Article Text
PDF
Statistics from Altmetric.com
Editor—DiGeorge syndrome (DGS, MIM 18840) and velocardiofacial syndrome (VCFS, MIM 192430) are associated with interstitial deletions of chromosome 22q11.2 and are considered to be phenotypic variations of the same underlying genetic defect. Studies of patients with diagnoses of DGS or VCFS have estimated that between 68% and 88% have 22q11.2 microdeletions detectable by fluorescence in situ hybridisation.(FISH).1-3 Most of the deletions arise de novo, but data collected from a large group of patients estimated that 28% of deletions are inherited.4 Of subjects with a deletion, 90% have the same approximately 3 Mb region deleted, 7% have a nested 1.5 Mb deletion, and in other rare cases, unique deletions and translocations have been detected.5
The common clinical features associated with DGS/VCFS are congenital heart malformation, abnormal facies, thymic hypoplasia, cleft palate, and hypocalcaemia. However, despite the vast majority of patients diagnosed with DGS/VCFS having the same region of 22q11.2 deleted, there is wide phenotypic variation. The combination of symptoms can be so severe that the patient dies in the neonatal period or so mild that the condition is diagnosed only after the birth of a more severely affected child.4 6 7 It is estimated that deletions of 22q11.2 occur in 1 in 4000 live births, with 75% of patients harbouring the deletion having some form of congenital heart disease (CHD).4 7 Conotruncal heart defects most commonly found in DGS/VCFS patients with 22q11.2 deletions are interrupted aortic arch (IAA) type B, truncus arteriosus (TA), and tetralogy of Fallot (TOF).4 7 8 Furthermore, congenital conotruncal cardiac defects account for around 50% of cardiac malformations seen in the neonatal period9 and approximately 50% of patients with conotruncal cardiac malformations have been found to have deletions at 22q11.2.10-13
It is therefore reasonable to suspect DGS or VCFS in patients presenting with specific conotruncal heart defects. As part of a study into DGFS/VCFS and immune function, we obtained blood from patients with heart defects commonly seen in DGS/VCFS undergoing corrective cardiac surgery in the neonatal period. Information about deletion status was not available at the time of corrective surgery.
Of particular interest was a male child born to consanguineous parents, who was found to have truncus arteriosus type II and interrupted aortic arch type B. His calcium status was normal, facial appearance was not abnormal, and although total T cell numbers were within normal ranges, CD8+ T cell numbers were low. A diagnosis of VCFS was made.
Methods
Blood was collected at surgery when the child was 7 days old. Deletion analysis was performed by FISH on metaphase chromosomes from peripheral blood lymphocytes using probe D22S75 (N25, Oncor). Two signals were obtained and the child was deemed not to harbour a 22q11.2 deletion. The child died aged 4 months. A year later, a female child was born to the same family. Her diagnosis was truncus arteriosus type II, but without an interrupted aortic arch. She was noted to have an abnormal facial appearance, with a small mouth, simple, cupped, low set ears, and sagging cheeks. Calcium status was normal and a diagnosis of VCFS was made. Blood was collected at corrective surgery when the child was 20 days old. FISH showed that there was no detectable deletion at 22q11.2. Parental metaphase chromosomes were also examined by FISH and no deletions were detected. Since DGS has also been associated with deletions of 10p13-14,14 chromosomes from both sibs and their parents were tested with probes SD10p1 and SD10p49 (kindly provided by Professor Peter Scambler). Again, signals were detected on both copies of chromosome 10 in all cells of both sibs and their parents.
Since conventional FISH probes had not detected deletions at 22q11.2 or 10p13-14, it was possible that one of the parents carried a microdeletion that was passed on to the affected children. Also, the pedigree of this family showed that there was a relatively high degree of consanguinity (fig 1). We genotyped the family for a series of polymorphic markers across the DiGeorge chromosomal region (DGCR) in order to identify whether one of the parents carried a small deletion that had not been detected by routine FISH, or whether there was a region of homozygosity at 22q11.2 in this consanguineous family that might be associated with CHD. Both affected children, their older unaffected sister, both parents, and both paternal grandparents were genotyped with the markers D22S420, D22S427, D22S1638, D22S941, D22S944, D22S264, D22S311, D22S306, and D22S42515 and a restriction fragment length polymorphism in the catechol-O-methyltransferase (COMT) gene.16 We also genotyped a novel dinucleotide repeat polymorphism (currently designated NLJH1) within cosmid clone 119f4 (AC004461).17 PCR primers for the marker are: forward, 5′-GTAATCCTGGGCA TC TATC-3′, reverse, 5′-TTCCAGTCTTT GGCTATTAC-3′.
Pedigree of the family with features of DGS/VCFS. The boxed area delineates the region of interest within the DGCR, which is coloured grey for VI.2, VI.3, and VI.4 where the region is homozygous.
Results
The initial genotyping was informative for all markers tested (fig1). The two affected sibs, VI.2 and VI.3, were found to be homozygous for three markers, NLJH1, D22S941, and D22S944. This region of homozygosity maps within both the commonly deleted 3 Mb region and the 1.5 Mb nested deletion found in a minority of DGS/VCFS patients. The region of homozygosity segregated with the presence of congenital heart defects associated with DGS/VCFS. Since the parents were heterozygous for all markers from D22S1638 to D22S264, there was no evidence to suggest that a microdeletion, undetectable by FISH, was carried on one of the parental chromosomes. To confirm that there was no microdeletion on chromosome 10, the affected sibs, the unaffected older sib, their parents, and paternal grandparents were also genotyped for markers D10S189, D10S547, and D10S191 on 10p13-14. The parents were informative for all three markers, and the affected sibs were discordant for all three loci (data not shown).
Homozygosity for the region that includes NLJH1, D22S941, and D22S944 is one interpretation of these data. However, there has been a previous report of DGS/VCFS in a sib pair as a result of germline mosaicism in the maternal chromosomes18 so we determined whether there might be hemizygosity at these loci in the affected sibs. Five FISH probes were generated from BAC clone 77h2 (AC000052), cosmid 49c12 (AC000079), cosmid 81h (AC000086), cosmid 91c (AC000091), and PAC p158l19 (AC006547). 77h2 maps between D22S427 and D22S420, and p158l19 maps distal to COMT, both regions for which the affected sibs are heterozygous. 49c12, 81h, and 91c map within the apparently homozygous region.17 Signals were detected on both copies of chromosome 22 in the younger affected sib for all five probes, ruling out germline mosaicism as the mode of transmission in this family (data not shown).
Allele frequencies were estimated from the six chromosomes of IV.8, IV.1, and V.1 (fig 1), who are all related, but do not appear to share any haplotypes. Linkage analysis performed on the haplotypes shown in fig 1 resulted in a multipoint lod score of 1.84. Ordinarily, a lod score of 3 is considered necessary to prove linkage at a given locus. However, where the link between locus and phenotype has been established beyond doubt in previous analyses using other families, there is said to be a prior probability of linkage which lowers the required threshold for proof in new families with the same phenotype. Further, this result was not obtained after testing many loci throughout the genome and selecting the most significant, but was the result of segregation analysis at the only two known DGS/VCFS loci. These data therefore provide strong evidence of recessively inherited DGS/VCFS in this family.
A fourth child was born to the family (VI.4, fig 1) after the analysis had been done for the other members. Genotyping showed that the child shared the extended haplotypes from D22S427 to D22S425 seen in the affected sibs. Chest x ray and echocardiogram were performed at 6 months and the aortic arch was found to be left sided. The facial appearance was assessed as normal by a clinical geneticist. However, we were not able to rule out an aberrant right subclavian artery. At 6 months, the child was clinically well, but too young for us to assess many of the other possible features of DGS/VCFS, such as hypernasal speech, learning difficulties, or psychiatric illness. Ethical approval and parental consent had been obtained for chest x ray, echocardiogram, and physical examination only, and so we were not able to assess immune cell numbers. The decision of the parents was that, unless clinically necessary, there should be no further investigations of the youngest child, and so we are currently not able to perform a more in depth evaluation.
Discussion
At first analysis, it might appear that the birth of the fourth child confounds our hypothesis that the region of homozygosity is associated with a DGS/VCFS phenotype. However, data from both mouse models and studies of DGS/VCFS patients show that even when the deletion is present, phenotypic variation is broad and penetrance can be incomplete. Studies have suggested that, despite the wide phenotypic variation, all subjects with a 22q11.2 deletion have some feature of DGS/VCFS.4 7 However, such studies are performed on the basis that all subjects included in the study carry the deletion, and testing for the deletion will have been as a result of the presence of some feature of DGS/VCFS. It remains a possibility that there are perfectly healthy subjects with undetected 22q11.2 deletions in the general population, but to test such a hypothesis would require screening an extremely large cohort of people with no features of DGS/VCFS for deletions of 22q11.2.
The region syntenic to human 22q11.2 is found on mouse chromosome 16 and studies have shown a high degree of gene conservation, although with some changes in gene order.19-22 Recently, engineered mouse models have begun to dissect out the genes that might be important in DGS/VCFS.23-25 Mice heterozygous for a deletion encompassing most of the region homologous to the human 22q11.2 DGS/VCFS deletion were generated. A total of 30% of heterozygous embryos and 18% of adult heterozygous mice presented with the cardiovascular abnormalities most frequently found in DGS/VCFS patients. Furthermore, the defects seen in the mice heterozygous for the deletion could be overcome by genetically complementing the deletion,23 supporting the idea that haploinsufficiency of a gene or genes within the deleted region is responsible for the phenotype. Mice carrying smaller deletions within the mouse DGCR have also been generated, but these show none of the cardiovascular abnormalities seen in DGS/VCFS.24 25 Taken together, data from the mouse models help narrow the region in which candidate genes for DGS/VCFS are likely to reside to that bounded byComt proximally andUfd1l distally. The homologous region on human chromosome 22q11.2 contains the same genes in the same order, but with the orientation reversed. This overlaps the region for which the two affected children in this family are homozygous. The maximum region of homozygosity is from D22S1638 proximally toCOMT distally, and is approximately 1 Mb; the minimum region is from NLJH1 proximally to D22S944 distally and is approximately 570 kb.
Current evidence supports the notion that when one copy of 22q11.2 is deleted, the DGS/VCFS phenotype occurs as the result of haploinsufficiency of a gene or genes that have been deleted. It is possible that in this family the associated haplotype carries a mutation that leads to a minor reduction in the product of a gene or genes in the region. Such a mutation might then have no discernible effect in heterozygotes, but homozygosity could lead to functional haploinsufficiency.
It is of particular interest that in the mouse model where cardiovascular abnormalities were found and in which there would have been a common genetic background, there was phenotypic variability and incomplete penetrance.23 Phenotypic variation in DGS/VCFS is well documented, and in familial cases where the affected members carry identical 22q11.2 deletions, phenotypic variation can be wide.4 6 7 Moreover, there are now several published reports of monozygotic twins with the deletion who are discordant, particularly for congenital heart disease.26-29
Further mouse models have recently been developed, refining the genes that are important in developing DGS/VCFS, which implicateTbx1 in the aetiology of the condition, particularly the cardiovascular malformations.30-32Our data are consistent with these models, as the human homologue,TBX1, is within the region for which the affected members of this family are homozygous.
In summary, we present evidence from a consanguineous family with two children presenting with conotruncal cardiac defects and features of DGS/VCFS and a region of homozygosity at 22q11.2. Further analysis of the homozygous region may facilitate the identification of the genes that are involved in cardiac morphogenesis.
Acknowledgments
We are grateful to the family for their participation and continued interest in the study. We wish to thank Professor Peter Scambler for constructive discussions and Professor Chris Inglehearn for advice and critically reading the manuscript. Chromosome 22 clones for FISH analysis were kindly provided by the Clone Resources Group, The Sanger Centre, Wellcome Trust Genome Campus, Hinxton, Cambs, UK (clonerequest{at}sanger.ac.uk). JH is a Wellcome Trust Research Career Development Fellow.
References
Footnotes
-
↵* Present address: Oxagen Ltd, Abingdon, UK