBS region.
The contrast between phenotypes associated with deletion and those associated with duplication of the WBS region, as well as the accompanying changes in gene expression, suggests that genes within this region are dosage sensitive. These observations also suggests that perturbation can have a dramatic, negative effect on both language development and visuospatial construction ability.
The duplication or deletion encompasses 26 to 27 genes, many of which can be ruled out as important contributors to the WBS phenotype on the basis of the correlation of genotype with phenotype in persons with atypical deletions.21–24 The minimal critical interval that must be deleted in classic WBS spans the region between the gene encoding elastin and the common distal breakpoint and encompasses just nine genes. Included within the WBS minimal critical interval are three general transcription factor 2 I (GTF2I) genes, encoding a family of proteins that are predicted to possess some functional redundancy.25 In a simple or more complex model, alteration of expression of a single gene — or in the case of the GTF2I family, possibly a combination of genes — might lead to a distinctive language-impairment phenotype that overlaps with that observed in Patient 1. Our discovery of the WBS duplication provides both the long-sought-after reference case for ascertainment of other patients and an entry point toward the discovery of new genes required for the normal acquisition and expression of language. The association between expressive-language delay and duplication of the WBS region also has implications for the molecular diagnosis of language delay.
Supported by grants from the Canadian Institutes of Health Research and the Sick Kids Foundation (both to Dr. Osborne), the Spanish Ministries of Health and Science and Technology (to Dr. Pérez-Jurado), Genome Canada/Ontario Genomics Institute and Genome Spain (to Drs. Pérez-Jurado, Scherer, and Osborne), the National Institute of Neurological Disorders and Stroke (NS35102, to Drs. Mervis and Morris), and the National Institute of Child Health and Human Development (HD29957, to Dr. Mervis). Dr. Scherer is a Canadian Institutes of Health Research investigator and an international scholar of the Howard Hughes Medical Institute, and Dr. Osborne is a Canadian Institutes of Health Research scholar.
We are indebted to the patient and his family for their participation in our study and to the Centre for Applied Genomics at the Hospital for Sick Children, Toronto, for technical assistance.